Respective Place of Venetoclax and MCL1 BH3 Mimetics in Multiple Myeloma Treatment

Targeting anti-apoptotic proteins of the BCL2 family by BH3 mimetics is a new promising therapeutic approach in multiple myeloma (MM). The specific BH3 mimetic targeting BCL2, BCLXL or MCL1 trigger apoptosis and exploit the dependency on these different anti-apoptotic proteins to kill tumor cells. Because MM is mostly considered as dependent on MCL1, the recent clinical availability of MCL1 BH3 mimetics underlines an urgent need to better define patients that would benefit from a MCL1 targeted therapy. In the present study, we used a BH3 mimetic toolkit that includes venetoclax, A1155463 and A1210477, which target BCL2, BCLXL and MCL1 respectively to define dependencies/co-dependencies in a large cohort of 60 myeloma patients (21 at diagnosis and 39 at relapse). Alternatively, MCL1 dependency was confirmed using the S63845 MCL1 inhibitor in MM patient samples. Mononuclear bone marrow/blood cells were treated overnight with the respective BH3 mimetic and cell death was specifically measured in the tumor cell population. Primary MM cells dependencies were stratified using PCA analysis in three groups as highly dependent, intermediately dependent or not dependent. Our study demonstrated that half of patients at diagnosis were BCL2 dependent while only 10% were BCLXL dependent. The dependence on BCL2 or BCLXL was not significantly different between samples at diagnosis and relapse. Strikingly, we found that the MCL1 dependency was 33% at diagnosis while it was 69% at relapse, suggesting a significant increase in MCL1 dependency during the disease progression (p=0.01). Besides, 36% of overall patients showed co-dependencies on BCL2/MCL1. We also identified primary MM cells that did not depend on any of the three pro-survival molecules, both at diagnosis and relapse.

Among this cohort of MM patients, 47 samples were further analyzed for the presence of recurrent translocations (t(11;14), t(6;14), t(4;14) and t(14;16)) allowing the analysis of dependencies in the different subgroups; these recurrent translocations lead to the overexpression of CCND1, CCND3, MMSET and MAF oncogenes, respectively. We found that BCL2 dependency was significantly higher in CCND1 subgroup (83%) compared to all other subgroups (20%, p=0.008). We also confirmed the BCL2/BCLXL mRNA ratio as a valuable biomarker to define BCL2 dependence (p=0.0001). At diagnosis, MCL1 dependency was absent in patients not harboring the common recurrent translocations while at relapse 6 out 9 patients not harboring the recurrent translocations were MCL1 dependent, indicating an increase of MCL1 dependency at relapse in this subgroup (p=0.03).

Mechanistically, we demonstrated that BAK is crucial for cell death induced by MCL1 mimetic A1210477, according to the protection of cell death observed by BAK knock-down and the complete disruption of MCL1/BAK complexes upon A1210477 treatment, observed in MM cell lines and also in a patient sample. Interestingly, this complex was also dissociated in A1210477 resistant cells but free BAK was simultaneously recaptured by BCLXL, supporting the role of BCLXL in A1210477 resistance. Thus, BCLXL may act as a sink to bind freed pro-apoptotic proteins from MCL1 and limits MM cell death triggered by the specific targeting of MCL1.

In conclusion, our study highlights the potential clinical use of BH3 mimetics in MM treatment guided by the practical ex-vivo testing of myeloma cell dependencies using the BH3 toolkit. This strategy could be used to identify the respective and tailored use of venetoclax, MCL1 BH3 mimetics or their combination in myeloma treatment.