Immune responses in multiple myeloma: role of the natural immune surveillance and potential of immunotherapies

MDSC are heterogeneous immature myeloid cells which are characterized by a potent ability to suppress anti-tumor immune responses mediated by T cells and NK cells [68, 69]. Under pathological conditions including cancer, perturbation of normal differentiation of myeloid cells leads to generation of MDSC, which is triggered by persistent exposure to tumor microenvironment-derived soluble factors such as stem-cell factor, GM-CSF, prostaglandins, IL-6, and VEGF [70]. MDSC are subsequently recruited into tumor site or lymphoid tissues in response to CCL2 [71], CXCL5 [72], and S100 proteins [73]. Initially, MDSC were identified in tumor-bearing mice as Gr-1⁺CD11b⁺ cells [74, 75]. Phenotypically, MDSC can be divided into granulocytic subset (CD11b⁺Gr-1^highLy6G⁺Ly6C^low G-MDSC) or monocytic subset (CD11b⁺Gr-1^midLy6G⁻Ly6C^high MO-MDSC) [68]. Accordingly, two subsets of MDSC possess different suppressive mechanisms: G-MDSC chiefly use reactive oxygen-species (ROS) such as hydrogen-peroxidase, whereas MO-MDSC use inducible nitric oxide synthase (iNOS) and arginase. Hydrogen-peroxide and iNOS-derived peroxynitrite inhibit T-cell receptor signal transduction [68], whereas arginase sequestrates l-arginine that is required for T cell proliferation [76], both of which dampen T cell-mediated anti-tumour immune responses in a cell-to-cell contact dependent manner. In addition to their direct immunosuppressive activities, MDSC are capable of inducing Tregs. Though the exact mechanism has not been fully understood, diverse molecules are reported to be implicated in the cross-talk between MDSC and Treg cells including arginase [77], CD40 [78] or cytokines (TGF-β, IFN-γ and IL-10) [79]. Moreover, MDSC stimulate tumor angiogenesis through secretion of MMP-9 or direct differentiation into CD31⁺ endothelial cells [80]. Thus, MDSC have multifaceted pro-tumor functions in tumor microenvironment. Recently, several studies have shown that MDSC are important players in myeloma-infiltrated immune microenvironments. In ATLN and DP42 murine myeloma models, both the proportion and the absolute number of G-MDSC and MO-MDSC in BM were significantly increased as early as 1 week after inoculation, and thereafter gradually decreased due to progressive expansion of myeloma cells [81]. Similar results were reported by another group in 5TMM models [82, 83], suggesting that the expansion of MDSC is an early event in MM. Proinflammatory S100 proteins play a pivotal role in the accumulation of MDSC in solid tumors [73]. Notably, S100A9-deficient mice showed prolonged survival compared to wild type mice after inoculation of OVA-expressing DP42 cells, which was associated with a reduction of MDSC in BM and an increase in OVA-specific CD8⁺ T cells. Furthermore, the survival benefit in S100A9-deficient mice was abrogated by antibody depletion of CD8⁺ T cells or adoptive transfer of MDSC [81], demonstrating that MDSCs dampen CD8⁺ T cell-dependent anti-myeloma immune responses, leading to myeloma progression.

In humans, Brimnes et al. [84] firstly reported that newly diagnosed MM patients have increased frequencies of CD14⁺ MO-MDSC in peripheral blood compared to healthy donors. On the contrary, recent studies showed that CD14⁻CD15⁺ G-MDSC, but not CD14⁺ MO-MDSC are increased in myeloma patients [81, 85, 86], while both subsets show the same level of suppressive activity against autologous T cells [86, 87]. In addition to their immunosuppressive activities, MDSC can directly stimulate proliferation of myeloma cells. Görgün et al. [85] showed that co-culture with MDSC marked enhanced proliferation of myeloma cells in vitro. Importantly, in this study, co-culture of myeloma cells and peripheral blood mononuclear cells (PBMC) from healthy donors was able to induce generation of MDSC, providing evidence for bidirectional interaction between myeloma cells and MDSC. Moreover, Favaloro et al. [86] reported that MDSC from MM patients can markedly induce Treg after co-culture with PBMC. Thus, MDSC contribute to immunosuppressive, tumor-favoring environment, providing a potential target in myeloma therapy.

