Background: The selective mTORC1 inhibitor everolimus (RAD001/Affinitor®) has recently been approved, in combination with exemestane, for treatment of advanced hormone receptor positive breast cancer after demonstrating significantly improved progression free survival for patients having progressed on a previous non-steroidal aromatase inhibitor. However, despite these improved outcomes many tumors go on to develop resistance to everolimus. Alterations in PI3K/AKT signaling, either through activating mutations in PIK3CA or loss of PTEN, have been shown to influence responses to everolimus treatment in vitro (Weigelt et al., 2011). In this study we used cell line xenograft models of everolimus resistance to assess the potential of a pan-PI3K inhibitor (BKM120) or a p110a-specific inhibitor (BYL719) to reverse resistance either as single agents or in combination with anti-ER therapy (Fulvestrant).

Materials and Methods: Everolimus resistant xenografts were developed from three originally everolimus sensitive ER+ cell lines carrying distinct genetic alterations that confer aberrant PI3K/AKT signaling; ZR75-1 (PTEN null), MCF7 (PIK3CA mutant) and KPL1 (wild-type). Mice were treated by daily oral everolimus (10 mg/kg) until tumors began to progress on treatment (∼ 35 days of daily treatment). Mice that developed everolimus resistant tumors were randomized into six treatment groups (n = 6) as follows: continued on everolimus (10 mg/kg), fulvestrant (5 mg/wk), BKM120 (35 mg/kg), BYL719 (50 mg/kg), fulvestrant + BKM120 or BYL719.

Results: All three cell line xenografts were initially responsive to everolimus. ZR75-1 xenografts were allowed to progress on everolimus until they reached a mean tumor volume greater than that on day zero, at which point mice were randomized into treatment groups. Tumor selected to continue on everolimus gained a mean tumor volume of 550 mm3 over the six-week duration of study, demonstrating resistance to everolimus. These tumors showed a partial response to Fulvestrant (28% TGI), BYL719 alone (40% TGI) and BYL719 combined with fulvestrant (61% TGI). In contrast, over 100% tumor growth inhibition (tumor regression) was observed with both single agent BKM120 or BKM120 plus fulvestrant. This is consistent with our in vitro findings that PTEN null tumor cells might be less sensitive to inhibition by BYL719 yet sensitive to BKM120. Data from the ongoing PIK3CA mutant and wild-type KPL1 models will inform whether these molecular subtypes of ER+ breast cancer are more sensitive to the pan-PI3K inhibitor or the p110α-specific inhibitor. Comprehensive pharmacodynamic analyses of these tumor tissues may identify the additional molecular determinants responsible for resistance to everolimus and help guide the design of future clinical studies.

Discussion: These pre-clinical data indicate that targeting PI3K either alone or in combination with fulvestrant overcomes resistance to everolimus in ER+ breast cancer.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr PD1-5.