Polymethylation Scores for Prenatal Maternal Smoke Exposure Persist Until Age 15 and Are Detected in Saliva
Background: Prenatal maternal smoking has negative implications for child health. DNA methylation signatures can function as biomarkers of prenatal maternal smoking. However little work has assessed how DNA methylation signatures of prenatal maternal smoking vary across ages, ancestry groups, or tissues. In the Fragile Families and Child Wellbeing study, we tested whether prenatal maternal smoking was associated with salivary polymethylation scores for smoking in participants. We assessed the consistency of associations at ages 9 and 15, their portability across participants from African, European, and Hispanic genetic ancestries and the accuracy of exposure classification using area under the curve (AUC) from receiver operating curve analyses. Results: We created saliva polymethylation scores using coefficients from a meta-analysis of prenatal maternal smoke exposure and DNA methylation in newborn cord blood. In the full sample at age 9 (n=753), prenatal maternal smoke exposure was associated with a 0.52 (95%CI: 0.36, 0.67) standard deviation higher polymethylation score for prenatal smoke exposure The direction and magnitude of the association was consistent when stratified by genetic ancestries. In the full sample at age 15 (n=746), prenatal maternal smoke exposure was associated with a 0.46 (95%CI: 0.3, 0.62) standard deviation higher polymethylation score for prenatal smoke exposure, and the effect size was attenuated among the European and Hispanic genetic ancestry samples. The polymethylation score was reasonably accurate at classifying prenatal maternal smoke exposure (AUC age 9=0.77, P value<0.001, age 15=0.77, P value<0.001). The polymethylation score showed higher classification accuracy than using a single a priori site in the AHRR gene (cg05575921 AUC=0.74, P value=0.03; age 15=0.73, P value=0.01). Conclusions: Prenatal maternal smoking was associated with DNA methylation signatures in saliva samples, a clinically practical tissue. Polymethylation scores for prenatal maternal smoking were portable across genetic ancestries and more accurate than individual DNA methylation sites. DNA polymethylation scores from saliva samples could serve as robust and practical clinical biomarkers of prenatal maternal smoke exposure.