Dual Blockade of c-MET and the Androgen Receptor in Metastatic Castration-resistant Prostate Cancer: A Phase I Study of Concurrent Enzalutamide and Crizotinib

Agents targeting androgen receptor (AR) signaling continue to be the cornerstone of therapy for patients with castration-resistant prostate cancer (CRPC). Enzalutamide is a potent AR antagonist which blocks the binding of the AR to testosterone and inhibits its nuclear translocation and subsequent DNA binding and coactivator recruitment (1). It significantly improves radiographic progression-free survival (PFS) and overall survival (OS) in both chemotherapy-naïve and post-docetaxel CRPC, as well as in the metastatic hormone-sensitive setting when given with initial androgen-deprivation therapy (ADT; refs. 2–4). As such, enzalutamide is now frequently administered in either the hormone-sensitive or castration-resistant settings.

Although novel AR inhibitors, such as enzalutamide, have significantly improved outcomes, most patients eventually progress due to the emergence of castration resistance and androgen insensitivity through various mechanisms, one of which may be c-MET activation (5). c-MET is a tyrosine kinase receptor encoded by the MET protooncogene, which is activated in a paracrine manner by its ligand hepatocyte growth factor/scatter factor. Increased c-MET signaling promotes tumor cell migration, invasion, proliferation, and angiogenesis through increased focal adhesion kinase, Ras/Raf/MEK/ERK, and PI3 kinase/Akt signaling (6–11). Preclinical data suggest that c-MET expression is inversely correlated with AR activation. Androgen-insensitive, AR-negative prostate cancer cell lines demonstrated significantly higher c-MET expression compared with AR-positive cell lines (7). Similarly, castration induced c-MET expression in murine models (12). Verras and colleagues provided mechanistic evidence for this inverse relationship by demonstrating that AR activation inhibited c-MET expression in a promoter- and ligand-dependent manner and that AR inhibition was associated with an adaptive increase in c-MET expression (10). In turn, increased c-Met signaling can inhibit AR expression, which further potentiates androgen insensitivity and resistance to AR-targeted therapy (13). In preclinical models, combined treatment with an AR antagonist and a c-MET inhibitor was more effective in inhibiting tumor growth than either agent alone (14, 15). These findings highlight the potential role of increased c-MET expression as a mechanism of resistance to AR inhibitor, enzalutamide. These data informed our hypothesis that treatment with a potent c-Met inhibitor, crizotinib, in combination with enzalutamide could target this adaptive increase in c-MET signaling and delay disease progression.

To test this hypothesis, characterizing the pharmacokinetics profile of concurrent enzalutamide and crizotinib administration and identifying the recommended phase II dose (RP2D) of crizotinib was critical. Hepatic metabolism mediated by hepatic CYP3A4 is an important route of elimination for crizotinib (16–18). Enzalutamide is a strong inducer of CYP3A4, which was recognized as having the potential to increase crizotinib clearance (19). However, the degree to which it could affect clearance of agents metabolized by CYP3A4 (e.g., crizotinib, docetaxel, and cabazitaxel) is unknown. Subsequent pharmacokinetics analyses with respect to enzalutamide drug–drug interactions have also shown that crizotinib-induced alterations in the pharmacokinetics behavior of enzalutamide would not be anticipated as CYP2C8, the major enzyme responsible for its hepatic metabolism, is not subjected to significant inhibition by crizotinib (19). Thus, given the significant potential for a drug–drug interaction between enzalutamide and crizotinib, we conducted a phase I study investigating the safety, tolerability, pharmacokinetics, and preliminary antitumor efficacy of concurrent enzalutamide and crizotinib in patients with metastatic CRPC.

