Epigenome-Wide Pooled Analysis Of Methylation In Children Related To Air Pollution Exposure At Birth: The Medall Study

Abstract

Introduction: Air pollution exposure has been increasingly associated with adverse health outcomes in children, including airway disease. Underlying mechanisms may involve epigenetic modifications such as DNA methylation. A few studies have reported differential methylation at specific loci in relation to ambient air pollution exposure, whereas studies based on epigenome-wide methylation data are limited. Methods: Within the framework of Mechanisms of the Development of ALLergy (MeDALL) Consortium, we investigated exposure to traffic-related nitrogen dioxide (NO2), in relation to blood DNA methylation by pooling data from four European birth cohorts (N=766) in which DNA methylation was measured in paired samples (either at 0 and 4/5 years or at 4 and 8 years), using the Infinium HumanMethylation450 BeadChip. All participating cohorts assessed NO2 exposure at birth address by means of land-use regression modelling according to a uniform exposure protocol (within the ESCAPE project). The associations were analyzed with robust linear regression. Results: In the analysis of the association between NO2 exposure at birth and cord blood DNA methylation levels (N=286), 4 CpGs mapped to 3 genes (CYP7B1, ANO1, LIN7A) appeared below the Bonferroni adjusted significance threshold and 18 CpGs below the false discovery rate (FDR), located within 11 gene regions, none of which have previously been implicated in relation to air pollution exposure. These associations were largely robust to estimated cell type adjustment using Houseman's method. NO2 exposure at birth was also associated with differentially methylated CpGs in 4-year-olds (N=766), although without clear overlap with those found in cord blood. Discussion: Our analysis revealed a set of genes which represent novel associations of methylation changes in relation to NO2 exposure. We will further update the current results with analyses based on additional datasets including older children, and different exposure windows.