Cellular and vaccine immunotherapy for multiple myeloma

Evasion of the immune system is now recognized as a hallmark of cancer pathogenesis.¹  Many of the immune evasion mechanisms identified in other malignancies have been implicated in the pathogenesis of multiple myeloma (MM). These include upregulation of cell-surface ligands that inhibit T-cell²  and natural killer (NK)-cell function,³  skewing of innate immune cells such as immature myeloid cells or macrophages toward phenotypes that suppress T-cell responses,^4-6  and other complex interactions in the bone marrow (BM) microenvironment that suppress antimyeloma immunity.^7-9  Cellular immunotherapy approaches entail infusion of either autologous or allogeneic immune effector cells, usually after ex vivo enhancement, to overcome inadequate endogenous cellular antimyeloma immunity. Vaccine approaches aim to stimulate endogenous antimyeloma T-cell responses by administering myeloma antigens in novel contexts (eg, with potent adjuvants) outside the BM microenvironment.

Although cellular and vaccine immunotherapies are being investigated for many cancers, certain clinical and biologic features of MM may render it particularly amenable to these modalities. Because MM resides primarily in BM, there are fewer barriers to trafficking of infused cellular therapies to their targets compared with malignancies that manifest primarily as tumors. Additionally, standard myeloma therapies can induce minimal disease states in many patients, which may provide a window of immune competence for vaccines to stimulate antimyeloma responses. Finally, the depletion of normal plasma cells is expected to be well tolerated; thus, it is not necessary to identify target antigens that distinguish neoplastic from non-neoplastic plasma cells. These factors make it likely that some of the promising cellular and vaccine immunotherapies currently in clinical and preclinical investigation will demonstrate clinical efficacy.

Here, we will review the major cellular and vaccine immunotherapies being evaluated in clinical trials. We will also discuss how these therapies might integrate and synergize with existing treatment paradigms.

Both autologous and allogeneic HSCT have been evaluated extensively in MM. Autologous SCT (ASCT) with high-dose melphalan conditioning confers an overall survival (OS) benefit as consolidation after initial therapy compared with continued standard-dose therapy^10-12 ; thus, autologous transplantation is a standard-of-care for eligible patients, although it is not considered curative. In contrast, allogeneic transplantation may be curative in a subset of patients, but its use is currently limited to select high-risk cases, ideally in a clinical trial, due to the absence of a consistent OS benefit in comparative studies.^13,14  Efficacy of both autologous and allogeneic transplantation is likely mediated in part by immunomodulatory effects of the conditioning regimen and graft. In autologous transplantation, the lymphocyte content of the autograft has been linked to higher posttransplant circulating lymphocytes counts, which in turn have been associated with improved progression-free survival (PFS) and OS.^15-17  Emergence of myeloma-specific immune responses after both autologous and allogeneic transplantation is associated with more favorable outcomes.^18-22  These observations suggest that the posttransplant period is a window of opportunity to stimulate clinically relevant antimyeloma immunity. Accordingly, HSCT serves as a platform for many of the cellular and vaccine approaches described later.

Reports of clinical responses to DLI after allogeneic SCT (allo-SCT) provide direct evidence of an allogeneic cellular immunotherapeutic effect.^23-25  Allogeneic transplantation with a CD34-selected (ie, T-cell–depleted) graft followed later by low-dose DLI may induce antimyeloma immunity without risk of graft-versus-host disease (GVHD).²⁶  A trial is ongoing to augment the specificity and efficacy of DLI by selective expansion and infusion of donor T cells specific for the WT1 tumor antigen (#NCT01758328).

