Dissecting transcription of the 8q24-MYC locus in prostate cancer recognizes the equilibration between androgen receptor direct and indirect dual-functions

Abstract

Abstract Background：Androgen receptor (AR) activation and repression dual-functionality only becomes known recently and still remains intriguing in prostate cancer (PCa). MYC is a prominent oncogene functionally entangled with AR signaling in PCa. However, AR regulatory mechanisms on MYC gene transcription remains unclear. Methods：Bioinformatics analysis of androgen-mediated RNA-Seq and MYC ChIP-Seq datasets are used for AR and MYC transcriptional networks. AR ChIP-qPCR analysis are programed to find AR binding sites (ABSs) which regulate MYC transcription. 3C-qPCR and 3C-ddPCR analyses affirmed androgen-dependent MYC-Pro-P10 interaction. CRISPR/Cas9-mediated double genomic knock-out (KO) strategy is used to show that P10-KO slightly lessened androgen-elicited MYC transrepression. Results：Here we recognize AR bi-directional transcription mechanisms by exploiting the prominent 8q24-MYC locus conferring androgen hyper-sensitivity. At ~25Kb downstream of the MYC gene, we identified an undefined ABS, P10. By chromatin analyses, we validated androgen-dependent spatial interaction between P10 and MYC-Promoter (MYC-Pro) and temporal epigenetic repression of these MYC-proximal elements. We next designed a CRISPR/Cas9-mediated double genomic knock-out (KO) strategy to show that P10-KO slightly lessened androgen-elicited MYC transrepression in LNCaP-AR cells. In similar genomic editing assays, androgen-mediated MYC repression became slightly deepened upon KO of P11, an ABS in the PVT1 gene locus highly enriched in AR-binding motifs and peaks. We also investigated multiple ABSs in the established PCAT1 super-enhancer that distally interacts with MYC-Pro for transactivation, with each KO pool consistently shown to relieve androgen-elicited MYC repression. In the end, we systemically assessed androgen effects in the 8q24-MYC locus and along PCa genome to generalize H3K27ac and BRD4 re-distribution from pioneer factors (FoxA1 and HoxB13) to AR sites. Conclusion：Together, we reconciled these observations by unifying AR dual-functions that are mechanistically coupled to and equilibrated by co-factor redistribution.