It is now established that tumor tissues are abundantly infiltrated by TAMs, which support tumor progression by angiogenesis, matrix remodeling and potent immunosuppression [88, 89]. In general, higher level of TAMs correlates with poor prognosis in many types of solid tumors [90] as well as hematological malignancies including MM [91]. In terms of ontogeny of TAMs, both tissue-resident macrophages and recruited macrophages coexist in tumor microenvironment [92]; however, recent studies have clarified that TAMs are phenotypically distinct from residential macrophages, and originate from circulating Ly6C⁺ inflammatory monocytes [71, 93, 94]. Colony-stimulating factor-1 receptor (CSF1R) signaling and CCL2-CCR2 interaction are implicated in the recruitment of monocytes to tumor microenvironment [71, 95] where differentiation and functional maturation of TAMs are regulated by Notch signaling and environmental factors such as hypoxia [94, 96, 97]. It remains unknown whether myeloma-associated macrophages originate from circulating monocytes or BM residential precursors; however, massive BM infiltration by CD68⁺ macrophages are observed in active MM patients, but not in MGUS patients or healthy donors [16, 98], indicating that macrophages represent a pivotal cellular component of the myeloma microenvironment. Myeloma-associated macrophages contribute to myeloma pathology by at least three different ways.

Firstly, myeloma-associated macrophages support myeloma growth through cytokine production. Importantly, myeloma-associated macrophages highly express IL-1 and TNF-α [99, 100], both of which stimulate production of IL-6 from mesenchymal stem cells (MSCs). Additionally, myeloma-associated macrophages secrete the anti-inflammatory cytokine IL-10, another growth factor for myeloma cells [100, 101]. Until recently, it remained unclear whether or not myeloma-associated macrophages produce cytokines in response to specific PAMPs and/or endogenous DAMPs in the myeloma microenvironment. Hope et al. firstly showed that myeloma-associated macrophages contribute to the inflammatory milieu through toll-like receptor (TLR)-2/6-mediated recognition of its proteoglycan agonist, versican. Furthermore, they also found that genetic ablation of tpl2 (Cot/MAP3K8), a downstream effector of TLRs, delays myeloma progression in Vk*myc transgenic mice [100], highlighting the importance of this pathway.

Another important function of myeloma-associated macrophages is angiogenesis and vasculogenesis. Angiogenesis within the myeloma microenvironment is amplified by a positive feedback loop of proangiogenic factors including VEGF, basic fibroblast growth factor (bFGF), TNF-α, and IL-6 [102]. In addition to secretion of these proangiogenic factors, myeloma-associated macrophages contribute to angiogenesis through a commitment toward an endothelial phenotype. Scavelli et al. [98] reported that exposure to VEGF and bFGF convert myeloma-associated macrophages into cells which are functionally and phenotypically similar to endothelial cells, leading to formation of capillary-like structures.

Lastly, myeloma-associated macrophages support myeloma cells in a cell-to-cell contact dependent manner. Zheng et al. [16] reported that myeloma-associated macrophages protect myeloma cells from caspase-dependent apoptosis, which confers resistance to chemotherapy. Notably, IL-6 was dispensable in this mechanism. Instead, macrophage-mediated myeloma survival was depend on interaction between P-selectin glycoprotein ligand-1 (PSGL-1)/selectins and ICAM-1/CD18 which transmit survival signaling including Src, Erk1/2 and c-myc [62].

The importance of NK cells for the control of myeloma progression has been demonstrated using NK cell-depleting antibodies in various mouse MM models [105, 106]. Furthermore, several groups established the ability of human NK cells to kill MM cell lines [107‐109]; and cytotoxic activity of autologous NK cells against patient-derived myeloma targets has also been reported [108, 110].

A particularity of NK cells is their ability to sense cells that have down-regulated MHC class I molecules. Human NK cells express various combinations of killer cell immunoglobulin–like receptors (KIR) that deliver negative signals upon binding to MHC class I molecules, thus preventing reactivity against normal healthy cells [111]. MHC class I down-regulation is frequently observed in cancer cells. Accordingly, early stages myeloma cells express low levels of MHC I, and are readily recognized by NK cells [108].