NK cells are large granular lymphocytes capable of directly killing transformed or infected cells, and stimulating other components of the immune system in an antigen-independent manner.²⁷  NK-cell activity is regulated by a complex interplay of activating and inhibitory signals emanating from target cells. Key activating factors are downregulation of major histocompatibility complex (MHC) class I (ie, “missing self”), as occurs in some malignancies and viral infections, and the presence of bound antibody, including antibodies used as cancer therapies. The efficacy of existing MM therapies including the immunomodulatory drug class (ie, thalidomide and its analogs) and the recently approved monoclonal antibody (mAb) elotuzumab, may partially be due to augmentation of endogenous NK-cell function.^28-32  For cellular therapy applications, autologous or allogeneic NK cells can be expanded ex vivo by coculture with artificial antigen-presenting cells (APCs) engineered to express ligands and cytokines that selectively expand NK cells.^33,34  Szmania et al recently reported results from 8 patients with refractory, high-risk MM who received either autologous or haploidentical allogeneic NK cells after ex vivo expansion. Significant in vivo proliferation of the infused NK cells were observed, but there were no objective myeloma responses.³⁵  Studies continue with this approach both in patients with refractory and asymptomatic MM (#NCT01313897 and #NCT01884688). NK cells from allogeneic sources may be more promising than autologous sources due to the dysfunction of endogenous NK cells in MM patients and the potential to select donors based on HLA profiles that would predict NK-cell alloreactivity.³⁶  Allogeneic NK cells can now be expanded from umbilical cord blood units,³⁷  and a study is ongoing in which expanded NK cells derived from umbilical cord blood units are infused after ASCT (#NCT01729091). Because NK cells are relatively short-lived in vivo, they may be most useful in combination with other therapies to deepen or prolong responses.

T cells can be identified in the BM of premalignant monoclonal gammopathies that react to neoplastic plasma cells.³⁸  In patients with active MM, myeloma-specific T cells can also be recovered, but proliferation and cytokine production is only observed after extended ex vivo expansion and activation.^38,39  These observations have led to clinical trials in which T cells are activated and expanded ex vivo from MM patients’ BM and re-infused as a cellular therapy. This use of activated MILs parallels an approach first developed for melanoma using tumor-infiltrating lymphocytes. The ability of tumor-infiltrating lymphocytes to induce durable clinical remissions and albinism in melanoma patients serves as one of the seminal demonstrations of the therapeutic potential of cellular immunotherapy.⁴⁰  MILs may be especially well suited for adoptive immunotherapy because MILs are enriched for long-lived memory T-cell phenotypes and express high levels of CXCR4, which would be expected to facilitate trafficking to BM.⁴¹  MILs for MM have been evaluated in 2 clinical trials, 1 completed⁴²  and 1 ongoing (#NCT01858558), in which activated/expanded MILs are infused on day 3 after ASCT. The first study of 25 patients demonstrated the basic feasibility of this approach and that stimulation of myeloma-specific immune responses, which were durable to at least 1 year posttransplant, correlated with superior clinical outcomes.⁴²  The ongoing study is a randomized trial that will evaluate the specific contribution of MILs to PFS and myeloma-specific immune responses.

The ex vivo expansion process used to manufacture MILs entails culture with microbeads bearing immobilized anti-CD3 and anti-CD28 antibodies. This technique⁴³  was initially evaluated clinically as a method to reconstitute immunity in HIV patients⁴⁴  and in cancer patients after autologous hematopoietic cell transplantation.⁴⁵  The development of this and other techniques⁴⁶  for massive ex vivo expansion of T cells was a critical breakthrough that enabled the T-cell therapies described here.

The MILs approach described earlier and the vaccine approaches described later rely on the endogenous T-cell repertoire. A limitation of this approach is that cancer antigens are often self-antigens, and T cells with high affinity for self-antigens are rare in the endogenous repertoire due to thymic selection. Even if tumor-specific neoantigens or cancer-testis antigens are present, the frequency of T cells specific for these antigens may be too low in the endogenous repertoire to generate clinically significant responses, particularly when the disease burden is high. Advances in techniques for genetically modifying lymphocytes can overcome these barriers by modifying large populations of polyclonal T cells to artificially confer heritable specificity for a single antigen of choice. In addition, genes can be inserted for co-receptors or regulatory elements that enhance T-cell activation upon antigen recognition, which may overcome mechanisms that operate in the tumor microenvironment to suppress T-cell responses. Retroviral vectors, based either on gammaretroviruses or lentiviruses, are the most commonly used vectors in recent and ongoing clinical trials of genetically modified T cells for cancer immunotherapy. Follow-up of patients who received retrovirus-transduced T cells >10 years ago for treatment of HIV have suggested the long-term safety of this approach.⁴⁷  In recent and ongoing trials, 2 main approaches have been used to genetically confer specificity for tumor antigens, both of which are being evaluated in MM:

