ANDROGEN RECEPTOR SPLICE VARIANT-V7: A Head’s Up — Its presence renders both Zytiga and Xtandi ineffective.
Ed Weber, M.D., Editor
November – December, 2014

Abstract 5001 presented at the 2014 ASCO Annual Meeting and the follow-up full data publication (Atonarakis and Luo et al. (NEJM Sept 2014) significantly advance our predictions as to who will and who will not respond to Xtandi and Zytiga: “AR-V7 and Resistance to Enzalutamide and Abiraterone in Prostate Cancer.” Their findings would serve to guide men positive for AR-V7 away from taking these drugs. These men have a zero percent likelihood of response. Knowing this ahead to time would avoid both a period of ineffectual therapy and also needlessly meeting expensive co-pays.

What is splice variant AR-V7? What consequences follow from its presence?

In the normal state the gene encoding the template for the full androgen receptor (AR) includes instructions for building the “ligand binding domain,” the target pocket for ligands testosterone (T) and dihydrotestosterone (DHT) as well as the part of the AR where flutamide, bicalutamide and enzalutamide bind. In splice variant V7 this information is absent (“spliced out’) and the result is an incomplete androgen receptor lacking this essential domain. This absence eliminates the possibility of T and DHT from interacting with the AR and promoting prostate cancer growth. By itself this might be beneficial, considering that the goal of androgen deprivation therapy is to impede the interaction of T and DHT with the androgen receptor. But there is a hitch: the variant AR-V7 is continuously self-activating (“constitutively active”) and can promote cancer growth in a ligand- independent manner.

In the absence of this ligand binding domain there is no site for for enzalutamide to block. Decreasing serum and intraprostatic androgens with Zytiga also becomes irrelevant. Both drugs are rendered ineffective.

How ineffective? The Atonarakis study analyzed circulating tumor cells (CTC) in patients with metastatic castrate-resistant prostate cancer (CRPC) for the presence or absence of the AR-V7 variant. Thirty-one men were treated with Xtandi and 31 with Zytiga. Analysis focused on PSA response, PSA progression-free survival, and overall survival.

The test for AR-V7 can also be performed on tissue from the prostate and metastatic sites, and the study showed good concordance between the AR-V7 status in CTC and tissue from these other sites

Men treated with Xtandi who were AR-V7 positive pre-treatment had a 0% PSA response compared to a 53% response in AR-V7 negative men. Correspondingly, in the men treated with Zytiga who were AR-V7 positive, the PSA response was also 0% compared to 68% for AR-V7 negative men.

The majority of men treated with Xtandi had previously been treated with Zytiga (65%), whereas 13% of those in the Zytiga cohort had previously been treated with Xtandi. This is likely to be the reason that a higher proportion of patients in the Xtandi cohort were AR-V7 positive.

The AR-V7 positive men in both groups had shorter PSA progression-free survival and overall survival.

The prevalence of AR-V7 appears to increase with the duration of androgen suppression.

The technical analysis for this truncated AR variant, AR-V7, was conducted at Johns Hopkins, which currently is the only institution performing the test. After isolation of the cell’s messenger RNA, identification of AR-V7 is made with an antibody (the “Adnatest’). It is possible that other methods of detection ( i.e. the “EPIC” antibody) may reveal that this splice variant is more widely prevalent.

AR-V7’s presence at diagnosis is associated with a more aggressive disease, suggesting that it may become a useful prognostic biomarker of adverse outcome.

The prevalence of AR-V7 in men with CRPC is 10 – 20 times greater than is found in hormone-naive men. Atonarakis noted that conversion from negative to positive can occur during the period of androgen suppressive treatment as was seen in the Xtandi group. No patients who were initially positive converted to negative, suggesting that patients are likely to remain positive over time.

In earlier work Luo found AR-V7 in 11% of hormone naive men rising to 70% in men with CRPC (Cancer Research, 2009). A paraphrase of remarks by Dr. Evan Yu, Associate Professor University of Washington, at the 2014 Annual ASCO meeting, stated: [AR-V7 was found in] “11.6% of patients who had received neither drug, 25.0% in those who had received only enzalutamide, 51.2% in those who had received only abiraterone, and 66.7% in those who received both.”

It is important to be clear that this study does not establish that AR-V7 causes the resistance to enzalutamide and abiraterone. It does however show that AR-V7 is associated with an unknown factor or mechanism which is responsible. Future work may elucidate the root cause and provide a method to subvert or bypass this adverse splice variant.

BOTTOM LINE: (From Atonarakis) – “If this finding is substantiated by others, it is possible that AR-V7 could be used as a biomarker to predict resistance to enzalutamide and abiraterone and to facilitate treatment selection.”