Concise Review: Defining and Targeting Myeloma Stem Cell-Like Cells


  • Masahiro Abe,

    Corresponding author
    1. Department of Medicine and Bioregulatory Sciences, University of Tokushima Graduate School of Medical Sciences, Tokushima, Japan
    • Correspondence: Masahiro Abe, M.D., Ph.D., Department of Medicine and Bioregulatory Sciences, University of Tokushima Graduate School of Medical Sciences, 3–18-15 Kuramoto-cho, Tokushima 770–8503, Japan. Telephone: +81-88-633-7120; Fax: +81-88-633-7121; E-mail:

    Search for more papers by this author
  • Takeshi Harada,

    1. Department of Medicine and Bioregulatory Sciences, University of Tokushima Graduate School of Medical Sciences, Tokushima, Japan
    Search for more papers by this author
  • Toshio Matsumoto

    1. Department of Medicine and Bioregulatory Sciences, University of Tokushima Graduate School of Medical Sciences, Tokushima, Japan
    Search for more papers by this author


Multiple myeloma (MM) remains incurable despite recent advances in the treatment of MM. Although the idea of MM cancer stem cells (CSCs) has been proposed for the drug resistance in MM, MM CSCs have not been properly defined yet. Besides clonotypic B cells, phenotypically distinct MM plasma cell fractions have been demonstrated to possess a clonogenic capacity, leading to long-lasting controversies regarding the cells of origin in MM or MM-initiating cells. However, MM CSCs may not be a static population and survive as phenotypically and functionally different cell types via the transition between stem-like and non-stem-like states in local microenvironments, as observed in other types of cancers. Targeting MM CSCs is clinically relevant, and different approaches have been suggested to target molecular, metabolic and epigenetic signatures, and the self-renewal signaling characteristic of MM CSC-like cells. Stem Cells 2014;32:1067–1073


The introduction of thalidomide, lenalidomide, and bortezomib has significantly improved the response rates and survival outcome in patients with multiple myeloma (MM), and these agents have been integrated into MM treatment. However, MM still remains incurable despite the implementation of novel anti-MM agents, high-dose chemotherapies, as well as immunotherapies, indicating the strong need for continued investigation to clarify the mechanisms of drug resistance and develop innovative strategies. The presence of cancer stem cells (CSCs) or cancer-initiating cells have been demonstrated in various types of cancers and regarded as a predominant cause of drug resistance [1-5]. Likewise, MM CSCs have been postulated and are considered to contribute to disease relapse through their drug-resistant nature. Thus, MM CSCs become among the most important targets in the treatment for MM. However, MM CSCs are still conceptual in many senses, because of the lack of sensitive and reliable methods to identify them. Therefore, there have been a conflicting controversy in defining and characterizing MM CSCs or CSC-like cells. In this review, we will discuss three topics, focusing on defining MM CSC-like cells, niches for MM CSC-like cells, and how to target clonogenic MM cells.

Defining MM CSC-Like Cells

Distinct clonotypic subpopulations have been extensively studied on their potential to develop MM disease. Therefore, we will first discuss the MM-initiating potential of clonotypic B cell and non-B cell plasma cell populations in MM in the subsections Clonotypic B Cells in MM and Clonotypic Plasma Cell Subpopulations in MM, respectively, which has been in conflicting controversy. Next, the characterization of side population (SP) cells in MM will be described in the subsection SP Cells, since SP cells are enriched for MM CSCs and often used as the surrogate of MM CSCs. Lastly, we will mention recent observations on the interconvertible states between distinct clonotypic cells in MM in the subsection Interconversion Between Differentiated and Undifferentiated Clonotypic Cells in MM.

Clonotypic B Cells in MM

MM plasma cells exhibit somatically hypermutated immunoglobulin gene sequences without intraclonal variations, which suggests that MM is clonally originated from a postgerminal center B cell or a more differentiated cell [6-8]. Besides clonal plasma cells, B cell populations in the peripheral blood and bone marrow in patients with MM have been demonstrated to contain clonal cells harboring a tumor cell-specific immunoglobulin gene sequence with somatic hypermutations in complimentarity determining region three or immunoglobulin VDJ rearrangement [9-13]. Thus, MM cells are phenotypically and functionally heterogeneous populations. Clonotypic B cells have been demonstrated to be able to give rise to monoclonal immunoglobulin-secreting plasma cells in vitro [14-16]. By analogy with normal postgerminal B cell differentiation into immunoglobulin-secreting plasma cells, a hierarchical model from such clonotypic B cells has been suggested in the development of MM.

