Single-cell transcriptome analysis reveals stem cell-like subsets in the progression of Waldenström’s macroglobulinemia

To the editor,

Waldenström’s macroglobulinemia (WM) is an uncommon lymphoproliferative disorder characterized by monoclonal immunoglobulin M protein production and bone marrow infiltration by lymphoplasmacstetytoid cells [1]. It remains an incurable disease. IgM monoclonal gammopathy of undetermined significance (MGUS) is the premalignant condition and is associated with the risk of developing WM [2]. MYD88^L265P is a recurrent mutation found in more than 90% of WM patients, and this genetic event is detectable in at least half of IgM MGUS patients [3]. However, the precise cellular landscape and the mechanisms of progression from IgM MGUS to WM remain unclear. Single-cell RNA sequencing (scRNA-seq) provides us a powerful approach to explore tumor heterogeneity and identify rare malignant cells [4]. Anti-CD38 monoclonal antibodies have been introduced into the therapeutic arsenal for plasma cell diseases [5]. Therefore, we analyzed CD38 + immune microenvironment together with B cells and plasma cells to illustrate the first cellular landscape of IgM MGUS and WM in single cell resolution.

We isolated bone marrow CD19 + and CD19-CD38 + cells from WM patients(n = 3), IgM MGUS patients(n = 3) and healthy donors(n = 3) and employed the 10 × Genomics platform to perform single-cell transcriptomic sequencing. After quality control, our dataset contained a total of 73,024 cells. We used uniform manifold approximation and projection (UMAP) to visualize the cell superclusters and identified clusters including B cells (CD19, MS4A1, and CD79A), plasma cells (SDC1, CD38, and PRDM1), CD3 + CD20 + CD19 + cells, T cells (CD3D, CD8A, and CD4) and NK cells (NCAM1, GNLY, and NKG7) based on the expression of canonical lineage markers and cluster-specific markers (Fig. 1A and Additional file 2: Figure S1A, B). We found that healthy donors had the highest percentage of B cells (72.7% in healthy donors, 19.0% in IgM MGUS patients, 60.5% in WM patients; p < 0.001) and that IgM MGUS and WM patients had more NK and T cells than healthy controls (0.06% and 17.7%, respectively, in healthy donors; 24.1% and 20.5% in IgM MGUS patients; 8.5% and 24.0% in WM patients; p < 0.001) (Fig. 1B). We also noticed that CD3 + CD20 + (CD19 +) cells were present in both IgM MGUS and WM patients (Additional file 2: Figure S1C–E).

First, we investigated the cell clusters correlated with WM pathology including B cells and plasma cells. We identified pre-B cells, pro-B cells and mature B cells based on the progenitor cell marker CD34, pre-BCR genes IGLL1 and VPREB1, and B-lymphoid lineage surface markers MME, CD38, FCER2 and CD27. We compared the B-cell gene expression profiles after removing the early B cells (Additional file 2: Figure S1F). Compared with those of IgM MGUS patients, mature B cells of WM patients showed upregulation of HES1, GADD45B, NEAT1, DUSP22, RGS1, RGS16, and PIM1 (Fig. 1C and Additional file 3: Table S1). Light chain restriction and higher copy number variants (CNV) levels were found in B cells of WM patients (Fig. 1C, D). SCENIC analysis showed that FOS, FOSB, EGR1 and JUNB were downregulated in B cells from WM and IgM MGUS patients (Fig. 1E). REL, a proto-oncogene promoting the survival and proliferation of B lymphocytes, was downregulated in B cells of WM patients, and IRF8 and ELF1 were upregulated in B cells of IgM MGUS patients (Fig. 1E). Nevertheless, plasma cells of WM patients, IgM MGUS patients and healthy donors were colocalized in one cluster, which might indicate their similar gene expression patterns. The significant differentially expressed genes (DEGs) across the groups are highlighted in scatter plots (Fig. 1F) and showed in Additional file 4: Table S2. In particular, we noticed that the expression levels of part of genes decrease in IgM MGUS patients and then rise again when the disease progresses to WM. We assumed that the regulation might be correlated with the progression from IgM MGUS to WM.

Next, we further determined the characteristics of the rare CD3 + CD19 + cells. Trajectory analysis revealed that CD3 + CD20 + (CD19 +) cells and pre-B cells were located at the same end of the branched structure and were then directed toward mature B cell and plasma cell fates, respectively (Fig. 2A). This inferred developmental trajectory suggested that CD3 + CD20 + (CD19 +) cells may be tumor stem cell-like subset. Using flow cytometry, we validated that CD3 was expressed in a fraction of CD19 + cells (mean: 3.69%, range: 0.28%-15.10%) of bone marrow samples from 8 WM patients (Fig. 2B, Additional file 2: Figure S2A and Additional file 5: Table S3). We compared the gene expression of CD3 + CD20 + CD19 + cells between WM patients and IgM MGUS patients (Additional file 6: Table S4). Genes including IGKC, JCHAIN, AREG, RGS1, IGHM were significantly upregulated in CD3 + CD20 + CD19 + cells of WM patients. In addition, targeted gene sequencing (TGS) was performed on CD3 + CD19 + and CD3-CD19 + cells from 6 WM patients (Additional file 7: Table S5). MYD88 were proved to be the most frequently mutated gene in this cell subset (Fig. 2C). And we inferred that MYD88 mutation might be the early events in tumorigenesis by variant allele fraction (VAF) analysis (Fig. 2D, E). Additional subclonal hits, such as CXCR4 and MAP2K1 mutations, could be acquired during tumor progression.

We observed higher proportions of T lymphocytes and NK cells in WM patients; thus, CellChat was employed to explore the cell–cell communication between immune cells and CD3 + CD20 + (CD19 +) cells in WM patients. We observed that the CXCR4-CXCL12 ligand-receptor pair was enriched in the interaction between CD3 + CD20 + cells and B cells (Additional file 2: Figure. S2B, C). The CCL signaling pathway were highly expressed in CD8 + T cell/NK-cell—CD3 + CD20 + cell interactions (Additional file 2: Figure. S2B, C). In addition, IL2 signaling, as well as TGF β signaling, were also involved in communication networks between CD3 + CD20 + cells and T/NK cells (Additional file 2: Figure. S2B, C).

In conclusion, our study provides comprehensive insights into mechanisms of progression from IgM MGUS to WM. We identified the rare CD3 + CD19 + cell subpopulation in WM patients. It will be interesting to explore novel therapeutic strategies targeting rare potential cells with “stemness” in future.