Prostate cancer (PC) is among the top five leading malignancies causing cancer mortality worldwide. Increased prevalence and declined mortality rate have led to an increment in follow-ups with a significant rise in the economic burden. In the last few years, several new options for metastatic disease have been approved leading to clinicians to have multiple choices of therapy sequences. However, not all patients initially respond and most of them eventually develop resistance. The ability of tumour cells to reprogram themselves and survive despite the blocked targets (tumour cell plasticity) may accounts for the absence of durable responses. In addition, the fact that initial treatments may affect the potential benefit of subsequent treatments highlights the need to discover predictive biomarkers for optimal treatment selection, allowing early changes in treatment for non-responding patients. This multidisciplinary project aims to elucidate the molecular mechanisms responsible for tumour cell plasticity associated to treatment response in metastatic PC. Patient-derived xenografts (PDX) in vivo models will be used, since they preserve molecular heterogeneity and therapeutic response observed in the clinic. They will be treated under different regimens. Tumour tissues will be analysed by RNA sequencing before and after treatment. Bioinformatics analysis will be used to establish molecular signatures of treatment response that will be validated in liquid biopsy (circulating tumour cells and cell free DNA) from metastatic PC patients undergoing different treatments. PDX-PC will provide a better understanding of the molecular evolution of the disease that will contribute to design more specific and individualized treatments. In the setting of non-curative therapy, the implementation of such molecular findings at clinical practice may help to guide treatment decisions, improve outcomes, and prevent unnecessary side effects and costly therapies in men with metastatic PC.