Clonotypic CD19+ B cells isolated and enriched from a portion of patients with MM showed the ability to form a new tumor in in vivo xenograft models, which implies the existence of cells with the capacity of self-renewal in clonotypic CD19+ B cells [14, 17]. Matsui et al. [18] have meticulously dissected clonotypic B cells in MM in terms of their capability of clonogenicity or self-renewal. They found that CD19+CD27+CD138 cells with a memory B-cell phenotype isolated from the peripheral blood of patients with MM were able to engraft in Nonobese diabetic/severe combined immunodeficiency (NOD/SCID) mice and give rise to mature CD138+ MM plasma cells. CD19+ B cells isolated from the engrafted mice were further confirmed to form MM tumorous lesions in secondary recipients, indicating self-renewal potential. Clonogenic MM progenitor cells can also be detected with in vitro colony assays. Hamburger and Salmon [19] have reported that more than 86% of tumor samples from MM patients were capable of colony formation at a frequency of 1 in 1,000–100,000 cells. Matsui et al. also found that CD138CD34 cells were able to form colonies in vitro with morphologically and phenotypically mature CD138+ plasma cells exhibiting intracellular immunoglobulin light-chain restriction that matched the patients' MM [14]. These results implicated that MM-initiating cells with tumorigenic progenitor nature exist in clonotypic postgerminal center B cells in MM, which may differentiate and reconstitute the bulk of MM cells. In addition, Boucher et al. [15] reported that light chain restricted B cells successfully grew colonies regardless of CD27 expression. They also demonstrated the colony formation by CD34+CD19+ immature B cells besides CD34CD19+ mature B cells, but not by early noncommitted CD34+CD19 cells from MM bone marrow samples, suggesting that even early undifferentiated clonotypic B cells may also have MM-initiating capacity.

Furthermore, clonotypic B cells appear to play a role in disease progression or recurrence in MM. The frequency of circulating clonotypic B cells has been demonstrated to increase in parallel with disease progression [20]. Clonotypic B cells are detected not only at diagnosis or at relapse but also persist even in patients with MM achieving a complete response after treatments. Therefore, clonotypic B cells appear to be a reservoir for MM-initiating cells to cause relapse.

Clonotypic Plasma Cell Subpopulations in MM

Because of the heterogeneous cellular composition of clonotypic cells in MM, phenotypically different clonotypic subpopulations in patients with MM have been extensively studied for their clonogenic potential. Despite the above persuasive evidence with clonotypic B cells, a number of studies have demonstrated robust clonogenic potential in non-B cell plasma cell subpopulations in MM cells. Yaccoby et al. successfully recapitulated human MM disease by intraosseous transplantation of CD382+ plasma cells from the majority of primary MM bone marrow specimens in SCID mice implanted with human fetal bone fragments (SCID-hu mice) [21] or with rabbit femurs (SCID-rab mice) [22]. In these mice, the implanted human or rabbit bone fragments foster the primary human MM specimens to induce MM growth within the bone fragments, with bone destruction and M protein secretion to the circulation. Furthermore, plasma cells recovered from the xenograft models were successfully transferred to secondary and tertiary recipients to produce MM disease with the typical manifestations. In contrast, the plasma cell-depleted bone marrow cells did not develop MM disease in the SCID-hu models up to 60 weeks after inoculation [21].

The clonogenic nature of MM plasma cells has been further studied in terms of CD138 negativity in SCID-rab mice. CD138CD19CD38[2]+ plasma cells isolated from three out of nine patients with MM were successfully engrafted to develop MM disease in SCID-rab mice [23]. CD138+ plasma cells from four out of nine patients were also able to produce MM in the mice, although more slowly than CD138 cells. Similar to the previous observations by Yaccoby et al. [21], however, the SCID-rab mice did not permit the engraftment of CD19+ B cells. More recently, Kim et al. [24] meticulously examined the clonogenic potential of fractionated bone marrow cells from patients with MM using more severely immunocompromized RAG2−/− and NSG mice. They found that patients' B cell and plasma cells successfully engrafted in these mice with human fetal bone grafts but not without the human bone, and that only the fraction of fully differentiated CD138+, CD38[2]+ MM cells was able to engraft and serially transfer the disease to secondary recipients. In their experimental models, fully differentiated MM plasma cells are suggested to enrich for long-lived and tumor-initiating cells whereas B cells do not.

