Checkpoint inhibition in myeloma

The fascinating history of cancer immunotherapy over more than the past century has been consistently marked by attempts to augment immunity against malignancy through various strategies designed to stimulate the immune system into antitumor action. From William Coley’s “toxins” until the relatively recent past, efforts have been virtually entirely dedicated toward inciting the immune system out of its apparent ignorance or passivity into action against cancer. Although advancement in the field has brought increased sophistication and precision to this approach (along with some periodic, but notable success, as well), the fundamental intent of invigorating the immune system into action via activating mechanisms (eg, immunostimulatory cytokines, tumor vaccines, allogeneic stem cell transplantation, adoptive cell transfer) has remained the principal goal.

More recently, however, a complementary approach to cancer immune therapy has led to a veritable revolution across the entire field—that of releasing the immune system from inhibition. Rather than conceiving of the immune system merely in an idling mode, awaiting an otherwise apparently absent signal to activate and eradicate cancer cells, a more complete understanding of the complicated relationship between immunity and cancer has unlocked an increasing number of promising novel targets for therapeutic intervention. The intent of this review is to summarize briefly the underlying biology of immune checkpoints in cancer and the rapidly emerging role of these pathways in myeloma, summarize recent and ongoing clinical trials of immune checkpoint inhibitors in myeloma, and consider ongoing challenges and future translational strategies to improve checkpoint immunotherapy in myeloma.

Although the adaptive and innate arms of the immune system exert effects through independent, complementary mechanisms, both systems are regulated through the expression of a large, complex, and expanding family of inhibitory receptors. The basic biology and relevance to myeloma of 3 of these inhibitory receptors, cytotoxic T lymphocyte antigen 4 (CTLA-4), programmed death 1 (PD-1), and killer immunoglobulin-like receptors (KIRs) will be highlighted here.

The first inhibitory receptor, CTLA-4, was functionally characterized more than 20 years ago in a murine model in which CTLA-4–deficient mice developed a severe lymphoproliferative syndrome and widespread inflammation.¹  Prior studies had already indicated that CTLA-4 binds the same ligand, B7-1, as CD28, an important costimulatory axis for T cells. Taken together, these findings were among the very first to elucidate the critical importance of regulatory feedback in maintaining immune balance and homeostasis. An early study suggested that CTLA-4 expression may be increased in about one-half of myeloma cases² ; however, very recent data suggest that CTLA-4 expression specifically on clonal T cells isolated from patients with myeloma seems lower than on nonclonal T cells.³  Although the presence of clonal T cells has been shown to correlate with improved survival in myeloma in a number of studies,^3-5  these T-cell clones may behave in a “senescent” manner, rather than being “anergic” or “exhausted” (as described in other malignancies).³  This important distinction may carry implications for translational therapy discussed in more depth later. Other recent data indicate that CTLA-4 appears to be overexpressed (with FOXP3) in bone marrow samples taken from newly diagnosed patients with myeloma, suggesting the local accumulation of CD4⁺ regulatory T cells in the tumor microenvironment.⁶  In addition, 2 independent studies have suggested a link between CTLA-4 polymorphisms and monoclonal gammopathy of undetermined significance (MGUS) and myeloma,^7,8  at least implicating a possible role for genetic variation in this immune checkpoint and susceptibility to plasma cell dyscrasias.

PD-1 was initially discovered in 1992; several years later, it was shown that PD-1–deficient mice developed autoimmune, lupus-like systemic inflammation with varied manifestations related to the murine background studied.^9,10  These findings demonstrated a similar immunoregulatory role for PD-1 as CTLA-4 but with important, subtle differences between the systems. First, unlike the expression of CTLA-4, which appears restricted to T-cell subsets, PD-1 is expressed on activated B cells and natural killer (NK) cells as well as T-cell subsets.¹¹  Second, the murine phenotypes of PD-1 deficiency suggest the pathway regulates tissue-specific immune responses rather than global T-cell activation.¹²  PD-1 recognizes 2 distinct cognate ligands that are different from CTLA-4—PD-L1 and PD-L2. These ligands are induced in the setting of inflammation, and this interaction mediates immune regulation and protects against autoimmunity during an ongoing immune response.¹¹ 

