Co-Targeting MCL-1 and BCL-2 Is Highly Synergistic in BH3 Mimetic- and Venetoclax/Hypomethylating Agent-Resistant and TP53 Mutated AML

Venetoclax (VEN), a highly selective BCL-2 inhibitor with limited single-agent activity in AML, has shown encouraging efficacy in combination with hypomethylating agents (HMA). Nevertheless, patients relapse and have limited treatment options. Like BCL-2, MCL-1 plays critical roles in the survival of AML cells and AML stem/progenitor cells. MCL-1 is also a known resistance factor to VEN. Preclinical studies have demonstrated that combined inhibition of BCL-2 and MCL-1 is highly effective in VEN-resistant AML cells. Diverse mechanisms contribute to the resistance to VEN, and likely also to BH3 mimetics targeting MCL-1 that are currently under clinical development in AML. The effectiveness of co-targeting BCL-2 and MCL-1 in the setting of various resistance mechanisms has not been fully explored.

We investigated combined inhibition of BCL-2 and MCL-1 in AML cells resistant to apoptotic stimuli through various mechanisms and demonstrate that co-inhibition of BCL-2 with VEN and MCL-1 with AMG176 synergistically targets AML cells that exhibit intrinsic or acquired resistance to BH3 mimetics in vitro and in vivo. We generated AML cells with acquired resistance to VEN (VEN-R) or AMG176 (AMG-R) by exposing the cells to increased doses of the drug and we also generated the cells genetically overexpressing BCL-2, MCL-1, or BCL-2A1. We found that both VEN-R and AMG-R MV4-11 cells expressed increased levels of MCL-1, BCL-2, and BCL2A1, but decreased BAX. Although BCL-XL levels decreased in AMG-R MV4-11 cells, BAK, PUMA, and BID levels were also markedly lower in these resistant cells compared to the parental controls. VEN or AMG176 as single agents had diminished activity against AML cells with acquired resistance not only to VEN, but also to AMG176 and AML cells genetically overexpressing MCL-1, BCL-2, or BCL2-A1. In addition, we found that TP53 mutated primary AML cells expressed low levels of BAX and that Molm13 cells acquired a TP53 mutation (R248W) expressed lower levels of BAX and were more resistant to VEN, consistent with clinical observations, and they were also more resistant to AMG176. However, when VEN and AMG176 were combined, synergy was observed (combination index < 1). We next treated AML patient samples and found that combined inhibition of BCL-2 and MCL-1 was synergistic in primary AML cells and stem/progenitor cells obtained from patients with various cytogenetics/mutations, including TP53 mutations, and from patients resistant to/relapsed from VEN- or VEN/HMA-based therapy, even when AML cells were co-cultured with bone marrow-derived mesenchymal stromal cells that mimic the bone marrow microenvironment. To demonstrate potential clinical relevance, we developed a PDX model from a clinically-acquired VEN/HMA resistant AML patient and treated the PDX-bearing mice with VEN, AMG176, and the combination. Remarkably, the combination of VEN and AMG176 demonstrated strong antileukemia activities, markedly diminished not only AML blasts but also AML stem/progenitor cells, as determined by CyTOF analysis, and significantly extended survival (median 336 vs 126 d for controls, P < 0001), while VEN (129 d) alone and even AMG176 (131 d) alone had minimal efficacy. Several mice in the combination group survived over 400 d and died probably from old age with only minimal residual leukemia.

In conclusion, we demonstrate the alteration of multiple BCL-2 family proteins contributes to BH3 mimetic resistance that can be overcome by combined inhibition of MCL-1 and BCL-2. The striking effectiveness of co-targeting BCL-2 and MCL-1 in AML resistance to a BH3 mimetic via various mechanisms or to VEN/HMA suggests broad clinical applications of this strategy, and warrants clinical evaluations.