Mitochondrial Priming Is Heterogenous in t(4;14) Multiple Myeloma and Can be Induced By Inhibition of NSD2

Introduction: Therapies for the treatment of multiple myeloma (MM) are primarily based on targeting plasma cell biology, while precision medicine approaches have been disappointing. One exception has been venetoclax in t(11;14) MM. This success is based on the dependence of t(11;14) MM cells on BCL2 for survival. In previous studies we found that all 11 t(4;14)-positive MM cell lines and four out of five patient samples tested were resistant to venetoclax. In this study we set out to better define the dependence of t(4;14) MM on members of the BCL2 family and the role of NSD2 in this process.

Methods: BH3 profiling was performed on 11 MM lines [10 t(4;14), one t(11:14)], and on KMS11 cells with a deletion in the NSD2 non-translocated allele (NTKO, NSD2 high) or the translocated allele (TKO, NSD2 low). Dynamic BH3 profiling (DBP) was performed by treating KMS11 or KMS11-NTKO cells for six days with 1 μM of the NSD2 inhibitor NSD2i-1(Excenen), followed by BH3 profiling. Cell death studies were performed by drug treatment for 24-48 h followed by Annexin V/propidium iodide staining and flow cytometry. For patient samples, cells were stained with Annexin V/anti-CD38 and -CD45. RNAseq was performed eight days post treatment with NSD2 inhibitor.

Results: BH3 profiling of t(11;14)-positive KMS12PE demonstrated a pattern consistent with venetoclax sensitivity, with strong priming as indicated by cytochrome (cyto) C release induced by broad spectrum peptides, BIM, BID and PUMA, little to no release with the MCL1 binding peptide MS1 and little effect with HRK that binds specifically to BCLXL. In contrast the BCL2/BCLXL binding peptide, BAD, indicated priming. Since HRK did not induce cyto C release, the effect of the BAD peptide can be attributed to BCL2 priming. We next performed BH3 profiling on 10 t(4;14) cell lines. Based on response to broad spectrum peptides we observed a range of overall priming with OPM2, H929 and JIM1 showing the greatest priming (≥60% cyto C release with 100 nM BIM peptide). KMS34, KMS18 and KMS26 were the least primed (≤15% release with 100 nM BIM). Heterogeneity was also observed in the response to the more specific peptides. H929, JIM1 and XG7 were highly primed on MCL1(≥59% release with MS1 peptide) while KMS34 and OPM2 were BCLXL primed (≥67% release with HRK peptide). The remaining lines (KMS18, KMS28, KMS26, KMS11) did not show strong priming on a single BCL2 family member, however combining peptides demonstrated priming, indicating co-dependence on more than one BCL2 family member. To determine the role of t(4;14), we performed BH3 profiling on KMS11 NTKO and TKO cells. NTKO cells displayed a similar pattern of dependency as KMS11 parental cells. In contrast TKO cells showed increased overall priming as evidenced by cyto C release by BIM. This priming appears to be on BCLXL and possibly BCL2 as evidenced by increased cyto C release induced by BAD and HRK peptides. Consistent with these findings, TKO cells were found to be more sensitive to venetoclax, the BCL2/BCLXL inhibitor AZD4320 and the BCLXL PROTAC DT2216. To further validate these findings, we performed DBP on KMS11 and KMS11 NTKO to determine the effect of NSD2 inhibition on mitochondrial priming. Consistent with differences in priming between TKO and NTKO, NSD2 inhibition resulted in an increase in cyto C release by BIM (P=0.0574), BAD (P=0.0012) and HRK (P=0.0026) peptides, indicating a shift to BCLXL and possibly BCL2 dependence. To determine the potential mechanism for these changes, RNAseq analysis was performed following NSD2 inhibition in KMS11 cells. After eight days of treatment there was a total of 626 genes with significant changes in gene expression (FDR ≤ 0.01, fold change ≥1.5). BCL2L1, which encodes BCLXL was significantly upregulated (1.586 fold, FDR=1.7x10^-25). Finally, we have previously demonstrated the value of ex vivo testing of venetoclax sensitivity in predicting clinical response. We have now tested 15 t(4;14) samples and surprisingly 5 of them demonstrated ex vivo sensitivity to venetoclax. Importantly the difference in sensitivity could not be explained by differences in chr1 gain/amp.

Conclusions: Mitochondrial priming is heterogenous in t(4;14) MM which likely explains the lack of responses to venetoclax. However, inhibition or loss of NSD2 results in increased priming and sensitivity to BCLXL and possibly BCL2 inhibition raising the possibility of combination therapy for targeting this high-risk form of myeloma.