Eligible patients were 18 years of age or older, had histologically confirmed prostate adenocarcinoma, Eastern Cooperative Oncology Group (ECOG) performance status <2, radiographic evidence of metastatic disease, serum testosterone levels <50 ng/dL, and evidence of disease progression based on rising PSA per the Prostate Cancer Working Group 2 (PCWG2) criteria or radiographic progression according to the RECIST 1.1 criteria (20, 21). There was no limit on the number of prior lines of therapy, and patients could have received docetaxel, ketoconazole, estrogens, abiraterone, or other antiandrogens, including prior enzalutamide, for CRPC. Patients with a small cell component on pathologic evaluation, brain metastases, or prior history of seizures or any medical or neurologic conditions that may predispose to seizures were excluded from the study. Patients were maintained on gonadotropin-releasing hormone agonist or antagonist therapy unless prior orchiectomy or intolerant. Concomitant use of strong CYP3A or CYP2C8 inhibitors, strong or moderate CYP3A inducers, or CYP2C8, CYP3A4, CYP2C9, and CYP2C19 substrates with narrow therapeutic indices was prohibited. The Dana-Farber Cancer Institute Institutional Review Board (Boston, MA) approved the protocol and informed written consent was obtained from all participants prior to enrollment in the study. The study was conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonisation of Good Clinical Practice Guidelines (NCT02207504).

A standard 3+3 dose-escalation design was used to determine the MTD of crizotinib in combination with enzalutamide. Consecutive patients were enrolled on increasing dose levels of crizotinib (dose level 1, 250 mg daily; dose level 2, 200 mg twice daily; and dose level 3, 250 mg twice daily; Supplementary Fig. S1) in combination with a fixed standard dose of enzalutamide (160 mg daily). The primary objective of the study was to establish a RP2D based on safety, tolerability, and pharmacokinetics analysis. The primary endpoint of the study was dose-limiting toxicity (DLT) of crizotinib in the first 28 days on crizotinib and enzalutamide. Secondary endpoints included the pharmacokinetics profile of both drugs in combination, incidence of adverse events (AEs) and laboratory abnormalities, and tolerability of the combination. Tolerability was defined as toxicity that results in study drug discontinuation or dose reduction that would not have been mandated by the protocol such as a DLT. The exploratory endpoints that assessed antitumor efficacy of the combination included PFS (defined as time from initiation of treatment to systemic or radiographic progression or death or censured at last follow-up), time to radiographic progression (rTTP), and time to progression including radiographic or PSA progression (TTP) and PSA response rate (defined as PSA decline from baseline to week 12). In addition, we investigated changes in markers of bone turnover such as serum bone-specific alkaline phosphatase (BALP) and serum C‐terminal telopeptide (CTX) from baseline to on-therapy with enzalutamide and crizotinib.

Toxicities were graded according to the Common Terminology Criteria for Adverse Events (CTCAE) version 4.0 (22). DLTs were assessed on the basis of commonly used prespecified criteria (Supplementary Table S1) and assessed during the first cycle (28 days) of therapy. The crizotinib dose level with DLTs in fewer than two of six patients was considered the MTD and was expanded such that a minimum of 12 patients were treated at the MTD, and a total of 24 patients would be enrolled in the whole study.

For exploratory antitumor efficacy endpoints, PSA was measured during each cycle (every 28 days), and imaging of the chest, abdomen, pelvis, and bone was performed every 12 weeks. Treatment was continued until evidence of confirmed radiographic progression per RECIST 1.1 or PCWG2 criteria, intolerable AEs, or withdrawal of patient consent. For the exploratory biomarker analysis of changes in CTX and BALP, blood samples were collected at baseline (cycle 1 day 1), on therapy (day 1 of cycles 2 and 4), and at the end of study visit.

