Integrated network model provides new insights into castration-resistant prostate cancer

Castration-resistant prostate cancer (CRPC) is the main challenge for prostate cancer treatment. Recent studies have indicated that extending the treatments to simultaneously targeting different pathways could provide better approaches. To better understand the regulatory functions of different pathways, a system-wide study of CRPC regulation is necessary. For this purpose, we constructed a comprehensive CRPC regulatory network by integrating multiple pathways such as the MEK/ERK and the PI3K/AKT pathways. We studied the feedback loops of this network and found that AKT was involved in all detected negative feedback loops. We translated the network into a predictive Boolean model and analyzed the stable states and the control effects of genes using novel methods. We found that the stable states naturally divide into two obvious groups characterizing PC3 and DU145 cells respectively. Stable state analysis further revealed that several critical genes, such as PTEN, AKT, RAF, and CDKN2A, had distinct expression behaviors in different clusters. Our model predicted the control effects of many genes. We used several public datasets as well as FHL2 overexpression to verify our finding. The results of this study can help in identifying potential therapeutic targets, especially simultaneous targets of multiple pathways, for CRPC.