Abstract Study question Does human embryo culture in different IVF culture media lead to DNA methylation alterations in IVF offspring? Summary answer Genome-wide analyses identified no significant DNA methylation differences between culture medium groups in IVF children (neonates or 9-year olds) from two culture media studies. What is known already During in vitro fertilisation (IVF) treatments, embryos undergo preimplantation development in an artificial environment, while concurrently undergoing epigenetic reprogramming. Adversity during this period, such as peri-conception calorie restriction, has been linked to persistent DNA methylation aberrations and increased risk of cardiometabolic disease. Early environmental adversity is suspected in IVF offspring as they are born with lower birthweights and show increased risk of cardiometabolic dysfunction in adulthood as compared to their naturally-conceived counterparts. This is further supported by the observation from two culture media trials (MEDIUM0 and MEDIUM1) that embryo culture in different culture media leads to differences in birthweight. Study design, size, duration We recruited singleton offspring from two IVF culture media trials. The MEDIUM0 study, a pseudo-randomized trial comparing G3 (Vitrolife) and K-SICM (Cook), was conducted from 2003-2006. At the 9-year follow-up, saliva was collected (cohort-A). The MEDIUM1 study, a multi-center randomized controlled trial comparing G5 (Vitrolife) and HTF (Lonza), was conducted from 2010-2012. Umbilical cord blood (UCB) was collected at birth (cohort-B). Participants/materials, setting, methods DNA methylation was analysed in 120 saliva samples (65 G3, 55 Cook) and 106 UCB samples (47 HTF, 59 G5) using the Infinium MethylationEPIC array (Illumina). Mixed effects linear models, correcting for (gestational) age, sex, sample composition and batch effects alongside maternal age, pregnancy complications and IVF centre for cohort-B, were implemented at single or aggregated sites. Methylation outliers were defined as values over three interquartile ranges below or above 25th and 75th percentiles respectively. Main results and the role of chance 111 of the 120 saliva samples (60 G3, 51 Cook) and 105 of the 106 UCB samples (47 HTF, 58 G5) passed our quality control criteria. We filtered sites on sex chromosomes, and based on quality, proximity to single-nucleotide polymorphisms, and proportion of missing values, leaving 650,000-700,000 of the 850,000 sites included on the EPIC array for our analyses. To account for heterogeneity in the cellular composition of our samples we estimated their cell compositions using a reference-based approach. First, we investigated individual CpG sites, finding no differentially methylated sites in either cohort after correction for multiple testing (false discovery rate adjusted p. value threshold < 0.1). Sites were then aggregated into regions based on their allocations to genes, promoters and CpG islands. No differentially methylated regions were identified in either cohort. A targeted analysis of DNA methylation of imprinting genes showed no differentially methylated sites or regions. To examine the contribution of stochastic epigenetic alterations we quantified the number of methylation outliers per sample. Although this revealed a predominance of hypomethylation outliers, there was no difference in the total number or distribution of DNA methylation outliers between the two culture media groups of cohort-A and cohort-B. Limitations, reasons for caution This analysis is currently limited by the lack of comparison to a naturally-conceived control group. As such, we cannot yet conclude whether IVF embryo culture, in any medium, is associated with DNA methylation aberrations. Additionally, given the large number of comparisons, we may lack power to detect small differences. Wider implications of the findings Although there are disparities in birth weight and childhood growth after embryo culture in different media, we observed no DNA methylation alterations preserved postnatally. Whether DNA methylation of these individuals deviates from that of naturally-conceived individuals will be determined in the near future. Trial registration number MEDIUM1: NTR 1979 /NL1866 (Netherlands Trial Registry)
Read full abstract