Q & A With Emmanuel S. Antonarakis, MBBCh: AR-V7 and Resistance to AR-Targeting Agents

Emmanuel S. Antonarakis, MBBCh

Emmanuel S. Antonarakis, MBBCh

Androgen receptor splice variant-7 (AR-V7) is a truncated form of the AR that does not have the ligand-binding domain. Detection of AR-V7 in circulating tumor cells (CTCs) can be used to predict resistance to AR-targeting agents, such as abiraterone or enzalutamide. Overall survival results of a study looking at AR-V7 and resistance to these agents were presented at the 2014 ESMO Congress.

To gain further insight into this study,Targeted Oncologyinterviewed the lead author, Emmanuel S. Antonarakis, MBBCh, an assistant professor at Johns Hopkins Medicine.

What exactly is AR-V7?

AR-V7 is an abnormal form of the androgen receptor, which is a splice variant — it’s missing part of the protein that includes the ligand-binding domain. It’s a truncated and shorter version of the androgen receptor – it is abnormal because it retains the N-terminal but it’s missing the C-terminal. This is important because the C-terminal is the part of the AR molecule that contains the ligand-binding domain. It’s also the part of the molecule that anti-androgens such as enzalutamide bind to, as well as androgens, such as testosterone.

Why was this study conducted?

In 2008, these androgen receptor splice variants were discovered in the laboratory, in part by my collaborator Dr. Jun Luo and others. We hypothesized that because these androgen receptors are missing the ligand-binding domain but are still constitutively active as transcription factors, patients receiving enzalutamide or abiraterone would be resistant to these drugs. We predicted that if these patients had AR-V7 splice variants, the drugs would have nowhere to bind and therefore they would not be active. We set about to test this hypothesis prospectively in 30 patients receiving standard-of-care enzalutamide and 30 patients receiving standard-of-care abiraterone. Prior to starting treatment, we actually took a blood sample, checked the CTCs for the presence or absence of AR-V7, and then tried to figure out what that meant in terms of their prognosis or their response/resistance to these two drugs.

What was found?

The findings were very interesting. First, I’ll discuss the enzalutamide-treated patients. Enzalutamide binds directly to the ligand-binding domain which is missing on AR-V7.  We looked at a total of 31 patients and about 39% of them (12/31) had the AR-V7 detected in their CTCs. Every patient that had AR-V7 did not obtain a PSA response to enzalutamide. Among the patients who received the drug, their PSA never went down and their best response was a continued increase in their PSA. In addition, progression-free survival was worse in AR-V7-positive patients compared to AR-V7-negative patients.

We also looked at 31 patients treated with abiraterone. Similarly, 19% of patients (6/31) were found to have AR-V7 in their CTCs. All the patients that had AR-V7 in their CTCs demonstrated resistance to abiraterone. In other words, their PSA never went down. In addition, progression-free survival was inferior in AR-V7-positive compared to -negative patients.

What we presented at the ESMO 2014 meeting this year is the overall survival results from both cohorts. We found that survival is worse in patients treated with enzalutamide who had AR-V7 detected in their CTCs compared with those patients who did not have any measurable AR-V7. Likewise, in patients receiving abiraterone, survival was inferior in patients that were AR-V7-positive compared with those that were AR-V7-negative.

We then conducted a multivariable analysis to adjust for some other factors; we wanted to see whether the presence of AR-V7 was independently prognostic after adjusting for those factors. We found that it was: the multivariable model showed that the presence of AR-V7 was an independently worse prognostic factor, and that survival in those patients that were AR-V7-positive was inferior even after appropriate adjustment.

The other factor that was also significant in the multivariable model was whether or not the patient had previously received abiraterone or enzalutamide. In the enzalutamide cohort, patients who had previously received abiraterone experienced worse survival. In the abiraterone cohort, those that had received enzalutamide prior — their survival was worse. Those were the only two factors in the multivariable model that portended a worse survival: (1) AR-V7 presence as well as (2) prior treatment with enzalutamide/abiraterone.

Can you put this in context? What does this mean for a community oncologist who treats patients with prostate cancer?

We believe this finding is important because enzalutamide and abiraterone are both widely available drugs in the U.S. and Europe. The drugs are effective in about 80% of people. In those patients who achieve responses, the responses can be quite prolonged, often longer than one year or even two years sometimes. We’ve known for a while that about 20% of patients have no response to these drugs at all. Our goal was to try to identify who those 20% of patients might be. Looking at the broader picture, we think we might have identified one potential mechanism of resistance to both drugs. Our thinking is that patients who test positive for AR-V7 prior to initiating therapy with these two drugs might receive an alternative therapy — perhaps a chemotherapy or new agent on a clinical trial. If these findings are confirmed and validated in larger studies, patients who do test positive for AR-V7 can be steered toward other therapies that do not inhibit the androgen receptor pathway.

What are the next steps for this research?

This research has opened up a can of worms of next steps. The next steps are quite tedious but very important. We want to perform this test in a CLIA-certified laboratory. The next step will be to see if other institutions across the country and across the world can perform this test in their laboratories. We want to prospectively validate this in a large, multicenter study and see whether the prognostic utility of AR-V7 holds up. Finally, we’d like to identify drugs that might inhibit or degrade AR variants that might have activity in these patients who test positive for AR-V7. Such drugs are in development.