These results suggest that MM plasma cells including CD138 and CD138+ cells have the potential to propagate MM clones in vivo in the absence of CD19+ B cells. However, a number of caveats against these results should be taken into account, including the difference among experimental animal systems, conditions used by different researchers to evaluate clonogenicity, and patients with different disease stages and risk factors. Nevertheless, MM CSC-like cells appear to be present in MM plasma cells, when analyzed in these human or rabbit bone-bearing immunocompromized mice.

Paino et al. [25] examined the presence of CD20+ cells in a panel of MM cell lines. A very small population of CD20dim+ cells (0.3%) were detected only in the RPMI8226 cell line. However, CD20dim+ RPMI8226 cells were not essential for CB17-SCID mice engraftment and showed lower self-renewal potential than the CD20 RPMI-8226 cells. Trepel et al. [26] demonstrated using their novel tracing approach that clonotypic B cells represented a very rare population in MM cells from patients with MM. These results may also support the presence of MM CSCs rather in MM plasma cells.

SP Cells

The SP phenotype is characteristic of stem cells in various normal tissues [27]. SP cells are identified by their ability to efflux Hoechst 33342 dye, a substrate for the ATP-binding cassette (ABC) transporter ABCG2, also known as breast cancer resistance protein1, suggesting that SP cells have high ABC transporter activity. SP cells have been also demonstrated as a drug-resistant fraction in many cancers and are considered to contain their CSCs [27-31]. In MM, the distinct fraction of SP cells has been detected in both MM cell lines and primary MM cells [32]. Matsui et al. [18] have found the clonogenic property and drug resistance in SP fractions of MM cells and reported a relationship of the SP fractions with CD138 phenotype and cellular quiescence in a cell cycle status.

Jakubikova et al. reported conflicting data in terms of the relationship of SP fractions to a CD138 phenotype and cellular quiescence. They observed marked difference in the percentage of SP cells among MM cell lines and primary MM cells as well as heterogeneity in a CD138 expression status within SP cells [33]. They tested MM cell lines with distinct CD138 and CD138low subpopulations as well as CD138+ cells and found that SP fractions were mainly present in CD138low and CD138+ populations with the lack of correlation between SP and CD138 or CD138low subpopulations. They also confirmed that SP cells contain clonogenic cells with high proliferation rates and greater tumorigenic potential in xenograft models compared with main population (MP) cells. In addition, SP cells sorted from the OPM1 MM cell line generated a significant fraction of MP cells in in vitro cultures, whereas MP cells produced mainly MP cells, suggesting that SP cells can generate a significant proportion of cells with MP phenotype.

We also sorted SP and MP fractions from MM cells and examined for their phenotypes and clonogenic activity. We found that SP fractions of the authentic MM cell line RPMI8226 contained CD138+ and CD138low phenotypes, but MP fractions were mainly in CD138+ phenotype [34], as previously shown by Jakubikova et al. [33]. In terms of cell cycle status, however, our results demonstrated that SP cells were mainly in a G0/G1 phase and had less proliferating ability than MP cells, which are consistent with the previous results by Matsui et al. [18]. The reasons of these discrepancies are largely unknown, but experimental procedures and conditions in different groups may affect the results. Also, there may be a possibility that a part of cells in SP fractions differentiate into MP cells with CD138+ phenotype during the experimental procedures. In normal hematopoiesis, quiescent slowly dividing hematopoietic stem cells generate cell progeny hierarchically organized into rapidly dividing transit amplifying and further differentiated cells. We should analyze and characterize transit amplifying cells in MM CSC-like cells or SP cells besides true MM CSCs, which may affect the heterogeneity of SP cells among MM cell samples. Anyhow, because SP cells merely represent a functionally distinct drug resistant cell fraction with enhanced ABCG2 activity, further clarification of the genuine component of stemness in SP cells and their heterogeneity is highly needed.

