Associations between per- and polyfluoroalkyl substances (PFAS), DNA methylation and gene expression from background exposed Norwegian women (2003–2006)

Abstract Per- and polyfluoroalkyl substances (PFAS) are a large group of widespread synthetic substances associated with several adverse health effects. A potential mechanism underlying these effects is epigenetic modifications, such as DNA methylation. However, evidence on associations between PFAS exposure and DNA methylation among adults is limited, and few studies have investigated the potential influence of PFAS-related methylation changes on gene expression. In this study, we investigated associations between PFAS exposure and DNA methylation in blood. Furthermore, the associations between these methylation differences and gene expression patterns were explored to assess potential functional effects. In addition, associations between PFAS concentrations and estimated white blood cell composition were investigated. This study included women (n = 269) from The Norwegian Women and Health Study (NOWAC), with available blood data on PFAS exposure, DNA methylation, and gene expression. We investigated epigenome-wide associations between PFAS exposure and DNA methylation at individual CpG sites using multivariable linear regression analyses. In a sensitivity analysis, the main models were additionally adjusted for smoking status. Furthermore, we assessed any associations between PFAS exposure and gene expression of genes mapped to CpG sites nominally associated with PFAS, also using multivariable linear regression. Finally, we examined associations between PFAS exposure and estimated white blood cell composition using multivariable linear regression models. Concentrations of PFUnDA were associated with DNA methylation at 21 unique CpG sites, comprising 17 positive and 4 negative associations. Sensitivity analyses did not substantially alter these findings. Furthermore, methylation levels at these sites did not follow a dose-dependent pattern, and the expression of genes mapped to these CpG sites was not significantly associated with exposure levels. No significant associations were observed between other PFAS (PFOA, PFNA, PFHxS, PFHpS, PFOS, and br-PFOS) and DNA methylation or gene expression. Associations between PFHxS, PFHpS, and PFOS with DNA methylation and gene expression were observed at the nominal significance level ( p value < 0.001), but these associations were modest in magnitude, lacked evidence of dose-dependence, and did not remain significant after correction for multiple testing. No statistically significant associations were observed between PFAS exposure and estimated white blood cell composition. Our findings suggest that PFUnDA may influence DNA methylation at specific CpG sites in blood of Norwegian women, though these epigenetic patterns were not accompanied by corresponding patterns in gene expression. Overall, there was limited evidence linking other PFAS to DNA methylation or gene expression.