Although normal plasma cells do not express PD-1 ligands, myeloma cells, like many cancers, appear to express PD-L1.^13-19  Data suggest that PD-L1 expression correlates with the genetic subtype of myeloma, being more highly expressed in cases with hyperdiploidy than nonhyperdiploid disease.¹⁵  PD-L1 expression is further increased by stimulation of myeloma cells with interferon-γ, Toll-like receptor ligands, and interleukin-6.^13,14  In addition, PD-L1 is expressed on additional cellular subsets in the bone marrow microenvironment of myeloma, in particular on myeloid-derived suppressor cells and plasmacytoid dendritic cells.^15,17  PD-L1 expression is associated with greater proliferative capacity and reduced susceptibility to antimyeloma therapy in PD-L1⁺ myeloma cell lines compared with PD-L1⁻ lines as well as higher expression of the antiapoptotic protein Bcl-2.¹⁴  PD-L1 expression seems to be particularly increased in the setting of relapsed/refractory disease^14,19  and associated with the presence of greater tumor burden in the marrow and increased serum lactate dehydrogenase levels.¹⁴  Recently, soluble levels of PD-L1 in the serum of newly diagnosed patients with myeloma have been shown to be higher than that observed in healthy donors.²⁰  Of interesting in this study, higher soluble PD-L1 levels at diagnosis were associated with both poorer overall response to treatment and shorter progression-free survival (but not overall survival, as has been shown previously in diffuse large B-cell lymphoma).²⁰ 

In a complementary manner, PD-1 appears to be expressed on T cells in myeloma in at least 5 studies,^15-19  including a murine model.¹⁸  PD-1⁺ T cells in myeloma have been variously described as functionally “senescent”³  or “exhausted,”¹⁸  an important distinction again discussed in more depth later. Four of these reports suggest that NK cells in myeloma may also express PD-1.^15-18 

Preclinical studies have shown that PD-L1 is induced on myeloma cells in the presence of bone marrow stromal cells; however, this effect appears to be attenuated with PD-1/PD-L1 interruption.¹⁵  PD-1 and/or PD-L1 blockade also appeared to enhance T- and NK-cell cytotoxicity against myeloma targets.¹⁵  Lenalidomide seemed to reduce the expression of PD-L1 on myeloma cells, further enhancing effector-cell cytotoxicity against myeloma targets.^15,16  Inhibition of the PD-1/PD-L1 pathway has also been shown to be associated with improved survival in a myeloma murine model.¹⁸  Taken together, these findings have raised significant interest in therapeutic targeting of the PD-1/PD-L1 axis in myeloma.

KIRs are a family of cellular receptors that mediate both activating and inhibitory function and are expressed principally on NK cells (but also on some T-cell subsets) that recognize major histocompatibility class I molecules on candidate target cells.²¹  NK cells are generally inhibited from killing unless the balance of activating signals vs inhibitory signals received favors a cytotoxic response such as in the case of a missing or mutated inhibitory KIR ligand on a target cell that also displays an activating ligand.^21,22  Disrupting the KIR-ligand relationship to promote NK-cell alloreactivity was first shown to improve outcomes for patients with acute myeloid leukemia (AML) undergoing haploidentical, T-cell–depleted hematopoietic stem cell transplantation.²³  Malignant plasma cells express surface ligands for inhibitory KIR that have been shown to contribute to NK-cell immune dysfunction in myeloma.^24-27 

The most common inhibitory KIRs (KIR2DL1, KIR2DL2, and KIR2DL3) are generally expressed on NK cells (as well as on some T cells) in myeloma,²⁷  and 1 recent publication suggests that the expression of KIR2DL1 is, in fact, greater in myeloma compared with in the healthy setting.²⁸  A small study suggests that patients with myeloma have a higher prevalence of KIR2DS4, a KIR that lacks function in many settings, than healthy controls.²⁹  Another study has suggested a link between KIR3DS1 expression and progression-free survival in myeloma.³⁰ 