• Transduced T-cell receptors (TCRs): In this approach, T cells are transduced with a TCR specific for a tumor antigen. This approach can be used to target both intracellular and cell-surface antigens. Any given TCR, however, is restricted to use in patients with specific HLA haplotypes, because the target is recognized in the context of MHC-I on target cells.

• Chimeric antigen receptors (CARs): In this approach, T cells are transduced with a chimeric receptor comprised of an extracellular portion that directly binds a cell-surface target on cancer cells and an intracellular portion consisting of signaling domains that govern T-cell activation. The extracellular antigen-binding region is typically derived from the variable regions of a mAb. The intracellular component typically consists of the CD3-ζ signaling domain in tandem with the signaling domain from co-stimulatory receptors such as CD28 or 4-1BB. Unlike TCRs, CARs are not HLA-restricted, but they generally can only be used to target cell-surface proteins.

Both TCR- and CAR-transduced T cells are promising approaches to cellular immunotherapy for cancer. In hematologic malignancies, anti-CD19 CAR T cells have led to promising early results in the treatment of patients with refractory B-cell malignancies.⁴⁸  We will focus here on recent and ongoing efforts to evaluate these techniques in MM.

NY-ESO-1 and MAGE-A3 are cancer-testis antigens upregulated in subsets of MM and other human cancers.^49-52  TCRs were developed against both NY-ESO-1 and MAGE-A3 based on affinity-enhancement of wild-type low-affinity TCRs identified in cancer patients. A phase 1 study was undertaken in which subjects were selected based on expression of the target antigen and carriage of the appropriate HLA type for each TCR. Autologous T cells were transduced with a lentiviral vector encoding the TCR and expanded ex vivo using anti-CD3/anti-CD28 beads. In this study, like the MILs study discussed earlier, T cells were infused shortly after ASCT. This design allowed the high-dose melphalan conditioning to serve as lymphodepleting chemotherapy, which stimulates the production of cytokines that enhance proliferation of the infused T cells.⁵³  This design also reflected the perception at the time, based on the experience with allo-SCT, that cellular immunotherapies would be most effective if administered in conditions of low disease burden, although subsequent studies with anti-CD19 CAR T cells in acute lymphoblastic leukemia and chronic lymphocytic leukemia demonstrated that genetically modified T cells could induce remissions in patients with high burdens of refractory disease.⁴⁸  The combination with autologous transplant in a small, single-arm study made it impossible to definitively determine whether the affinity-enhanced TCRs exerted a direct antimyeloma effect. Patients who received the NY-ESO-1–specific TCR and whose disease progressed when modified T cells were still detectable in vivo exhibited downregulation of NY-ESO-1 expression in their myeloma cells, suggesting that the modified T cells exerted a selective pressure.⁵⁴  Autologous GVHD of the skin and gastrointestinal tract developed in 3/20 subjects on this study and was felt to be a general effect of activated T-cell infusion posttransplant, as has previously been reported,⁵⁵  rather than an off-target effect of the TCR. An ongoing phase 1/2 study is evaluating the NY-ESO-1–specific TCR in relapsed/refractory MM (#NCT01892293).

The portion of this study using the MAGE-A3–specific TCR was halted after 2 patients ( 1 MM patient and 1 melanoma patient treated in another trial) developed fatal cardiotoxicity that was determined to be due to cross-reactivity of the TCR against titin, a protein expressed in cardiac myocytes.^56,57  This toxicity developed despite prior studies with MAGE-A3 vaccines that proceeded safely, highlighting the greater potency of genetically enhanced cellular therapies compared with other antigen-specific immunotherapy approaches such as vaccines and mAbs. Evaluation of these unfortunate toxicities led to the development of tools to predict off-target toxicity that should enhance the safety of engineered TCRs in the future.

