Putative Prostate Cancer Risk SNP in an Androgen Receptor‐Binding Site of the Melanophilin Gene Illustrates Enrichment of Risk SNPs in Androgen Receptor Target Sites

ABSTRACT Genome‐wide association studies have identified genomic loci, whose single‐nucleotide polymorphisms (SNPs) predispose to prostate cancer (PCa). However, the mechanisms of most of these variants are largely unknown. We integrated chromatin‐immunoprecipitation‐coupled sequencing and microarray expression profiling in TMPRSS2‐ERG gene rearrangement positive DUCaP cells with the GWAS PCa risk SNPs catalog to identify disease susceptibility SNPs localized within functional androgen receptor‐binding sites (ARBSs). Among the 48 GWAS index risk SNPs and 3,917 linked SNPs, 80 were found located in ARBSs. Of these, rs11891426:T>G in an intron of the melanophilin gene (MLPH) was within a novel putative auxiliary AR‐binding motif, which is enriched in the neighborhood of canonical androgen‐responsive elements. T→G exchange attenuated the transcriptional activity of the ARBS in an AR reporter gene assay. The expression of MLPH in primary prostate tumors was significantly lower in those with the G compared with the T allele and correlated significantly with AR protein. Higher melanophilin level in prostate tissue of patients with a favorable PCa risk profile points out a tumor‐suppressive effect. These results unravel a hidden link between AR and a functional putative PCa risk SNP, whose allele alteration affects androgen regulation of its host gene MLPH.


Significant P-values are highlighted in bold; borderline significant P-values are underlined
Supp. Figure S1. Dynamics of AR binding to the PSA enhancer. DUCaP cells were stimulated with 1 nM R1881 for the indicated time points, harvested and subuected to ChIP isolation of AR bound DNA. Rabbit IgG was used as antibody control. A: PCR with AR-precipitated samples were used to visualize the dynamics of AR recruitment to its target gene PSA. M: DNA ladder. B: QPCR was used to quantify the androgen stimulated AR enrichment to the binding site relative to control vehicle-treated samples (Contr). Values in (B) represent the mean values and standard deviations of at least three independent experiments.
Supp. Figure S2. Characterization of AR binding sites. A: Illustration of MACS calculated P values of the AR binding peaks. B: Illustration of the distance of ARBSs to the transcription start sites (TSSs) of their adjacent genes. The upper and lower panels represent the distance of ARBSs to the TSS of the primary AR target genes as defined in legend to Figure 1 (upper panel) and to the TSS of all other genes (lower panel), respectively. C: Illustration of the distance of the identified 3 putative ARE motifs to the TSS of the adjacent genes.
Supp. Figure S3. Simulation analysis of ChIP-seq peaks. Number of reported GWAS PCa risk SNPs and their proxies with simulated AR binding sites. A: All proxy SNPs with linkage disequilibrium (ld) of r 2 ≥0.2 with the reported SNPs were included and the simulation of ARBSs was repeated 10,000 times to determine random co-localization. B: All proxies with ld of r 2 ≥0.5 were included. C: All proxies with ld r 2 ≥0.7 were included. The vertical red line indicates the actual number of SNPs in the identified ARBSs and demonstrates the enrichment of risk SNPs in AR binding sites.
Supp. Figure S4. Id structure of SNPs in the melanophilin gene. The PCa risk index SNP rs7584330 and the potential PCa risk SNPs rs11891429 and rs2294884 are indicated in relation to the melanophilin gene and the AR ChiP-seq peaks. From top to bottom: refseq gene, ARBS peaks, IgG peaks (control), potential PCa risk SNPs and Id structure.
Supp. Figure S5. AR binding site in the 7 th intron of the MLPH gene. A: The DNA sequence of the MLPH gene ARBS is shown. The putative ARE motifs 2 and 3, the rs11891426 SNP within the motif 2 ARE element, as well as the primers used for cloning and sequencing of the ARBS are indicated. B: Illustration of MLPH ARBS reporter gene constructs used. The wild-type construct containing the SNP rs11891426 major T allele was designated as MLPH-wt, the constructs with the SNP site Rs11891426 mutated to the minor G allele as MLPH-G. The mutated motif 3 or motif 2 ARE reporter vectors were designated as MLPH-M3-mut or MLPH-M2-mut, respectively.
Supp. Figure S6. Correlation of MLPH mRNA expression to the SNP genotypes and to AR expression in PCa tissue specimens. Expression of MLPH and AR transcripts was analysed by real-time RT-PCR in RNA samples extracted from benign and malignant areas of frozen radical prostatectomy specimens (n=77). The relative transcript expression levels were calculated in relation to the house keeping gene TBP and further calibrated to an inter-assay and inter-plate control sample (cDNA mix of 3 benign and 3 malignant tissue sections). A-D: Scatter plot of MLPH mRNA expression stratified according to the genotypes of rs11891426, rs2292884 and rs7584330 (A), tissue types and genotype category of rs11891426 (B), rs2292884 (C) and rs7584330 (D). E: Benign (Ben) and cancer (Ca) tissue types. F: Scatter plot of MLPH versus AR mRNA demonstrating the correlated expression of AR and MLPH mRNAs in prostate tissue samples. Welch's corrected unpaired t-test was used for the analysis of differences between the groups. * P < 0.05, ** P < 0.01 and *** P < 0.001.