Genomic strategy for targeting therapy in castration-resistant prostate cancer.

Purpose: Despite treatments which lower circulating androgens, advanced prostate cancers often maintain androgen receptor (AR) signaling. The variable response to secondary hormonal manipulations in men with castrate-resistant prostate cancer (CRPC) creates a compelling need for strategies to individualize therapy based on the molecular features of each patient's tumor.

Methods: A transcription-based AR activity signature was developed from an androgen-sensitive prostate cancer cell (LNCaP) and tested on independent data sets of prostate cancer cell lines and human tumors to assess its precision and accuracy in detecting AR activity. The AR signature was applied to multiple sets of prostate specimens to determine how AR activity changes with hormone therapy and progression and oncogenic pathway analysis was used to identify biologic pathways correlating with AR activity.

Results: A robust AR signature accurately predicts AR activity in multiple prostate cancer cell lines, has minimal variation between replicate samples, and accurately reflects an individual's hormone status and intraprostatic dihydrotestosterone levels. The AR signature finds AR activity to be high in local, untreated prostate tumors and decreased in prostate tissue after neoadjuvant hormone therapy and in CRPC. Heterogeneity of AR activity exists along the spectrum of prostate cancer progression and decreasing predicted AR activity correlates with increasing predicted Src activity and sensitivity to dasatinib (Src-targeting kinase inhibitor).

Conclusion: A transcription-based AR signature can detect AR activity within individual prostate cancer specimens and has the potential to help individualize and improve care for patients with CRPC.