Evaluation of exposure-response-safety relationship of model-informed low-dose 500 mg abiraterone acetate in prostate cancer patients

Prostate cancer is a common cancer among men worldwide. Large-scale clinical studies of the 1,000 mg daily dosing of abiraterone acetate (AA) have confirmed its antitumor efficacy in patients with metastatic hormone-sensitive prostate cancer (mHSPC) or metastatic castration-resistant prostate cancer (mCRPC), regardless of their cancer's response to androgen deprivation therapy (ADT) or treatment duration. However, this dosage was indirectly justified based on the absence of dose-limiting toxicities (DLTs) in prior phase I dose-escalation trials, where a plateau in the increase of upstream steroids relating to secondary mineralocorticoid excess was observed at doses greater than 750 mg and up to 2,000 mg daily [1, 2]. Notably, prostate specific antigen (PSA) levels declined at all investigated doses (250 to 1,000 mg) [1, 2].

Cytochrome P450 17A1 (CYP17A1) is involved in both adrenal and de novo intratumoural androgen biosynthesis. We previously identified that abiraterone targeted CYP17A1 via a two-step binding mechanism [3]. Our subsequent pharmacokinetic/pharmacodynamic (PK/PD) simulations found that both the 1,000 mg and 500 mg doses of AA achieved comparable > 80% apparent target CYP17A1 enzyme occupancy and equipotent reduction of downstream plasma dehydroepiandrostenedione-sulfate (DHEA-S) levels, despite the difference in systemic exposure of abiraterone [3]. In addition, we developed physiologically-based pharmacokinetic (PBPK) models for AA and abiraterone via a middle-out approach [4], which enabled the prospective prediction of abiraterone systemic exposure at different doses.

Our research group participated in a global phase II study that demonstrated a 250 mg dose of AA taken with a low-fat meal achieved comparable PSA metrics to the standard 1,000 mg AA dose taken in a fasting state in patients with CRPC [5]. However, the fat content of food could significantly impact the relative bioavailability of abiraterone [4], and controlling food intake poses a challenge in outpatient settings and during the long-term use of abiraterone. By analyzing the PK data, we observed that the systemic exposure of a lower dose of 500 mg of AA (fasted) is comparable to that of a 250 mg dose of AA with a low-fat meal. Furthermore, our modeling studies revealed that 500 mg AA is promising in achieving optimal antitumor efficacy, and diminishing mineralocorticoid-related adverse outcomes simultaneously. In addition, patients will pay less with a half-reduced dose. Currently, data on the administration of 500 mg AA in prostate cancer patients remains insufficient. To address this gap, we conducted a proof-of-concept phase I study in mCRPC and mHSPC patients newly initiated on 500 mg once daily AA. Simultaneous PBPK/PD simulations of the low-dose AA were performed to further support the unique relationship between systemic exposure and pharmacological response of abiraterone.

The clinical cohort study was conducted at National University Hospital (NUH), Singapore. The study was approved by the Domain Specific Review Boards of National Healthcare Group, Singapore, and was conducted in accordance with Declaration of Helsinki. Between November 2021 and September 2023, 7 men with mHSPC and 2 men with mCRPC were enrolled for the final analysis (median age, 72 years; range, 65 to 90). Enrolled patients were initiated on 500 mg AA once daily for 12 weeks, plus oral prednisolone 5 mg twice daily for mCRPC and 5 mg once daily for mHSPC. After this period, patients were reverted to the standard 1,000 mg dose due to ethical considerations and were followed up with routine clinical visits. The primary objectives were to determine the PK of abiraterone, as well as evaluate the pharmacological response, i.e. percentage change in PSA from baseline to 12 weeks, and safety of 500 mg AA treatment. Secondary objective was the measurement of endocrine biomarkers (testosterone, androstenedione, DHEA-S, and cortisol) to further assess the pharmacological response of our low-dose AA treatment. Details of study design and data analysis are described in Supplementary Materials. Characteristics of enrolled patients are detailed in Supplementary Materials. Population-based ADME simulator Simcyp (version 23, Sheffield, UK) was utilized for simultaneous PBPK/PD simulations of abiraterone PK and time-dependent CYP17A1 enzyme occupancy. Our model-informed 500 mg and clinically approved 1,000 mg AA dosages were simulated. Details of modeling workflow are provided in Supplementary Materials.

