Abstract P2-16-21: A randomized phase I trial of nanoparticle albumin bound paclitaxel (nab-paclitaxel, Abraxane®) with or without mifepristone for advanced breast cancer

Up to 40% of breast cancers have been shown to express the glucocorticoid receptor (GR), and activation of the GR is associated with poor prognosis in ER-negative breast cancer. We hypothesize that GR activation in breast cancer cells initiates anti-apoptotic signaling contributing to chemotherapy resistance. Based on our compelling preclinical data demonstrating that GR antagonism with mifepristone (mif) increases paclitaxel (pac) induced breast cancer cell death in vitro and in vivo, we conducted the first clinical trial of the combination of anti-GR therapy and chemotherapy in patients (pts) with metastatic breast cancer (MBC). Because the combination of mif and nab-paclitaxel (nab) had not previously been administered to pts, and because nab can result in cumulative neurotoxicity, a randomized, placebo-controlled, phase I design was used. Mif is known to inhibit CYP2C8, an enzyme involved in the metabolism of pac, thus plasma pac levels were monitored to evaluate for significant changes in clearance. A 3+3 dose escalation scheme was planned, with an initial nab dose of 100mg/m2 weekly and a mif dose of 300 mg/d for 2 days (day prior to and of nab infusion). For each dose level (DL), patients were randomized 3:2 to mif:placebo (p) for cycle 1 (C1). After C1, pts randomized to p were crossed over to receive mif at their assigned DL for the duration of study treatment. DL1 was deemed intolerable due to neutropenia, so the nab dose was decreased to 80 mg/m2 for DL2, and dose escalation of mif beyond 300 mg was halted. Serum cortisol and ACTH levels—biomarkers of effective GR blockade—were measured before and after mif treatment. Archival tumor was collected to determine tumor GR expression, and peripheral blood lymphocytes (PBLs) were collected to evaluate GR-target gene expression after mif. 9 pts were enrolled. Median age was 56 yrs (range 47-74). Median number of prior therapies for MBC was 2 (range 0-3). 8 pts had triple-negative breast cancer (TNBC), and 1 had ER+ disease. 8 of 9 pts had recurred after taxane-based therapy. 1 pt had a CR, 4 pts a PR, and 4 pts POD. 4 of 5 pts who responded to nab plus mif had previously relapsed after taxane-based therapy. 4 pts were treated at DL1 (2 mif, 2 p). Both pts randomized to mif experienced a dose-limiting toxicity (DLT) of neutropenia during C1. 5 pts were treated at DL2 (3 mif, 2 p), and 2 of the 3 pts randomized to mif for C1 experienced a DLT (neutropenia). Plasma pac levels were consistent with delayed clearance of nab when co-administered with mif for most pts. All pts were found to have a 2-fold or greater rise in their serum cortisol levels, demonstrating effective adrenal GR inhibition. Mif delays clearance of nab in most pts, and results in DLT. Given the inter-patient variability in delay of nab clearance by mif, co-administration produces unpredictable toxicity. However, given the responses seen in taxane-pretreated TNBC pts, GR antagonism is a promising approach for treating aggressive TNBCs. As neither carboplatin nor gemcitabine are metabolized, mif is unlikely to affect clearance of these agents. As they are widely used for advanced TNBC, both agents represent attractive chemotherapy partners for future clinical investigation. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P2-16-21.