In addition, myeloma cell recognition by NK cells involves various activating receptors including NKG2D, DNAX accessory molecule (DNAM-1 or CD226) and the natural cytotoxicity receptors (NCRs) NKp46, NKp30, NKp44 [107, 108]. NKp46 is certainly a key receptor for NK cell recognition of malignant plasma cells because its inhibition strongly reduced NK cell-mediated killing of all the myeloma cell lines so far tested [107]. Human NKG2D binds to MHC class I related chain A and B (MICA/MICB) and to UL16 binding proteins (ULBP1-6). ULPB1-3 has been detected on some myeloma cell lines while high levels of MICA were observed on BM-derived MM cells [108]. The NK cell-mediated killing of MICA-expressing myeloma cells was found to be NKG2D-dependent [107]. Moreover, nectin-2 (CD112) and the poliovirus receptor (PVR, CD155), the two known DNAM-1 ligands, are heterogeneously expressed on malignant plasma cells. Indeed, a study including 12 MM patients revealed CD155 and/or CD112 expression on all but two samples [107]. Blocking DNAM-1 inhibited the in vitro killing of CD155-expressing myeloma cell lines [107]. Importantly, the role of DNAM-1 in controlling myeloma progression has been investigated in vivo, in Vk*myc transgenic mice that spontaneously develop MM [105]. In this study, DNAM-1^+/+, DNAM-1^+/− and DNAM-1^−/− Vk*myc mice were monitored for disease development and survival over 800 days. Mice lacking DNAM-1 exhibited higher levels of serum monoclonal protein and succumbed earlier to MM. Although this work highlights the importance of DNAM-1 in MM immunosurveillance, the role of DNAM-1 for NK cell-mediated control of MM growth is still to be demonstrated because, akin to NKG2D, DNAM-1 is not a NK cell-specific receptor but is also expressed on T cells.

Most studies have focused on the cytolytic activity of NK cells against MM cells. However, little is known about NK cell-derived IFN-γ in this context. Interestingly, IFN-γ-deficient mice injected with MM cell lines show shorter survival associated with higher tumor burden, when compared with WT mice [105]. IFN-γ not only stimulates innate and adaptive immune responses [112], but it also inhibits the in vitro proliferation of myeloma cells [113] and interferes with the RANKL signaling pathway to decrease osteoclastogenesis [114]. Thus, IFN-γ production by NK cells may significantly reduce MM pathology and this pathway would require further investigation.

An early report described increased numbers of CD56⁺CD3⁻ NK cells in the BM and blood of newly diagnosed myeloma patients [104]. Subsequent studies confirmed that patients with MGUS or active myeloma present elevated numbers of circulating NK cells [115]. Surprisingly, patients with higher numbers of NK cells at diagnosis were found to have worse prognoses [104]. In fact, increased NK cell numbers may be seen as an unsuccessful attempt of the immune system to control myeloma cell expansion. It is now well established that NK cell activity is largely compromised in MM patients since various mechanisms contribute to impair NK cell recognition and killing of myeloma cells.

Immune escape of cancer cells involves two mechanisms: the immunoediting of tumor cells and the suppression of immune functions [116]. Both phenomena have been observed in MM. Interestingly, NK cell receptor ligands on myeloma cells are progressively edited during myeloma progression, outlining the role of NK cell control in the early stages of the disease and suggesting that impairment of NK cells responses may constitute a major event in promoting MGUS progression to MM. Indeed, malignant plasma cells or myeloma cell lines derived from early-stage MM/MGUS patients exhibit higher levels of MICA or Fas than plasma cells obtained from patients with active disease or cell lines derived from late-stage pleural effusions [108, 117, 118]. In addition, two studies observed a down-regulation of MHC class I molecules on the surface of plasma cells from early stages but not late stages MM patients [108, 117]. Of note, these results contrast with a third study that reported opposite observations i.e. an up-regulation of MHC class I molecules on BM plasma cells from MGUS patients compared with healthy donors and MM patients [119]. Still, myeloma cell-lines established from the BM are sensitive to NK cell-mediated lysis whereas cell-lines generated from pleural effusions from the same donor are resistant [108]. Likewise, increased degranulation of NK cells in the BM of MM-bearing mice could only be observed at early time points of disease development [106]. MICA shedding from the surface of malignant plasma cells generates soluble MICA that may contribute to altered NKG2D expression and defective NK cell functions [118]. While decreased NKG2D expression on NK cells from MM patients has been confirmed by another study, the role of soluble MICA in this process has been questioned [120]. In addition to NKG2D, other activating receptors showed reduced expression on NK cells from active myeloma patients. These include DNAM-1, 2B4/CD244 as well as the low affinity Fc receptor CD16 [107, 121]. Therefore, altered expression of activating receptors is likely to contribute to myeloma cell escape from cancer immunosurveillance. A recent study indicated that skewed chemokine levels hinder NK cell trafficking to the BM during the early asymptomatic stages of the disease [106]. This may represent another mechanism that contributes to myeloma cell escape from NK cell control.

