With the rapid development of new therapeutic agents, better strategies are needed to select patients with high risk or relapsed acute lymphoblastic leukemia (ALL) for experimental therapy. We developed a large-scale drug response profiling platform that is based on an established ALL co-culture system on bone marrow stromal cells, with the advantage of standardized conditions in serum-free medium. Comparing the profiles of 61 ALL samples that were enriched for relapsed and refractory ALL, we identified informative patterns of drug sensitivity and resistance comparing between clinical and cytogenetic ALL subtypes. We detected strong and selective activity for therapeutic agents that could be used in the clinic. These include a BCP-ALL with particular sensitivity to SMAC mimetics and a T-cell ALL subsets with very high sensitivity to dasatinib or topotecan. Here we focus on identification of ALLs subtypes that are critically dependent on BCL2, based on their response patterns to the BCL2-specific BH3-mimetic venetoclax (ABT-199). A subset of BCP-ALL, including MLL-AF4 and TCF3-HLF positive ALL cases were highly sensitive to venetoclax. As reported by others recently, we also detected a smaller subgroup of T-ALL that were highly sensitive to BCL2 inhibition and confirm that these are not restricted to early thymic precursor ALL. Sensitivity to venetoclax in vitro correlated with anti-leukemic activity in ALL xenografts in vivo. Moreover combination testing on our platform indicated synergistic activity of venetoclax with conventional anti-leukemic agents such as dexamethasone and vincristine, and with the new class BRD4 inhibitors. In vivo, combination of venetoclax with dexamethasone and vincristine completely prevented disease progression of very aggressive TCF3-HLF ALL cases compared to a transient delay in disease progression that we observed in the control arms with single agent venetoclax or dexamethasone + vincristine treatment. Based on our results, we propose that drug activity profiling should be evaluated prospectively in clinical phase I and II settings as a tool to develop more personalized approaches for patients with resistant disease that are at urgent medical needs.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.