Human exposure to flame-retardants is associated with aberrant DNA methylation at imprinted genes in sperm

Emerging evidence suggests that early exposure to endocrine disrupting chemicals has long-term consequences that can influence disease risk in offspring. During gametogenesis, imprinted genes are reasonable epigenetic targets with the ability to retain and transfer environmental messages. We hypothesized that exposures to organophosphate (OP) flame-retardants can alter DNA methylation in human sperm cells affecting offspring's health. Sperm and urine samples were collected from 67 men in North Carolina, USA. Urinary metabolites of a chlorinated OP, tris(1,3-dichloro-2-propyl) phosphate, and two non-chlorinated OPs, triphenyl phosphate and mono-isopropylphenyl diphenyl phosphate, were measured using liquid-chromatography tandem mass-spectrometry. Sperm DNA methylation at multiple CpG sites of the regulatory differentially methylated regions (DMRs) of imprinted genes , , , , , , , , , , and was quantified using bisulfite pyrosequencing. Regression models were used to determine potential associations between OP concentrations and DNA methylation. We found that men with higher concentrations of urinary OP metabolites, known to originate from flame-retardants, have a slightly higher fraction of sperm cells that are aberrantly methylated. After adjusting for age, obesity-status and multiple testing, exposure to mono-isopropylphenyl diphenyl phosphate was significantly related to hypermethylation at the DMRs. Exposure to triphenyl phosphate was associated with hypermethylation at the DMR; and tris(1,3-dichloro-2-propyl) phosphate exposure was associated with altered methylation at the and DMRs. Although measured methylation differences were small, implications for public health can be substantial. Interestingly, our data indicated that a multiplicity of OPs in the human body is associated with increased DNA methylation aberrancies in sperm, compared to exposure to few OPs. Further research is required in larger study populations to determine if our findings can be generalized.