Blood analysis for the androgen splice variant 7 (AR-V7) in circulating tumor cells (CTCs) is now commercially available at Johns Hopkins. The possibility for a man to learn his AR-V7 status now directly focuses attention on the clinical importance of AR-V7 and how it might directly affect the choice between taxane (i.e. Docetaxel or Cabazitaxel) chemotherapy or androgen receptor inhibition with, say, Zytiga or Xtandi. This decision is commonly faced by many men whose PSAs are rising after primary treatment despite androgen deprivation therapy (ADT), i.e. men with “castration-resistant prostate cancer” (CRPC), non-metastatic or metastatic.

[An order form and instructions for sending a specimen to the Hopkin’s lab is available upon request to [email protected].]

Circulating tumor cells:

CTCs are a rare population of cells in blood which are in dynamic equilibrium with the totality of the cancer, reflecting the genome of the primary tumor but also expressing the mutational landscape of the progressively evolving heterogeneous tumor burden. Sophisticated techniques permit genomic analysis of single cells, or aggregated pools of cells; but also snippets of messenger RNA, and even circulating free DNA.

The identification and enumeration of CTCs has been well studied (usually by the CellSearch method) and is recognized to offer important prognostic and predictive information. However, this article will focus on the molecular characterization of CTCs, in particular AR-V7 and its clinical consequence.

“All roads lead to Rome.”

The androgen receptor (AR) is the common nexus of all hormonal interventions directed toward subverting the AR signaling that promotes prostate cancer growth. It is the ligand binding pocket of this lengthy protein into which testosterone (T) and dihydrotestosterone (DHT) fit to initiate the AR signaling to its many androgen response DNA targets. And it is this pocket that antiandrogens such as Zytiga and Xtandi block to inhibit such signaling.

The AR-V7 splice variant:

AR-V7 is not a naturally occurring gene but a modification of the basic AR gene and codes for a “splice-variant,” a truncated form of the full-length AR. This variant form (there are > 20 AR splice variants) lacks the crucial binding pocket for T and DHT. The AR-V7 gene is present at a low percentage early in the disease but is increasingly over-expressed as the disease becomes more aggressive.

Additionally, its over-expression is progressively induced as a result of sequential anti-androgen therapy, i.e. Zytiga and Xtandi. The subversive behavior of these variants is that they are “constitutively” active, meaning that they promote adverse AR signaling even though lacking the ligand-binding pocket.

AR-V7 is not a mutation, so it is not captured on the standard genomic tests, such as Guardant360, FoundationOne Medicine, or Caris Life Sciences, which all report on the important mutations in genes; genes such as BRCA1 & 2, in which mutations might lead to the protocol use of the parp inhibitor, Olaparib; or the retinoblastoma (RB) gene which might alert to the transition to neuroendocrine differentiation.

Androgen-receptor variant 7 messenger RNA (mRNA) can be detected in circulating tumor cells by a quantitative reverse-transcriptase-polymerase-chain-reaction assay [RT-PCR] or by an antibody directed at the AR-V7 protein.

Clinical Application: What Does a Positive Test for AR-V7 imply?

The lead introduction to this issue was presented by Antonarakis and his colleagues from Johns Hopkins in the New England Journal of Medicine, Sept 2014: “AR-V7 and resistance to enzalutamide [Xtandi] and abiraterone [Zytiga] in prostate cancer.”

Their study prospectively analyzed men with metastatic CRPC (mCRPC)
 31 treated with enzalultamide (39% positive for AR-V7 in CTCs) and
 31 abiraterone treated patients (19% positive).
 The result: The PSA response (defined as > 50% PSA decline from baseline) was

  • 0% v 58% for AR-V7 positive v negative men in the enzalutamide group, and
  • 0% v 68% in the abiraterone cohort, respectively.
  • Poorer outcomes by various measures reflected these differences.

In JAMA Oncol, Aug 2015, the Hopkins group reported on men with mCRPC starting taxane therapy and concluded the presence of AR-V7 was “not associated with primary resistance to taxane therapy.”

  • “Of 37 taxane-treated patients enrolled 17 (46%) had detectable AR-V7 in CTCs.” The PSA response to taxane therapy in AR-V7-positive vs AR-V7-negative men was 41% vs 65%; P=.19, i.e., not considered statistically significant.
  • This 41% PSA response in AR-V7 positive patients is to be compared to the 0% PSA response to Zytiga or Xtandi in the men reported by the Hopkin’s group in the 2014 article (see above). However, “… in AR-V7 negative men, taxanes and enzautamide and abiraterone may have comparable efficacy.”

