PLoS ONE · 2009 · 79 citations · 47 references
Life SciencesMolecular BiologyOncologyGene ExpressionGene Expression AnalysisTranscriptional RegulationCancer GenomicsCancer TranscriptomicsHormone ReceptorsNuclear ReceptorsProstate CancerSteroid Hormone SignalingAndrogen Receptor ExpressionAndrogen Receptor RegulationAndrogen ReceptorsExpression ProfilingLigand-independent ActivationLigand-independent Receptor ActivationLigand-independent SignalingTumor Suppressor Pathways
BACKGROUND: The androgen receptor (AR) plays important roles in the development of male phenotype and in different human diseases including prostate cancers. The AR can act either as a promoter or a tumor suppressor depending on cell types. The AR proliferative response program has been well studied, but its prohibitive response program has not yet been thoroughly studied. METHODOLOGY/PRINCIPAL FINDINGS: Previous studies found that PC3 cells expressing the wild-type AR inhibit growth and suppress invasion. We applied expression profiling to identify the response program of PC3 cells expressing the AR (PC3-AR) under different growth conditions (i.e. with or without androgens and at different concentration of androgens) and then applied the newly developed ChIP-seq technology to identify the AR binding regions in the PC3 cancer genome. A surprising finding was that the comparison of MOCK-transfected PC3 cells with AR-transfected cells identified 3,452 differentially expressed genes (two fold cutoff) even without the addition of androgens (i.e. in ethanol control), suggesting that a ligand independent activation or extremely low-level androgen activation of the AR. ChIP-Seq analysis revealed 6,629 AR binding regions in the cancer genome of PC3 cells with an FDR (false discovery rate) cut off of 0.05. About 22.4% (638 of 2,849) can be mapped to within 2 kb of the transcription start site (TSS). Three novel AR binding motifs were identified in the AR binding regions of PC3-AR cells, and two of them share a core consensus sequence CGAGCTCTTC, which together mapped to 27.3% of AR binding regions (1,808/6,629). In contrast, only about 2.9% (190/6,629) of AR binding sites contains the canonical AR matrix M00481, M00447 and M00962 (from the Transfac database), which is derived mostly from AR proliferative responsive genes in androgen dependent cells. In addition, we identified four top ranking co-occupancy transcription factors in the AR binding regions, which include TEF1 (Transcriptional enhancer factor), GATA (GATA transcription factors), OCT (octamer transcription factors) and PU1 (PU.1 transcription factor). CONCLUSIONS/SIGNIFICANCE: Our data provide a valuable data set in understanding the molecular basis for growth inhibition response program of the AR in prostate cancer cells, which can be exploited for developing novel prostate cancer therapeutic strategies.
47
Pattern Recognition and Machine Learning
Journal of Electronic Imaging · 2007 · 22.1K citations
High-Resolution Profiling of Histone Methylations in the Human Genome
Artem Barski, Suresh Cuddapah, Kairong Cui et al. · Cell · 2007 · 6.9K citations · Full text
Ewan Birney, Shane Neph, Evelyn J. Cheung et al. · Nature · 2007 · 5.3K citations · Full text
Pattern Recognition and Machine Learning
Radford M. Neal · Technometrics · 2007 · 4.7K citations
Genome-Wide Mapping of in Vivo Protein-DNA Interactions
David S. Johnson, A Mortazavi, R Myers et al. · Science · 2007 · 3K citations