AR-FL mRNA Correlates With Response to AR-Targeted Therapies

Emmanuel S. Antonarakis, MD

Circulating tumor cell mRNA of full-length androgen receptor (AR-FL) correlated with detection of the AR-V7 variant and with response to AR-targeted therapy for castration-resistant prostate cancer (CRPC), according to results of a prospective clinical study.

Higher levels of AR-FL mRNA correlated with declining prostate-specific antigen (PSA) progression, progression-free survival (PFS), and overall survival (OS). AR-FL mRNA levels were also associated with detection of AR-V7—previously shown to predict lack of response to abiraterone acetate (Zytiga) and enzalutamide (Xtandi)—and with a lower PSA response in patients treated with either of the AR-targeting drugs. These were reported during the 2017 Genitourinary Cancers Symposium.

“Full-length AR copy number is prognostic for clinical outcomes in patients treated with abiraterone and enzalutamide,” said study author Emmanuel S. Antonarakis, MD, associate professor of oncology and urology at Johns Hopkins Medicine. “In addition to AR-Vs, AR-FL quantification could serve as another molecular biomarker of abiraterone/enzalutamide sensitivity.”

Taken together, the 2 biomarkers could provide a basis for tailored CRPC treatment, he added. Tumors with only wild-type AR may be treated with conventional AR-directed therapy. Aberrant AR-driven tumors would require treatment with next-generation AR-directed therapies that are currently in various stages of development. Finally, AR-independent tumors (AR null/low) would warrant consideration of non—AR-directed therapies, such as a taxane, immunotherapy, radium-223 dichloride (Xofigo), olaparib (Lynparza), or enrollment in a clinical trial.

The emergence of liquid biopsy technology has allowed real-time investigation of the AR across its entire axis. Studies of the AR-V7 variant have shown that detection in circulating tumor cells (CTCs) may be predictive for outcomes in patients treated with abiraterone or enzalutamide. In contrast, no data on the potential predictive or prognostic value of AR-FL mRNA quantification from CTCs had been presented, said Antonarakis.

Previous studies involving analysis of ctDNA showed that the presence of AR amplification or mutation was associated with inferior outcomes in patients treated with enzalutamide or abiraterone, as compared with wild-type AR. Continuing that line of investigation, Antonarakis and colleagues conducted a prospective analysis of AR-FL in CTCs obtained from men initiating treatment with abiraterone or enzalutamide for CRPC.

Using a quantitative RT-PCR assay, investigators examined the relationship between baseline AR-FL status and PSA response, PSA progression-free survival (PSA-PFS), clinical/radiographic PFS, and OS. They developed multivariable models to adjust for AR-V7 status, PSA level, Gleason score, number of prior therapies, prior abiraterone/enzalutamide use, prior taxane therapy, presence of visceral metastases, and ECOG score.

Antonarakis and colleagues used a CTC-based AR-FL assay, including selection and detection kits. The selection kit permitted isolation and enrichment of CTCs and immunomagnetic capture via epithelial and tumor-associated antigens. The detection kit permitted mRNA purification and use of multiplex RT-PCR probes for PSA, prostate-specific membrane antigen, EGFR, and actin. Investigators developed custom primers for AR-FL and AR-V7.

For the analysis, investigators enrolled 202 men prior to initiation of AR-targeted therapy. Patients were separated by AR-FL status into 3 categories: AR-FL negative or undetectable (n = 97, 48%), AR-FL positive below the median (n = 52, 26), and AR-FL positive above the median (n = 53, 26%).

Comparison of AR-FL status by AR-V7 status showed that AR-V7—negative status by CTC (n = 113, mean 10.5 copies) was associated with a median AR-FL level of 1.4 copies, whereas AR-V7–positive CTCs (n = 36, mean 136.9 copies) were associated with a significantly higher AR-FL level (median 35.5 copies; P <.001). Further analysis confirmed a significant association between AR-FL and AR-V7 (P < .001).

AR-FL level had a significant association with PSA response. Patients who had at least a 50% reduction in PSA level from baseline (n = 104) had a mean AR-FL value of 6.7 copies (median value 0). Patients who did not have at least a 50% reduction in PSA level in response to treatment had a mean AR-FL value of 55.4 copies (median 3.9; P < .001).

PSA response rate (at least 50% decline from baseline) differed significantly across the 3 previously established AR-FL cutoffs (P < .001): AR-FL negative, 62% (60 of 97); AR-FL median, 28% (15 of 53).

The other outcomes of interest also differed significantly by AR-FL status (negative, median). The median values for PSA-PFS were 9.6, 6.2, and 2.5 months, respectively (P <.001). Values for PFS were 11.1, 8.7, and 3.2 months (P < .001), and median values for OS were 33.3, 18.0, and 11.3 months (P < .001).

“CTC-based liquid biopsy allowed full-length AR mRNA quantification,” said Antonarakis. “Higher AR-FL levels correlated with AR-V7 positivity. Higher AR-FL also correlated with inferior clinical outcomes and generally remained significant in multivariable models. AR-V7 remained independently prognostic for all outcomes in multivariable analyses.”

Silberstein J, Luber B, Wang H, et al. Clinical significance of AR mRNA quantification from circulating tumor cells (CTCs) in men with metastatic castration-resistant prostate cancer (mCRPC) treated with abiraterone (Abi) or enzalutamide (Enza). J Clin Oncol. 2017;35 (suppl 6S; abstr 132).

<<<

View more from the 2017 Genitourinary Cancers Symposium