B-cell maturation antigen (BCMA), a tumor necrosis factor receptor family cell-surface protein, is a very promising cell-surface target that is nearly exclusively expressed on plasma cells.⁵⁸  In addition to CAR T cells, BCMA is also being evaluated as a target for antibody-drug conjugates and bispecific antibodies. The National Cancer Institute published promising preclinical data with anti-BCMA CAR T cells^59,60  and presented early results of a phase 1 clinical trial at the 2015 American Society of Hematology Annual Meeting showing striking antimyeloma responses in patients with refractory disease who were treated at the highest dose levels.⁶¹  Cytokine release syndrome, an inflammatory response resembling septic shock or macrophage activation syndrome, was reported in patients who developed antimyeloma responses, similarly to what has been reported with anti-CD19 CAR T cells in acute lymphoblastic leukemia, chronic lymphocytic leukemia, and non-Hodgkin lymphoma (NHL).⁶²  Two additional clinical trials with anti-BCMA CAR T cells are ongoing but have not yet reported results (#NCT02546167 and #NCT02658929).

Anti-CD19 CAR T cells have been investigated in a pilot clinical trial in combination with ASCT. CD19 is not commonly expressed on MM plasma cells, but several lines of evidence suggest that targeting CD19 in MM may prevent progression after other effective therapies by eliminating rare myeloma cells with cancer stem-cell capabilities that exhibit a B-cell rather than a plasma-cell phenotype. Early results from this study were reported in a case report of a patient with refractory myeloma and poor response to prior autologous transplant, but who experienced a durable complete response when a second transplant was followed by anti-CD19 CAR T cells.⁶³ 

Table 1 lists other CAR T-cell targets being investigated preclinically and in early-stage clinical trials.^59,61,64-69  CARs targeting κ light chain, Lewis Y-antigen, and NKG2D ligands are in early phase clinical trials; these trials are not exclusively for MM patients, and few MM patients have been treated. Trials with CS1-directed CARs are likely to open soon. CS1 is a promising target against which a mAb therapy, elotuzumab, has already demonstrated efficacy. A challenge with CS1-directed CARs is CS1 expression on some activated T cells, raising the possibility the CAR T cells may exhibit cytotoxicity against each other. This so-called CAR T-cell fratricide may be addressable with gene-editing techniques to knock out the CS1 gene in the CAR-transduced cells. CD138 and CD38 have also been considered as CAR targets, but expression of CD138 on epithelia and CD38 on other essential hematopoietic lineages raises the potential for off-target toxicity.

The autologous cellular immunotherapies discussed earlier are personalized therapies that require patient-specific manufacturing, which poses several challenges. The time required for manufacturing is typically 2 to 4 weeks; in patients with highly refractory myeloma, new disease complications such as renal failure may develop during this waiting period. Patient-specific factors and prior therapies may affect the composition and health of the lymphocytes harvested for manufacturing, which in turn may affect the potency and toxicity potential of the manufactured product. These challenges can be addressed by preparation of the product using allogeneic cells obtained from healthy donors. Cellular therapies prepared from allogeneic sources could be manufactured in advance, cryopreserved, and available “off the shelf,” thus eliminating the waiting periods required for manufacturing autologous products. The cellular composition of cells manufactured from normal donors should also be easier to standardize because variability introduced by the underlying malignancy and prior therapies is eliminated. Use of allogeneic sources, however, introduces additional challenges of GVHD and rejection of the cells by the recipient; these may be at least partially preventable by further genetic modifications that downregulate expression of the endogenous TCRs and MHC.