The sampling schedule was designed to define the plasma concentration–time profiles of both drugs over a single dosing interval for the initial dose given on day 1 of cycle 1 and the dose given on day 1 of cycle 2. It was assumed that steady-state pharmacokinetics for continuous once- or twice-daily administration was achieved prior to the beginning of the second cycle of therapy for both drugs based upon previously reported investigations of their clinical pharmacokinetics (23, 24). Blood samples (4 mL) were collected in Vacutainer tubes with spray-coated K₂EDTA before dosing, at 0.5, 1.0, 2.0, 4.0, 6.0, and 8.0 hours after dosing on day 1, and before dosing on day 2 of cycles 1 and 2. Samples were also obtained before dosing on day 15 of cycles 1 and 2 and day 1 of cycle 3. Blood collection tubes were centrifuged to harvest the plasma, which was stored in cryovials at −80°C. The concentrations of crizotinib and enzalutamide in the plasma samples were concurrently determined by reversed-phase high-performance liquid chromatography with tandem mass spectrometric detection. The analytic method was adapted from methods that had been reported previously for the independent determination of each drug in human plasma (25, 26). The analytic method was validated and applied to the analysis of study samples as recommended in the FDA Guidance for Industry, Bioanalytical Method Validation, May 2018 (https://www.fda.gov/media/70858/download). At the lowest concentration included in the calibration curves, which was 1.0 ng/mL for crizotinib and 10.0 ng/mL for enzalutamide, interday accuracy was within 2% of the nominal concentration and the precision was ≤3.5% for both analytes. Interday accuracy ranged from 97.5% to 102.7% and the precision ranged from 4% to 8.1% for all other calibration standards for both analytes.

The plasma concentration–time data were analyzed by noncompartmental methods using model 200 for extravascular drug input in WinNonlin Professional version 5.0.1 (Pharsight Corp). The minimum (trough) steady-state concentration of each drug in plasma (C_min^ss) was calculated as the geometric mean of its assayed concentration in all samples collected just prior to dosing in cycles 2 and 3. Area under the plasma concentration–time curve from the time of dosing to the time of last sample collected prior to administration of the next dose (AUC_τ) was estimated using the linear-log trapezoidal method. For patients receiving crizotinib twice daily, the dosing interval was assumed to be 12 hours and the crizotinib concentration in the predose sample collected on day 2 was assumed to be equivalent to the concentration at the end of the dosing interval. Total oral clearance (CL/F) was calculated as the dose divided by AUC_τ^ss. The accumulation factor for repeated dosing (X) was calculated by dividing AUC_τ for the dose given in cycle 2 by that for the dose on day 1 of cycle 1. Time values and the accumulation factor is reported as the arithmetic average ± SD. Geometric means were calculated for all other pharmacokinetics variables (27). The jackknife technique was used to estimate the SD of geometric means (28).

Patients were evaluable for the DLT assessment if they received >85% of the planned dosing and completed the assessment period. The number of patients experiencing a DLT at each dose level in the first 28 days of study (cycle 1) was summarized with the method of Atkinson and Brown (29), which was used to construct a two‐stage exact binomial confidence interval (CI). Toxicities were summarized as number and percent of patients, overall, and according to maximum grade experienced. The time to event distributions for PFS, rTTP, and TTP were estimated using the Kaplan–Meier method. PSA response rates were summarized descriptively. The levels of bone turnover biomarkers (BALP and CTX) at each timepoint were summarized as median, interquartile range (IQR), and range (minimum and maximum). The differences in levels between timepoints were analyzed using the Wilcoxon signed-rank test.

Between December 2014 and January 2016, 24 patients were enrolled in the study, which included 16 patients treated in the dose escalation phase (250 mg daily: n = 3; 200 mg twice daily: n = 6; and 250 mg twice daily: n = 7). Seven patients were treated in the 250-mg twice-a-day cohort, because one patient had inadequate dosing to confirm the absence of DLT. An additional eight patients were enrolled in the dose expansion phase at the MTD. In the overall cohort, the median age at enrollment was 70 years (range, 53–80 years), and 46% (n = 11) of patients had Gleason ≥8 disease (Table 1). Most common sites of metastases were bone (88%; n = 21) and lymph nodes (54%; n = 13), while 29% (n = 7) of patients had visceral metastases. Most patients (54%; n = 13) had received prior systemic chemotherapy, with the median duration of chemotherapy being 4.1 months (range, 2.0–11.7 months). Seventy percent (n = 17) had progressed on prior abiraterone, of whom seven had also received prior enzalutamide. Three patients were naïve to docetaxel, abiraterone, and enzalutamide, and five patients had received all three agents prior to study.

