Synthetic folic acid supplements are consumed by 50% of US adults. High folic acid intake can result in CpG island hypermethylation. Recent studies suggest high intake may contribute to carcinogenesis particularly in those persons unable to metabolize folic acid efficiently such as persons homozygous for the 19bp deletion polymorphism in the dihydrofolate reductase(DHFR) gene (rs70991108). Homozygotes (19% of US adults) have less efficient conversion of synthetic folic acid into biological folate. Women homozygous for the DHFR19bp deletion (DHFR19bpdel) are at increased risk for having a child with retinoblastoma, a neuro-ectodermally derived early childhood cancer that has served as a model for understanding carcinogenesis. This risk appears elevated among women homozygous for DHFR19bpdel who consumed folic acid containing prenatal supplements in the first trimester. In retinoblastoma, 37% of disease-causing mutations in the CpG-rich gene, RB1, are due to deamination of methylated cytosines in RB1, while an additional 12% are caused by methylation of the RB1 promoter. We hypothesize that these methylation related RB1-inactivating changes may be induced by less functional genetic variations in folate metabolism and excess exposure in early retinal development. Murine models suggest particular sensitivity to folic acid exposure in early neural development. In humans, periods of dietary inadequacy in early gestation are associated with hyper-methylation at meta- stable epialleles. Pilot data from mutation assays and methylation arrays from our ongoing case-series study of newly diagnosed retinoblastoma (mean age 23 months) in Central Mexico, EpiRbMx, suggest that women who have the less efficient DHFR 19bpdel and high folic acid prenatally have higher risk of a child with a methylation related RB1-inactivating mutation or differentially hyper-methylated RB1 promoter directly resulting in tumor formation. Maternal DHFR19bpdel homozygosity combined with high prenatal folic acid intake may lead to RB1 hyper-methylation in developmentally susceptible cells, thereby contributing to carcinogenesis. Based on the results of our pilot data, we propose to use existing samples and data from 337 cases in our ongoing EpiRbMx study (founded in 2003) including pre-treatment measures of mother and child plasma nutrients, diet intake, genotype, and tumors, to test our hypothesis of an association between maternal DHFR 19bpdel and prenatal folic acid exposure (using plasma and intake measures) and 1) methylation related RB1 mutations, specifically in somatic (non-germline) RB1 mutations, and 2) methylation of the RB1 promoter. We will further examine the impact of DHFR and prenatal folic acid exposure on differential methylation in metastable epialleles to document their effect on the developing epigenome. Given the high frequency of RB1 mutations among poor prognosis adult tumors such as small cell lung cancer(54%) or glioblastoma(11%) and the high proportion (19%) of adults with less efficient folic acid metabolism, the potential translational impact of our study is high, as our results can inform public health policy and evaluation of synthetic folic acid exposure.
Prior work in our cohort of children with retinoblastoma suggests that women homozygous for DHFR19bpdel are at increased risk to have a child with retinoblastoma, particularly when taking prenatal folic acid supplements. We will examine how maternal and child DHFR and folic acid intake may be associated with methylation and associated mutations in retinoblastoma tumor DNA. Retinoblastoma offers us a unique model for examining how folic acid may affect carcinogenesis, of particularly significance given that other studies suggest that elevated folic acid intake may contribute to carcinogenesis in breast, colon and lung cancer
