Abstract
The treatment armamentarium for metastatic castration-resistant prostate cancer (mCRPC) has increased significantly over the past several years. Approved drugs associated with improved survival include androgen pathway-targeted agents (abiraterone acetate and enzalutamide), chemotherapeutics (docetaxel and cabazitaxel), an autologous vaccine (sipuleucel-T) and a radiopharmaceutical (radium-223). Abiraterone acetate, a prodrug of abiraterone, inhibits the CYP17A enzyme, a critical enzyme in androgen biosynthesis. Abiraterone has regulatory approval in mCRPC in both chemotherapy-naïve patients and in the post-docetaxel setting based on results from two randomized phase III studies. In the COU-AA-302 trial, abiraterone demonstrated significant improvement in the coprimary endpoints of radiographic progression-free survival and overall survival, as well as in a number of secondary endpoints including time until initiation of chemotherapy, time until opiate use for cancer-related pain, prostate-specific antigen progression-free survival and decline in performance status. Abiraterone is well-tolerated, although adverse events associated with this agent include abnormalities in liver function testing and mineralocorticoid-associated adverse events. This review evaluates the use of abiraterone in mCRPC prior to the use of chemotherapy.
Introduction
Prostate cancer is the second most commonly diagnosed malignancy in men around the world with an estimated 1,111,700 new cases and 307,500 deaths per year [Torre et al. 2015]. Metastatic prostate cancer is characterized by a period during which suppression of serum testosterone with androgen deprivation therapy (ADT) is sufficient to control disease. Unfortunately, this period is followed by transition to castration resistance, during which progression occurs despite continued suppression of testosterone. This is referred to as metastatic castration-resistant prostate cancer (mCRPC). Formerly, this disease state was known as hormone-refractory prostate cancer. As it is now known that androgen receptor (AR) signaling remains critical to disease progression in castration-resistant disease, this term is no longer used [Scher and Sawyers, 2005]. That prostate-specific antigen (PSA) rises in the setting of progression in mCRPC exemplifies this point as PSA is an androgen-regulated gene. The clinical relevance of targeting androgen signaling in mCRPC is demonstrated by the survival advantage of both abiraterone acetate and enzalutamide in this disease state.
Docetaxel was the first agent found to prolong survival in mCRPC and gained regulatory approval for this indication in 2004 [Petrylak et al. 2004; Tannock et al. 2004]. In recent years there has been a rapid increase in drugs available to treat this disease with the approvals of cabazitaxel (2010), sipuleucel-T (2010), abiraterone (post-docetaxel 2011, chemotherapy-naïve 2012), enzalutamide (post-docetaxel 2012, chemotherapy-naïve 2014) and radium-223 (2013). In the case of abiraterone and enzalutamide, approvals following docetaxel use were initially granted based on trials in this population; these trials were COU-AA-301 and AFFIRM, respectively [De Bono et al. 2011; Scher et al. 2012]. Subsequent trials in chemotherapy-naïve mCRPC were conducted for abiraterone (COU-AA-302 [Ryan et al. 2013]) and enzalutamide (PREVAIL [Beer et al. 2014]) leading to extension of approvals to this population. This review focuses on the use of abiraterone in chemotherapy-naïve mCRPC.
Abiraterone: rationale for use in mCRPC and mechanism of action
Following suppression of serum testosterone with ADT, intraprostatic androgen levels remain at 20–30% of pre-ADT levels and are sufficient to maintain expression of androgen-responsive genes such as PSA and AR [Page et al. 2006; Mostaghel et al. 2007]. Sources of persistent intratumoral androgens include uptake of adrenal androgens [dehydroepiandrosterone (DHEA) and androstenedione] and de novo synthesis of androgens by prostate cancer cells. CYP17A is critical to androgen biosynthesis in the adrenal glands and its expression has been identified in metastatic lesions from CRPC [Efstathiou et al. 2012]. Thus, CYP17A is a rational target to address persistent androgen synthesis in mCRPC.
