Examining the epigenetic mechanisms of childhood adversity and sensitive periods: A gene set-based approach

Abstract

Background: Sensitive periods are developmental stages of heightened plasticity when life experiences, including exposure to childhood adversity, have the potential to exert more lasting impacts. Epigenetic mechanisms, including DNA methylation (DNAm), may provide a pathway through which adversity induces long-term biological changes. DNAm shifts may be more likely to occur during sensitive periods, especially within genes that regulate the timing of sensitive periods. Here, we investigated the possibility that childhood adversity during specific life stages is associated with DNAm changes in genes known to regulate the timing and duration of sensitive periods.

Methods: Genome-wide DNAm profiles came from the Avon Longitudinal Study of Parents and Children (n = 785). We first used principal component analysis (PCA) to summarize DNAm variation across 530 CpG sites mapped to the promoters of 58 genes previously-identified as regulating sensitive periods. Gene-level DNAm summaries were calculated for genes regulating sensitive period opening (n_genes = 15), closing (n_genes = 36), and expression (n_genes = 8). We then performed linear discriminant analysis (LDA) to test associations between seven types of parent-reported, time-varying measures of exposure to childhood adversity and DNAm principal components. To our knowledge, this is the first time LDA has been applied to analyze functionally grouped DNAm data to characterize associations between an environmental exposure and epigenetic differences.

Results: Suggestive evidence emerged for associations between sexual or physical abuse as well as financial hardship during middle childhood, and DNAm of genetic pathways regulating sensitive period opening and expression. However, no statistically significant associations were identified after multiple testing correction.

Conclusions: Our gene set-based method combining PCA and LDA complements epigenome-wide approaches. Although our results were largely null, these findings provide a proof-of-concept for studying time-varying exposures and gene- or pathway-level epigenetic modifications.

Keywords: ALSPAC; Childhood adversity; Epigenetics; Plasticity; Sensitive periods.

Figure 1.. Flowchart of analyses.

After adjusting for technical variables (Step 1), variation among CpG sites annotated to the promoter of a given gene was summarized at the gene level into two principal component (PC) scores (Step 2). PCs were then adjusted for confounders (Step 3). Finally, at the gene set level, associations between time-varying exposures to childhood adversity and PCs of all genes within a gene set were assessed using a linear discriminant analysis (Step 4). The sensitive period most strongly associated with the PCs was selected based on canonical R², as highlighted in orange in this example.

Figure 2.

Variances of CpGs located in promoters of three sets of sensitive period genes, as explained by the first two principal components (PCs). The number of CpGs annotated to the promoter of each gene is noted in parentheses. The red dashed line indicates 50% of the variance. For 53 of 58 genes, variance explained by the first two PCs exceeded 50