Background
Despite recent improvements in treatment strategies, including rituximab, proteasome inhibitors, bendamustine, and BTK-inhibitors, Waldenström macroglobulinemia (WM) remains an incurable disease [
1,
2]. Our understanding of the pathogenesis and immune escape mechanisms of WM is still limited [
3]. The clinical onset of WM is often characterized by the development of anemia and progressive tumor infiltration, highlighting the importance of tumor infiltration in disease development and progression [
4]. Genetic analyses have demonstrated recurrent mutations of the myeloid differentiation primary response-88 (MYD88) gene in more than 87.5% of WM patients in our lymphoma center of Blood Disease Hospital, CAMS [
5]. However, MYD88 mutations are neither specific nor sufficient for the pathogenesis of WM and can be detected in IgM monoclonal gammopathy of undetermined significance (MGUS) as well as in other B cell lymphomas [
5‐
7]. The mechanisms underlying the pathogenesis as well as the cellular origin of WM remain poorly understood [
8]. Recently, the immune microenvironment has emerged as a promising therapeutic target in hematologic malignancies [
9], with the reciprocal influence underlying the co-evolution of tumor cells and immune cells. However, studies on the immune microenvironment in WM are currently limited to the evaluation of only a few parameters [
10‐
12]. An in-depth exploration of the immune cell dysfunction and the co-evolutionary with WM cells may reveal opportunities for effective intervention.
The aberrant T-cell marker expression on tumor cells of WM was revealed in our previous study [
13]. However, the molecular characteristics and the underlying function of these WM cells have not been fully investigated. High-throughput single-cell RNA sequencing (scRNA-seq) is a powerful approach for studying heterogeneous tissues and kinetics cellular processes [
14,
15]. Here, scRNA-seq analysis of BM mononuclear cells (BMNCs) from five newly diagnosed WM patients (NDWMs) and six healthy donors (HDs) were studied, and the heterogeneity of WM cells and non-malignant cells in the BM microenvironment was investigated. The aberrant expression of T cell markers was confirmed at the protein level in WM-B cells and WM-plasma cells by multicolor flow cytometry analysis. Pseudotime-ordered analysis elucidated that CD19
+CD3
+ malignant cells were presented at an early stage of B cell differentiation. Colony formation assay further identified that CD19
+CD3
+ malignant WM-B cells acted as potential WM precursors. Additionally, our results elucidated that the immunosuppressive states of T cells and dysregulated natural killer (NK) cells are highly correlated with WM cell infiltration in the BM microenvironment. Notably, exhausted cytotoxic CD8
+ T cells with diverse stages were discriminated in the WM BM according to their distinct transcriptional characteristics. The precursor (pre-) exhausted CD8
+ T cells with co-expression of exhaustion and cytotoxic markers were identified, and they were more responsive to immune therapies than the terminally exhausted CD8
+ T cells. Altogether, our study facilitates further understanding of the biological heterogeneity of tumor cells and immunosuppressive microenvironment in WM. These data may have implications for the development of novel immunotherapies, such as targeting pre-exhausted CD8-T cells in WM.
Discussion
The opportunity to study entire transcriptomes in detail using scRNA-seq has fueled many important discoveries in the pathogenesis of malignancy, including WM [
30]. Despite recent advances in the genetics of WM tumor cells, current information about the immune milieu in WM as well as the cell origin of this malignancy remain poorly understood [
31]. There is a very high prevalence of MYD88
L265P in patients with WM, providing a genetic marker of the disease [
32,
33]. However, many IgM MGUS patients with MYD88
L265P do not develop a B cell malignancy [
34]. Progression from MYD88
L265P to malignant lymphoma may be fundamentally driven by the emergence of cooperating genetic alterations and the tumor microenvironment [
35]. Our recent report indicates that MYD88
L265P is a recurrent mutation in WM patients in China [
36], and all patients included in present study harbored MYD88
L265P mutation. Here, scRNA-seq analysis provides a comprehensive single-cell transcriptomic atlas to characterize cellular ecosystems in WM BM. We firstly delineated a novel model for the ecosystem of WM, wherein tumor cells and immune cells co-evolve kinetically, and clarified an aberrant immune suppressive milieu. These would provide novel insights into the disease development and progression.
