Targeting dormant tumor cells to prevent recurrent breast cancer: a randomized phase 2 trial.

Breast cancer recurrence may arise from dormant disseminated tumor cells (DTCs) that persist in bone marrow and other sites. Clinically, DTCs are independently associated with breast cancer recurrence and death. Preclinical studies in mouse models identified autophagy and mammalian target of rapamycin (mTOR) signaling as critical mechanisms of tumor dormancy and escape. We subsequently tested the effects of transient versus chronic inhibition of autophagy with chloroquine or hydroxychloroquine (HCQ) and mTOR signaling with rapamycin (RAPA) or everolimus (EVE) on residual tumor cell (RTC) burden and recurrence-free survival (RFS). In mice harboring dormant RTCs, inhibition of mTOR alone or in combination with autophagy inhibition decreased RTC burden and improved RFS in a duration-dependent manner. RTC number was strongly and inversely correlated with RFS, suggesting that RTC reduction mediated an improvement in RFS. To translate findings clinically, we performed a randomized phase 2 trial (CLEVER) of HCQ, EVE or their combination in breast cancer survivors within 5 years of diagnosis who had detectable DTCs on bone marrow aspirate. Primary endpoints were feasibility and safety; secondary endpoints included DTC reduction/clearance and RFS. In total, 51 DTC patients initiated HCQ (n = 15), EVE (n = 15) or HCQ + EVE (n = 21). Treatment was feasible and tolerable; only one patient discontinued early for grade 3 toxicity. At 42 months median follow-up, landmark 3-year RFS for HCQ, EVE and HCQ + EVE was 91.7%, 92.9% and 100%, respectively, and was greater in those who cleared DTCs versus those who did not (hazard ratio (HR) = 0.21 (95% confidence interval 0.01-3.4)). Posterior probabilities were 98-99.9% that three cycles of HCQ, EVE or HCQ + EVE led to reduced or undetectable DTCs compared to observation alone, with estimated DTC reductions of 80%, 78% and 87%, respectively. These findings provide proof-of-concept that targeting dormant RTCs with HCQ, EVE or their combination in breast cancer survivors or mouse models depletes minimal residual disease, warranting a definitive human randomized controlled trial. ClinicalTrials.gov registration: NCT03032406 .