Placental DNA Methylation in Relation to Maternal Periconceptional Prenatal Vitamin Use and Child Outcomes in the Marbles Prospective Autism Study

Background:  Placental tissue, usually discarded at birth, is a potential rich source of for epigenetic biomarkers that the interface of genetic risk and in utero exposures in autism spectrum disorders (ASD). In addition, maternal use of prenatal vitamins containing the methyl donor folic acid could alter placental methylation in persistent ways that influence neurodevelopment, especially during a period of dynamic methylation and reprogramming around conception. Compared with other human tissues, placenta contains partially methylated domains (PMDs) that are more similar to oocytes and pre-implantation stages of development.

Objectives: This study was designed to identify regions of differential DNA methylation in placenta from a prospective ASD study in high-risk families. We also studied the relationship between maternal prenatal vitamin use and DNA methylation.

Methods: MARBLES (Markers of Autism Risk in Babies-Learning Early Signs) study involves families with at least one child with ASD so subsequent children were at significantly higher risk (nearly 18%) of having another autistic child and were planning another pregnancy. Mothers were interviewed about prenatal vitamin use during a new pregnancy. Placentas were collected for the younger siblings who were followed until they were 3 years old and clinically diagnosed with ASD or typical development (TD). MethylC-seq was performed on DNA isolated from 20 ASD and 21 TD male placentas using Illumina next-generation sequencing on HiSeq 2000 machine with one sample per lane using single-end 100 bp sequencing. We identified differentially methylated regions (DMR) using the DMR finder approach based on bsseq R package.

Results: Two DMRs showed significant differences after FDR correction between ASD and TD selected by DMR finder based on MethyC-seq data and validated by pyrosequencing (Fig. 1). One DMR that was hypomethylated in ASD compared with TD showed a positive association between prenatal vitamins taken during the first pregnancy month and percent methylation. The other DMR was hypermethylated and its methylation tended to be negatively associated with prenatal vitamin intake. Both DMRs were also associated with Mullen Scales of Early Learning on four subscales categories (Visual Reception, Fine Motor, Receptive Language and Expressive Language) and the Early Learning Composite (Fig. 2).

Conclusions: This relatively small study of DNA methylation differences in placental samples from the MARBLES prospective study identified two high confidence DMRs that could be useful in assessing risk for ASD at birth and determining the impact of maternal prenatal vitamin usage on ASD occurrence in offspring.