Nuclear-specific AR-V7 Protein Localization is Necessary to Guide Treatment Selection in Metastatic Castration-resistant Prostate Cancer

Abstract

Background: Circulating tumor cells (CTCs) expressing AR-V7 protein localized to the nucleus (nuclear-specific) identify metastatic castration-resistant prostate cancer (mCRPC) patients with improved overall survival (OS) on taxane therapy relative to the androgen receptor signaling inhibitors (ARSi) abiraterone acetate, enzalutamide, and apalutamide.

Objective: To evaluate if expanding the positivity criteria to include both nuclear and cytoplasmic AR-V7 localization ("nuclear-agnostic") identifies more patients who would benefit from a taxane over an ARSi.

Design, setting, and participants: The study used a cross-sectional cohort. Between December 2012 and March 2015, 193 pretherapy blood samples, 191 of which were evaluable, were collected and processed from 161 unique mCRPC patients before starting a new line of systemic therapy for disease progression at the Memorial Sloan Kettering Cancer Center. The association between two AR-V7 scoring criteria, post-therapy prostate-specific antigen (PSA) change (PTPC) and OS following ARSi or taxane treatment, was explored. One criterion required nuclear-specific AR-V7 localization, and the other required an AR-V7 signal but was agnostic to protein localization in CTCs.

Outcome measurements and statistical analyses: Correlation of AR-V7 status to PTPC and OS was investigated. Relationships with survival were analyzed using multivariable Cox regression and log-rank analyses.

Results and limitations: A total of 34 (18%) samples were AR-V7-positive using nuclear-specific criteria, and 56 (29%) were AR-V7-positive using nuclear-agnostic criteria. Following ARSi treatment, none of the 16 nuclear-specific AR-V7-positive samples and six of the 32 (19%) nuclear-agnostic AR-V7-positive samples had ≥50% PTPC at 12 weeks. The strongest baseline factor influencing OS was the interaction between the presence of nuclear-specific AR-V7-positive CTCs and treatment with a taxane (hazard ratio 0.24, 95% confidence interval 0.078-0.79; p=0.019). This interaction was not significant when nuclear-agnostic criteria were used.

Conclusions: To reliably inform treatment selection using an AR-V7 protein biomarker in CTCs, nuclear-specific localization is required.

Patient summary: We analyzed outcomes for patients with metastatic castration-resistant prostate cancer on androgen receptor signaling inhibitors and standard chemotherapy. Patients with circulating tumor cells that had AR-V7 protein in the cellular nuclei were very likely to survive longer on taxane-based chemotherapy, and tests unable to distinguish where the protein is located in the cell are not as predictive of benefit.

Keywords: AR-V7; Circulating tumor cells; Liquid biopsy; Prostate cancer.

Fig. 1

AR-V7 protein can be localized to the nucleus and/or cytoplasm of circulating tumor cells (CTCs). Example images show (A) individual CTCs with AR-V7 nuclear-specific localization, (B) CTC with predominantly cytoplasmic AR-V7 staining and (C) CTC cluster with varying AR-V7 expression. (D) Sample-level positivity for these CTCs. DAPI = 4′,6-diamidino-2-phenylindole; CK = cytokeratin.

Fig. 2

Nuclear-specific AR-V7 localization is required for specificity of PTPC for patients on ARSi therapy. Waterfall plots of the percentage change in PSA at 12 wk on therapy, stratified by the number of previous lines of therapy. Each bar represents an individual patient. (A) Pre-ARSi AR-V7 status according to nuclear-specific localization only. (B) Pre-ARSi AR-V7 status according to nuclear-agnostic localization. (C) Pre-taxane AR-V7 status according to nuclear-specific localization only. (D) Pre-taxane AR-V7 status according to nuclear-agnostic localization. PSA = prostate-specific antigen; PTPC = post-therapy PSA change; ARSi = androgen receptor signaling inhibitor. * Longer than 12 wk (see the text).

Fig. 3

Nuclear-specific AR-V7 localization improves prognostication of overall survival (OS) for patients on ARSi therapy. OS is shown for patient samples stratified by pre-ARSi AR-V7 status (n = 128) determined according to (A) nuclear-specific localization and (B) nuclear-agnostic localization and stratified by pre-taxane AR-V7 status (n = 63) determined according to (C) nuclear-specific localization and (D) nuclear-agnostic localization. ARSi = androgen receptor signaling inhibitor.

Fig. 4

Nuclear-specific AR-V7 localization is necessary for prediction of a treatment-specific reduction in risk of death. Individual covariates were tested for additive power in predicting outcome using a Cox proportional hazards model. The p values are the result of compensating for the other factors listed. The interaction of therapy and AR-V7 status was further investigated using a multivariable Cox proportional hazards model. The forest plot shows the hazard ratio and 95% confidence interval (CI) for (A) a cohort evaluated for AR-V7 positivity utilizing nuclear-specific localization and (B) a cohort evaluated for positivity utilizing nuclear-agnostic localization. ARSi = androgen receptor signaling inhibitor; PSA = prostate-specific antigen; phos. = phosphatase; LDH = lactate dehydrogenase.