Exposure of Human Prostaspheres to Bisphenol A Epigenetically Regulates SNORD Family Noncoding RNAs via Histone Modification

Abstract

Bisphenol A (BPA) is a ubiquitous endocrine disruptor exerting lifelong effects on gene expression in rodent prostate cancer (PCa) models. Here, we aimed to determine whether epigenetic events mediating the action of BPA on human prostaspheres enriched in epithelial stem-like/progenitor cells is linked to PCa. We performed genome-wide transcriptome and methylome analyses to identify changes in prostaspheres treated with BPA (10 nM, 200 nM, and 1000 nM) or estradiol-17β (E2) (0.1 nM) for 7 days and validated changes in expression, methylation, and histone marks in parallel-treated prostaspheres. BPA/E2-treatment altered expression of 91 genes but not the methylation status of 485,000 CpG sites in BPA/E2-treated prostaspheres. A panel of 26 genes was found repressed in all treatment groups. Fifteen of them were small nucleolar RNAs with C/D motif (SNORDs), which are noncoding, small nucleolar RNAs known to regulate ribosomal RNA assembly and function. Ten of the most down-regulated SNORDs were further studied. All 10 were confirmed repressed by BPA, but only 3 ratified as E2-repressed. SNORD suppression showed no correlation with methylation status changes in CpG sites in gene regulatory regions. Instead, BPA-induced gene silencing was found to associate with altered recruitments of H3K9me3, H3K4me3, and H3K27me3 to 5'-regulatory/exonic sequences of 5 SNORDs. Expression of 4 out of these 5 SNORDs (SNORD59A, SNORD82, SNORD116, and SNORD117) was shown to be reduced in PCa compared with adjacent normal tissue. This study reveals a novel and unique action of BPA in disrupting expression of PCa-associated SNORDs and a putative mechanism for reprogramming the prostasphere epigenome via histone modification.

Figure 1.

Supervised clustering of transcript expression in PPrEC prostaspheres treated with BPA/E2 for 7 days. Each column represents an individual analysis comparing expression levels in PPrEC prostaspheres treated with 0.1nM E2 (E2), 10nM BPA (B10), 200nM BPA (B200), 1000nM BPA (B1000), and vehicle only as a control. Heatmap shows that 15 of 26 down-regulated genes (vertical blue line) are SNORDs. The repression was most consistent among prostaspheres from 3 donors treated with 200nM BPA. Green indicates down-regulation and red indicates up-regulation.

Figure 2.

SNORD expression in BPA-treated prostaspheres. PPrEC prostaspheres from 3 donors were treated as in Figure 1 to determine SNORD expression by qPCR. Trefoil factor 1 (TFF1) represents an estrogen-regulated gene. Data (mean ± SEM) are expressed as relative fold change normalized against endogenous GAPDH and compared with control of each experimental set. *, P < .05 vs control by one-way ANOVA and Tukey test (n = 3).

Figure 3.

Dose-response effect of BPA/E2 on NPrEC prostasphere growth. NPrEC prostaspheres treated as in Figure 1 were cultured in ultralow attachment dishes for 7 days. The size of prostaspheres increased after BPA/E2 exposure. ***, P < .001 vs control by one-way ANOVA and Tukey test (n = 4).

Figure 4.

Expression of SNORDs and their histone marks in NPrEC prostaspheres. NPrEC prostaspheres were treated as in Figure 1 and analyzed for SNORD expression by qPCR. ChIP assay was performed on active marks H3K4me3 and H3K27me3 and repress mark H3K9me3. Input denotes the amount of DNA sample loaded for the assay. Data (mean ± SEM) are expressed as relative fold change as normalized against endogenous GAPDH and compared with control of each experimental set. *, P < .05 by one-way ANOVA and Tukey test (n = 4).

Figure 5.

Expression of nonhistone mark-associated SNORDs in NPrEC prostaspheres. NPrEC prostaspheres treated as in Figure 1 were analyzed for SNORD expression by qPCR. Expression of these SNORDs was not associated with histone marks (data not shown). TFF1 represents an estrogen-regulated gene. Data (mean ± SEM) are expressed as relative fold change normalized against endogenous GAPDH and compared with control of each experimental. *, P < .05 by one-way ANOVA and Tukey test (n = 4).

Figure 6.

SNORD suppression was associated with prostate tumors. Expression of SNORD58A, SNORD59A, SNORD82, SNORD116-21, and SNORD117 was confirmed by qPCR analysis and compared between the prostate tumors and their matched nonmalignant prostate tissues. Data (mean ± SEM) are expressed as relative fold change normalized against endogenous GAPDH. *, P < .05; **, P < .01; ***, P < .001 by paired t test (n = 23).