CYP17A is responsible for the conversion of pregnenolone and progesterone to the androgen precursors of testosterone, DHEA and androstenedione (Figure 1). CYP17A has two sequential enzymatic activities; the 17α-hydroxylase function is critical to cortisol synthesis whereas the C17,20-lyase activity is required for androgen biosynthesis.
Steroid biosynthesis pathways and mechanism of action of abiraterone acetate. CYP17 has two sequential enzymatic activities. The 17α-hydroxylase activity is critical to production of cortisol. The C17,20-lyase activity is required for synthesis of testosterone and estradiol. Thus, inhibition of C17,20-lyase function is the desired function of abiraterone as it is this enzymatic activity that is responsible for androgen production.
Ketoconazole is an antifungal agent that inhibits multiple cytochrome P450 enzymes including CYP17A. Ketoconazole is an active drug in prostate cancer, but is poorly tolerated and is not proven to prolong survival. Inadequate durability of response is associated with a rise in serum androgens and progression of disease [Small et al. 2004]. The activity of ketoconazole in prostate cancer provided the stimulus to develop abiraterone, a more potent and selective inhibitor of CYP17A. Abiraterone inhibits the hydroxylase and lyase activities resulting in suppression of serum androgens and estrogens (Figure 1). This is accompanied by a decrease in cortisol, which leads to an increase in adrenocorticotropic hormone (ACTH) production by the anterior pituitary gland. ACTH induces an increase in mineralocorticoids (deoxycorticosterone and corticosterone). Co-administration of corticosteroids suppresses ACTH-induced mineralocorticoid production, thereby minimizing mineralocorticoid-related side effects of abiraterone.
Early studies of abiraterone
The first paper to describe abiraterone acetate was published in 1994 [Barrie et al. 1994]. In this article, a number of inhibitors of CYP17A were described. Among the most potent was abiraterone acetate (CB7630), the active metabolite of which (abiraterone or CB7598) was shown to suppress serum testosterone in mice with a compensatory increase in luteinizing hormone (LH). The first in-human dose escalation study of abiraterone was published a decade later [O’Donnell et al. 2004]. In non-castrate patients diagnosed with prostate cancer, abiraterone was shown to rapidly decrease serum testosterone to castrate levels. However, the compensatory increase in LH led to an increase in serum testosterone to non-castrate levels in some men. For this reason, future development of the drug was done in conjunction with ADT. It was noted that, while cortisol levels were unchanged following abiraterone, adrenocortical suppression lead to impaired response to exogenously administered ACTH.
A subsequent phase I study was conducted in chemotherapy-naïve mCRPC [Attard et al. 2008]. A dose of 1000 mg daily of abiraterone was selected given a plateau in pharmacodynamic effect at this dose. The mineralocorticoid receptor antagonist, eplerenone, was used to mitigate side effects related to secondary mineralocorticoid excess. Dexamethasone was used if mineralocorticoid side effects persisted despite eplerenone. Abiraterone was well tolerated and showed activity at all dose levels. There was an increase in serum ACTH and in hormones upstream of CYP17A. There was durable decrease in serum testosterone, DHEA and androstenedione.
Phase II studies of abiraterone were conducted in chemotherapy-naïve and docetaxel-treated mCRPC [Danila et al. 2010; Ryan et al. 2011]. To mitigate mineralocorticoid-related adverse events, prednisone at 5 mg twice daily was given along with abiraterone 1000 mg daily in both of these studies. Phase III trials were then conducted in chemotherapy-naïve (COU-AA-302 [Ryan et al. 2013]) and in docetaxel-treated (COU-AA-301 [De Bono et al. 2011]) populations. While both trials met their primary endpoints and led to abiraterone approval in their respective disease states, this review focuses on COU-AA-302.