In the dose-escalation cohort, two patients (one at 200 mg twice a day and one at 250 mg twice a day) experienced a DLT of grade 3 alanine aminotransferase (ALT) elevation. On the basis of the DLT assessment, the MTD of crizotinib was the full standard dose of 250 mg twice daily, which was the dose level selected for the subsequent expansion cohort (n = 8), where two DLTs were observed, resulting in an overall DLT rate of 16.7% with a two-stage exact binomial 90% CI, 7%–36%. Overall, the most common treatment-related AEs (TRAE) of any grade (Table 2) were fatigue (50%; n = 12), transaminitis (38%, n = 9), nausea (33%; n = 8), and vomiting, constipation, and diarrhea (21%; n = 5 each). Grade ≥3 TRAEs occurred in 25% of patients (n = 6). Grade 3 TRAEs included increased ALT (n = 2), fatigue (n = 1), and hypertension (n = 1). Grade 4 TRAEs (n = 2) included one cardiac event (QT prolongation, ventricular fibrillation, and cardiac arrest all in the same patient), as well as one deep vein thrombosis/pulmonary embolism in one patient. No grade 5 events occurred. Dose modifications were required in 54% (n = 13) of patients; 11 patients required dose holds, while two patients experienced both dose reductions and dose holds.

The median duration of study treatment was 3 months (IQR, 2–6 months). The most common reasons for treatment discontinuation were disease progression (n = 17), withdrawal of consent (n = 2), lack of apparent clinical benefit per treating investigator (n = 1), and new cardiac comorbidity requiring administration of a cardiac drug that had the potential for study drug interactions (n = 1). Treatment was discontinued because of intolerable grade 1 fatigue and intolerable grade 1 vomiting in one patient each treated at the MTD. At the time of data cutoff (July 2018), one patient, who was enzalutamide and docetaxel naïve at study entry, remained on treatment with time on study of 36 months (Fig. 1).

PSA response (≥50% reduction) was observed in 38% (n = 9) of patients at any crizotinib dose level and in 33% (n = 5) of patients in the MTD cohort, while 75% (n = 18) of patients experienced any degree of PSA reduction (Fig. 2). PSA response rate was 71% (n = 5/7) and 24% (n = 4/17) among the patients naïve to and refractory to abiraterone and/or enzalutamide, respectively. Objective response rate per RECIST 1.1 was 12% (n = 3) in the overall cohort, with a median PFS of 5.5 (95% CI, 2.8–21.2) months (Table 3). Among patients treated with prior abiraterone and/or enzalutamide (n = 17), the median PFS was 2.8 months (95% CI, 1.9–27.6). Enzalutamide-experienced patients (n = 7) had poor outcomes with a PSA response rate of 14% and a median PFS of 2.1 months (95% CI, 1.7–5.6). At the time of data cutoff, all patients were alive.

