Preclinical Activity of Novel MCL1 Inhibitor AZD5991 in Multiple Myeloma

We and others have previously demonstrated that MM is often dependent on MCL1 or co-dependent on MCL1 and BCLXL or BCL2 for survival. Therefore, drug development targeting MCL1 has been a top priority. Here we report on AZD5991, a specific small molecule inhibitor of MCL1. We treated 17 MM cell lines with increasing concentrations of AZD5991 for 24 h and measured Annexin V staining to determine the IC50s. Nine of the cell lines tested were highly sensitive to AZD5991 with IC50 values below 100 nM, 6 lines exhibited an intermediate sensitivity (IC50 100-1000 nM), and only 2 cell lines tested were resistant (IC50 >1000 nM). Six of the highly sensitive lines are t(11;14) and sensitive to venetoclax suggesting co-dependence on BCL2 and MCL1 for survival.

We also determined the effect of the bone marrow microenvironment on the response of MM cell lines to AZD5991. We reported that IL-6 protects MM cell lines and patient samples from apoptosis by making the cells more MCL1 dependent. Based on this, we predicted IL-6 would have little to no effect on AZD5991-induced cell death. We treated 12 cell lines with AZD5991 in the presence of 1 ng/mL IL-6 or 10% Hs5 conditioned medium (CM) for 24 h and found that only 3/12 and 2/12 lines were protected from apoptosis in the presence of IL-6 and CM, respectively. Interestingly, when co-cultured with the stromal cell line Hs5, 7/11 lines tested were protected from AZD5991-induced cell death, suggesting cell-cell contact is influencing the response. This is in contrast to ABT-737 and venetoclax where cell-cell contact provided no additional protection than CM.

Mechanistically apoptosis induced via MCL1 inhibition is not dependent on BIM expression as is the case with BCLXL and BCL2 inhibition. KMS26 and LP1 MM cell lines contain a bi-allelic deletion of BIM and we have reported their resistance to ABT-737. However, both cell lines respond to AZD5991 with IC50 values in an intermediate sensitivity range. Co-immunoprecipitation (CoIP) studies were employed to determine the protein bound to MCL1 that could be promoting apoptosis upon release. We found NOXA and BAK bound in KMS26 and LP1 and both were released from MCL1 in response to AZD5991. Additionally, CoIPs performed on cell lines expressing BIM showed NOXA, BIM, and BAK bound to MCL1 and released following treatment. To further investigate we used CRISPR-cas9 to generate MM cell lines lacking expression of NOXA, BAK, BAX, or BIM. In KMS26 and LP1, deletion of NOXA and BAX had little effect on AZD5991-induced cell death while the BAK deletion significantly inhibited apoptosis in both cell lines. Similar results were observed in the BIM expressing cell line OCI-My5, with no protection from AZD5991-induced apoptosis in the NOXA and BAX edited lines, significant protection in the BAK-deleted line, and an intermediate degree of protection in the BIM knockout line. In KMS18, BIM deficiency had a minimal effect on apoptosis following MCL1 inhibition, however both BAX and BAK were required for AZD-induced cell death.

Additionally, we have tested 41 samples from 37 patients for sensitivity to AZD5991. Samples were treated with increasing concentrations to determine IC50 values in the same manner as the MM cell lines. The samples segregated into 4 groups based on IC50. The most sensitive group (N=3) had an IC50 below 10 nM. The largest group had an IC50 range of 50-114 nM (N=26). The last two cohorts were more resistant with a range of 500-916 nM (N=10) and 2 samples with an IC50 over 1300 nM. Since MCL1 is on 1q21, a frequently amplified region in MM, we determined if 1q21 gain was associated with sensitivity. For the 35 samples where FISH data were available, 18 had 1q21 gains by FISH while 17 were negative. There is a trend for the 1q21 gain cohort to be more sensitive (P=0.0573), with only 2/18 having an IC50 above 109 nM. In contrast for the 1q21 negative 7/17 were in the resistant groups. Thus 1q21 may be a marker of sensitivity to MCL1 inhibitors.

The data reported here demonstrate that AZD5991 is effective at inducing apoptosis in MM and can overcome soluble microenvironment resistance factors that influence the response to venetoclax. This appears to be due to differential requirements for pro-apoptotic factors for BCL2 and MCL1 inhibition and suggests an underappreciated complexity in the role of BCL2 and MCL1 in cell survival. Finally these findings also suggest that 1q21 gain may be a marker for AZD5991 sensitivity. A clinical trial is currently ongoing in myeloma.