Plasma PK of abiraterone at week 2 with available samples from seven patients was analyzed. Observed and simulated plasma concentrations of abiraterone are illustrated in Figure 1A-B. Corresponding PK parameters are provided in Supplementary Materials. PK sampling up to 6 h post-dose was implemented due to ethical considerations, and PBPK simulation for the 24-h PK of abiraterone was utilized as the proxy for further evaluation. Simulated plasma concentrations of abiraterone recapitulated our clinical observations from 0 to 6 h post-dose (Figure 1A). 24-h PK profiles revealed that the systemic exposure of 500 mg AA was comparable with previous results from mCRPC patients under the same dose [1, 2], and was approximately half of that previously observed or simulated with a 1,000 mg dose of AA [4].

Proportion of patients achieving a significant decrease in PSA at early post-treatment stage (usually within 12 weeks) has been frequently used as a hallmark for measuring response to a variety of prostate cancer therapies [6, 7]. Decline in PSA at week 12 was observed in all our nine patients (Figure 1C). Seven patients (78%) demonstrated decrease in PSA levels of ≥ 50% at any visit (Figure 1D). In brief, 6 mHSPC patients achieved PSA decline of ≥ 50% at week 4, and further decreased to ≥ 80% at week 12 (Figure 1D). One mCRPC patient achieved PSA decline of 91.1% at week 12, and another patient exhibited substantial PSA decline at week 8 (90.4%) but a rebound at week 12 (22.1%) (Figure 1D). The PSA rebound might possibly be associated with germline mutations in the DNA damage repair gene, ataxia-telangiectasia mutated (ATM), detected in his cancer tissues. These mutations have been found to be associated with attenuated responses to androgen receptor (AR)-targeted therapy [8].

Low-dose AA was safe and well tolerated during the 12-week treatment. Adverse events (AEs) of any cause occurred in 6 out of 9 patients. Details of safety profile are provided in Supplementary Materials. Hypokalemia was previously the only grade 3 or grade 4 AE in the 500 mg AA dosage groups [1, 2]. Consistently, hypokalemia was the most common AE in our study (Supplementary Materials). Other reported AEs were not observed in our study.

Baseline levels of testosterone, androstenedione, DHEA-S and cortisol were similar between 2 types of patients in our study. Circulating testosterone levels at baseline were in the castrate range (median, 14.62 ng/dL; range, 3.57 to 43.15) in all 9 patients (Figure 1E). From Visit 1 onwards, decline of levels of the 4 steroids were well correlated, demonstrating substantial suppression by low-dose AA (Figure 1E-H). The observations were also consistent with findings from previous studies on 1,000 mg AA therapy, which reported that suppression of downstream steroids of CYP17A1 was correlated with PSA decline at week 12 [6, 9, 10]. Therefore, our endocrine profiles broadened the evidence in supporting the pharmacological response of 500 mg AA by including 7 mHSPC patients in addition to 2 mCRPC patients.

Our clinical observations were substantiated via PBPK/PD modeling of daily doses of both our low-dose 500 mg and clinically approved 1,000 mg AA. CYP17A1 enzyme occupancy remained above 80% despite a 50% reduction in systemic exposure to abiraterone after 2 weeks of 500 mg AA treatment (Figure 1B). In addition, free CYP17A1 continued to be slowly released from tight binding, while abiraterone has been eliminated systemically.

Our preliminary research involved a small cohort of patients with mCRPC and mHSPC in the treatment group only, and evaluations were conducted over a relatively short period of 12 weeks. Despite this limitation, our proof-of-concept phase I study and PBPK/PD modeling results underscored the pharmacological response of the low-dose regimen, and were consistent with our hypothesis that a lower dose of AA is promising for achieving optimal antitumor efficacy, reducing adverse outcomes, and alleviating financial burdens simultaneously. A long-term, large-scale, controlled clinical trial is essential to further evaluate and confirm the clinical efficacy of low-dose AA therapy.

No author has an actual or perceived conflict of interest with the contents of this article.

This work is funded by Joseph Lim Boon Tiong Urology Cancer Research (Grant: A-0002678-01-00). The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

The study was approved by the Domain Specific Review Boards of National Healthcare Group, Singapore (2020/00258), and was conducted in accordance with Declaration of Helsinki. All patients provided written informed consent.

Clinicaltrials.gov: NCT06193993, 2023-Dec-05.

Singapore HSA: No.: 2020/00258, 2020-Oct-30.