5-hydroxymethylcytosine (5hmC), formed through the ten-eleven translocation (TET) methylcytosine dioxygenase mediated oxidation of 5-methylcytosine (5mC) at cytosine-phosphate-guanine (CpG) dinucleotides, is believed to mainly serve as an intermediate in the DNA demethylation pathway, though recent evidence suggests that 5hmC may also play a functionally relevant role. We have conducted an epigenome-wide association study (EWAS) to assess the association between placenta 5hmC, obtained through parallel bisulfite and oxidative bisulfite modification of DNA and array-based assessment, and newborn birthweight in the Rhode Island Child Health Study (RICHS). We also assessed whether the removal of 5hmC signal impacts the observed results from traditional epigenome-wide studies that rely on BS modification-based (combined 5mC and 5hmC) assessment alone. We identified 5hmC at one CpG in the CUBN gene to be significantly associated with birthweight (FDR < 0.05) and demonstrate that expression of that gene was also associated with birthweight. Comparison of 5hmC+5mC and 5mC EWAS effect estimates reveal a strong correlation (r = 0.77, p < 0.0001). Our study suggests that traditional assessment of 5mC through bisulfite modification alone provides an accurate assessment of CpG-specific DNA methylation for EWAS studies but was unable to provide evidence of widespread associations between placental 5hmC and birthweight.
Keywords: birthweight; epigenome-wide association study; hydroxymethylation; placenta.