Mean values of the steady-state pharmacokinetics parameters for crizotinib and enzalutamide at each dose level are summarized in Table 4. The mean C_min^ss of crizotinib increased progressively from 20.7 ± 1.8 ng/mL for patients in dose level 1 (250 mg daily) to 50.3 ± 31.1 ng/mL in dose level 2 (200 mg twice daily) and 70.2 ± 36.4 ng/mL in dose level 3 (250 mg twice daily). For historical comparison, the mean C_min^ss of crizotinib was reported as 303.6 ng/mL (n = 44) and 285.0 ng/mL (n = 331) in two clinical trials in which the drug was given alone at a dose of 250 mg twice daily to patients with advanced non–small cell lung cancer with normal renal function (30). The mean (± SD) AUC_τ^ss for crizotinib 250 mg twice a day given in combination with enzalutamide 160 mg daily in the 12 patients with evaluable pharmacokinetics data in dose level 3 was 1,000 ± 476 ng•h/mL. The population pharmacokinetics estimate of the geometric mean AUC_τ^ss for crizotinib was 3,817 ng•h/mL when given as a single agent at the same dose and schedule to patients with cancer with normal renal function (30). The administration of crizotinib in combination with enzalutamide in our study resulted in an approximately 74% decrease in systemic exposure to crizotinib. In contrast, crizotinib did not appear to affect the plasma pharmacokinetics of enzalutamide. Its mean apparent oral clearance was similar at each dose level when combined with crizotinib and in very good agreement with the 0.61 ± 0.20 L/hour mean apparent oral clearance of single-agent enzalutamide in 93 patients with prostate cancer (31).

Median serum BALP and CTX levels at baseline were 18.8 μg/mL (IQR, 12.8–26.9) and 1.8 ng/mL (IQR, 0.6–2.8), respectively. There was no significant change in serum BALP and CTX levels from baseline (cycle 1 day 1) to on-therapy (day 1 of cycles 2 and 4) or at the end of the study (Supplementary Fig. S2).

We report the results of the first study to explore concurrent c-MET and AR inhibition with crizotinib and enzalutamide in patients with metastatic CRPC. We observed a clinically significant pharmacokinetics drug–drug interaction resulting in a 74% decrease in steady-state crizotinib levels when administered concurrently with enzalutamide. The toxicity spectrum of the combination of full doses of enzalutamide and crizotinib was consistent with their known toxicities when administered as monotherapies. No RP2D was endorsed for crizotinib in combination with standard-dose enzalutamide given the significant drug–drug interaction, lack of apparent efficacy above that expected with enzalutamide alone, and the potential for worsening gastrointestinal and liver toxicities if higher doses of crizotinib were explored.

The current treatment of metastatic CRPC relies on the sequential use of agents targeting AR signaling, such as enzalutamide and abiraterone, or taxane-based chemotherapies (32). Recent studies have demonstrated a significant survival benefit with the addition of AR-targeted agents, such as abiraterone acetate, enzalutamide, or apalutamide, to ADT in the hormone-sensitive setting (4, 33, 34). Despite the survival improvement seen with earlier utilization of these agents, almost all patients eventually progress and develop castration-resistant disease. The relatively low response rates and PFS with sequential use of these agents suggest the emergence of shared mechanisms of cross-resistance (35).

Increased c-MET signaling might be one such mechanism of adaptive resistance. Several prior preclinical studies have highlighted the role of c-MET signaling in prostate cancer progression. However, the VEGF and MET inhibitor, cabozantinib, previously failed to improve OS in CRPC, albeit in a heavily pretreated patient population, the majority of whom (>90%) had progressed on ≥3 prior lines of therapy, including docetaxel and abiraterone acetate or enzalutamide (36). A possible explanation for the lack of survival benefit to c-Met inhibition monotherapy is that a significant proportion of patients with CRPC continue to have intact AR signaling, which can downregulate c-MET expression (37). Indeed, preclinical studies indicate that concurrent, stringent AR inhibition might be necessary for the upregulation of c-MET and sensitivity to c-MET inhibitors (14). To investigate the role of concurrent c-MET and AR inhibition, we conducted this trial where patients were treated with both a potent c-MET inhibitor, crizotinib, and an effective second-generation AR inhibitor, enzalutamide.

