Human exposure to the ubiquitous environmental carcinogens polycyclic aromatic hydrocarbons (PAHs) occurs predominantly through the diet, tobacco smoke, and air pollution. While the genotoxic effects of these compounds have been well characterised, it was hypothesised that DNA methylation changes induced by PAHs could be a potential mechanism of their carcinogenicity. This study aimed to identify DNA methylation changes associated with PAH exposure in a mouse model and in human cohorts. Reduced representation bisulphite sequencing was carried out on lung tissue from Benzo[a]pyrene exposed mice. Additionally, PAH8 exposure from air and dietary sources estimated from land use regression models and food frequency questionnaires were used with data from Illumina Infinium HumanMethylation450 in EPIC-Italy (Training subset: N = 493; Testing subset: N = 208) and EPIC-NL (N = 132) cohorts.
Several differentially methylated CpG sites (Treated vs Untreated: N = 430; p < 0.05, Δ methylation > 25%), 500 b.p. windows (Treated vs Untreated: N = 1780; p < 0.05, Δ methylation > 25%), and probes (Air PAH8 exposure: N = 204; Dietary PAH8 exposure: N = 171; Combined air and dietary PAH8 exposure: N = 274; FDR q < 0.05) were identified in the analyses carried out. Although there were little to no overlaps between mouse and human studies at the CpG or gene level, in both the mouse and human analyses significantly fewer changes than expected by chance occurred at promoter regions. Additionally, the three human EWAS showed that different routes of PAH exposure may have different effects on DNA methylation, and when these exposures were combined, the methylation changes observed represented the separate exposures. These observations require further validation, but the results suggest that PAH-DNA adduct formation, which does not occur in a gene-specific manner, could be driving DNA methylation changes.