Though BM is a primary lymphoid organ, BM also functions as a secondary lymphoid organ where T cells responses are initiated [125]. In this context, efficient uptake and processing of circulating tumor-associated antigens by BM CD11c⁺ DCs is critical for the priming of T cell-mediated anti-tumor immune responses. Many studies concluded that DCs from MM patients have impaired T-cell stimulation capacities, whereas contradictory results exist regarding the frequency and phenotype of DCs [126‐129]. Several soluble factors including IL-6, TGF-β and IL-10 seem to be involved in the impairment of DC functions [127, 128]. Recently, Leone et al. reported that DCs accumulate in BM during the MGUS-to-MM progression. In this study, DCs purified from MGUS/MM patients were able to engulf apoptotic myeloma cells, cross-present them and activate tumor-specific CD8⁺ T cells whereas CD28–CD80/86 interaction between live myeloma cells and DCs down-regulated expression of proteasome subunits in myeloma cells [130]. This mechanism may enable myeloma cells to evade CD8⁺ T-cell killing in spite of efficient T cell priming.

pDCs, the other major subset of DCs, are also involved in myeloma pathology. pDCs play pivotal roles for generation of normal plasma cells and antibody responses through secretion of type I IFN and IL-6 [131]. Chauhan et al. [132] showed that numbers and frequency of BM pDCs are increased in MM patients and that pDCs confer growth, survival, chemotaxis, and drug resistance in myeloma cells.

Mouse models support an instrumental role of cytotoxic CD8 T cells in MM immunosurveillance [105]. Several pieces of evidence indicate that myeloma cells express tumor antigens able to trigger T cell responses. Analysis of the T cell receptor (TCR) variable gene repertoire revealed clonal expansions of CD8 T cells in MGUS and early stage MM patients that probably reflect chronic stimulations with myeloma-derived antigens [133]. Tumor-specific T cells able to lyse autologous myeloma cells can be generated from the blood or BM of myeloma patients using myeloma lysate-pulsed DCs [134, 135]. Nonetheless, T cells freshly isolated from MM patients fail to recognize autologous tumor cells and to secrete IFN-γ, suggesting that they probably do not exert a strong anti-myeloma activity in vivo [134]. Conversely, freshly isolated T cells from the BM of MGUS patients produce IFN-γ when stimulated in vitro with DCs loaded with autologous tumor cells [136]. These data suggest that the anti-myeloma activity of tumor antigen-specific T cells is lost during the progression from MGUS to active myeloma. Noteworthy, T cells reactive against the embryonal stem cell-associated antigen SOX2 have been detected in MGUS but not MM patients [137]. SOX2 was reported to be expressed in a progenitor fraction of myeloma cells and anti-SOX2 T cell immunity correlates with a favorable outcome. In addition, patients that survived more than 10 years present expanded cytolytic T cell clones that, unlike the majority of MM patients, respond to stimulation by proliferating and producing IFN-γ [138]. Interestingly, T cells isolated from MGUS or MM patients are activated by DCs loaded with autologous but not allogeneic tumor lysates [134‐136]. This indicates that T cell responses against MM are specific of each myeloma clone and differ from one patient to another. The antigenic properties of the variable region of the secreted monoclonal protein (idiotope) have been extensively studied [139]. Unfortunately, idiotype-specific responses are usually hinder by several tolerance mechanisms, including the deletion of high avidity idiotype-specific T cells [140]. In addition to idiotopes, general tumor antigens are shared among MM cells from different patients. Those include NY-ESO-1, MAGE-A3, Muc-1, sperm protein 17, PRDI-BF1 and XBP-1 and CD138 [139]. Adoptive transfer of T cells engineered to express a high affinity TCR for a myeloma-specific antigen represents an attractive therapy. As a example, the infusion of NY-ESO-1-specific engineered T cells recently showed promising results in a phase I/II clinical trial [141].

