Meta description: AR-V7 in circulating tumor cells predicts resistance to AR-signaling inhibitors in mCRPC. A liquid biopsy guide to the validating cohorts, assays, and limits.

AR-V7 in Circulating Tumor Cells: A Predictive Biomarker of AR-Signaling Inhibitor Resistance in mCRPC

Many men with metastatic castration-resistant prostate cancer (mCRPC) who start an androgen-receptor (AR)–signaling inhibitor — enzalutamide or abiraterone — will not respond, and a liquid biopsy marker called AR-V7 helps explain why. AR-V7 is a splice variant of the androgen receptor that lacks the ligand-binding domain — the very part of the receptor these drugs target. Because it's missing that domain, the truncated receptor is constitutively active: it drives AR signaling without needing androgen and without a pocket for the drug to bind. State the mechanism plainly and the clinical hypothesis follows naturally — tumors carrying AR-V7 should be poor candidates for drugs designed to shut off ligand-dependent AR signaling.

The clinical question

In mCRPC, enzalutamide and abiraterone are mainstays, but resistance is common and frequently primary rather than acquired. The practical problem for the clinic is sorting, before or early in treatment, which patients are unlikely to benefit from another AR-targeted agent. AR-V7 was proposed as a blood-based answer to that question: a predictive biomarker that could identify a resistant phenotype non-invasively, using a tube of blood instead of a metastatic biopsy.

The analyte and the assay

The analyte lives in circulating tumor cells (CTCs) — intact tumor cells shed into the bloodstream. AR-V7 has been measured two ways. The original work detected AR-V7 messenger RNA from CTCs enriched from whole blood using reverse-transcription PCR [1]. A later analytically validated platform detected AR-V7 protein localized to the nucleus of individual CTCs by an immunofluorescence assay, scoring a cell as positive only when the protein was present in the nucleus — a design meant to capture functionally active receptor rather than incidental expression [2].

What a positive result establishes, and what it doesn't, matters. A positive AR-V7 result identifies a subgroup with a lower probability of responding to AR-signaling inhibitors; it's a probabilistic predictive signal, not a deterministic one. It also can't say anything when no CTCs are detected — the assay requires tumor cells in the sample to interrogate. A negative result in a CTC-poor specimen is uninformative, not reassuring.

The evidence

The founding study was a prospective observational cohort of 62 men with mCRPC beginning either enzalutamide (31 patients) or abiraterone (31 patients), with AR-V7 mRNA measured in CTCs [1]. The association was striking. Among enzalutamide-treated men, none of the AR-V7–positive patients (0 of 12) achieved a PSA response, versus 53% of AR-V7–negative patients. In the abiraterone cohort, no AR-V7–positive patient (0 of 6) had a PSA response, versus 68% of negative patients. AR-V7 positivity also tracked with shorter PSA progression-free survival and shorter overall survival. The signal was strong, but the numbers of positive patients were small — a dozen here, a handful there — which is exactly why the field didn't stop at this report.

A pivotal refinement came from a study using the nuclear-localized AR-V7 protein immunofluorescence assay across 161 men with mCRPC, examined in the context of both AR-signaling inhibitors and taxane chemotherapy [2]. This is where the biomarker earned its keep. The finding that genuinely reshaped how AR-V7 should be interpreted was its treatment specificity: nuclear AR-V7–positive men had better overall survival on taxane chemotherapy than on AR-signaling inhibitors, whereas AR-V7–negative men did better on AR-signaling inhibitors [2]. In other words, AR-V7 doesn't merely flag a worse prognosis — it distinguishes between two therapeutic classes, which is the difference between a marker that tells you a patient is sick and one that helps characterize which drug class the biology is likely to resist.

Prospective, blinded validation followed in the PROPHECY study, a multicenter trial of 118 men with high-risk mCRPC starting abiraterone or enzalutamide, with AR-V7 assessed by two independent assays — the CTC mRNA test and the nuclear protein immunofluorescence test [3]. Baseline AR-V7 detection was associated with shorter progression-free and overall survival after adjustment for standard prognostic variables. For the protein (Epic) assay, the adjusted hazard ratio for overall survival was 3.14 (95% CI, 1.60–6.14), and for the mRNA (modified Johns Hopkins) assay the adjusted overall-survival hazard ratio was 1.90 (95% CI, 1.11–3.24) [3]. Specificity for predicting failure to achieve a PSA response was high across assays, while sensitivity was comparatively low — meaning a positive result is informative, but a negative result does not guarantee response. Agreement between the two assays was modest, a point with direct practical consequences.

Where it stands

The evidence supports AR-V7 in CTCs as a predictive biomarker of resistance to AR-signaling inhibitors in mCRPC, now backed by prospective, blinded, multi-assay validation [1,2,3]. Several caveats bound its use:

• Assay dependence. The mRNA and protein assays are not interchangeable; they showed only modest concordance and produced different hazard ratios in head-to-head prospective testing [3]. A result is meaningful only in reference to the specific platform that generated it.
• CTC dependence. The test interrogates tumor cells captured from blood. In patients with few or no detectable CTCs, AR-V7 status may be indeterminate rather than negative — an important distinction that limits applicability in lower–disease-burden states [1,2,3].
• Low sensitivity. Because sensitivity for predicting non-response is limited, AR-V7 negativity does not promise benefit from AR-signaling inhibitors [3].
• Scope of the evidence. The cited cohorts were modest in size and focused on abiraterone and enzalutamide; broad generalization beyond those settings isn't established by these data [1,2,3]. Interpretation belongs within the full clinical picture, and this article is educational — it is not guidance for any individual's care.

The most useful forward-looking implication is methodological: AR-V7 showed that a single-cell, protein-localization liquid biopsy readout could do something a prognostic marker can't — point toward one treatment class over another — and that's the standard the next generation of CTC-based predictive assays will be measured against.

References

1. Antonarakis ES, et al. AR-V7 and resistance to enzalutamide and abiraterone in prostate cancer. N Engl J Med. 2014. PMID 25184630.
2. Scher HI, et al. Association of AR-V7 on circulating tumor cells as a treatment-specific biomarker with outcomes and survival in castration-resistant prostate cancer. JAMA Oncol. 2016. PMID 27262168; doi:10.1001/jamaoncol.2016.1828.
3. Armstrong AJ, et al. Prospective multicenter validation of androgen receptor splice variant 7 and hormone therapy resistance in high-risk castration-resistant prostate cancer: the PROPHECY study. J Clin Oncol. 2019. PMID 30865549; doi:10.1200/JCO.18.01915.