The close correlation between toxicity and efficacy observed with use of CAR T cells for other malignancies presents an additional challenge; the brief experience reported with anti-BCMA CAR T cells by the National Cancer Institute suggests this challenge will apply to MM. In addition to variability in product composition, disease burden at the time of therapy likely influences the magnitude of in vivo T-cell proliferation and toxicity. Over time, clinical trials will need to evaluate biomarkers, personalized dosing algorithms, and use of adjunctive therapies to fine-tune response and toxicity in a patient-specific fashion.

Cancer vaccines aim to stimulate endogenous immune responses against malignant cells. In MM, vaccines are generally considered to have the greatest potential in combination with other effective therapies that have immunomodulatory mechanisms of action; in this setting, vaccines may increase the likelihood of response or improve response duration. The vaccine approaches discussed here are all being clinically evaluated as combination therapies. Key aspects of vaccine design are the selection of antigen (or source of antigen) and a mechanism to enhance antigen presentation.

Many cancer vaccine strategies aim to augment presentation of tumor antigens by dendritic cells, the most potent APCs in the immune system. Dendritic cell presentation of tumor antigens can be augmented by differentiating/activating dendritic cells ex vivo in the presence of tumor antigen,⁷⁰  an approach used in a US Food and Drug Administration-approved vaccine for prostate cancer, sipuleucel-T.⁷¹  In hematologic malignancies and other settings where cancer cells are readily accessible, the cancer cells themselves can serve as an antigen source, which has the potential to stimulate multiplex immunity to known cancer antigens, previously unidentified antigens, and patient-specific neoantigens. One approach to stimulating dendritic cell presentation of antigens from autologous cancer cells is to physically fuse the cancer cells with the dendritic cells.⁷²  In a phase 2 study, an autologous myeloma/dendritic-cell fusion vaccine was administered to MM patients after ASCT. Vaccination was associated with emergence of myeloma-specific CD4 and CD8 T-cell responses and deepening of posttransplant myeloma responses.⁷³  An ongoing clinical trial is evaluating this fusion vaccine in combination with a PD-1 inhibitor (#NCT01067287),⁷⁴  based on promising preclinical results with this combination.⁷⁵  A multisite randomized phase 2 study of the myeloma/dendritic-cell fusion vaccine in combination with lenalidomide maintenance therapy after ASCT is ongoing through the Blood and Marrow Transplant Clinical Trials Network (#NCT02728102).

A second dendritic-cell–based approach in clinical evaluation in MM is electroporation of Langerhans-type dendritic cells with messenger RNAs encoding the myeloma antigens CT7, MAGE-A3, and WT-1 (#NCT01995708). The basis for this study is preclinical results suggesting that Langerhans-type dendritic cells induce more potent T-cell responses than other dendritic-cell subsets,⁷⁶  and that messenger RNA electroporation stimulates maturation and activation of Langerhans-type dendritic cells.⁷⁷  In this study, Langerhans-type dendritic cells are differentiated from stem cells collected for ASCT, and the vaccines are administered posttransplant. Results have not yet been reported.

Granulocyte-macrophage colony-stimulating factor (GM-CSF) has been extensively investigated as a vaccine adjuvant, based on its ability to recruit and activate dendritic cells and other APCs.⁷⁸  This concept has been validated clinically with the recent Food and Drug Administration approval of talimogene laherparepvec (also called T-Vec) for treatment of melanoma. Talimogene leherparepvec is a herpes simplex virus modified to carry the GM-CSF gene to augment antigen presentation and lyse tumor cells; it is injected directly into tumors but can induce regression of distant, uninjected tumors, indicating induction of systemic antitumor immunity.^79,80  GVAX (Aduro Biotech) is a vaccine platform in which cells are injected after genetic modification to express GM-CSF.⁸¹  An MM GVAX is being evaluated in which cells from 2 myeloma cell lines are admixed with K562 cells modified to express GM-CSF. In a small study, durable PFS and emergence of myeloma-specific immune responses were reported in patients who had been stable with low-levels of detectable MM (“near-complete response”) on lenalidomide-containing regimens who received serial myeloma GVAX injections in combination with lenalidomide.⁸² 