Helper T cells play pivotal roles in adaptive immune responses and imbalanced polarization of CD4 T cell responses could largely impact on MM growth. Several reports describe a deregulated cytokine network in MM but not all of them agree on the nature of the changes. An early study established that T cells from MGUS patients stimulated with autologous monoclonal IgG are more efficient producers of IL-2 and IFN-γ when compared with idiotype-reactive T cells from late MM patients [142]. In the same line, increased IL-4 production by T cells from MM patients indicated that a shift toward Th2 polarization emerges with disease progression. This hypothesis is supported by another study describing decreased levels of IFN-γ and increased levels of IL-10 and IL-4 in the serum of 62 myeloma patients compared with 50 healthy donors [143]. IL-6 production by T cells may contribute to decreased Th1 responses in MM patients [144]. However, elevated Th1/Th2 ratios in the blood of MM patients in initial diagnosis and refractory phase have also been reported [145, 146] and high percentages of IFN-γ producing T cells were observed in MM patients [147]. Further work is needed to determine whether these discrepancies could account for variations in Th1/Th2 polarization during the course of the disease or may be explained by differences between BM and peripheral blood. Likewise, the role of Treg responses in MM is still unclear. An initial study demonstrated decreased FoxP3-expressing Tregs in spite of elevated percentages of CD25⁺CD4⁺ cells in the blood of MGUS and MM patients [148]. This group suggested that CD25⁺ T cells from MGUS or MM patients failed to suppress the proliferation of PBMC stimulated with anti-CD3 and concluded that MM Tregs were dysfunctional. However, one caveat in this assay is the intrinsic defect in the proliferation of PBMC isolated from MM patients. It was later established that CD4⁺CD25^hi Tregs from MM patients are as efficient as Tregs from healthy donors at suppressing allogeneic T cell proliferation [149]. Results are still conflicting regarding Treg frequencies that are alternatively described as increased [138, 149, 150] or reduced [147, 148] in MM patients. Interestingly, patients with high peripheral Treg frequencies show reduced survival [150]. Th17 cells are a pro-inflammatory subset of CD4 T cells that produces IL-17 and IL-22. IL-6 plays a pivotal role in dictating the balance between Tregs and Th17 cells [151]. Treg/Th17 ratios were reportedly increased in MM patients, albeit lower in patients with long-term survival [138]. Yet, increased proportions of IL-17-producing CD4 T cells and increased serum concentrations of the Th17-associated cytokines IL-1β, IL-6, IL-17, IL-21, IL-22 and IL-23 have been observed in MM patients, when compared with healthy controls [48, 147, 152]. IL-17 might contribute to MM pathology as it induces the proliferation of MM cell lines in vitro [48] and promotes MM-associated bone lesions [53]. Finally, increased frequency of IL-22 and IL-13 double-producing T cells has been detected in the blood and BM of relapsed and late stage MM patients [153]. These Th22 cells are likely to sustain MM pathology since IL-22 favors the proliferation and resistance to drug-induced cell death of some MM cell lines and IL-13 indirectly promotes MM cell survival through the activation of BMSCs.

NKT cells are characterized by the expression of both T cell and NK cell receptors. NKT cells recognize glycolipid antigens presented by the MHC-class I-like molecule CD1d and exert strong anti-tumor responses through direct cytotoxicity or release of pro-inflammatory cytokines, including IFN-γ [154]. Despite its absence on MM cell lines, CD1d is expressed by primary myeloma cells [155]. Lisophosphatidylcholine has been identified as a NKT cell ligand expressed on plasma cells from MM patients [156]. Interestingly, frequencies of lisophosphatidylcholine-recognizing NKT cells are dramatically increased in MM patients. Lisophosphatidylcholine stimulates IL-13 production by NKT and thus probably favors angiogenesis and tumor-promoting inflammation. Actually, NKT cells from MM patients are dysfunctional and unable to produce IFN-γ when stimulated with the glycolipid α-galactosylceramide [157]. Of note, similarly to conventional T cells, NKT cells isolated from MM patients can be rescued in vitro; and APC-stimulated NKT cells efficiently lyse primary autologous myeloma targets as well as CD1d-transfected MM cell lines [155, 157].

MR1-restricted mucosal associated invariant T (MAIT) cells are another type of invariant T cells that, similarly to NKT cells, have simplified patterns of TCR expression and respond immediately to antigen stimulation [158]. Albeit abundant in humans (5 % of total blood T cells), MAIT cells have not been investigated in the context of MM.

