Unraveling the Heterogeneity of Multiple Myeloma Cells By Single-Cell RNA Sequencing Analysis

Multiple myeloma (MM) is a malignancy of clonal plasma cells with identical variable-diversity-joining (VDJ) region recombination of immunoglobulin loci (called repertoire) and extensive genome or transcriptome heterogeneity. Even though many new treatment modalities have been developed and the prognosis of patients has improved significantly, most cases of MM remain incurable. Recent evidence suggests that non-genetic cell plasticity and changes in cell state underlie the therapy resistance and tumor relapse, but the mechanisms remain largely unknown. Here, we have attempted to understand comprehensive architecture of the entire tumor cell populations using repertoire clonality as fingerprint. First, we subdivided bone marrow samples of primary MM at different stages (n=8) into 11 fractions based on the known surface antigens of MM and performed bulk RNA sequencing (bulk RNA-seq). Analysis of the repertoire revealed that tumor cells were also present in some minor fractions other than the main fraction (Lin-/CD19-/CD38++/CD138+) as the stage progressed. Importantly, cells with clonal repertoire were detected in CD138 negative fraction (Lin-/CD19-/CD38++/CD138-) in most samples (Monoclonal gammopathy of undetermined significance (MGUS) / Smoldering multiple myeloma (SMM) = 4.15±2.28%, Primary MM = 6.78±3.59%, Relapse MM = 7.8±5.82% [mean±95%Cl], n=24). The CD138 positive and negative cells were also morphologically different, suggesting that they were composed by cells of distinct characteristics. Indeed, comparison of the transcriptome data revealed 296 differentially expressed genes (DEGs) between these fractions (250 up-regulated and 46 down-regulated in CD138 negative fractions). Of interest, several pathways related to H3K4 methylation were positively enriched in CD138 negative fraction (p <0.001), which included KMT2A, KMT2B, KMT2C, KMT2D, and ASH1L, suggesting that differential methylation of histone H4 accounts for the heterogeneity of MM. In order to achieve a high-resolution and comprehensive evaluation of the entire tumor cell populations, we performed single-cell RNA sequencing (scRNA-seq) and single-cell VDJ targeted sequencing (scVDJ-seq) concurrently on Lin-/CD38++ sorted cells of 8 bone marrow samples from 7 patients including pre- and post-treatment status samples. This method allowed us to define normal cells and MM cells at the single-cell level by clonal repertoire sequence independent of transcriptome characteristics. We divided 18,031 MM cells into 23 clusters after integration. We first defined CD138 positive and negative MM gene signature using bulk RNA-seq data and then evaluated the scRNA-seq data. Notably, the CD138 axis appeared to clearly subdivide main MM populations into those with CD138 positive and negative MM gene signature. We also found that a part of clusters, which showed CD138 negative MM gene signature, tended to remain at high proportions after treatment. These populations were characterized by low expression levels of MHC class-I components and TNFRSF17/BCMA as well as high expression of MCL1, MALAT1 and NEAT1, which have been previously implicated in treatment resistance, and were present in all samples. H3K4 methylation related genes were also up-regulated in these populations. Furthermore, we found several clusters with unique and interesting characteristics independent of the CD138 axis. Trajectory and velocity analysis suggested that all cell populations transit to each other. Taken together, our results establish the heterogeneity of MM cells by the CD138 axis, which may be characterized by epigenetic plasticity. They also highlight CD138 negative population as a potential cause of treatment resistance and relapse. By using information of repertoire, we have achieved a comprehensive and accurate single-cell analysis of MM cell diversity that was not limited by cell surface antigens or transcriptome characteristics. We are currently analyzing near minimal residual disease (MRD) level samples using the same scRNA-seq methods and the functional and epigenetic properties of CD138 positive and negative MM cells.