Background: The enzyme methylenetetrahydrofolate reductase (MTHFR) irreversibly commits folate to the end point of methylation in one carbon metabolism. Two single nucleotide polymorphisms (677C->T, 1298A->C) in the MTHFR gene have been shown to result in a thermolabile enzyme with reduced activity. The 677TT and 1298CC genotypes have each been associated with increased risk for neural tube defects, and folate supplementation is currently recommended during pregnancy to reduce such complications. Deficient one carbon metabolism cycling leading to aberrant DNA methylation, has been proposed as a mechanism for the observed association. We evaluated the association of MTHFR genotype and DNA methylation in the placenta; the organ at the maternal-fetal interface which is responsible for folate acquisition from maternal circulation during gestation.

Methods: Term samples were collected from chromosomally normal pregnancies at BC Women’s Hospital. 30 samples of three MTHFR genotype combinations: 677CC & 1298AA (n=10, low risk), 677TT & 1298AA (n=10, 677 high risk) and 677CC & 1298CC (n=10, 1298 high risk) were assessed for genome-wide DNA methylation using the Illumina Infinium HumanMethylation450 BeadChip (450k array). Subset-quantile within array normalization was applied for probe-level normalization and ComBat batch correction was applied to remove variability due to position and chip. Technically biased and XY chromosome probes were removed from analyses, resulting in a filtered dataset of 442,348 probes.

Results: Using the filtered dataset, unsupervised hierarchical clustering did not group samples by MTHFR genotype, and there was no significant difference by group in average 450k array DNA methylation. Linear models were fit on a probe-by-probe level to test for differential DNA methylation by MTHFR genotype while adjusting for fetal sex and gestational age. Using a lenient threshold of false discovery rate (FDR) <0.18, only one CpG site was significant in the comparison of the 1298 high risk group to low risk group (FDR <0.02), and none in the comparison of the 677 high risk group to low risk group. Further exploration of the data will be conducted including subdividing the analysis based on CpG island density and gene functional regions. However, our preliminary analysis suggests little effect of MTHFR genotype on DNA methylation in placentae from this Canadian population.