The goal of this application is to develop mouse models for prostate cancer using transgenic and gene knockout technologies that create perturbations in signal transduction pathways relevant to the pathogenesis of the human disease. Based on work from this group and others, we will focus on the IGF, PTEN/AKT and TGFf3RII signaling pathways. Mice containing activating or inactivating mutations in genes which regulate these pathways will be constructed singly and in combinations to develop mouse models which recapitulate the human disease. In parallel, we will develop technologies to evaluate and improve upon these models based on recent accomplishments of our group in identifying novel prostate-specific genes and in imaging gene expression in vivo using positron emission tomography (PET). Specifically, we will test an alternative prostate- specific promoter which targets expression to the basal rather than secretory epithelial cells, as we believe transgenes targeted to this compartment of the prostate gland may create better models. To address the significant problem of detecting metastasis in these animals (particularly to bone), we will create transgenic animals expressing PET reporter genes to allow selective, high-resolution imaging of prostate cancer cells in living mice. Finally, we will use recombinase-expressing adenoviruses to selectively activate or inactivate floxed versions of signaling pathway genes in a developmentally controlled fashion by surgical injection into the prostate gland. These studies will provide insight into the molecular pathways responsible for human prostate cancer and will generate clinically relevant mouse models. Furthermore, this group will provide novel technologies for imaging of gene expression that will be a resource for other cancer models in the consortium, particularly in defining metastatic patterns and in assessing response to therapies.