Postdoctoral Associate University of Calgary, Alberta, Canada
Introduction: The BCL2 inhibitor venetoclax has shown promising results in multiple myeloma (MM) patients harboring the t(11;14). However, among these patients the responses are not universal, suggesting that several factors influence its response. We have previously demonstrated that MM cells adapt to its selective pressure by transitioning from B cell-like to a mature plasma cell-like transcriptional and epigenetic signatures. In this study, we used MM primary samples and cell lines to functionally validate some of the key factors involved in venetoclax resistance.
Methods: Serial BM aspirates were collected from 15 relapsed MM patients; 6 harboring t(11;14) and 9 without the translocation, prior to initiation of therapy and at the time of relapse. All t(11;14) patients were treated with venetoclax. Chromatin accessibility and mRNA profiling of isolated CD138+ MM cells were performed using Chromium Single Cell ATAC and 3' Reagent Kits (10x Genomics) respectively. Pairwise sequencing was performed on NexSeq 500 platform. Cell Ranger, Seurat and ArchR were used for sample demultiplexing, barcode processing, single-cell 3' gene counting and data analysis. Overexpression of MCL1, NOXA, and RUNX1 were induced with lentivirus transduction.
Results: By comparing the epigenome and transcriptome of pre- and post-venetoclax samples from t(11;14) patients, we identified MCL1 as the major event responsible for venetoclax resistance. As such the overexpression of MCL1 in the venetoclax-sensitive KMS12BM induced resistance to venetoclax through the shifting of BIM loading from BCL2 to MCL1. Furthermore, the MCL1 upregulation was associated with downregulation of the apoptotic sensitizer NOXA and low NOXA/MCL1 ratio. Of interest, ectopic expression of NOXA in the venetoclax-resistant U266 led to loss of MCL1 binding to BIM, confirming the important role of NOXA in priming BCL2 dependency. Furthermore, based on scATAC seq data that showed an enrichment of RUNX1 binding in venetoclax-resistant cells we have decided to overexpress RUNX1 in KMS12BM. As expected, we observed a significant reduction in venetoclax-induced cell death in these cells when compared with WT through downregulation of BCL2. Of note, we also observed an enrichment of RUNX1 binding in non-t(11;14) cells when compared with t(11;14). As such, treatment with the RUNX1 inhibitor AI-10-49 in the non-t(11;14) OPM2 cells resulted in overexpression of NOXA and loss of MCL1 binding to BIM confirming the dual effect of RUNX1 on MCL1 dependence by repressing NOXA and BCL2.
Conclusions: Taken together these data underline the complexity of the mechanisms involved in venetoclax resistance and showed that the loss of BCL2 dependency can be due to NOXA downregulation and upregulation of MCL-1 and RUNX-1. Therefore, the use of agents that can prime BCL2-dependency through upregulation of NOXA and shifting BIM loading to BCL2 (RUNX1 inhibitors) could be explored in combination with venetoclax in MM patients who acquire MCL1 dependence following treatment.
