Glycosylation Single-Cell Transcriptomic Profiling Decodes Driver Mechanism and Genetic Characteristics of Circulating Plasma Cells in Multiple Myeloma

Introduction

Multiple myeloma (MM) is a hematological malignancy characterized by the infiltration of clonal plasma cells in the bone marrow (BM), with heterogeneity in biological characteristics and outcomes. Circulating plasma cells (CPCs) migrating from BM can be detected in the peripheral blood (PB) of virtually all newly diagnosed multiple myeloma (NDMM) patients. The quantification of CPCs is a powerful prognostic factor that redefines high-risk MM. However, the biological features of CPCs and the mechanisms of malignant plasma cell extravasation from BM to PB remain unclear. In this study, we elucidated the heterogeneity and genetic characteristics of CPCs and explored the underlying mechanism of CPC formation in MM.

Methods

We comprehensively analyzed transcriptomes and glycoproteomics in paired BM and PB samples from 4 NDMM patients using glycosylation single-cell RNA sequencing (scRNA-seq). Subsequently, we included paired samples from an additional 10 NDMM patients for further analysis by CyTOF.

Result

Based on glycosylation scRNA-seq analysis, we elucidated the specific transcriptome characteristics and glycan profiles of CPCs and bone marrow plasma cells (BMPCs) on a single-cell level. Genes associated with cell migration (EMP3, AHNAK), cellular adhesion (TYROBP, CD44), angiogenesis (LGALS1), and osteoclastogenesis (ANXA2) were upregulated in CPCs. Conversely, MPO, which facilitates MM progression, was upregulated in BMPCs. Therapeutically relevant target genes (BCMA and FcRH5) were downregulated in CPCs while drug-resistance genes were overexpressed. The glycosylation abundance of BMPCs was higher than CPCs.

To further identify the signatures and investigate the heterogeneity of MM cells, we categorized BMPCs and CPCs into 14 distinct cell clusters (C1-C14). Remarkably, only the C12 subcluster was present in both BM and PB with a relatively higher glycosylation burden. RNA velocity analysis revealed that C12 exhibited a strong directional flow toward other subclusters. Survival analysis based on the MMRF CoMMpass dataset (n = 858) showed that patients with high C12 had an inferior OS compared to low C12 (p < 0.0001). These findings indicate that C12 is the CPC-initiating cell cluster, essential for CPC production and a prognostic factor for survival. Next, we identified a total of 11 gene modules by hotspot analysis. We found that module 9, which highly expressed 21 genes (including PTTG1, CDC20, STMN1, RRM2, and HMGB1), was shared by both BMPCs and CPCs. KEGG enrichment analysis showed genes in module 9 were enriched for the cell cycle. Differentially expressed gene analysis between C12 in BM and PB showed that SPP1 was upregulated in medullary CPC-initiating cells. The high expression of the module 9 gene set in BM is a prerequisite for the production of CPCs, and on this basis, upregulation of SPP1 is the key to PB migration of MM cells. Furthermore, CyTOF analysis substantiated the presence of a subpopulation in CPC-initiating cells, which exhibited a characteristic MM stem cell-like phenotype of CD19+CD27-CD138-.

To enhance our comprehension of CPC trafficking, we have concentrated on the immune microenvironment, intending to provide additional explanations for CPC dissemination. Mean Spearman correlation analysis revealed that MM cells positively correlated with T/NK cells (R = 0.87, p < 0.001). Subgroup composition of T/NK cells showed that BM demonstrated a trend toward a higher proportion of proliferating T cells. Gene signature scores were higher in proliferating T cells estimated by UCell based on annotated genes from the glycolysis, oxidative phosphorylation, hypoxia, and TCA cycle pathways, indicating that these metabolic pathways were highly active in the proliferating T cells. CellChat analysis disclosed that CPC-initiating cells showed the highest interaction numbers with proliferating T cells, including through the pivotal SPP1 signaling pathway.

Conclusion

In this study, we employed multi-omics techniques to precisely analyze the heterogeneity of CPCs from multiple perspectives of the transcriptome and proteome. Our results demonstrated that the circulation of MM was driven by specific CPC-initiating cells with unique genetic profiles and interaction with immune microenvironment cells.

No relevant conflicts of interest to declare.