Bone marrow NK cell profile predicts MRD-negativity in multiple myeloma patients treated with daratumumab-based therapy
Description
Single-cell transcriptomics was performed to investigate the bone marrow NK cell compartment of myeloma patients at diagnosis (n=19) and healthy controls (n=5). We show reduced proportion of cytotoxic NK cells in a subset of MM patients at diagnosis, which correlated with decreased cytokine production and NK cell degranulation against MM cells in the presence of the anti-CD38 antibody daratumumab. In line with these findings, a low proportion of CD16+ bone marrow NK cells at diagnosis was associated with a reduced likelihood of achieving MRD-negativity post-consolidation in patients treated with daratumumab, bortezomib, thalidomide and dexamethasone in conjunction with autologous stem cell transplantation in the CASSIOPEIA trial. These findings highlight the impact of the bone marrow NK cell compartment on therapeutic outcomes in MM patients receiving immunotherapy with CD38-targeting antibodies.
Files
Steps to reproduce
Important notices NK cells were identified in silico from a CD38-positive and CD38-negative immune dataset previously generated at Myeloma Research Rotterdam (De Jong et al. Nat Immunol. 2021 Jun;22(6):769-780) which was further supplemented by 6 NDMM patient samples. The CD38pos.zip folder contains CD38-positive dataset, while the CD38neg.zip folder contains the CD38-negative dataset. For both folders seperate subfolders exist for controle bone marrow and NDMM BM. Within the source subfolder there are subfolders per patient which contain the filtered_feature_bc_matrix output directories for the immune dataset. We first processed each individual patient separately prior to combining data from multiple patients; low-quality cells were excluded. Datasets from individual patients were then merged by integration in Seurat. Please note that sample MM10 was excluded for the CD38-negative dataset. NK cells were identified based upon transcription of KLRF1, KLRD1, NKG7 and GNLY; these cells were then subsetted. We checked well-defined marker genes for each cluster to identify potential contaminating cell populations such as T cells (CD8A, CD8B, CD4), B cells (MS4A1, CD19, VPREB1), plasma cells (SDC1, LAMP5, SLAMF7) and myeloid cells (LYZ, CD14, FCGR3B, ELANE, FCER1A, CD1C). One cluster contained an admixture of cell types and was removed prior to subsequent analysis. Of note, we refrained from further analyzing the CD38-negative NK cell dataset due to the limited number of cells (average 202 cells/patient). A basic script is included for R analyses. More details regarding the scripts generated during this study can be found at https://github.com/MyelomaRotterdam/Tahri-et-al.2025.git With any questions, please contact us at t.cupedo@erasmusmc.nl or s.tahri@erasmusmc.nl