Interconversion Between Differentiated and Undifferentiated Clonotypic Cells in MM

CD138+ differentiated MM plasma cells or MP fractions with CD138+ mature plasma cell populations have been demonstrated to possess a clonogenic capacity, although lower than CD138−/low immature MM cells [23] or SP fractions [33]. It is hard to explain the clonogenic capacity of differentiated MM cells only with the notion of a one-way hierarchical model from MM CSCs to differentiated MM cells. Similar to the prediction in the biology of cancers of the transition between differentiated and dedifferentiated states [35-37], transition of MM plasma cells into MM-initiating cells can be speculated. Recently, Chaidos et al. [20] elegantly addressed a phenotypic and functional interconvertible state between CD138+ and CD138 cells. They identified a distinct subpopulation with CD19CD38[2]+CD319+CD138 immunophenotype in clonotypic cells in both the bone marrow and circulation of patients with MM, and termed as pre-plasma cell. They tested directions of phenotypic transitions within clonotypic subpopulations in the bone marrow samples of 30 patients with MM by dynamic mathematical models that were constructed assuming a forward CD19+ cell → pre-plasma cell → CD138low → CD138+ plasma cell differentiation potential. A CD138+ plasma cell to pre-plasma cell transition was identified in the analysis while the transitions of pre-plasma cell to CD19+ cells, CD138low plasma cells to CD19+ cells, and CD138+ plasma cells to CD19+ cells were not supported. In addition, both pre-plasma cells and CD138low plasma cells were identified in the bone marrow of mice receiving CD138+ plasma cells. Together with the observations with the forward transition of clonotypic CD19+ cells and/or CD138low cell to differentiated CD138+ MM plasma cells [14, 15, 17, 18], bidirectional transitions between differentiated and undifferentiated MM clonotypic cells are suggested in patients with MM (Fig. 1). Because the inoculation of phenotypically differentiated MM cells alone could form MM in xenograft models, which is transplantable to secondary and tertiary recipients, the above findings further imply the existence of phenotypic as well as functional plasticity between MM CSCs and non-CSC populations.

Figure 1.

Localization of distinct clonotypic cells in MM and transitions between MM progenitors and differentiated cells. MM cells are phenotypically and functionally heterogeneous populations. Clonotypic B cells have been demonstrated to be able to give rise to monoclonal immunoglobulin-secreting plasma cells. Besides clonotypic B cells, phenotypically distinct MM plasma cell fractions have been demonstrated to possess a clonogenic capacity. A distinct subpopulation termed as pre-PC with CD19CD38[2]+CD319+CD138 immunophenotype in clonotypic cells is identified in both the bone marrow and circulation of patients with MM [20]. Bidirectional transitions between pre-PC and differentiated MM cells have been demonstrated. When human CD138+ mature MM cells were injected via tail veins, the spleen and liver contained predominantly immature clonotypic cells, namely pre-PC and CD138low plasma cells, rather than CD138+ mature MM cells, although CD138+ mature MM cells composed the majority of clonotypic cells in the bone marrow. The existence of phenotypic as well as functional plasticity between MM CSCs and non-CSC populations is also suggested. The immunophenotypes of the indicated cell populations are as follows: clonotypic B cell, CD19+CD138; Pre-PC, CD19CD38[2]+CD319+CD138; and MM cell, CD382+CD138+. Abbreviations: BM: bone marrow; CSC, cancer stem cell; MM, multiple myeloma; PB, peripheral blood; pre-PC: pre-plasma cell.

Niches for MM CSC-Like Cells

Bone provides a unique microenvironment for myeloma cell growth and survival, including niches to foster clonogenic MM cells. MM cells stimulate bone resorption by enhancing osteoclastogenesis, while suppressing bone formation by inhibiting osteoblastic differentiation from bone marrow stromal cells, leading to extensive bone destruction with rapid loss of bone [38, 39]. Osteoclasts induced by MM cells are not only bone-resorbing cells but also produce multiple growth and survival factors for MM cells, including tumor necrosis factor (TNF) family cytokines, BAFF and APRIL, and interleukin (IL)-6 to enhance MM cell growth and survival in concert with bone marrow stromal cells, thereby forming a vicious cycle between tumor progression and bone destruction [39, 40]

