Dynamic rewiring of the androgen receptor protein interaction network correlates with prostate cancer clinical outcomes.

The androgen receptor (AR) is a ligand-inducible transcription factor, a member of the nuclear receptor superfamily, which plays an important role in the development and progression of prostate cancer (CaP). The transformation to CaP has been linked to several somatic AR gene mutations and changes in AR protein complex formation, which in turn increase the potential activity of the receptor. Thus, to address the mechanism of AR-mediated neoplastic transformation, we developed in vitro methodology to isolate and characterize, via mass spectrometry, AR complexes of three AR genetic variants: wild type-AR, and two somatic gain-of-function AR prostatic mutants (T877A-AR and 0CAG-AR isoforms). To fully characterize the significance of our large raw data set, we employed a sophisticated systems biology approach to create an integrative protein-interaction network profile for each AR isoform. Our comparative analysis identified subnetwork cluster profiles for AR isoforms (WT, T877A, and 0CAG) that segregated AR isoforms on the basis of androgen stimulation conditions and mutant aggressiveness. Furthermore, results from additional correlative gene microarray analysis studies of all three AR isoform (WT, T877A, 0CAG) subnetwork clusters were assessed and found to be significantly enriched in tumor versus normal prostate tissues. We also identified two AR-interaction clusters, containing 21 and 30 proteins, respectively, that showed unfavourable prognosis outcome of recurrent cancers, on the basis of PSA, Gleason score and combined PSA/Gleason score. In conclusion, we have characterized a large panel of novel AR-interacting proteins, through a combined proteomics/systems biology screen, that are of clinical relevance and could potentially serve as novel markers for diagnosis and prognosis of CaP.