A randomized phase I trial of nanoparticle albumin-bound paclitaxel with or without mifepristone for advanced breast cancer
The primary objective of this study was to determine the safety and tolerability of the combination of the GR steroidal antagonist mifepristone and nab-paclitaxel. We also sought to confirm that systemic GR antagonism was occurring with the mifepristone dose used in this trial. Dose-limiting neutropenia (as specified in the protocol) was observed at both doses of nab-paclitaxel; therefore, we halted the study and analyzed the pharmacokinetic data for a potential drug–drug interaction. The novel randomized design we used allowed us the opportunity to understand how mifepristone impacted nab-paclitaxel clearance in the same patient, as those patients who were randomized to placebo for cycle 1 crossed over to receive mifepristone for cycle 2. Data from the pharmacokinetic studies suggest that for some patients, there appeared to be a delay in paclitaxel clearance when co-administered with mifepristone (see Fig. 2a, b). For the remaining patients, there was no clear evidence of a drug–drug interaction, and paclitaxel concentrations 24 h after administration were consistent with levels observed in initial phase I studies of nab-paclitaxel (Nyman et al. 2005; Ando et al. 2012).
The neutropenia observed in this phase I trial, while dose-limiting by protocol definition, was safe and manageable, and the majority of patients who experienced neutropenia continued to receive study treatment with dose delays, dose modifications, and/or the institution of growth factor support. Most of the five patients who received growth factor support received 2–3 days of once daily filgrastim, with rapid recovery of their counts. Of the nine patients studied, four did not experience neutropenia, and thus did not require dose modifications or growth factor support. Of the five who did experience neutropenia, one discontinued study treatment due to persistent neutropenia despite growth factor support and dose modification (patient 4), with the other four remaining on study therapy with the assistance of growth factor support. Of note, patient 4 continued to have neutropenia on nab-paclitaxel even after the discontinuation of mifepristone, and had to discontinue nab-paclitaxel therapy completely due to persistent neutropenia despite dose reduction and the institution of growth factor support. Of the other four, only one also required a dose reduction for the management of neutropenia (patient 5), but she was able to stay on study therapy for 10 cycles, before coming off for neuropathy.
To explore if there was a pharmacodynamic interaction between mifepristone and nab-paclitaxel (leading to a decreased neutrophil count), we compared the log-fold reduction in ANC for those randomized to mifepristone for the first cycle of therapy to those randomized to placebo. While we did see a greater log-fold reduction in ANC when mifepristone was given compared to placebo, this correlation was borderline significant. However, some cases of neutropenia observed in this trial did not appear to be the result of delayed nab-paclitaxel clearance. The greater log-fold reduction seen when mifepristone was administered with nab-paclitaxel as compared to placebo is suggestive of a pharmacodynamic interaction between mifepristone and nab-paclitaxel leading to neutropenia.
We hypothesize that some cases of neutropenia observed in this study may be related to the enhancement of chemotherapy-induced neutrophil apoptosis in the setting of a GR antagonist, and not delayed clearance of nab-paclitaxel. Paclitaxel levels observed in this trial were consistent with those levels observed in other trials using similar doses of nab-paclitaxel, further suggesting a pharmacodynamic mechanism for the neutropenia, rather than a simple pharmacokinetic increase in paclitaxel levels. Glucocorticoids have been shown to protect neutrophils from apoptosis (Kato et al. 1995). In the setting of GR antagonists, it is possible that neutrophils will be more susceptible to cell death, particularly in the setting of apoptosis-inducing cytotoxic therapies such as nab-paclitaxel. As such, combining GR antagonism with chemotherapy in general may require growth factor support. As some patients did not develop neutropenia, it is likely that dose modifications and growth factor support will need to be individualized based on observed toxicities.
In an exploratory analysis evaluating tumor GR and AR expression, we observed that tumors with strong GR expression were typically AR-negative, and tumors with strong AR expression were typically GR-negative. Expression of BR is negatively regulated by AR signaling in prostate cancer (Xie et al. 2015), and our data suggest that this this is likely the case in breast cancer as well. Six patients had tumors that were GR-positive (five were strongly GR-positive, and one was weakly GR-positive), and all six of these tumors were triple-negative at the time of recurrence (two of these patients initially presented with ER-positive disease, but tumors had converted to TNBC at the time of disease recurrence). Interestingly, of the five patients with GR strongly-positive tumors, four patients had a response to therapy (two CRs and two PRs, all confirmed), and one had SD (unconfirmed). The patient with GR weakly-positive disease progressed rapidly. These observational data support our hypothesis that GR-positive TNBC may benefit from the addition of a GR antagonist to chemotherapy.
In conclusion, nab-paclitaxel plus mifepristone appears to be a tolerable regimen, with a primary toxicity of neutropenia. While neutropenia was a protocol-defined DLT, the combination of mifepristone and nab-paclitaxel at both dose levels studied was tolerable, and the neutropenia was easily managed. It also appears that the interaction between mifepristone and nab-paclitaxel is not necessarily pharmacokinetic in nature, but rather pharmacodynamic. Thus, given the ease of management of the neutropenia and the promising efficacy of the combination, we are proceeding with a randomized phase II trial of nab-paclitaxel with mifepristone versus placebo in patients with advanced, GR positive, triple-negative breast cancer. We will use a nab-paclitaxel dose of 100 mg/m2 given 3 weeks on followed by 1 week off, with a mifepristone dose of 300 mg given the day before and the day of each nab-paclitaxel dose. As neutropenia was not observed in all patients, even at the higher dose level used, we will use the standard dose of 100 mg/m2, the standard dose used in the metastatic setting.