Our previously published study reported the aberrant T-cell marker expression on tumor cells of WM by gene expression profiling [
13]. The current study further confirms the finding of CD19
+CD3
+ malignant WM cells at single-cell resolution and validates at the protein level by flow cytometry analysis. Besides characterizing the three subgroups of canonical malignant cells in WM including B cells, plasma cells (PC), and plasmacytoid lymphocytes, two novel tumor cell subpopulations were identified at single-cell resolution with distinct transcriptomic profiles including CD19
+CD3
+ and CD138
+CD3
+ cells. This is the first time identified CD138
+CD3
+ tumor cells in WM. The CD138
+CD3
+ tumor cells also expressed high levels of IGHM, CD27, and XBP1, but were absent of CD19, CD22, CD24, and CD10 expression. Consistent with Catherine et al
. recent report that MYD88
L265P mice exhibited very early increased IgM PC differentiation [
37]. Continuous MYD88 activation is associated with expansion and the transformation of IgM-differentiating plasma cells. Additionally, our data demonstrated that canonical malignant plasma cells and CD138
+CD3
+ cells mainly presented in WM patients with low malignant B cell infiltration, and were almost absent in high-malignant B cell-infiltration ones. These findings support that a large number of malignant B cells will interfere with plasma cell differentiation in WM microenvironment, and further causes the de-differentiation of plasma cells. Tumor cell architecture involves in the malignant transformation, and the plasma cell differentiation was impaired in patients with higher malignant B cell infiltration.
In agreement with the flow cytometric pattern described previously, less mature CD138
+PAX5
+ plasma cells were significantly more abundant in WM than in marginal zone lymphoma (MZL) or plasma cell myeloma (multiple myeloma or MM). Conversely, more mature CD138
+IRF4
+ cells were rare in WM relative to MZL and myeloma [
38]. Compared with malignant plasma cells (WM-PC), malignant B cells (WM-BC) present the survival advantage in WM. Therefore, our data support the notion that the optional treatment for WM is needed to rapidly eliminate malignant B cells [
39].
CD3 and CD19 are the characteristic surface markers of mature T lymphocytes and B lymphocytes in human, respectively. Rizwan et al
. recently reported a special subset of immune cells that characteristically dual expresses key lineage markers of both B and T cells (CD19
+CD3
+ cells) in type 1 diabetes patients, which proposes stimulating autologous CD4-T cells and may contribute to autoimmunity [
40]. Additionally, CD19
+CD3
+ cells were also discovered in many types of cancer and would be a potential novel tumor immune marker as our previous study reported [
13,
41]. Although the etiology of WM is unknown, recent research advances have all implicated autoimmune and chronic inflammatory conditions in the causation of the disease [
42‐
44]. In this study, this malignant subpopulation was confirmed by the expressions of clonotypic kappa or lambda light chain/IGHM/CD22/CD27 and negative for CD10, CD24, and CD38 by scRNA-seq. This immunophenotype matches memory B cells (smIgM
−/+/CD10
−/CD19
+/CD20
+/CD27
+/CD38
low+/CD45
+) [
8]. Further analysis demonstrated that CD19
+CD3
+ malignant cells present an early stage of B cell differentiation by pseudotime analysis compared with CD19
+CD3
− canonical B cells. Colony formation assay supported the progenitor cell features of CD19
+CD3
+ malignant cells. These findings propose that CD19
+CD3
+ cells may act as potential WM precursors. Tracing the cell of origin is one of the major directions of WM research [
45,
46], as its identification would allow us to understand the development of the disease and uncover potential therapeutical targets [
47]. Kaushal et al. recently reported that CD19
+CD10
+ pre-B cells and CD19
−CD34
− pro-B cells harbor the phenotypic and molecular signatures of the malignant Waldenström clone [
30]. We speculate the specific molecular signatures of WM, including MYD88 mutation, +4, 6q-, +12, and +18q would be already harbored in CD19
+CD3
+ cells, the potential progenitor cells of WM. Due to the small proportion of CD19
+CD3
+ and CD138
+CD3
+ among malignant cells, the cell origin and the pathophysiological functions including MYD88 state of them were not analyzed in the present study, which would be further investigated in our future study.