The original Hopkin’s study group of 62 was expanded to 202 men with mCRPC and reported on at the 2016 ASCO Annual Meeting in Abstract 5012.

  • They subcategorized the PSA response into three groups: men in whom CTCs were not identified; CTC+ AR-V7 negative men; and CTC+ AR-V7 positive men.
  • PSA responses were 75%, 52%, and 14%, respectively.
  • When either enzalutamide or abiraraterone were used first-line, those who were CTC+ AR-V7+ had a 27% PSA response; whereas when either drug was used secondarily the response was 5%.
  • Their finding showed, as have many other studies, that identifying CTCs (usually expressed as > or < than 5 CTC/7.5cc blood in the CellSearch system) conferred a poorer prognosis independent of AR-V7 status. In the Hopkins study. “CTC+AR-V7+ pts were more likely to have Gleason >8, M1 [metastatic] disease at diagnosis, higher PSA, prior Abi/Enza use, prior taxane [Docetaxel chemotherapy] use” and a poorer performance status.

In Abstract 183 (ASCO 2017, GU Cancer Symposium) by the Hopkins group their premise was “The AR-V7 splice variants may confer resistance to AR-directed therapies but not taxane chemotherapies” [Docetaxel and Carbazitaxel].

  • They queried 100 clinicians who had submitted blood for AR-V7 analysis from men with mCRPC and provided patient histories.
  • The data confirmed that the percentage of AR-V7 increases through sequential therapy.
    • Patients who had had neither prior Zytiga nor Xtandi were positive in 29%;
    • Those who had either of those drugs were positive 39%, and
    • If both drugs had been used, 62% were AR-V7 positive.
    • Finding: Clinicians most often chose taxane therapy in patients found to be AR-V7 positive.

Studies on this clinically important issue are ongoing. However, the observation of the adverse effect of response to hormone therapy in the circulating tumor cells of AR-V7 positive men was further confirmed by Scher et al. from Sloan-Kettering, JAMA Oncology, June 2016: “The results validate CTC nuclear expression of AR-V7 protein in men with mCRPC as a treatment-specific biomarker that is associated with superior survival on taxane therapy over ARS [androgen receptor signaling] – directed therapy in a clinical practice setting.”

Of interest, the analytic method used for the Scher study was different from the Hopkin’s technique. The Scher assay method was based on a rabbit monoclonal antibody staining of nuclear AR-V7 protein. This test will become commercial available later this year from EPIC/Genomic Health, Inc. It is easily conceivable that the two techniques measure slightly different forms of AR-V7 and might give different percentages of AR-V7 when tested on the same population. It is essential to know more about this. Dr. Antonarakis (personal communication) regards both tests as “extremely robust” in AR-V7 detection. Dr. Antonarakis indicated that a prospective head-to-head study is ongoing comparing the two methods in 120 patients who will each have both tests. Trial results are expected in 6 – 9 months.

So … Is assessing a man’s AR-V7 status test ready to bear the weight of clinical decisions?

… Maybe, … depending on the clinical situation. Clearly the application of this important test requires, and will get, additional validation. Validation studies will provide information about the prevalence of AR-V7 in different clinical situations and its effect on treatment outcome. This will allow physicians to better guide men as to their best treatment options.

However, on the basis of what is currently known, two scenarios can be imagined. A management decision might be based on the findings of the Abstract 5012 cited earlier. Envision a man with a history of Gleason 8 cancer who, at a later period in his disease during ADT, exhibited early mCRPC with a rising PSA and a doubling time of 9 months. If his blood showed circulating tumor cells (an adverse prognostic sign by itself) and he tested AR-V7 positive, his likelihood of benefiting from Zytiga or Xtandi would be low. The choice of taxane chemotherapy could be considered as a more likely pathway to response. On the other hand, if no CTCs were found, he has a reasonable chance of responding to the addition of Zytiga or Xtandi. Of course, there will be situations in the murky middle-ground where guidance from additional studies will be helpful and clinical judgment will be required.

These are the types of difficult decisions that patients and physicians are likely to face in the near future.


Fully validated information about the clinical significance of CTCs positive for AR-V7 will become a valuable tool to guide prostate cancer management. The guidance in personalized treatment decisions available with AR-V7 testing, combined with the high-level accuracy in staging now possible with the PSMA based PET/CT total body scanning, and additionally with the emergence of potential treatment with 177-Lutetium PSMA targeted radioisotope therapy, taken all together, will have a marked positive effect on the management of prostate cancer.

For a discussion of PSMA PET/CT scanning see PCa Commentary #109 and for an introduction to PSMA targeted 177-Lutetium radioisotope therapy see PCa Commentary #110.