The spectrum of observed TRAEs was consistent with that observed with enzalutamide and crizotinib monotherapy in prior clinical trials. In the phase III AFFIRM trial investigating enzalutamide in the post-docetaxel setting, the most frequent AEs were fatigue (34%), diarrhea (21%), and arthralgias (21%) with significant cardiac events observed in 6% (3). In the phase III study investigating crizotinib in treatment-naïve, anaplastic lymphoma kinase-ALK-positive non–small cell lung cancer, the most common AEs associated with crizotinib monotherapy were diarrhea (61%), vomiting (46%), constipation (43%), and increased ALT/aspartate aminotransferase (AST, 36%; ref. 38). Cardiac events have been observed in 5% of patients on crizotinib monotherapy (39). In comparison, the most common TRAEs in our study included fatigue (50%), transaminitis (38%), nausea (33%), constipation (21%), and diarrhea (21%). There was no correlation between toxicity and the pharmacokinetics profile as evidenced by the subtherapeutic C_min^ss (19.2–69.8 ng/mL) and AUC^ss (453.3–1,137.4 ng•h/mL) even in patients who experienced a grade ≥3 TRAE. This finding is consistent with the prior phase I dose-escalation study evaluating crizotinib doses ranging from 50 to 300 mg twice a day where grade 3 ALT elevations were observed even at lower dose levels (200 mg daily) and gastrointestinal AEs (nausea and vomiting) occurred irrespective of dose levels (40). The higher rate of fatigue observed with the combination is likely due to subadditive effects of enzalutamide and crizotinib plus inclusion of a heavily pretreated and advanced population of patients with progressing CRPC (3, 38). One patient in our study experienced a significant cardiac event comprised of QT prolongation, ventricular fibrillation, and cardiac arrest. This rate of significant cardiac events is similar to prior published results with either enzalutamide or crizotinib monotherapy at 5%–6% (3, 38, 39). Although the majority required dose modification because of AEs, only two patients (8%, both at the MTD) discontinued treatment because of toxicity (intolerable grade 1 fatigue and intolerable grade 1 vomiting); a similar treatment discontinuation rate was reported in prior enzalutamide (8%) and crizotinib (5%) monotherapy trials (3, 38).

Although enzalutamide's effect on CYP3A4 metabolism was known at the time of design of the study (19), the extent of its impact on crizotinib pharmacokinetics was unknown. Hence, investigating their pharmacokinetics drug–drug interactions was an important objective of this clinical trial. Interestingly, in subsequent studies, enzalutamide was shown to have similar effects on the plasma pharmacokinetics of docetaxel and cabazitaxel, although the magnitude of the effect was considerably less than we observed with crizotinib (12% and 22% decrease in docetaxel and cabazitaxel exposure, respectively; refs. 41, 42). Given the relatively smaller pharmacokinetics interaction seen with cabazitaxel, a phase I/II trial of enzalutamide in combination with cabazitaxel is currently ongoing (NCT02522715). In contrast to these studies, we observed a significantly higher apparent oral clearance of crizotinib at steady state than would be expected from the prior clinical trial of crizotinib monotherapy in patients with advanced non–small cell lung cancer (30). Concomitant use of agents that are strong inhibitors of CYP3A4/5 and CYP2C8 and moderate to strong inducers of CYP3A4/5 activity were prohibited in this study. Thus, the observed decrease in the systemic exposure of crizotinib across dose levels is consistent with enzalutamide-mediated induction of hepatic CYP3A4 activity. Given the significant decrease in crizotinib exposure seen in our study, further investigation of this combination at the approved doses is not planned. It is possible that further dose escalation of crizotinib beyond 250 mg twice a day could achieve therapeutic concentrations. However, given the incidence of grade ≥3 TRAEs even with the subtherapeutic crizotinib concentrations seen in our study, the toxicity would likely be prohibitive. Our study highlights the importance of evaluating the potential for pharmacokinetics drug interactions in the development of novel combination therapeutic strategies.

