Placental DNA methylation as mediators of assisted reproduction and parental metabolic risk on offspring phenotype: a multiethnic cohort study

Alterations in genomic imprinting are posited as a mechanism for effects of assisted reproduction technologies (ART) and in vitro fertilization (IVF) on long term offspring phenotype. Our previous work has investigated the role of fetal cord tissue DNA methylation in mediating these relationships (Huang, et al. 2021. DOI: 10.1038/s41467-021-25899-4). To study the extent to which fetal-derived placental DNA methylation may mediate these relationships, we conducted a nested cohort study within the Growing Up in Singapore Towards healthy Outcomes (GUSTO) prospective parent-offspring study. We assayed DNA methylation in 408 bulk placental tissue samples arising from 200 placenta obtained at the time of live singleton birth. We oversampled on the basis of conceptions via IVF (ART group) and spontaneous conceptions amongst couples where the father had one or more cardiometabolic risk factors (advanced age, obesity, hypertension, or diabetes history; SC_HIGHRISK group). We then took a random sample of all other spontaneous conceptions (SC_RANDOM). For each of 200 placentas, one sample each was taken from the fetal-facing and maternal facing side of the (fetal-derived) placenta. An additional 8 technical replicated were assayed resulted in the 408 total samples. After QC and elimination of replicates, DNA methylation data were available from 196 unique fetal-facing and 195 maternal-facing placental bulk tissue samples. These repeated measures were used to reduce residual technical and unmeasured confounding biases in downstream analyses. Following delivery of placenta, bulk tissue samples were sampled from across the respective aspects of the placenta, washed, aliquoted, and immediately frozen at -80 C by standard protocols. Samples were later crushed, thawed, extracted, and assayed according to protocols outlined below. Overall, samples arose from N = 200 placenta distributed as: 68 ART group; 43 SC_HIGHRISK group; 89 SC_RANDOM group. After QC, DNA methylation data were available for 66 ART; 42 SC_HIGHRISK; and 88 SC_RANDOM subjects. Principle components analyses showed no strong associations between DNA methylation and technical covariates with the exception of placental side. Cell types were imputed from a single-cell reference using the planet R package.