P-060: Subclone-specific microenvironmental impact and drug response in refractory multiple myeloma revealed by single cell transcriptomics

Background Relapsed/refractory multiple myeloma (RRMM) is characterized by a high inter- and intratumor heterogeneity and a complex interplay of myeloma cells with the bone marrow microenvironment (BME). Accordingly, there is an urgent need to dissect subclone structure, transcriptional heterogeneity and cellular interactions to unravel the molecular mechanisms underlying drug resistance in RRMM. Methods Single cell RNA sequencing (scRNA-seq) of ~210,000 cells from bone marrow aspirates sorted into CD138+ and CD138– fractions was conducted for 20 heavily pretreated RRMM patients. RRMM subclones were called from the scRNA-seq data based on a copy number aberration (CNA) analysis that was confirmed by interphase fluorescence in situ hybridization and whole genome sequencing. From the scRNA-seq data the composition and abundance of immune cell types in the BME was determined. Interactions of myeloma cells with the BME were characterized by an analysis of the correlated expression of ligand-receptor pairs. Selected findings from the scRNA-seq analysis were validated by flow cytometry. Results Subclones with distinct chromosomal aberrations were reliably identified at the single cell level based on CNAs inferred from scRNA-seq data. The analysis revealed a subclonal 1q-gain (+1q) in 10/20 samples, for which a gene expression signature of recurrently upregulated genes was derived. These +1q subclones frequently expanded during treatment. Furthermore, RRMM cells shaped an immune suppressive BME by upregulation of inflammatory cytokines and close interplay with the myeloid compartment. RRMM cells appear to reprogram the BME by upregulation of inflammatory cytokines and ligands of inhibitory receptors expressed on T and NK-cells. Specifically, the immune cell compartment of RRMM patients displayed an accumulation of PD1+ γδ T-cells and myeloid populations compared to healthy donors or early disease. In addition, we observed a depletion of hematopoietic progenitors in patients with high expression of genes associated with inflammatory signaling in the BME. Upon treatment with immunomodulatory drugs, reprogrammed plasmacytoid dendritic cells expanded. Focusing on patients with +1q, we found an enrichment of a rare M2-like tumor associated macrophage population while GZMB+ NK effector cells were depleted in these patients, indicating that +1q has a distinct effect on the BME in RRMM. Conclusions Our study resolves transcriptional features of subclones in RRMM and mechanisms of microenvironmental reprogramming. The insight gained in our study on the evolution of RRMM heterogeneity and its bone marrow milieu will support the development of novel treatment approaches and potentially guide clinical decision making.