By analogy with the formation of a cancer premetastatic niche by the migration and seeding of hematopoietic cells to metastatic sites [41], MM cell-induced cell types in MM bone lesions, namely osteoclasts, vascular endothelial cells, and bone marrow stromal cells with defective osteoblastic differentiation, appear to play an important role in creating a beneficial cellular environment for MM cell growth and survival as a feeder for MM cells. Such a microenvironment can be construed as a “MM niche.” In addition to the growth advantage that MM cells have acquired through genetic alterations and clonal selections, the cellular microenvironment skewed in MM endows MM cells with an additional growth and survival potential. MM cells preferentially reside in the bone marrow microenvironment, and recent preclinical and early clinical studies with new agents such as proteasome inhibitors and immunomodulatory drugs indicated that targeting not only MM cells but also their interactions with the host bone marrow microenvironment might improve the outcome of MM treatment [42], suggesting a role for the bone marrow microenvironment in the survival of MM cells and their progenitors. The bone marrow microenvironment appears to be hypoxic; especially, the bone–bone marrow interface is markedly hypoxic to provide the hypoxic endosteal niche [43]. Besides the direct cellular interaction, hypoxia created in tumor niches triggers adaptive responses to facilitate tumor progression. Tumorigenic or tumor-initiating MM plasma cells with the CD45+ immature phenotype have been demonstrated to be selected in the native hypoxic bone marrow microenvironment in 5T2MM mouse MM models [44].

As reported for SCID-hu or SCID-rab models in which human MM cells were injected into bone marrow cavities of subcutaneously implanted human or rabbit bones, MM growth was restricted to the implanted bones, suggesting that human or rabbit bones serve niches for MM CSCs in these experimental models [21-23]. Furthermore, Yaccoby et al. [45] demonstrated that in cocultures with osteoclasts, differentiated MM cells became plasmablastic along with reduced expression of CD38 and CD138 and developed resistance to dexamethasone-induced apoptosis. MM cells constitutively overexpress C-X-C chemokine receptor 4 (CXCR4), and bone marrow stromal cells produce a large amount of stromal cell-derived factor 1 (SDF-1). The SDF-1-CXCR4 axis plays an important role in MM cell homing and accumulation to the bone marrow [46]. Similar to hematopoietic stem cells, MM CSCs may lodge their niche in the bone marrow through the SDF-1-CXCR4 interaction to maintain their stemness. Blockade of the SDF-1-CXCR4 interaction with specific antagonists, such as AMD3100 [46] and TN14003 [47], demonstrated to reduce MM xenografts, which may perturb the MM CSC-niche interaction and sensitize MM CSCs to chemotherapeutic agents.

Interestingly, Chaidos et al. [20] found that when human CD138+ mature MM cells were injected via tail veins, the spleen and liver contained predominantly immature clonotypic cells, namely CD19CD38[2]+CD319+CD138 pre-PC and CD138low plasma cells, rather than CD138+ MM cells, although CD138+ mature MM cells composed the majority of clonotypic cells in the bone marrow (Fig. 1). Taken together, these results suggest dedifferentiation of mature MM cells in vivo and preferential localization of immature MM clonotypic cells or MM progenitors in a certain microenvironment. Further studies are needed to explore the localization of MM CSC-like cells and the mechanisms of their tumorigenicity and plasticity upon interaction with their preferential microenvironment.

Targeting Clonogenic MM Cells

The most important property of CSCs is that they are resistant to chemotherapeutic agents. Various approaches have been proposed to target molecular, metabolic and epigenetic signatures, and the self-renewal signaling pathway characteristic of MM CSC-like cells (Fig. 2).

Figure 2.

Targeting multiple myeloma cancer stem cell (MM CSC)-like cells. Targeting MM CSC-like cells or side population cells is clinically relevant to improve therapeutic efficacy. Different approaches have been preclinically investigated to target molecular, metabolic and epigenetic signatures, and the self-renewal signaling pathways characteristic of MM CSC-like cells. Abbreviations: ATP, adenosine triphosphate; ADP, adenosine diphosphate; C3B3: anti-HLA class I Fv diabody; 3BrPA, 3-bromopyruvate; BCRP1: breast cancer resistance protein 1; HDAC, histone deacetylase; HKII, hexokinase II; MCT1, monocarboxylate transporter1; SMO: smoothened.

