Abstract P5-08-06: PI3K blockade enhances the antitumor activity of eribulin in PIK3CA-mutant eribulin-resistant tumor xenografts

Abstract

Abstract Eribulin is a recently approved microtubule-targeting agent (MTA) for the management of heavily pretreated metastatic breast cancer (BC) patients. Constitutive PI3K/Akt/mTOR survival pathway activation, either via mutational activation of the PI3K catalytic subunit (PIK3CA) or via inactivation of the tumor suppressor PTEN, may confer resistance to MTAs. Hence, we hypothesized that PI3K-pathway activation limits the antitumor activity of eribulin in HER2-negative BC and that PI3K inhibition enhances the efficacy of this chemotherapeutic agent. The predictive value of PIK3CA mutation or PTEN loss towards eribulin response was interrogated in vitro, using a panel of fourteen HER2-negative BC cell lines, and in vivo using six tumor models from cell-line or patient-derived tumors. While PIK3CA mutation did not seem to be predictive in vitro, the PIK3CA-mutated xenograft models underwent tumor progression upon single-agent eribulin therapy (Table 1). In the absence of concomitant PIK3CA mutation, PTEN loss was neither predictive in vitro nor in vivo. Moreover, eribulin induced PI3K-pathway activation in tumor xenografts, a potential escape mechanism to MTA-therapy. To validate the negative predictive value of PIK3CA mutation, BC tumors from patients treated with eribulin in the neoadjuvant and in the metastatic setting are being queried, as the pharmacodynamic activation of the PI3K-pathway upon eribulin treatment. Class I pan-PI3K (BKM120) or PI3K-a-specific (BYL719) inhibitors were used in vitro to block the PI3K-pathway concomitantly with eribulin treatment, resulting in enhanced antiproliferative and proapoptotic activity. Strikingly, in PIK3CA mutant xenograft models, while eribulin alone exhibited limited antitumor activity compared to PIK3CA-wild type models, co-administration of a PI3K inhibitor induced marked tumor regression (BKM120 data in Table 1, BYL719 data will be reported). Moreover, addition of the PI3K inhibitor at progression with eribulin single-agent also resulted in tumor regression. Of note, PIK3CA-wild type models also exhibited increased antitumor activity with the combined therapy compared to single-agent treatments. The precise mechanism by which the combination of eribulin and a PI3K-targeting agent results in tumor regression is currently under investigation, embracing both the induction of mitotic catastrophe in tumor cells and the regularization of the tumor vasculature. These results support the clinical development of therapeutic regimens combining PI3K-inhibitors to the approved MTA eribulin and might be predictive of clinical benefit both in the PIK3CA-mutant and -wild type breast cancer population. Table 1. Percentage change in tumor volume.StatusPIK3CA mutantPIK3CA mutant and PTEN-lowPTEN-lowWTModel/ TreatmentMCF7LPDX44CAL51MDA-MB-468PDX88PDX98Eribulin39 ± 9683 ± 51364 ± 127-47 ± 14-26 ± 34-36 ± 33BKM120501 ± 169116 ± 111119 ± 73-9 ± 23128 ± 3787 ± 112Eribulin + BKM120-70 ± 27-80 ± 1051 ± 46-75 ± 12-94 ± 5-65 ± 13Patient- (PDX) and cell line-derived tumor xenografts were treated with eribulin mesylate (0.1mg/kg, 3IW) and/or BKM120 (27.5 mg/kg, 6IW) for 26-31 days.. IW, in week. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P5-08-06.