Two completed studies have tested vaccines against the cancer-testis antigen MAGE-A3 administered after ASCT. In the first study of 27 subjects, a vaccine comprised 2 HLA-A2–restricted epitopes from MAGE-A3 administered with GM-CSF, the TLR-3 agonist poly-ICLC, and montanide. This vaccine was combined with infusion of autologous T cells expanded with anti-CD3/anti-CD28 beads, as described earlier, based on prior studies showing that posttransplant infusion of activated autologous T cells could significantly augment vaccine-specific immune responses if administered prior to pretransplant T-cell collection and after posttransplant T-cell infusion.^83-85  Subjects exhibited MAGE-A3–specific T-cell responses as expected, but, as a small, single-arm study, no conclusion could be drawn about efficacy.⁸⁶  The second study used a vaccine comprised of full-length recombinant MAGE-A3 and an alternative adjuvant (AS15); this 13-patient study also administered the vaccine posttransplant and demonstrated antigen-specific T-cell responses and loss-of-target antigen expression in patients who progressed, suggesting vaccine-induced selective pressure against myeloma cells expressing the target.⁸⁷ 

Studies are ongoing in the post-autologous transplant setting and in patients with high-risk smoldering MM using a vaccine comprised of HLA-A2–restricted peptides derived from the MM antigens XBP1, CD138, and CS1, administered with montanide (PVX-410; Oncopep, Inc.) (#NCT01718899 and #NCT02700841). Memory CD8+ T-cell responses were reported in smoldering MM patients who received PVX-410 alone⁸⁸ ; the ongoing study combines PVX-410 with lenalidomide.

MM, like all mature B-cell malignancies, is characterized by the production of a unique, clone-specific immunoglobulin, often referred to as the “idiotype.” As a myeloma-specific protein, anti-idiotype immunity may be clinically beneficial. Studies of anti-idiotype vaccines for NHL have demonstrated immunogenicity and clinical efficacy.^89-91  Favorable results were reported in a small, single-arm study in which stem cell donors for MM patients undergoing allo-SCT were vaccinated with an idiotype vaccine derived from the recipient’s MM prior to stem cell harvesting.⁹²  A randomized phase 2 study is ongoing in which MM patients receive vaccines consisting of idiotype protein conjugated to the adjuvant “keyhole limpet hemocyananin” and GM-CSF after ASCT (#NCT01426828). This study incorporates pretransplant priming and posttransplant infusion of autologous T cells, as described in the MAGE-A3/montanide study discussed earlier, to enhance the anti-idiotype immune response. Results have not yet been reported.

Although cellular and vaccine immunotherapies for MM remain investigational, the progress and multitude of clinical trials evaluating these approaches is encouraging. Although we have focused on cellular and vaccine approaches here and discussed them as separate modalities, there is strong rationale for combination immunotherapy approaches, especially those that combine antigen-specific approaches such as vaccines with nonspecific immune activators like checkpoint inhibitors. Evaluation of combination therapies poses challenges for clinical trial design. Dissecting the individual effects of components of combined therapies often requires randomized studies, which increases the cost and complexity of clinical trials. This challenge highlights the need for cooperative multisite studies, such as the ongoing Blood and Marrow Transplant Clinical Trials Network study of the myeloma/dendritic-cell fusion vaccine discussed earlier, and encouragement of patients to participate in clinical research. The relatively long expected survival of MM patients also poses challenges for evaluating therapies that are most likely to be effective in settings of low disease burden. This challenge highlights the need for well-validated surrogate end points, such as minimal residual disease status, to expedite evaluation of these promising therapies. Overall, the progress discussed here suggests that MM is quite amenable to cellular and vaccine immunotherapies and that these modalities are likely to become part of the expanding array of therapies available to patients afflicted with this challenging disease.

Edward A. Stadtmauer, Perelman Center for Advanced Medicine, 12 South Tower, 3400 Civic Center Blvd, Philadelphia, PA 19104; e-mail: edward.stadtmauer@uphs.upenn.edu.