Our study was conducted in a heavily pretreated patient population of whom 54% had received prior chemotherapy and 71% prior abiraterone and/or enzalutamide. Nearly 30% had visceral metastases. Enzalutamide-refractory patients were included in this trial based on the hypothesis that c-MET upregulation with enzalutamide monotherapy might be a potential resistance mechanism, which could be targeted and potentially reversed with concurrent treatment with crizotinib. This inclusion decreased the likelihood of observing higher PSA or objective responses. Indeed, PSA response rates were significantly higher in patients naïve to abiraterone/enzalutamide at 71% compared with 24% in patients with prior exposure. Objective responses were only seen in the enzalutamide- and docetaxel-naïve patients (n = 3), which included two patients previously treated with abiraterone. Durable disease control ranging from 20 to 36 months was seen in five patients, which included three patients treated with crizotinib 250 mg twice daily and one patient enrolled on the 250-mg daily and 200-mg twice-daily dose levels each. All five patients were naïve to enzalutamide, but two had received prior abiraterone, and one had been treated with docetaxel. However, given the subtherapeutic steady-state crizotinib trough concentrations seen in these patients (C_min^ss ranging from 33.6 to 91.4 ng/mL), it is unlikely that crizotinib contributed significantly to enzalutamide's antitumor efficacy seen in these patients. Patients who had received prior enzalutamide and/or abiraterone had poor outcomes with PSA response rate of 24%, which is in-line with previously reported response rates with sequential use of these agents (35). Median PFS in the subset of patients was low at 2.8 months.

As an exploratory analysis, we also investigated changes in serum BALP and CTX levels as on-therapy biomarkers. Several prior studies have reported significantly higher c-MET expression in bone metastases from patients with CRPC compared with primary tumors, and its role in bone remodeling in the metastatic bone tumor microenvironment (43–45). In a prior phase II trial in metastatic CRPC, treatment with the VEGF and MET inhibitor, cabozantinib, resulted in improvement in bone pain and reduction in total alkaline phosphatase and plasma CTX levels (46). In our study, we did not find any significant change in these biomarkers during the course of treatment. The significant reduction in crizotinib exposure may explain the lack of change in our study.

The main limitations of this study in determining any preliminary efficacy of concurrent c-Met and AR inhibition were the subtherapeutic dosing of crizotinib due to the pharmacokinetics drug–drug interaction and the heterogenous extensively pretreated patient population. Patients were not selected on the basis of tumor expression of c-MET, which theoretically might better identify patients likely to benefit from concurrent AR and c-MET inhibition and is worthy of future study.

Concurrent c-MET and AR inhibition represents an attractive therapeutic strategy in CRPC and is supported by extensive preclinical evidence. Although DLT criteria prohibiting dose escalation to full dose of crizotinib and subsequent full dose expansion were not met, concurrent treatment with enzalutamide and crizotinib resulted in a clinically significant 74% reduction in crizotinib systemic exposure presumably attributable to enzalutamide's induction of CYP3A4. On the basis of the significant drug–drug interaction observed, limited efficacy outside of what would be expected of enzalutamide monotherapy, and gastrointestinal and hepatic toxicity likely consistent with crizotinib's known effects that could worsen if higher doses were investigated, no RP2D was endorsed and additional investigation of this combination is not planned. However, further exploration of this strategy earlier in the disease course with different c-Met- and novel AR-targeting agents that are less likely to have significant drug–drug interactions is warranted, and such a prospective trial in metastatic CRPC is in development.

