Associations between prenatal air pollution exposure and cord blood DNA methylation are modified by neighborhood socioeconomic deprivation
On this page:
Background: Prenatal environmental exposures may have lasting effects on child health, and socioeconomic stressors experienced during pregnancy may increase susceptibility to environmental insults. Fetal epigenetic programming via altered cord blood DNA methylation may reveal pathways by which prenatal air pollution and socioeconomic deprivation exert joint adverse effects on child health.
Objectives: To determine associations between period-specific ambient air pollution exposure during pregnancy and DNA methylation in umbilical cord blood cells, and to determine how neighborhood-level socioeconomic deprivation modifies associations between air pollutants and DNA methylation.
Methods: Among 341 eligible pregnant participants enrolled in the Newborn Epigenetics STudy (NEST) cohort, we geocoded addresses at enrollment and used publicly available national air pollution models to assign exposures to fine particulate matter (PM2.5), nitrogen dioxide (NO2), and ozone (O3) averaged over a 3-month preconception period, each trimester, and the duration of the pregnancy. We collected umbilical cord blood at delivery and evaluated epigenome-wide DNA methylation using the Illumina Infinium HumanMethylation450 BeadChip array. We derived a neighborhood deprivation index (NDI) reflecting several socioeconomic indicators at the Census-tract level. We fit separate linear regression models for the association between each period-specific air pollutant concentration and methylation beta values at each of 429,746 CpGs. We tested product interaction terms between Census-tract level NDI and each period-specific air pollution measure, and conducted models stratified by NDI above versus below the median if the interaction was significant at a false discovery rate (FDR) < 0.05.
Results: Over the full pregnancy period, median exposure to PM2.5 was 13 ug/m3, median O3 was 43 ppb, and median NO2 was 21 ppb. The population was diverse, with 50% of participants reporting their race/ethnicity as Black, 42% reporting White and 8% all other racial and ethnic groups. The mean (SD) NDI at enrollment was 0.52 (7.93). After covariate adjustment, there were significant associations (FDR<0.05) between full-pregnancy average PM2.5 and methylation at 1 CpG, between second trimester PM2.5 and methylation at 1 CpG, and between first trimester O3 and methylation at 7,489 CpGs. There were significant interactions between third trimester NO2 exposure and NDI for 20 CpGs, and between preconception NO2 exposure and NDI for 3 CpGs, two of which overlapped with the third trimester NO2 results. In models stratified by NDI for the 21 CpGs with significant interactions, the association between NO2 exposure and DNA methylation had opposite directions in the low NDI stratum and the high NDI stratum. Genes annotated to the differentially methylated CpGs were previously linked to diseases including asthma (SFRS8), coronary artery disease (AP2A2), epilepsy (DYNC1H1), and autoimmune disease (PRRC2A), and included a gene (KCNQ1) previously found to modify the association between air pollution exposure and fasting glucose.
Conclusions: We observed associations between preconception and prenatal exposure to ambient air pollution, and patterns of DNA methylation in cord blood in a predominantly urban, southern US cohort. Neighborhood-level socioeconomic factors during pregnancy may alter offspring susceptibility to the adverse developmental effects of prenatal exposure to traffic-related air pollution. Future research will assess the effects on early childhood cardiometabolic health.