Contrasting the effects of intra-uterine smoking and one-carbon micronutrient exposures on offspring DNA methylation

Abstract

Maternal smoking and micronutrient intake during pregnancy are two strong biological candidates for impacting the developing epigenome. The extent to which DNA methylation in offspring is modified by these intrauterine exposures has not been presented in parallel. In this review, we summarize human studies which have investigated genome-wide DNA methylation in the offspring in relation to maternal smoking and one-carbon micronutrient exposure during pregnancy. We contrast the primarily independent efforts for these two categories of exposure, and potential explanations for these differences. We emphasize methodological considerations such as power to detect methylation signals, exposure assessment, control of sources of variability, causal inference and the role of observed methylation changes in mediating downstream outcomes in the offspring.

Keywords: DNA methylation; epigenetic epidemiology; epigenome-wide association study (EWAS); folate; maternal smoking; micronutrients; one-carbon metabolism.

Figure 1.. Network visualization of individual studies investigating maternal smoking in relation to offspring DNA methylation.

Squares represent individual studies; numbers represent related PubMed IDs; circles represent gene regions; text represents annotated gene names; arrows represent the links between the studies and the gene regions, in other words, highlighting studies in which CpG sites have been identified in these gene regions at either Bonferroni significance or with replication/validation attempts; the size of the circle is proportional to the number of independent studies in which the gene regions have been identified. N.B. this network plot does not include the recent epigenome-wide association studies meta-analysis for maternal smoking (PMID 27040690).

Figure 2.. Network visualization of individual studies investigating maternal folate in relation to offspring DNA methylation.

Squares represent individual studies; numbers represent related PubMed IDs; circles represent gene regions; text represents annotated gene names; arrows represent the links between the studies and the gene regions, in other words, highlighting studies in which CpG sites have been identified in these gene regions at either Bonferroni significance or with replication/validation attempts; the size of the circle is proportional to the number of independent studies in which the gene regions have been identified.

Figure 3.. Comparison of design of a Mendelian randomization study and randomized controlled trial in the context of establishing a causal effect of maternal folate on DNA methylation.

Figure 4.. Mendelian randomization analysis to establish the causal effect of maternal smoking-associated DNA methylation change in relation to offspring birthweight, using genetic variants robustly associated with DNA methylation (meQTLs).

meQTL: Methylation quantitative trait loci.