Although novel therapies are often evaluated in cell lines or cell line-based xenografts, patient-derived xenografts (PDX) can provide highly effective pre-clinical models for drug testing. For breast cancer, primary tumours can be serially propagated in immunocompromised mice through orthotopic (mammary fat pad) transplantation. Such models offer significant advantages over cell line-based xenografts as they faithfully recapitulate the primary tumour. To date, most PDX models have been derived from triple negative tumours, with limited reports on their use in pre-clinical drug testing. Few ER-positive (luminal) models have been described.
Over-expression of the pro-survival protein BCL-2 is common in breast cancer, with elevated levels found in approximately 85% of luminal tumours. Small molecule inhibitors termed ‘BH3 mimetics’ that bind and neutralize BCL-2 pro-survival proteins have recently been described. These are showing considerable promise in early phase studies of lymphoid malignancies. We sought to establish whether it might be feasible to target luminal B tumours with combination therapy comprising endocrine therapy (tamoxifen) and a BH3 mimetic (ABT-737 or ABT-199).
A panel of 36 primary breast tumour xenografts (including 15 luminal tumours) were derived. Three BCL-2-positive luminal B models (23T, 315T, 50T), as determined by Ki67 and gene profiling, and a control BCL-2-positive, ER-negative model (838T) were selected for further study. Cohorts of mice bearing tumour xenografts were treated with either ABT-737 (which neutralises BCL-2, BCL-XL and BCL-W), tamoxifen or the combination. Tumour response and overall survival were significantly improved by combination therapy in all three ER-positive xenograft models, when compared to tamoxifen alone (p<0.005). Despite abundant BCL-XL expression in tumours, similar efficacy was observed with the potent selective BCL-2 inhibitor ABT-199 and tamoxifen (p<0.005), revealing that BCL-2 is a crucial target. Unexpectedly, both BH3 mimetics were found to counteract the side effect of tamoxifen-induced endometrial hyperplasia. In addition, we observed that BH3 mimetics synergize with dual PI3K/mTOR inhibitors in the induction of apoptosis. Notably, in the 315T model (which exhibited elevated pAKT), triple therapy with ABT-737, the PI3K/mTOR inhibitor PKI-587 and tamoxifen further augmented tumour response in vivo (when compared to ABT-737 and tamoxifen; p<0.004).
Collectively, our findings provide a rationale for clinical evaluation of BH3 mimetics in early phase studies in breast cancer. Here, BCL-2 protein or mRNA expression (as determined by immunohistochemistry or RT-PCR, respectively) could provide a suitable companion biomarker for patient selection.