The first genomic landscape study of metastatic castration-resistant prostate cancer (mCRPC) found a high frequency of mutation and/or deletion of TP53 (>50% of mCRPC patients) relative to primary prostate cancers (<5%). Changes in androgen receptor (AR), PTEN, and RB seen in other profiling efforts were also observed. We have recently found that TP53 loss, particularly in combination with RB pathway dysfunction, confers resistance to enzalutamide in prostate cancer cell line and xenograft models. Remarkably, this resistance occurs in the absence of restored AR pathway signaling. Instead, these drug-resistant cells display reduced expression of luminal epithelial markers (CK8, NKX3.1, AR) and acquire expression of basal epithelial and neuroendocrine markers (CK5, p63, synaptophysin). This profile of lineage plasticity is heterogeneous across the population of cells with TP53 and RB loss and reversible upon restoration of TP53 function. Similar to the well-documented effects of TP53 and RB disruption in enhancing the efficiency of generating induced pluripotent stem cells (iPSC) from fibroblasts using “Yamanaka” transcription factors, we postulate that loss of TP53/RB pathway function creates a state of increased plasticity that allows luminal prostate cancer cells to escape the growth suppressive effects of castration therapy by assuming characteristics of basal epithelial or neuroendocrine cells which no longer depend on AR signaling for growth. Parallel studies of CRPC patients, using circulating tumor cells (CTCs) from phlebotomy samples obtained from men prior to a change in systemic therapy in the 1st, 2nd and/or 3rd line setting, reveal quantifiable differences in morphological and molecular tumor heterogeneity within the same patient. Using the validated Epic platform, which enables serial characterization of multiple phenotypic and molecular features including morphology, protein expression, and genotype, we developed a measure of CTC heterogeneity within an individual patient that reveals some patients to have primarily clonal, homogeneous CTCs and others to have a markedly heterogeneous collection of CTCs. Our preliminary results show that a high CTC heterogeneity index is correlated with worse treatment response to AR-targeted drugs (abiraterone, enzalutamide) but not to taxanes (docetaxel, cabazitaxel). The primary goal of this project will leverage our finding, from preclinical prostate model systems, that TP53 and RB loss promote a state of epithelial lineage plasticity and resistance to AR therapy, in order to ask if TP53 and RB are molecular determinants of the high CTC heterogeneity observed in patients. The combination of mechanistic (Aim 1) and correlative clinical (Aim 2) studies in this project is designed to enable development of biomarkers that identify patients destined for a poor response to AR-targeted therapies.