COU-AA-302: a phase III study of abiraterone in chemotherapy-naïve mCRPC
COU-AA-302 study design and patient population
COU-AA-302 was an international, double-blind, placebo-controlled trial in mCRPC in which patients (n = 1088) were randomly assigned at a 1:1 ratio to receive abiraterone acetate (1000 mg by mouth daily) + prednisone (5 mg twice daily) or placebo + prednisone [Ryan et al. 2013] (Figure 2). Eligibility criteria stipulated patients to be chemotherapy-naïve with asymptomatic to minimally symptomatic mCRPC with disease progression evidenced by rising PSA by Prostate Cancer Working Group 2 (PCWG2) or radiographic progression according to modified RECIST criteria. Symp-tomatology was assessed using the Brief Pain Inventory-Short Form (BPI-SF) with a score of 0–1 classified as asymptomatic and a score of 2–3 classified as minimally symptomatic. Visceral metastatic disease was not allowed. Previous therapy with a first-generation antiandrogen (bicalutamide, flutamide or nilutamide) with progression following withdrawal was required. Previous therapy with ketoconazole was not allowed. Patients were stratified by Eastern Cooperative Oncology Group (ECOG) performance status of 0 versus 1. The coprimary efficacy endpoints were radiographic progression-free survival (rPFS) and overall survival (OS). Secondary endpoints included time to opiate use for cancer-related pain, time to use of chemotherapy, time to deterioration of performance status and time to PSA progression. Cycles were defined as 28 days of treatment. Disease assessments with serum PSA, computerized tomography (CT) or magnetic resonance imaging (MRI) of the chest, abdomen and pelvis and bone scan were performed at baseline, prior to cycles 3, 5 and 7, and then every third cycle. Treatment was continued until disease progression, defined by radiographic progression or unequivocal clinical progression without radiographic progression. Rising PSA in the absence of radiographic or clinical progression was not an indication to discontinue study treatment.
COU-AA-302 trial schema.
COU-AA-302 efficacy
A second interim analysis for OS took place after 43% of the anticipated 733 deaths had occurred [Ryan et al. 2013]. At the time of this analysis more deaths had occurred in the placebo group [186 of 542 (34%)] than with abiraterone [147 of 546 (27%)]. This represented a 25% reduction in the risk of death with abiraterone [hazard ratio (HR), 0.75; 95% confidence interval (CI), 0.61–0.93; p = 0.01]. This survival advantage did not meet the prespecified boundary for efficacy. However, at this interim analysis there was a clear benefit in rPFS favoring abiraterone (16.5 versus 8.3 months; HR, 0.53; 95% CI, 0.45–0.62; p < 0.001). The treatment effect of abiraterone was favorable for both OS and rPFS across all prespecified subgroups. Based on efficacy and tolerability, after the second interim analysis the data safety monitoring committee recommended unblinding the study with crossover of placebo-treated patients to abiraterone.
A subsequent interim analysis [Rathkopf et al. 2014] and a final analysis [Ryan et al. 2015] have since been published (Table 1). Notably, after a median follow up of 49.4 months and 741 deaths, despite early unblinding and subsequent therapy in 80% of the placebo group (67% of abiraterone group), the benefit in OS with abiraterone reached statistical significance (34.7 versus 30.3 months, HR, 0.81; 95% CI, 0.70–0.93; p = 0.0033). The final result for rPFS confirmed the advantage with abiraterone (16.5 versus 8.2 months; HR, 0.52; 95% CI, 0.45–0.61; p < 0.001). Secondary outcomes favored abiraterone including time to decline in performance status (12.3 versus 10.9 months; HR, 0.83; 95% CI, 0.72–0.94; p = 0.005), time to initiation of chemotherapy (26.5 versus 16.8 months; HR, 0.61; 95% CI, 0.51–0.72; p < 0.0001), time to opiate use (33.4 versus 23.4 months; HR, 0.72; 95% CI, 0.61–0.85; p < 0.0001) and time to PSA progression (11.1 versus 5.6 months; HR, 0.50; 95% CI, 0.43–0.58; p < 0.0001). Experimental endpoints also favored abiraterone. These included PSA decline by ⩾50% (68% with abiraterone versus 29% placebo), median time to decline in functional status measured as Functional Assessment of Cancer Therapy-Prostate (FACT-P) score (12.7 versus 8.3 months; HR 0.79; 95% CI, 0.67–0.93; p = 0.005) and objective response by RECIST (in the 438 patients with measurable disease) [36% versus 16%; relative risk (RR), 2.27; 95% CI, 1.59–3.25; p < 0.001].