Several therapeutic agents exert their anti-MM efficacy at least in part through the recovery or the augmentation of NK cell responses [28]. As previously mentioned, new drugs such as thalidomide, IMiDs and proteasome inhibitors potentiate NK cell-mediated killing of MM cells. Furthermore, compared with T and B cells, NK cells reconstitute early following autologous SCT and may contribute to the success of this therapy [196]. NK cells are also important mediators of the GvM effects after T cell depleted allogeneic SCT, especially in the case of KIR-ligand incompatibility i.e. when donor NK cells do not express inhibitory KIRs recognizing host MHC class I molecules. Reduced relapse rates have been observed in MM patients receiving KIR-ligand mismatched allogeneic transplants [197]; and infusion of KIR-ligand mismatched NK cells followed by autologous SCT achieved 50 % of near complete remission in advanced MM patients [198]. Furthermore, IPH2101 (1-7F9), an anti-inhibitory KIR antibody could restore NK cell responses in relapsed/refractory MM patients [199]. Phase I clinical trials indicate that IPH2101 is well tolerated when given as a single agent or in combination with lenalidomide, but its efficacy against MM has yet to be proven [199, 200]. Alternatively, reprogramming NK cells with chimeric antigen receptors (CARs) specific for MM antigen could increase their reactivity toward myeloma cells. Myeloma cells express high levels of CD138 and a pre-clinical study demonstrated the ability of CD138-specific CAR NK cells to markedly delay MM growth and prolong survival in a xenograft model [201]. Finally, monoclonal antibody (mAb) therapy targeting myeloma cells have demonstrated clinical efficacy when combined with bortezomib or lenamidomide [202]. CD16 on NK cells binds to the constant region of Ig and certainly plays a key role in mAb therapies by triggering antibody-dependent cellular cytotoxicity (ADCC) of mAb-coated tumor cells. Among the different mAbs tested, daratumumab targets CD38, an ectoenzyme commonly used as a marker of myeloma cells. Daratumumab administered as a single agent has demonstrated anti-myeloma activity in clinical trials [203]. Elotuzumab is another successful mAb [204]. Elotuzumab recognizes CS1 (SLAMF7), a glycoprotein universally expressed on MM cells. Elotuzumab activity appears to be dependent on NK cell-mediated ADCC [205]. Interestingly, NK cells express low levels of CS1. The binding of elotuzumab to CS1 directly promotes NK cell activity and thus contributes to enhanced anti-tumor effects [206].

Immune checkpoints such as programmed-death 1 (PD-1) and cytotoxic T-lymphocyte associated protein 4 (CTLA-4) down-regulate T cell responses and thereby maintain self-tolerance. The use of mAbs to disrupt the receptor-ligand interactions involved in these pathways has shown remarkable results in melanoma [207]. Immune checkpoint modulation also holds promise for the treatment of hematological malignancies [208]. The high levels of PD-1 observed on NK and T cells from MM patients together with the expression of PD-1 ligand (PD-L1) on MM cells [209, 210] strongly encouraged the investigation of PD-1 blockade in MM patients. Yet, a phase 1 clinical trial using anti-PD1 mAb nivolumab in MM patients yielded disappointing initial results, with no objective responses [208]. Noteworthy, disease remained stable in 18 of 27 patients, indicating that PD-1 blockade might still have an effect in MM and could be efficient in combination with other therapeutics. In vitro studies suggested that PD-1 blockade by pidilizumab (CT-011) would synergically combine with lenalidomide [210] or with a DC/Myeloma fusion vaccine [209]. Consequently, a phase 2 clinical trial is currently ongoing to assess the efficacy of a DC/tumor vaccine in conjunction with pidilizumab following autologous SCT (NCT011067287) [164]. Additionally, a preclinical study indicated that blocking PD-L1 together with other immune checkpoints (CTLA-4, LAG-3 or TIM-3) promotes the survival of MM-bearing mice following low dose total body irradiation [211]. Investigation of immune checkpoint inhibitors is currently booming and should eventually lead to the advent of efficient combinations strategies in MM.

An alternative to immune checkpoint blockade is the use of agonist mAbs directed again co-stimulatory molecules. Approaches using anti-CD137 mAbs has been shown to elicit potent T and NK cell-mediated responses in murine MM models [105, 212]. Clinical trials are presently ongoing to evaluate the safety and beneficial effects of two agonist anti-CD137 mAbs in cancer patients [213].

Additional strategies able to promote the immune-mediated elimination of myeloma cells include DC-based therapies and vaccines as well as T cell infusions [164]. Vaccination can be directed against a specific tumor antigen such as the idiotype protein but can also use full tumor lysates, apoptotic bodies or fusion between DCs and MM cells [214]. This second option allows the development of T cell responses directed toward the whole spectrum of tumor antigens and thus prevents a possible escape caused by the down-regulation of a single targeted antigen. DC-based therapies aim to foster the expansion of tumor-specific lymphocytes in vivo; an alternative strategy is the infusion of ex vivo expanded T cells. Similarly to NK cells, T cells can be engineered to express CARs, thereby allowing the specific targeting of myeloma cells [215].