The androgen receptor (AR) regulates male physiology and is Central to the ontogeny and progression of prostate cancer. Understanding AR's role strongly impacts human health by identifying genetic risk factors and new therapeutic targets. In normal as well as pathologic cell growth, AR directs divergent actions, such as proliferation or differentiation, through gene targets that vary with promoter, cell type and stage. These context- dependent and opposing functions of AR are evident in prostate cancer where AR initially suppresses tumor development but ultimately promotes oncogenesis. AR's dual nature is highlighted by germ-line and somatic mutations of the AR gene. We hypothesize that mechanisms underlying differential gene regulation by variant ARs may reveal more effective ways to combat pathological actions of the wild type AR. In order to use genetic tools, we converted the mouse AR gene to the human sequence by homologous recombination. We first tested the effect of a polymorphic glutamine (Q) tract associated with male phenotypic variation. Q tract length in humanized AR mice impacts the androgen axis, and affects initiation, progression and therapy response of prostate cancer in a transgenic model. Somatic AR mutations in these tumors evince selection by treatment. Moreover, this was substantiated in a select set of human patients. Mutations chosen for further study affecting the N-terminal domain influence ligand-dependent and -independent activation. The mutant ARs use multiple mechanisms that ultimately alter target gene selectivity to promote disease. We will use a multipronged genetic and genomic analysis to define upstream and downstream mechanisms by which a subset of Q tract variants and mutant ARs ("the variants") confer differential gene regulation. Our aims will define variant AR activity in novel cell and mouse models and identify the downstream effectors and upstream activators of aberrant function by Next Generation deep sequencing. In Aim I, a small set of variant ARs will be stably expressed in normal and cancer cells to view their distinct biological programs in vitro and in xenograft tumors. In Aim II, pathologic activity of variant ARs will be tested in mice with prostate- targeted transgenes, in mice predisposed to cancer by deletion of one PTEN allele, and in mice with varying androgen axis strength. In Aim III, gene expression programs directed by the variant ARs will be defined by high-throughput sequencing of transcripts (RNA-tag-Seq) to identify direct and indirect AR targets. Chromatin immunoprecipitation and high-throughput sequencing (ChIP-Seq) will probe differential promoter recognition and highlight coregulators directing different biological outcomes. Differences shared amongst AR variants, in addition to unique actions, will highlight critical downstream targets and upstream regulators of wild type AR. Combined genetic and genomic study of these AR variants will reveal aspects of wild type AR function so that in future we may selectively block disease-promoting while retaining disease-inhibiting effects of AR in diverse contexts. These results will have immediate applications and broad significance for future studies.