P-221: Plasma cell lineage and driver heterogeneity associated with multiple myeloma progression

Introduction: Multiple myeloma (MM) is characterized by aberrant clonal plasma cell growth in the bone marrow (BM). Despite advances in therapeutic options and identification of prognostic factors, patients still relapse. We aim to delineate tumor heterogeneity and map cell state transitions between B and plasma cells (PCs) to understand tumor intrinsic (genomic drivers) and extrinsic (BM microenvironment) changes facilitating MM progression.

Methods: We studied PCs from patients in the Multiple Myeloma Research Foundation (MMRF) CoMMpass study, integrating data from CD138-sorted and unsorted sequencing datasets to reveal the landscape of PCs in MM. BM Mononuclear cells were sorted into a CD138 enriched fraction for whole exome sequencing (WES), whole genome sequencing (WGS), and bulk RNA-seq to characterize the tumor genomic and transcriptome. We utilized single-cell (scRNA-seq) from 361 samples (263 at diagnosis) on the non-enriched remaining cells to characterize the BM microenvironment; We characterize these PCs, their unique features relative to bulk RNA expression, and interactions with genomic drivers. This dataset was further integrated with external unsorted datasets including normal BM (n=12), monoclonal gammopathy of undetermined significance (n=3), smoldering MM (n=12), and newly diagnosed MM (n=18).

Results: Within the scRNA-seq, we detect an average of 10% of PC cells across patients, totaling ~90,000 cells. By integrating findings from matched WES data, we mapped somatic drivers and copy number alterations to over 21 transcriptionally distinct plasma cell clusters. We identified a specific PC cluster with significantly more somatic mutations, including 64% non-driver mutations and another cluster with CCND1 mutations in 38.9% of its cells. Transcriptionally, CCND1 expression is higher in non-progressors than rapid progressors, highlighting the prognostic value of CCND1 alterations in MM. Relapsed patients (n=50) exhibited an enrichment of a PC cluster harboring DIS3 mutations, compared to baseline patients. Analyzing over 900 cases based on WES, we found TP53 (10.4% vs. 1.8%) and BRAF (9.4% vs. 4.8%) mutations more enriched in White than Black patients, a pattern also observed in scRNA-seq data. These findings suggest an epidemiological contribution to MM predisposition influencing prognosis and treatment outcomes. Finally, we integrated genomic alteration mapping and trajectory analysis, revealing two PCs exhibiting joint expression of B and plasma markers and genomic alterations, suggesting a potential early/precancer state.

Conclusions: In summary we identified distinct PC subclusters with unique transcriptional and mutational profiles. Further by incorporating external datasets, we traced the B to plasma cell trajectory and characterized novel pre-cancer states at single-cell resolution. Understanding the interplay between genomic drivers and the BM microenvironment will reveal novel targets and underpinnings of MM disease progression.