Abstract A1-02: Chip-Seq and RNA-Seq studies reveal novel insights into the interactive and transcriptomic architecture of the human prostate cancer

Abstract

Abstract Introduction: Among men, carcinoma of the prostate is the second most common cause of cancer death in the US, with 29,480 deaths per year [American Cancer Society, Cancer Facts & Figures 2014]. Once prostate cancer (PCa) has metastasized, controlling it is often unsuccessful. Current methods and markers used to predict PCa outcome are inadequate. Discovering a new PCa marker that promises to improve prognostication beyond what is achieved with other reported biomarkers becomes of utmost importance. The proposed research is aimed at understanding genome-wide binding of transcriptional repressor DAXX and its epigenetic partner DNMT1 in the PCa cell line PC3, and determine changes in gene expression levels between the wild-type (WT) and DAXX knock-down (K/D) PC3 cells. Methods: To determine the genome-wide distribution of DAXX in PC3 cells, a two-step cross-linking procedure (protein-protein & DNA-protein) preceded chromatin immunoprecipitation (ChIP) using anti-DAXX and anti-DNMT1 to collect chromatin bound to the respective antibody, and was followed by deep sequencing (ChIP-Seq). Non-immunoprecipitated (input) chromatin, subjected to the same treatment, served as a control. Massively parallel RNA sequencing (RNA-Seq) was used to investigate in an unbiased fashion the expression of different genes, comparing the WT and DAXX K/D expression patterns in PC3 cells. Genomic data analysis and visualization of ChIP-Seq and RNA-Seq analyses was performed as follows: ChIP-Seq sequencing reads were mapped to the human genome (GRCh37/hg19) using Bowtie2. Genome browser BedGraph tracks and read density histograms were generated using HOMER. Peak finding and annotation to the nearest RefSeq gene promoter was performed using HOMER. De novo motif discovery was also carried out using HOMER. RNA-Seq reads were aligned to the human genome (GRCh37/hg19) using STAR. Read densities were visualized by preparing normalized BigWig files using HOMER and uploading them to the UCSC Genome Browser. Gene expression was determined by identifying reads on the appropriate strand overlapping exons defined by RefSeq using HOMER. Differentially expressed genes were then identified using EdgeR. Functional Enrichment with Gene Ontology was performed using DAVID. Results: Using ChIP-Seq, 59,818 DAXX peaks were found in WT PC3 cells, compared to 42,934 in DAXX K/D cells. Importantly, DAXX peaks overlapped with DNMT1 peaks, and DNMT1 peaks were lost after DAXX knockdown (2,045 DNMT1 peaks in WT vs. 408 DNMT1 peaks in DAXX K/D). DNMT1 and DAXX bound to transcriptional activator motifs, presumably inducing their repression. Thus DNMT1 enrichment is dependent on DAXX, further corroborating previous findings that DAXX recruits DNMT1 to repress its targets. Furthermore, the distribution of gene transcription start site (TSS) to the nearest DAXX ChIP-Seq peak was analyzed. Regulated gene TSS are normally found closer to the ChIP-Seq peaks than non-regulated peaks. Autophagy genes were found to be closer to DAXX ChIP-Seq peaks than all other genes combined, implying they are more likely to be regulated by DAXX (with the difference being very significant: p<10-7). Using RNA-Seq, 2,716 genes were found to be induced and 2,329 repressed by DAXX K/D. Genes induced by DAXX K/D included those involved in translation and autophagy. Related to autophagy genes, ChIP-Seq and RNA-Seq data showed significant overlap. Specifically, among autophagy genes, DAXX presence was inversely correlated with expression of positive regulators of autophagy - DAPK1, DAPK3, ATG8, ATG18, ATG101, LC20, ATG3, ATG4, Beclin1, ATG7, and Deptor, whereas it was linearly correlated with negative regulators of autophagy – mTOR and Raptor. Conclusions: In the prostate cancer genome, DAXX co-localizes with DNMT1. DAXX represses autophagy gene expression, presumably because of its recruitment of DNMT1. DAXX can serve as a new marker and therapeutic target in prostate cancer. Citation Format: Lorena A. Puto, Chris Benner, Tony Hunter. Chip-Seq and RNA-Seq studies reveal novel insights into the interactive and transcriptomic architecture of the human prostate cancer. [abstract]. In: Proceedings of the AACR Special Conference on Translation of the Cancer Genome; Feb 7-9, 2015; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2015;75(22 Suppl 1):Abstract nr A1-02.