Epigenetic Intergenerational Transmission: Mothers’ Adverse Childhood Experiences and DNA Methylation

STUDY SYNOPSIS

Introduction Summary:

Individual differences in risk for mental disorders over the lifespan are shaped by forces acting before the individual is born— in utero, but likely even earlier, during the mother’s own childhood. The environmental epigenetics hypothesis proposes that sustained effects of environmental conditions on gene expression are mediated by epigenetic mechanisms. Recent human studies have shown that adversities in the early environment are correlated with DNA methylation in childhood.^1–5 However, these studies are limited by either 1) Candidate-gene approaches, which only look at specific genes and CpG sites, precluding a broader understanding of methylation across the genome, or 2) Methylome-wide approaches with small sample sizes, (e.g. <200), increasing both type I and type II errors. One recent exception is an methylome-wide analysis in 691 children followed longitudinally since birth. In this study, Dunn et al found that 38 CpG sites were differentially methylated in children at 7 years of age following exposure to adversities.⁶ However, clear evidence of childhood adversity-induced DNA methylation conserved across decades, into adulthood, and whether it is passed on to offspring, is lacking.

In the current study, we will test whether adverse experiences in mothers’ childhoods are correlated with DNA methylation in peripheral blood during pregnancy and in cord blood samples from their newborn infants. We expect to find differentially methylated CpG sites in pregnant women and in their newborn infants associated with mothers’ adverse childhood experiences (ACEs), and we expect the differentially methylated regions to at least partially overlap between mothers and infants, indicating enduring and transmitted impacts of mothers’ ACEs on DNA methylation. As a secondary analysis, we will test women’s depression and anxiety symptoms during pregnancy as a possible mediator of the association between mothers’ ACE exposure and prenatal/neonatal DNA methylation, given that 1) ACEs have consistently been shown to be associated with perinatal mood and anxiety disorders,⁷ and 2) depression and anxiety have been correlated with DNA methylation in adults, ⁸ and prenatal anxiety and depression have been associated with DNA methylation in infant offspring. ^9,10 We expect the associations between mothers’ childhood adversities and DNA methylation in pregnancy and newborns to be partially but not fully mediated by mothers’ prenatal depression and anxiety symptoms, demonstrating a pathway from ACEs to DNA methylation above and beyond ACEs’ impact on prenatal mood symptoms.

Method Summary

Data are from the Avon Longitudinal Study of Parents and Children (ALSPAC) Accessible Resource for Integrated Epigenomic Studies (ARIES) sub-study.¹¹ The ALSPAC design included all women in the Avon Health District in the United Kingdom (UK) who gave birth between April 1, 1991 and December 31, 1992. The ARIES subsample provided blood samples during pregnancy and umbilical cord blood samples upon delivery. Women provided retrospective self-reports during pregnancy of whether they had experienced during childhood the ten ACEs studied in the flagship CDC-Kaiser ACE study. A cumulative score (0–10) will be used to measure ACEs. Genome-wide methylation status of over 485,000 CpG sites was measured using the Illumina Infinium®HumanMethylation450K BeadChip assay. To test whether mothers’ ACE exposure is associated with methylation in DNA from maternal peripheral blood during pregnancy and cord blood at birth, we will use generalized linear regression models. Beta values for the ~450,000 probes on the Illumina bead chip that pass quality control checks will be the outcome variables. The 0–10 ACE summary score will be the main predictor variable. The following variables will be covariates: maternal age at pregnancy, parity, maternal smoking during pregnancy, maternal education, pre-pregnancy BMI, gestational age at birth, cell type composition, and technical covariates identified in the epigenetics data quality check. The analysis with newborn DNA will also be conducted separately in male and female newborns, based on the hypothesis that male placentas, which typically have lower levels of X-linked O-linked N-acetylglucosamine transferase gene, are more likely to have epigenetic alterations with robust transcriptional responses to maternal prenatal stress.¹² In all analyses, the false discovery rate will be set at 0.05. To test for mediation, linear structural equation models will be used. Beta values at CPG sites found to be significant in the primary analyses will be the outcome variables. The independent variable and covariates will be the same as in the primary analyses. The Edinburgh Postnatal Depression Scale (EPDS) total score will be tested as a mediator.

Significance

This analysis tests an extended timeframe for the impact of childhood adversity: mothers’ childhood affecting DNA methylation in her newborn offspring, and could provide evidence for a novel biological pathway for the enduring impact of childhood adversity across generations. Understanding the biological mechanisms by which Adverse Childhood Experiences program future generations for mental health risk would expand the possibilities for designing and evaluating interventions to prevent and ameliorate the negative impacts of adversities on future generations’ mental health.