Efficacy results of COU-AA-302 [Ryan et al. 2015].
CI, confidence interval; OS, overall survival; PSA, prostate specific antigen; rPFS, radiographic progression-free survival.
COU-AA-302 safety
Abiraterone plus prednisone was well tolerated with grade 3 or 4 adverse events occurring in 54% of the treatment group and 44% of the placebo group (Table 2). Adverse events leading to treatment discontinuation occurred in 13% with abiraterone versus 10% with placebo, while adverse events leading to death occurred in 4% versus 3%, respectively. Adverse events occurring in ⩾ 15% of patients and more common with abiraterone included fatigue (40% versus 35%), arthralgia (29% versus 24%), constipation (24% versus 20%), nausea (24% versus 23%), hot flush (23% versus 18%), diarrhea (23% versus 18%), bone pain (21% versus 19%), pain in extremity (17% versus 16%), cough (18% versus 14%), musculoskeletal pain (16% versus 15%) and insomnia (15% versus 12%). Mineralocorticoid side effects were more common with abiraterone: fluid retention or edema (31% versus 24%), hypokalemia (19 versus 13%) and hypertension (24% versus 14%) (Table 3). Cardiac disorders occurred more commonly with abiraterone [all grade (23% versus 18%) and grade ⩾3 (8% versus 4%)]. Abnormalities in liver function tests were also more common with abiraterone; increase in alanine aminotransferase (ALT) [all grade (13% versus 5%) and grade ⩾3 (6% versus 1%)] and increase in aspartate transaminase (AST) [all grade (12% versus 5%) and grade ⩾3 (3% versus 1%)].
Adverse event data from COU-AA-302 [Ryan et al. 2015].
AE, adverse event.
Adverse events of special interest from COU-AA-302 [Ryan et al. 2015].
AE, adverse event; ALT, alanine aminotransferase; AST, aspartate transaminase.
The final analysis showed no increase in toxicity following prolonged use of abiraterone with prednisone [Ryan et al. 2015]. This is of particular importance for physicians who may have some concern regarding long-term exposure to corticosteroids.
Agent selection in chemotherapy-naïve mCRPC
Pain as a consideration
Four non-cytotoxic, regulatory approved agents (abiraterone, enzalutamide, sipuleucel-T and radium-223) are shown to prolong survival in chemotherapy-naïve mCRPC. Among these, all but radium-223 are approved for asymptomatic mCRPC. As the ALSYMPCA trial of radium-223 in mCRPC required patients to be symptomatic, regulatory approval requires pain if this agent is to be used [Parker et al. 2013]. Due to exclusion of patients with moderate to severe pain from registration trials in chemotherapy-naïve mCRPC, there is little known about benefit of abiraterone and enzalutamide in these patients. As sipuleucel-T is not associated with objective responses, this agent is not appropriate when pain control is needed. Thus, in the prechemotherapy setting in mCRPC, radium-223 is an appropriate consideration for symptomatic disease while registration trials for abiraterone and enzalutamide do not provide guidance when considering these agents in this population.
Docetaxel as a first-line agent
While abiraterone and enzalutamide are shown to delay the time until chemotherapy in mCRPC, there are patients for whom chemotherapy may be more appropriate as a first-line therapy. Unfortunately, there are as yet no prospective data to guide the decision of chemotherapy versus androgen pathway agents in the first-line setting. Factors to consider include degree of pain, pace of disease progression, presence of visceral metastatic disease, comorbidities, performance status, patient preference and duration of response to ADT. Patients with significant pain, aggressive disease kinetics and short duration of benefit to initial ADT may be more appropriate for chemotherapy in the first-line setting.
