Detection of differential DNA methylation in repetitive DNA of mice and humans perinatally exposed to bisphenol A

Abstract

Developmental exposure to bisphenol A (BPA) has been shown to induce changes in DNA methylation in both mouse and human genic regions; however, the response in repetitive elements and transposons has not been explored. Here we present novel methodology to combine genomic DNA enrichment with RepeatMasker analysis on next-generation sequencing data to determine the effect of perinatal BPA exposure on repetitive DNA at the class, family, subfamily, and individual insertion level in both mouse and human samples. Mice were treated during gestation and lactation to BPA in chow at 0, 50, or 50,000 ng/g levels and total BPA was measured in stratified human fetal liver tissue samples as low (non-detect to 0.83 ng/g), medium (3.5 to 5.79 ng/g), or high (35.44 to 96.76 ng/g). Transposon methylation changes were evident in human classes, families, and subfamilies, with the medium group exhibiting hypomethylation compared to both high and low BPA groups. Mouse repeat classes, families, and subfamilies did not respond to BPA with significantly detectable differential DNA methylation. In human samples, 1251 individual transposon loci were detected as differentially methylated by BPA exposure, but only 19 were detected in mice. Of note, this approach recapitulated the discovery of a previously known mouse environmentally labile metastable epiallele, Cabp(IAP). Thus, by querying repetitive DNA in both mouse and humans, we report the first known transposons in humans that respond to perinatal BPA exposure.

Keywords: Bisphenol A; DNA methylation; developmental origins of health and disease; environmental epigenomics; epigenetics; interindividual variation; next-generation sequencing; repetitive DNA; transposon.

Figure 1.

Flowchart for Experimental and Analysis Pipeline. Both humans and mice underwent MethylPlex enrichment prior to sequencing. Reads were Repeat Masked and used to identify group level differences, and aligned to reference genomes to identify individual differences in repeat DNA methylation. Asterisks denote significance, *P < 0.05, **P < 0.01.

Figure 2.

Percent CpG Enrichment by Exposure. Both human and mouse samples exhibited increased CpG content after MethylPlex enrichment. (A) The human low BPA group was slightly but significantly lower than the medium group (P = 0.01), and the high group (P < 0.05). (B) The mouse samples exhibited no significant differences in CpG enrichment by exposure group.

Figure 3.

Percent Reads by Repeat Class and Exposure. (A) A subset of classes in human was hypomethylated in the medium BPA group. (B) No classes were differentially methylated by exposure group in mice. Asterisks denote significance, *P < 0.05, **P < 0.01.

Figure 4.

Percent Reads by Repeat Family and Exposure. (A) A subset of repeat families was hypomethylated in the medium BPA group. (B) No families were differentially methylated by exposure group in mice. Asterisks denote significance, *P < 0.05, **P < 0.01.

Figure 5.

Individual Transposon Detection Pipeline. Both mouse and human reads were filtered for reads containing partial unique and partial repetitive sequence. The unique sequence was mapped to the reference genome and read count differences corresponding to DNA methylation differences were tested for significance with EdgeR.

Figure 6.

Individual Repeats Differentially Methylated by Exposure. (A) Number of human transposon insertions exhibiting significantly different read counts across BPA exposure groups. (B) Mouse transposon insertions exhibiting differential read counts across BPA treatment groups.