Multiple myeloma is a malignancy of long-lived plasma cells of the bone marrow that is rarely curable. Thus, despite recent advances in the development of new therapies, additional approaches are required. We investigated potential molecular vulnerabilities in the BCL2 family. Using the MMRF CoMMpass (NCT0145429) study (IA13), we determined the frequency of nonsynonymous coding mutations in the BCL2 family. Analysis of baseline samples from 982 patients revealed that mutations in the BCL2 family are relatively rare events. No mutations were observed in the 3 pro-apoptotic effector genes BAX, BAK1 and BOK. Similarly, in the BH3-only genes, mutations were rare with no mutations in BCL2L11 (BIM), BAD, BID, HRK and BMF and only single mutations in BBC3 (PUMA) and PMAIP1 (NOXA). In the anti-apoptotic BCL2 genes, mutations were also rare with no mutations in BCL2, BCL2L1 (BCLX) and BCL2L10 (BCLB). A single sample had a mutation in BCL2A1 (A1) while 2 samples had mutations in BCL2L2 (BCLW). Interestingly MCL1 was mutated in 10 baseline samples (1.02%) and the frequency of the mutations in these samples was high (median 0.391, range 0.066-0.531). Therefore, we further investigated the functional consequences of the MCL1 mutant alleles.
Of the 10 mutations detected, 1 was in the N-terminal region (G32R) and 4 were in the PEST domain in the N-terminal half of MCL1 that is associated with regulating protein stability (V140I, P142S, E149Q and E173K). An additional mutation was found in an uncharacterized region between the PEST and BH1 domains (L186F). We focused on the 4 mutations that lie within or near the functional BH1 (V249L and L267V) and within the BH3 (N223S, and R214Q) domains. Wild type (WT) MCL1 and the four mutant MCL1 constructs were introduced into a murine B-ALL cell line that has endogenous murine MCL1 flanked with LoxP sites and confirmed expression by western blot analysis.
Human MCL1 can replace murine MCL1 in this cell model, therefore we are determining if the myeloma-derived mutants of MCL1 can complement loss of mouse Mcl1 and will report on these findings. However, we began to functionally characterize these mutations by taking advantage of an anomaly in the development of inhibitors of human MCL1. To date, no inhibitor developed against human MCL1 is as effective against murine Mcl1. Thus, differences in the activity of these inhibitors reflects changes in dependence on the introduced human MCL1. We treated cells with increasing concentrations of the MCL1 inhibitor S63845 and measured cell death (Annexin V/PI) at 24 hours. As expected, cells where the empty vector was introduced were highly resistant to S63845-induced cell death (less than 20% at 1000 nM) while cells expressing the human WT MCL1 were significantly more susceptible to the MCL1 inhibitor across a concentration range of 100 to 1000 nM. The V249L, N223S and R214Q mutations mimicked the sensitivity of the WT MCL1 suggesting they did not alter MCL1 function in the cells. In contrast the L267V mutation resulted in a dose curve that was more similar to the empty vector control suggesting this mutation either resulted in loss of function or in an MCL1 molecule that could not be inhibited by the drug. Since drug binding can stabilize MCL1 by competing for E3 ligase binding, we determined the effect of S63845 on MCL1 protein levels. We found that with all the mutants, S63845 dramatically increased human MCL1 protein expression ruling out lack of drug binding. We next performed MCL1 co-immunoprecipitation assays, and found that BIM release correlated with S63845 sensitivity. In addition to not releasing BIM, the L267V did not effectively release NOXA and BAK after S63845 treatment. Taken together, the L267V mutation does not prevent the binding of S63845 to free MCL1, rather it blocks the ability of drug to displace pro-apoptotic proteins required to induce cell death.
We then tested another MCL-1 inhibitor in clinical trials, AZD-5991. The L267V mutation was completely resistant to AZD-5991-induced apoptosis despite evidence of drug binding. Interestingly the other 3 mutations also resulted in diminished killing activity when compared to cells expressing the WT MCL1, suggesting that these mutations may also influence drug function. Together these data suggest that myeloma-derived mutations in MCL1 may not necessarily influence MCL1 function, however they could alter responses to an emerging class of inhibitors where 3 drugs are currently in clinical trials.
Secrist:AstraZeneca: Employment. Cidado:AstraZeneca: Employment, Equity Ownership. Tron:AstraZeneca: Employment. Lonial:Amgen: Research Funding. Boise:Abbvie: Consultancy; AstraZeneca: Honoraria.
Asterisk with author names denotes non-ASH members.