Bisphenol A-associated epigenomic changes in prepubescent girls: a cross-sectional study in Gharbiah, Egypt

Abstract

Background: There is now compelling evidence that epigenetic modifications link adult disease susceptibility to environmental exposures during specific life stages, including pre-pubertal development. Animal studies indicate that bisphenol A (BPA), the monomer used in epoxy resins and polycarbonate plastics, may impact health through epigenetic mechanisms, and epidemiological data associate BPA levels with metabolic disorders, behavior changes, and reproductive effects. Thus, we conducted an environmental epidemiology study of BPA exposure and CpG methylation in pre-adolescent girls from Gharbiah, Egypt hypothesizing that methylation profiles exhibit exposure-dependent trends.

Methods: Urinary concentrations of total (free plus conjugated) species of BPA in spot samples were quantified for 60 girls aged 10 to 13. Genome-wide CpG methylation was concurrently measured in bisulfite-converted saliva DNA using the Infinium HumanMethylation27 BeadChip (N = 46). CpG sites from four candidate genes were validated via quantitative bisulfite pyrosequencing.

Results: CpG methylation varied widely among girls, and higher urinary BPA concentrations were generally associated with less genomic methylation. Based on pathway analyses, genes exhibiting reduced methylation with increasing urinary BPA were involved in immune function, transport activity, metabolism, and caspase activity. In particular, hypomethylation of CpG targets on chromosome X was associated with higher urinary BPA. Using the Comparative Toxicogenomics Database, we identified a number of candidate genes in our sample that previously have been associated with BPA-related expression change.

Conclusions: These data indicate that BPA may affect human health through specific epigenomic modification of genes in relevant pathways. Thus, epigenetic epidemiology holds promise for the identification of biomarkers from previous exposures and the development of epigenetic-based diagnostic strategies.

Figure 1

β-score distribution in the HumanMethylation27 BeadChip in the presence and absence of CpG islands in 1 randomly chosen sample. All samples display similar distributions as shown above. (A) The first histogram represents the β-score frequency of all probes. (B) The β-score frequency of the probes located within CpG islands is graphed. (C) The β-score frequency of the probes located outside of CpG islands is graphed.

Figure 2

Histogram of p-values (bin = 0.1) in methylation analysis using covariates including age, BMI, and log BPA in empirical Bayes modeling using LIMMA package in R software with BPA as a (A) dichotomous or (B) continuous variable.

Figure 3

The 200 most variable CpG sites from 46 samples were clustered using Recursively Partitioned Mixture Model (RPMM) for Beta and Gaussian Mixtures from R package, a model-based clustering algorithm that returns a hierarchy of classes, similar to hierarchical clustering and finite mixture models. (A) RPMM clustering divided 46 samples into 3 clusters (1, 2, and 3), containing 12, 15, and 19 samples, respectively. (B) More than half of the BPA-high samples were clustered within cluster 1, while BPA-low and –intermediate samples were more evenly distributed across all 3 clusters.

Figure 4

In the dichotomous model analysis, multiple regions in chromosome X were enriched with hypomethylated genes in BPA-high vs. BPA-low groups.

Figure 5

Differential methylation observed in genes known to be interacting with BPA from the Comparative Toxicogenomics Database (CTD). (A) Top candidate genes that undergo hypermethylation in BPA-high vs. BPA-low groups. (B) Top candidate genes that undergo hypomethylation in BPA-high vs. BPA-low groups. The reported expression change in genes with the asterisk did not inversely correlate with the methylation status.

Figure 6

Quantitative levels of methylation for HOXA10, BEX2, CXorf23, and DTL (N = 12 for Non-detect and 13 for High-BPA for HOXA10, CXorf23, and DTL, and N = 9 for Non-detect and N = 9 for High-BPA for BEX2).