CSCs exhibit high levels of ABC transporter activity and confer drug resistance besides their clonogenic or tumor-initiating capacity. Malignant cells increase their expression of glycolytic enzymes and glucose uptake to markedly enhance glycolysis (aerobic glycolysis; the Warburg effect), which leads to the production of a large amount of ATP and biomass such as nucleic acids and lipids essential for cell survival and division [48-51]. In parallel with enhanced glycolysis, ATP production by oxidative phospohorylation in the tricarboxylic acid cycle in mitochondria is suppressed through oncogenic alterations including the mutation of p53 [48-51]. The lactate analog 3-bromopyruvate (3BrPA) preferentially gets into malignant cells through monocarboxylate transporter1 overexpressed on them and inhibits glycolysis [52]. Because ABC transporter activity is dependent on ATP [53, 54], and because ATP production in cancer cells is largely dependent on enhanced glycolysis, inhibition of glycolysis by 3-BrPA was found to promptly and effectively suppress ATP production and ABC transporter activity in ABC transporter-expressing malignant cells and restored their susceptibility to anticancer drugs [55]. Although the relationship between ABCG2 expression and SP phenotype has been established, other ABC transporters are also involved in the SP phenotype with high ABC transporter activity and drug resistance. Therefore, the efficacy of strategies targeting a single transporter seems limited. In contrast, inhibition of glycolysis is thought to be able to simultaneously inactivate all types of ABC transporters in cancer cells including SP cells, because each transporter is dependent on ATP generated largely through enhanced glycolysis. We further examined the status of glycolysis in SP and MP cells from RPMI8226 cells. The SP cells exhibited the increased expression of genes involved in the glycolytic pathway including GLUT1, GLUT3, PDK1, and PFK2 and generated a larger amount of ATP and lactate per cell compared to MP cells [55]. These results suggest that glycolysis is highly accelerated in SP cells. Furthermore, the ability of RPMI8226 cells to form colonies was completely abolished by relatively low concentrations of 3BrPA. From these results, it is plausible that inhibition of glycolysis can disrupt their clonogenic capacity and drug resistant nature in SP cells.

We previously demonstrated that human leukocyte antigen (HLA) class I molecules are overexpressed in MM cells and that a recombinant single-chain Fv diabody against this molecule efficiently crosslinks HLA class I and induces Rho-mediated actin aggregation, which leads to MM cell death without effector function [56, 57]. Subsequently, we found that SP cells were resistant to anti-MM agents, the alkylator melphalan and the proteasome inhibitor bortezomib, but the newly engineered single-chain Fv diabody C3B3 could induce cytotoxicity in both SP cells and MP cells to a similar extent [34]. Moreover, C3B3 suppressed colony formation of SP cells and tumorigenesis of MM cells in vitro and in vivo. Crosslinking of HLA class I by C3B3 mediated disruption of lipid rafts and actin aggregation, which led to inhibition of gene expression of β-catenin and pluripotency-associated transcription factors such as Sox2, Oct3/4, and Nanog. Conversely, the knockdown of Sox2 and Oct3/4 mRNA reduced the proportion of SP cells, suggesting that these factors are essential in maintenance of SP fractions in MM cells. Although chemotherapeutic drugs seem to have limited effects on clonogenic MM cells or SP cells because of their detoxification potential, antibody-based agents can directly target cell surface molecules and have a potential to overcome drug resistance of SP cells with high drug efflux and detoxification activity.

Jakubikova et al. [33] reported that lenalidomide but not thalidomide decreased the percentage of SP cells in MM cells in a dose- and time-dependent manner. Interestingly, thalidomide as well as lenalidomide significantly decreased the percentage of SP cells in coculture with bone marrow stromal cells, suggesting these agents may target the interactions between CSCs and their surrounding microenvironment in the bone marrow. Lenalidomide also reduced colony formation of SP cells. Because both agents did not affect ABC transporter function in SP cells, precise mechanisms of the reduction of SP cells and their clonogenic potential remain to be clarified. In contrast, other studies showed that clonotypic CD19+ cells were resistant to lenalidomide while inhibiting the clonogenic growth of CD138+ MM cells [15, 18]. Boucher et al. [15] reported that the histone deacetylase inhibitor panobinostat but not lenalidomide was able to target the clonotypic CD19+ cells.

