Abstract B12: LSD1 mediates global AR transcription suppression in prostate cancer cells

Abstract

Abstract This project is to study the molecular basis of LSD1-dependent AR suppression in prostate cancer cells. Androgen receptor (AR) is highly expressed in prostate cancer (PCa) and plays a pivotal role in tumor growth. Patients generally respond to androgen deprivation therapy (ADT). However, the tumors invariably relapse, and these relapsed tumors (called castration-resistant prostate cancer, CRPC) are generally more aggressive and relatively resistant to current AR antagonist treatments. Significantly, these relapsed tumors express increased levels of AR mRNA and express multiple AR regulated genes, indicating that AR transcriptional activity has been restored. Despite the critical role AR plays in PCa development and progression to CRPC, the mechanisms that contribute to the consistent substantial increases in AR mRNA in CRPC are poorly understood. Using a VCaP xenograft model, we initially detected a rapid and substantial upregulation of AR mRNA upon castration, suggesting a negative feedback loop regulating AR mRNA levels, which may make a significant contribution to increasing AR mRNA in CRPC (Cancer Research 2009, 69:6027). In a recent published study (Cancer Cell 2011, 20:457), we reported that the agonist liganded AR binds to an enhancer in the second intron of the AR gene, and that increased AR gene expression and progression to CRPC is associated with increased H3K4 methylation and increased recruitment of other transcription factors to this site. Significantly, in contrast to AR function as a transcriptional activator on previously studied androgen responsive elements (AREs), we found that the agonist liganded AR functions as a direct transcriptional repressor at this site. The repression is due to AR recruitment of an H3K4 demethylase, LSD1. Interestingly, AR binding to this site and repression of AR gene expression occurs at slightly higher androgen levels than those required for classical androgen regulated genes, consistent with a negative feedback loop to regulate AR signaling. Pathway analyses of additional AR repressed genes showed marked enrichment for genes mediating DNA synthesis and cell cycle progression, while AR stimulated genes were associated with lipid/protein synthesis and cellular metabolism. Significantly, a set of these androgen repressed genes that are associated with proliferation are overexpressed in CRPC. Taken together, these results indicate that the agonist liganded AR functions as a transcriptional repressor on a subset of genes that enhance proliferation. We propose that ADT is effective because it initially suppresses all AR functions, but that the partial restoration of androgen levels and AR activity in CRPC cells may provide a strong growth advantage by stimulating cellular metabolism without downregulation of AR repressed genes that enhance cellular proliferation. Combining the expression array data and ChIP-seq analysis on AR in VCaP and VCaP-derived castration-resistant VCS2 cells, we have identified gene subsets with AR direct binding on the gene locus. Interestingly, while AR binding is highly enriched (∼2fold) in AR-activated gene subset, the binding is less enriched (1.4fold) in AR-suppressed gene subset. The lower enrichment could mean that fewer genes in the AR –repressed group are directly regulated by AR, but could also be in part technical and reflect somewhat weaker binding of AR to AR-repressed genes. To further elucidate the molecular basis for AR suppression function, we focused on the role of LSD1 in the global AR transcriptional repression as previous results showed that expressions of a number of AR-suppressed genes are dependent on LSD1 activity. LSD1 was initially identified in corepressor complexes and shown to function by demethylating mono- and dimethylated H3K4. However, it was subsequently shown to also function as a coactivator through demethylation of repressive mono- and dimethylated H3K9 when associated with AR. Our data indicate that the association with AR does not determine the coactivator versus corepressor function of LSD1, and that it is instead determined by properties of the element to which it is being recruited. Through genome wide ChIP-seq analysis of LSD1 in VCaP and LNCaP cells, we found that LSD1 binding is significantly enriched for both AR-activated and –repressed (both ∼1.5fold) genes, suggesting that both AR activation and repression are dependent on LSD1 activity. Furthermore, we searched the transcription factor-binding motif on LSD1 binding sites. While E2F1 is highly enriched for LSD1 binding in both activation and repression loci, the ZBTB (zinc finger and BTB domain containing) transcription repressor binding motif was only significantly enriched in repression loci. Among this family, PLZF and LRF are highly expressed in prostate cancer cells. Therefore, the current work has focused on these two factors and we propose that LSD1 mediates AR gene suppression through interaction with ZBTB transcription repressors. Citation Format: Changmeng Cai, Housheng He, Sen Chen, X. Shirley Liu, Myles Brown, Steven P. Balk. LSD1 mediates global AR transcription suppression in prostate cancer cells [abstract]. In: Proceedings of the AACR Special Conference on Advances in Prostate Cancer Research; 2012 Feb 6-9; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2012;72(4 Suppl):Abstract nr B12.