A. Tripathi reports personal fees from Foundation Medicine (advisory role) and Pfizer (advisory role), and grants from EMD Serono (research funding to institution), Aravive Inc. (research funding to institution), Bayer (research funding to institution), Clovis Oncology (research funding to institution), WindMIL Therapeutics (research funding to institution), and Corvus Pharmaceuticals (research funding to institution) outside the submitted work. J.G. Supko reports other from Astellas Pharma (fee for service) during the conduct of the study. M. Regan reports grants from Astellas/Pfizer [formerly Medivation, the study received funding and enzalutamide drug support from Astellas/Pfizer (formerly Medivation) and crizotinib drug only support from Pfizer] during the conduct of the study, and grants from Bayer, Novartis, Pfizer, TerSera Therapeutics, and Pierre Fabre, grants, personal fees, and nonfinancial support from BMS, grants and other from Ipsen [consulting (Institution)], nonfinancial support from Roche, and grants and nonfinancial support from AstraZeneca outside the submitted work. M.-E. Taplin reports personal fees from Pfizer (attended an advisory board) during the conduct of the study, and personal fees from AstraZeneca (advisory board), Bayer (advisory board), AbbVie (advisory board), Janssen (advisory board), Clovis (advisory board, DSMB), Arcus Biotechnology (advisory board), Myovant (member DSMB), UpToDate (author), and OncLive (panel member) outside the submitted work. A.D. Choudhury reports grants from NCI SPORE Career Development Award (genomic profiling of circulating tumor cells in a trial of combined crizotinib and enzalutamide) during the conduct of the study and grants from Bayer [research support (institution)] and personal fees from Bayer (advisory board), Dendreon (advisory board), and Clovis (advisory board) outside the submitted work. C. Yu reports other from Daiichi Sankyo, Inc. (employee) outside the submitted work. Z. Sun reports grants from City of Hope (grantee of NIH) during the conduct of the study. C.J. Sweeney reports grants from Astellas (IST support from trial), Bayer, and Dendreon, and other from Pfizer (drug supply) during the conduct of the study; grants and personal fees from Astellas, Pfizer, and Janssen, and personal fees from Lilly outside the submitted work; is listed as an inventor on an unlicensed patent for tristetraprolin prognostic marker in prostate cancer issued to Dana-Farber Cancer Institute: DFCA and Harvard School of Public Health; is a cofounder of Leuchemix, which owns and is licensee of a patent on dimethylaminoparthenolide as a treatment for cancer; and holds a patent for cabozantinib plus abiraterone issued to Exelixis. L.C. Harshman reports grants from Astellas/Pfizer (formerly Medivation, research funding and drug supply for the clinical trial) and Prostate Cancer Foundation (Young Investigator Award) during the conduct of the study, personal fees for consulting or advisory role from Genentech, Dendreon, Pfizer, Medivation/Astellas, Exelixis, Bayer, Kew Group, Corvus, Merck, Novartis, Michael J Hennessy Associates (Healthcare Communications Company and several brands such as OncLive and PER), Jounce, EMD Serrano, and Ology Medical Education, personal fees for support for research travel from Bayer and Genentech, and Surface Oncology (current employment), and grants for research funding to the institution from Bayer, Sotio, Bristol-Myers Squib, Merck, Takeda, Dendreon/Valient, Janssen, Medivation/Astellas, Genentech, Pfizer, and Endocyte (Novartis) outside the submitted work. No potential conflicts of interest were disclosed by the other authors.

A. Tripathi: Investigation, writing-original draft, writing-review and editing. J.G. Supko: Formal analysis, validation, investigation. K.P. Gray: Data curation, formal analysis. Z.J. Melnick: Data curation. M. Regan: Data curation, formal analysis. M.-E. Taplin: Investigation, writing-review and editing. A.D. Choudhury: Investigation, writing-review and editing. M.M. Pomerantz: Investigation, writing-review and editing. J. Bellmunt: Investigation, writing-review and editing. C. Yu: Investigation, writing-review and editing. Z. Sun: Conceptualization, investigation, writing-review and editing. S. Srinivas: Investigation, writing-review and editing. P.W. Kantoff: Supervision, investigation, writing-review and editing. C.J. Sweeney: Supervision, investigation, writing-review and editing. L.C. Harshman: Conceptualization, formal analysis, supervision, investigation, writing-review and editing.

The study received funding and enzalutamide drug support from Astellas/Pfizer (formerly Medivation) and crizotinib drug only support from Pfizer. L.C. Harshman also received support from the Prostate Cancer Foundation (2013 PCF Young Investigator Award).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.