Choosing abiraterone or enzalutamide
Both abiraterone and enzalutamide have a survival benefit and regulatory approval for use in chemotherapy-naïve mCRPC. As there are no randomized data to guide the decision between abiraterone and enzalutamide, the adverse event profiles and entry criteria of registration trials can be considered. In COU-AA-302, abiraterone was associated with grade ⩾3 increase in ALT and AST in 6% and 3% of patients, respectively [Ryan et al. 2015]. Therefore, in the setting of significant baseline hepatic disease, enzalutamide may be the preferred choice. Also, intolerance to corticosteroids would favor choosing enzalutamide for an individual patient. In PREVAIL, enzalutamide was not associated with an increased seizure risk (one event in both treatment groups), though in AFFIRM 0.6% of patients on enzalutamide had a seizure compared with no such events with placebo [Scher et al. 2012; Beer et al. 2014]. In the setting of a seizure disorder, or if a predisposing condition to seizures exists, abiraterone may be the better option.
COU-AA-302 excluded patients with visceral metastatic disease, whereas PREVAIL included such patients (198 of 1633 patients). Subgroup analysis of patients with visceral metastases in PREVAIL showed that enzalutamide was associated with a benefit in rPFS (median not yet reached versus 3.6 months; HR, 0.28; 95% CI, 0.16–0.49) [Beer et al. 2014]. Enzalutamide was not associated with a statistically significant increase in OS in the subgroup with visceral involvement, although there was a trend toward improved survival (27.8 months versus 22.8 months; HR, 0.82; 95% CI, 0.55–1.23). While visceral metastatic disease does not preclude benefit from abiraterone in patients treated with previous docetaxel [Goodman et al. 2014], such data do not exist in chemotherapy-naïve mCRPC. Therefore, if one were to rely strictly on chemotherapy-naïve mCRPC clinical trial data to choose between these two agents then enzalutamide may be favored.
Previous exposure to ketoconazole may also be used to guide this treatment decision. Ketoconazole was not allowed in either COU-AA-301 or COU-AA-302 and, notably, both AFFIRM and PREVAIL excluded patients having progressed while on ketoconazole. A small trial evaluated use of abiraterone in chemotherapy-naïve mCRPC patients who had previously been treated with ketoconazole [Kim et al. 2014]. PSA decrease by ⩾50% was numerically less (41%; 95% CI 26%–56%) than expected based on COU-AA-302 data. Thus, previous ketoconazole does not preclude benefit from abiraterone, but data suggest that previous ketoconazole may reduce the likelihood of benefit from this agent. The benefit of enzalutamide in patients having progressed on ketoconazole is uncertain.
Based on retrospective data, there is emerging concern that crossresistance exists between abiraterone and enzalutamide [Noonan et al. 2013; Cheng et al. 2015]. Prospective data are not yet available, although a clinical trial is underway that addresses this issue (see Clinical trial section below). At present, potential crossresistance should be considered when counseling patients with regard to sequence of available agents.
Biomarkers and prognostic models
Androgen-receptor splice variant 7
Ideally biomarkers will ultimately guide treatment choices in mCRPC. One recently described biomarker is AR splice variant 7 (AR-V7). AR-V7 is a truncated AR that lacks the ligand-binding domain, resulting in androgen pathway signaling in a ligand independent manner. Reverse transcriptase polymerase chain reaction (RT-PCR) was used to detect AR-V7 in circulating tumor cells in mCRPC patients prior to initiation of either abiraterone (n = 31) or enzalutamide (n = 31) [Antonarakis et al. 2014]. AR-V7 was detected in 39% and 19% of patients treated with enzalutamide and abiraterone, respectively. PSA response was observed in no patients with detectable AR-V7, compared with 53% and 68% of AR-V7 negative patients treated with enzalutamide or abiraterone, respectively. Detectable AR-V7 was associated with inferior survival. Based on these results, AR-V7 is undergoing further evaluation as a potential biomarker in mCRPC.