Cancer immune evasion is a major stumbling block in designing effective anticancer therapeutic strategies [
48]. Consistent with previous report by Beltra et al. we identified four exhausted phases of CD8
+ T cells, especially the precursor exhausted CD8
+ T cells in the immunosuppressive microenvironment of WM for the first time [
49]. Understanding the features of and pathways to T cell exhaustion has crucial implications for the success of checkpoint blockade therapies [
50]. Kallies et al. recently reported that these precursor exhausted T cells are responsible for the proliferative burst that generates effector T cells in response to immune checkpoint blockade targeting programmed cell death 1 (PD1), and increased pre-exhausted cell frequencies have recently been linked to increased patient survival [
51]. Our results showed malignant cells highly expressed immune checkpoint molecules CD47 and CD48, and strongly interacted with pre-exhausted T cells via CD47-LGALS9 and CD47-SIRPG molecules. Since pre-exhausted T cells re-engaged some effector biology and increased in response to immune checkpoint blockade [
49,
52,
53], our results support that CD47 would be a potential therapeutic target to reverse CD8
+ T cells cytotoxic dysfunction in WM. Much progress including ongoing clinical trials has been made in targeting CD47 for cancer immunotherapy in solid tumors and hematological malignancies [
54‐
56]. Since low-tumor-infiltration patients displayed a high level of pre-exhausted CD8
+ T cells compared to high-tumor-infiltration ones, we speculate that lower-tumor infiltration WM patients would respond well to immune therapies such as anti-PD-L1 antibodies according to Kallies et al. [
51] report. However, CD8-EOMES T cells, the terminally exhausted cells, were resistant to reinvigoration by PD-L1 blockade [
49,
52,
53]. Our data demonstrated that metabolic reprogramming of terminally exhausted CD8
+ T cells by IL-10 efficiently enhances anti-tumor immunity [
57]. Of note, we found that terminally exhausted CD8
+ T cells displayed a dysregulated metabolism, and the IL-10 signal pathway was notably enriched in them, indicating that reprogramming metabolic profiles may be essential for reactivating CD8
+ T cells in WM. Therefore, these findings support the potential clinical value of IL-10-related therapy in WM, especially for patients with high tumor infiltration. In addition, WM malignant cells strongly interacted with NK and CD8
+ T cells via the CD74-MIF axis. Blocking the CD74-MIF axis potentiates CD8
+ T cell infiltration and drives pro-inflammatory M1 conversion of macrophages in the tumor microenvironment [
58]. Thus, CD74 would be another potential therapeutic target for reversing the NK cell dysfunction in WM.
Although we described the aberrant status of immune cells in the WM microenvironment, including exhaustion of CD8+ T cells and functional depletion of NK cells, the fundamental mechanisms inherent to immune cell dysfunction need to be further clarified. Further studies are needed to unravel the underlying mechanisms and provide more potential strategies to reverse the immunosuppressive microenvironment in WM. This would help guide better-tailored therapy strategies and benefit the long-term controlling of WM.
In summary, in the present study, the biological heterogeneity of malignant cells and the altered dysfunctional states of immune cells were further uncovered in WM. This integrative analysis provides novel insights into the pathogenesis of WM and helps the development of novel therapies to benefit patients.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.