ASCO GU 2018: Transcriptomic Heterogeneity of Androgen Receptor Activity in Primary Prostate Cancer: Identification and Characterization of a Low AR-active Subclass

San Francisco, CA (UroToday.com) Dr. Spratt presented original work from a cross-institutional collaborative effort studying androgen receptor signaling and activity. The androgen receptor (AR) is a key player in the pathogenesis of prostate cancer (PC). We now know that dysregulation and/or mutation of the AR gene correlates with significant differences in PC-related outcomes. AR signaling is involved in many important pathways (DNA signaling, immune modulation, etc…). AR amplifications/mutations are common (and often actionable), and it is also known that AR-activity (ARA) may play an important role in PC pathogenicity. Genomic analyses show that ARA is highly variable in patients with localized PC. 

In order to understand the clinical and biological ramifications of differences in ARA, Dr. Spratt and colleagues analyzed 7 cohorts of patients with PC to understand the effect of ARA heterogeneity on PC outcomes. One cohort was prospectively collected from the Decipher database (>5k patients), and 6 other institutional (retrospective) cohorts were gathered totaling another ~1,000 patients. 

They found that there is a vast amount of AR heterogeneity in localized PC within these cohorts. Intriguingly, they demonstrated a poor correlation between AR expression and actual AR activity. PSA also had absolutely no correlation with ARA, so should not be used as a means to infer ARA clinically. Among these patients, low ARA was discovered in 10-15% of patients with localized prostate adenocarcinoma. 

Diving deeper, the investigators found that low ARA tumors portended poorer prognosis. First, they demonstrated changed immune-stimulatory properties, such as increases in effector immune cells and decreases in suppressor immune cells. They also increased PD-L1 signaling. Low ARA also associated with decreases in mutations to DNA repair pathways, though this characteristic makes these tumors perhaps more sensitive to the DNA-damaging effects of radiation. Finally, genomic analysis demonstrated that low ARA tumors appear to be more basal-like in nature and showed differentiation similar to neuroendocrine tumors, which portends poor outcomes.

Looking at clinical outcomes, they found that a decrease in ARA associated with increased metastatic potential (HR 2.71 for distant metastases), even after local therapy with radiation or surgery. As ARA increased, drug sensitivity also increased, especially to the standard hormonal and taxane-based therapies for advanced PC. Interestingly, this suggests that ARA may be useful in future trials as a biomarker for predicting sensitivity to hormonal therapy or taxane chemotherapy.

Our evolving understanding of AR-signaling pathways and ARA is giving us a far better understanding of the biology of aggressive PC. Low ARA activity is now known to associate with poor clinical outcomes, and ARA may be an important biomarker for treatment sensitivity in the future. Hopefully, our understanding of ARA will provide even more potential targets for advanced PC in the near future.


Presented by: Daniel Spratt, M.D.; University of Michigan

Written by: Shreyas Joshi, MD, Fox Chase Cancer Center, Philadelphia, PA at the 2018 American Society of Clinical Oncology Genitourinary (ASCO GU) Cancers Symposium, February 8-10, 2018 - San Francisco, CA