Telomerase activity is required for the maintenance of normal adult stem cells. Brennan et al. [58] examined the effects of telomerase inhibition on clonogenic MM growth. The long-term treatment with imetelstat, a specific inhibitor of the reverse transcriptase activity of telomerase, resulted in a significant reduction in telomere length and the inhibition of clonogenic MM growth both in vitro and in vivo. In addition, relatively short time treatment with the agent for 3 days also suppressed clonogenic growth while inducing the differentiation of MM CSC-like cells without affecting telomere length. Further studies are warranted to examine more detailed effects of telomerase inhibitors or telomerase reverse transcriptase inhibitors, as well as the mechanism of action of these agents in modifying the growth and differentiation of MM CSC-like cells.

Monoclonal antibody-based therapy has emerged as an important therapeutic strategy in patients with neoplastic diseases. Given that MM CSCs are contained in clonotypic B cells, anti-CD20 antibody may be able to target MM CSCs. Treatment with anti-CD20 antibody in combination with complement reduced clonogenic potential of MM cells in vitro, although the clinical efficacy of the humanized anti-CD20 antibody rituximab has been limited [59]. The monoclonal antibody-mediated ligation of CD44 eradicated CSCs in human acute myeloid leukemia-bearing animal models [60]. The strategy targeting CD44 may also be applicable to CSCs in MM.

Signaling pathways such as Wnt, Hedgehog, and Notch have been suggested to play a role in CSC maintenance and phenotype in various types of cancers [61]. In MM, Wnt [62], Hedgehog [63], and Notch [64] signaling has been shown as critical pathways for the growth of MM cells. Specifically, β-catenin, a key regulator of Wnt pathway is illegitimately activated in MM cells [62, 65]. The Oct4/Sox2/Nanog pluripotency axis has consistently been reiterated as the central pathway of pluripotency and self-renewal in embryonic stem cells [66, 67]. Because β-catenin also regulates expression of these pluripotency-associated genes [68, 69], Wnt/β-catenin pathway might become a novel target for MM cancer stem cells. Of note, Hedgehog signaling has been suggested to play a significant role in the maintenance of tumor stem cell compartment in MM [63]. The Hedgehog pathway activity appeared to be concentrated within MM CSCs and its inhibition by cyclopamine suppressed subsequent clonal expansion of MM CSCs without affecting mature MM cell growth [63]. Notch receptors and their ligand Jagged1 are expressed in MM cells [70]. The Notch pathway is active in MM cells, and ligand-induced Notch signaling appears to promote MM cell growth [70]. Notch also positively controls the SDF-1-CXCR4 axis to enhance MM cell migration, proliferation and resistance to apoptosis [71]. However, the precise roles of Notch signaling remain to be further clarified in clonogenic MM cells, especially in the context of tumor microenvironments.


Because MM remains incurable even after effectively debulking the tumor mass, the idea of MM CSCs has been proposed to represent functional properties of small numbers of cells that have innate resistance to chemotherapeutic agents and the ability to live long and regrow in local microenvironments. Recent insights into the biology of putative MM CSCs provide the important information for the design of efficient targeting strategies. However, a number of questions have been unanswered and MM CSCs are still conceptual. Like normal tissue stem cells, MM CSCs may not be a static population and survive as phenotypically and functionally different cell types via the transition between stem-like and non-stem-like states in local microenvironments. Given the interconversion between MM CSC-like cells or undifferentiated clonotypic cells and MM cells, we should target all types of clonotypic cells from MM CSCs to mature MM cells. We also need to clarify where MM CSCs localize and how they self-renew upon interaction with MM microenvironment. The establishment of sophisticated, refined, and humanized methods to recapitulate MM microenvironment, which supports MM CSC-like cell survival and self-renewal, is urgently needed. Further elucidation of the molecular mechanisms that keep MM clonogenesity and drive the transition between stem-like and non-stem-like states in local microenvironments will provide us with new approaches with crucial impact on the therapeutic paradigm in MM.


This work was supported in part by Grants-in-aid for Scientific Research (C) to M.A., and a National Cancer Center Research and Development Fund (21-8-5) to M.A. from the Ministry of Health, Labor and Welfare of Japan. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Author Contributions

M.A.: conception and design, financial support, manuscript writing, and final approval of manuscript, T.H.: conception and design and manuscript writing; T.M.: conception and design, manuscript writing, and final approval of manuscript.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.