Apoptosis is controlled through the dynamic interactions of the Bcl2 protein family, and cancers have evolved mechanisms to hijack this pathway to evade apoptosis, often by upregulating anti-apoptotic proteins (e.g. Mcl1, Bcl2). This survival adaptation creates a dependency that could be exploited therapeutically, which is why considerable effort has been made to develop small molecule inhibitors of the anti-apoptotic Bcl2 family proteins. This class of drug was clinically validated with the approval of venetoclax, a selective Bcl2 inhibitor, for the treatment of CLL. Venetoclax is undergoing evaluation in numerous other clinical trials for predominantly hematologic malignancies.

Despite impressive responses observed with venetoclax in CLL (ORR 79%), acquired resistance is beginning to emerge. Likewise, other hematologic indications are more intrinsically resistant to venetoclax, exhibiting much lower response rates in the respective Phase I clinical trials (AML = 38%, NHL = 44%, MM = 12%). Preclinical studies with venetoclax have reported increased levels of other anti-apoptotic proteins as a likely mechanism contributing to both de novo and acquired resistance. Therefore, combining cell death inducing agents that inhibit Mcl1 or BclxL could be a means of combating resistance.

Inhibition of cyclin-dependent kinase 9 (CDK9), which regulates transcription elongation, has been reported to reduce protein levels of genes with short-lived transcripts and proteins, such as MCL1. AZD4573 is a novel and selective CDK9 inhibitor that shows potent single agent activity, inducing cell death in vitro and tumor regressions in vivo in a diverse set of hematologic cancers (Cidado et. al., AACR Annual Meeting 2018). AZD4573 is currently being evaluated in a Phase I clinical trial for patients with hematological malignancies (NCT03263637).

This study evaluates whether combinations of AZD4573 with other cell death inducing agents could overcome de novo venetoclax monotherapy resistance. A panel of 12 AML and 6 DLBCL cell lines were treated for 6 h with venetoclax, AZD4573, and a selective BclxL probe compound (AZ'3202) either as single agents or in combinations and assayed for caspase activation. The Loewe model was used to calculate synergy scores to assess benefit over monotherapy, and combinations with scores >5 were deemed beneficial. Treatment with AZD4573+venetoclax and AZD4573+AZ'3202 resulted in beneficial combinations for 13/18 and 10/18 cell lines, respectively. On the other hand, venetoclax+AZ'3202 showed significant combination benefit in only two cell lines (NB4, SUDHL4), suggesting a primary dependency upon Mcl1 for most of these hematologic cancer cell lines. Interestingly, cell lines sensitive to single agent AZ'3202 (4 AML, 0 DLBCL) did not show any combination benefit when treated with AZD4573+venetoclax, highlighting the exquisite dependency of those four models upon BclxL and mutual exclusivity with Mcl1.

Cell lines benefitting from the AZD4573+venetoclax combination tended to fall into one of two categories: having single agent activity to either agent that is enhanced by the combination or having no single agent activity but the combination shifts the cell line into a responder. SUDHL4 cells were sensitive to AZD4573 (caspase activation EC50 = 18 nM) but not venetoclax (EC50 = 476 nM) while OCI-AML3 was insensitive to both (EC50 > 30 µM). In vitro biomarker kinetic analysis revealed an increase in Mcl1 levels (~2-fold) after 3 h of venetoclax treatment that was abrogated upon combination treatment, providing a mechanistic rationale for the combination benefit. Furthermore, when tested in an OCI-AML3 tumor xenograft study in mice, AZD4573 or venetoclax monotherapy exhibited minimal tumor growth inhibition (44% and 16%, respectively) while the combination led to tumor regressions (64%) with minimal effect on body weight. In a SUDHL4 tumor xenograft study, venetoclax monotherapy displayed minimal tumor growth inhibition (25%), but intermittent dosing of AZD4573 exhibited 94% tumor growth inhibition. Still, combination therapy demonstrated a clear benefit as it led to complete tumor regressions with 6/8 mice remaining tumor-free until the end of study (150 days). Together, this work presents supporting evidence that combining cell death inducing agents would be effective at overcoming de novo or acquired resistance associated with monotherapy treatments.


Cidado:AstraZeneca: Employment, Equity Ownership. Boiko:AstraZeneca: Employment. Proia:AstraZeneca: Employment. Woods:AstraZeneca: Employment. Tavana:AstraZeneca: Employment. San Martin:AstraZeneca: Employment. Tron:AstraZeneca: Employment. Shao:AstraZeneca: Employment. Drew:AstraZeneca: Employment.

Author notes


Asterisk with author names denotes non-ASH members.