Prognostic models
In patients treated with first-line chemotherapy, baseline factors of performance status, sites of metastatic disease, lactate dehydrogenase (LDH), opioid use, albumin, hemoglobin, PSA and alkaline phosphatase can be used to classify patients into good- and poor-risk groups [Halabi et al. 2014]. Similarly, a prognostic model for survival with use of abiraterone has been developed. This model, derived from COU-AA-301 data, includes performance status, LDH, alkaline phosphatase and sites of metastatic disease (specifically whether liver metastases are present) and time from initiation of ADT to beginning abiraterone (> versus ⩽36 months) [Chi et al. 2013]. Patients are classified into to good-, intermediate- or poor-risk groups. Although external validation of this model has been confirmed in prechemotherapy and postchemotherapy patients, the model performed less well in prechemotherapy patients suggesting need for refinements in this population [Ravi et al. 2014]. In chemotherapy-naïve patients, median survival for good-risk patients (n = 44) was 45.6 months versus 34.5 months in intermediate/poor-risk patients (n = 20) (HR 1.79; 95% CI, 1.02–3.17; p = 0.042).
Clinical trials in chemotherapy-naïve mCRPC
Clinical trials are exploring the combination of abiraterone with other agents and the sequence of available agents in chemotherapy-naïve mCRPC. Alliance 031201 [ClinicalTrials.gov identifier: NCT01949337] is a phase III trial evaluating abiraterone with enzalutamide versus enzalutamide alone in patients with asymptomatic or mildly symptomatic chemotherapy-naïve mCRPC. The primary outcome measure is OS. A similar study in this population is evaluating the AR antagonist JNJ56021927 (ARN-509) with abiraterone versus abiraterone plus placebo. The primary outcome is rPFS [ClinicalTrials.gov identifier: NCT02257736]. ERA 223 [ClinicalTrials.gov identifier: NCT02043678] is a phase III trial evaluating abiraterone with radium-223 versus abiraterone plus placebo in asymptomatic to minimally symptomatic chemotherapy-naïve mCRPC. The primary outcome measure is symptomatic skeletal event-free survival.
A randomized phase II trial sponsored by the British Columbia Cancer Agency is evaluating the sequence of abiraterone and enzalutamide in chemotherapy-naïve mCRPC [ClinicalTrials.gov identifier: NCT02125357]. This two-arm study evaluates both sequences. Patients will be transitioned to the second agent following PSA progression on the first with PSA response rate to the second agent as the primary endpoint.
Additional studies are evaluating (or have evaluated) abiraterone in earlier disease states including the neoadjuvant setting [Taplin et al. 2014], metastatic, hormone-naïve disease [ClinicalTrials.gov identifier: NCT01715285, NCT01957436, NCT00268476] and in non-metastatic CRPC [Ryan et al. 2014].
Conclusion
The COU-AA-302 trial demonstrated that abiraterone improves OS and rPFS in chemotherapy-naïve mCRPC with a consistent benefit across all subgroups analyzed. There was also an advantage in secondary endpoints including time to decline in performance status, time to opiate use and time to initiation of chemotherapy. Abiraterone is well tolerated, although several adverse events including those associated with mineralocorticoid excess (fluid retention, hypokalemia and hypertension) and liver function test abnormalities may occur. Based on the results of COU-AA-302, the approval for abiraterone was extended to include chemotherapy-naïve mCRPC.
Along with abiraterone, several agents are approved in chemotherapy-naïve mCRPC. The requirement for co-administration of corticosteroids with abiraterone and the adverse event profile of available drugs may guide selection of agents for individual patients. Ongoing studies are addressing issues of patient selection and the sequencing/combination of these agents. Addressing these issues is necessary to maximize the benefit of available agents. Biomarkers to guide precision medicine in mCRPC are badly needed and AR-V7 is the most promising identified thus far.
Abiraterone prolongs survival, forestalls the need for chemotherapy, and maintains the quality of life in asymptomatic or mildly symptomatic, chemotherapy-naïve mCRPC. The efficacy of this agent combined with a favorable toxicity profile and the convenience of oral administration solidify abiraterone as an important treatment consideration in this patient population.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest statement
The authors declare no conflicts of interest in preparing this article.