Identification of TDRD1 as a direct target gene of ERG in primary prostate cancer.

Molecular classification of ERG-rearranged prostate cancer clarifies the role of TMPRSS2-ERG in the development and progression of prostate cancer. The objective of our study was to identify direct ERG target genes in ERG-rearranged prostate cancer. Two independent cohorts of primary prostate cancer (Cohort A, n=48; Cohort B, n=31), a cohort of late-stage prostate cancer (n=51) and expression array data of a cohort of primary prostate tumors from a different institute (n=128) were analyzed for expression of genes that were coexpressed with ERG overexpression. By genome-wide expression analysis and Q-RT-PCR it was shown that the gene Tudor domain containing 1 (TDRD1) was by far the strongest correlated gene with ERG overexpression in both Cohort A and B. Expression array analysis of the patient cohort from a different institute showed a large overlap in genes that were positively correlated with ERG overexpression, including TDRD1. In late-stage prostate cancer, TDRD1 was also coexpressed with ERG overexpression, although a proportion of ERG-negative late-stage samples expressed TDRD1. TDRD1 expression was not associated with ETV1 overexpression. In the prostate cancer cell line VCaP, downregulation of ERG by shRNA lead to a lower expression level of TDRD1 and resulted in a decreased activity of the TDRD1 promoter. By mutation analysis we identified a functional ERG binding site in the TDRD1 promoter. Our findings show TDRD1 as the first identified upregulated direct ERG target gene that is strongly associated with ERG overexpression in primary prostate cancer.

Stepwise androgen receptor dimerization.

Androgen-regulated gene expression is a highly coordinated dynamic process mediated by androgen receptor (AR) ligand binding and DNA binding, and by specific AR protein-protein interactions. The latter include DNA-binding domain (D-box) interactions in AR homodimers, and the interaction of the FQNLF motif in the AR N-terminal domain and the coactivator groove in the ligand-binding domain (N/C interaction). We have studied these interactions in AR homodimerization using quantitative imaging techniques. We found that the initial cytoplasmic intramolecular AR N/C interaction after ligand binding is followed by a D-box-dimerization-dependent transition to intermolecular N/C interaction in a proportion of nuclear ARs. The consecutive steps leading to homodimerization are initiated prior to DNA binding. Our data indicate the presence of nuclear pools of both AR homodimers and monomers. On the basis of AR-regulated reporter assays we propose specificity in regulation of gene expression by AR homodimers and monomers mediated by AR domain interactions. Moreover, our findings elucidate important steps in the spatiotemporal organization of AR intra- and inter-molecular interactions.

Inhibition of androgen receptor functions by gelsolin FxxFF peptide delivered by transfection, cell-penetrating peptides, and lentiviral infection.

BACKGROUND: Prostate cancer (PC) growth is dependent on the androgen-androgen receptor (AR) axis. Because current androgen ablation therapies of PC lead to resistance, novel approaches to block AR activity are urgently needed. METHODS: We inhibited AR function beyond the level of hormone binding by blockade of the coactivator groove in the ligand-binding domain (LBD) using a high-affinity gelsolin FxxFF peptide. Following peptide selection, the effect of the gelsolin FxxFF peptide on AR functions was determined in Hep3B cells that were transiently transfected with pM-peptide expression vectors or were incubated with synthetic gelsolin FxxFF peptide coupled to the TAT cell-penetrating peptide. Lentiviruses expressing the gelsolin FxxFF peptide were used to study endogenous AR target gene expression in LNCaP cells. RESULTS: pM-Gelsolin FxxFF efficiently interfered with AR N/C interaction and specifically inhibited AR-regulated reporter gene activity. The peptide did not inhibit progesterone receptor (PR) and glucocorticoid receptor (GR) activity, nor constitutively active gene promoters. The peptide also specifically blocked in vitro interactions of AR LBD with peptides. Like the gelsolin FxxFF peptide expressed by an expression vector, synthetic TAT-gelsolin FxxFF peptide efficiently blocked AR N/C interaction and inhibited full-length AR-regulated reporter gene activity. It hardly affected PR and GR activity, but the effect on constitutively active promoters was variable. Lentiviral gelsolin FxxFF peptide inhibited expression of KLK2 and NDRG1, but hardly affected PSA and TMPRSS2. CONCLUSIONS: Our results show that the AR coactivator groove may function as a target to overcome therapeutic failure that arises during current androgen ablation therapies.

Functional screening of FxxLF-like peptide motifs identifies SMARCD1/BAF60a as an androgen receptor cofactor that modulates TMPRSS2 expression.

Androgen receptor (AR) transcriptional activity is tightly regulated by interacting cofactors and cofactor complexes. The best described cofactor interaction site in the AR is the hormone-induced coactivator binding groove in the ligand-binding domain, which serves as a high-affinity docking site for FxxLF-like motifs. This study aimed at identifying novel AR cofactors by in silico selection and functional screening of FxxLF-like peptide motifs. Candidate interacting motifs were selected from a proteome-wide screening and from a supervised screening focusing on components of protein complexes involved in transcriptional regulation. Of the 104 peptides tested, 12 displayed moderate to strong in vivo hormone-dependent interactions with AR. For three of these, ZBTB16/PLZF, SMARCA4/BRG1, and SMARCD1/BAF60a, the full-length protein was tested for interaction with AR. Of these, BAF60a, a subunit of the SWI/SNF chromatin remodeling complex, displayed hormone-dependent interactions with AR through its FxxFF motif. Vice versa, recruitment of BAF60a by the AR required an intact coactivator groove. BAF60a depletion by small interfering RNA in LNCaP cells demonstrated differential effects on expression of endogenous AR target genes. AR-driven expression of TMPRSS2 was almost completely blocked by BAF60a small interfering RNA. In summary, our data demonstrate that BAF60a directly interacts with the coactivator groove in the AR ligand-binding domain via its FxxFF motif, thereby selectively activating specific AR-driven promoters.