Several studies suggest that alteration of the KIR-ligand relationship facilitates myeloma disease progression as well.^31,32  These findings are reminiscent of the “cancer immunoediting hypothesis,” which posits that the immune system may control not only tumor “quantity” but also influence tumor “quality.”³³  This provocative idea came from an earlier observation that tumors that developed in immunodeficient mice had greater immunogenicity (“unedited”) than tumors in immunocompetent mice (“edited”). Related to myeloma, it appears that the expression of inhibitory KIR ligands increases on tumor cells in the transition from MGUS to overt myeloma (combined with decrease in NK-cell activating ligand expression) and that this imbalance contributes directly to an immunoediting process leading ultimately to loss of immune surveillance.³⁰  This idea is supported by more recent data suggesting that NK-cell licensing is subverted in the setting of myeloma, leading to less diverse and weaker affinity repertoires of KIR-ligand interactions.³²  As inhibitory KIR-ligand expression increases on myeloma cell targets, NK cells, in parallel, progressively lose capacity to recognize these cells and contributing to loss of immune surveillance.^24,31 

Experimental manipulation of the KIR-ligand relationship has suggested the importance of this axis in myeloma as well. For instance, bortezomib has been shown to downregulate expression of KIR ligands on myeloma cells, which increases susceptibility to NK-cell lysis.³⁴  Data from 2 preclinical studies indicate that KIR-ligand disruption with anti-inhibitory KIR monoclonal blocking antibodies enhances NK-cell cytotoxicity against myeloma targets either alone or in combination with lenalidomide and/or the anti-CD38 antibody, daratumumab.^25,35  In clinical studies, 1 report suggested (consistent with the observation in AML described previously) that KIR-ligand mismatch in allogeneic stem cell transplant in myeloma may protect against relapse.³⁶  Finally, results from another study suggest that patients with myeloma who receive allogeneic stem cell transplantation from donors with KIR haplotype B (which contains a greater number of activating KIR than haplotype A) have superior progression-free and overall survival (similar to what has been reported in AML as well).³⁷  Taken together, these findings suggest that the KIR-ligand system may be another promising, therapeutic target for checkpoint inhibition in myeloma.

Since the approval of ipilimumab as the first monoclonal antibody (mAb) checkpoint inhibitor (targeting CTLA-4) in 2011 for the treatment of melanoma, clinical research in the field of immune checkpoint inhibition for cancer has exploded. In the period from 2010 to 2015 alone, more than 40 new patents were filed internationally for novel immune checkpoint inhibitors and novel combinatory checkpoint-based therapies.³⁸  With the immune checkpoint biology described here in mind, many trials have begun to investigate the prospects of therapeutically targeting immune checkpoints in myeloma. This section will review a number of recent and ongoing clinical trials of immune checkpoint inhibition in myeloma.

Ipilimumab (Yervoy, MDX-010) is a fully human, immunoglobulin G₁ (IgG₁) anti-CTLA-4 mAb approved for the treatment of melanoma. Tremelimumab (formerly ticilimumab, CP-675 206) is a second, fully human IgG₂ mAb binding to CTLA-4. To date, there are no published clinical trial results with either ipilimumab or tremelimumab in myeloma specifically; however, there are some clinical data available and a number of ongoing and planned clinical trials. Ipilimumab was studied in the context of relapse of hematologic malignancy following allogeneic stem cell transplantation in which 6 of the 29 patients enrolled had myeloma.³⁹  Three objective responses were seen, but none occurred in myeloma. Ipilimumab is the subject of at least 3 active or planned trials in myeloma either in combination (NCT01592370) or after autologous (NCT02681302) or allogeneic (NCT01822509) stem cell transplantation. A study involving tremelimumab (NCT02716805) is also planned in the autologous transplant setting, as well.

Nivolumab (Opdivo, BMS-936558) is a human IgG₄ PD-1 mAb that is approved for the treatment of melanoma, squamous non–small cell lung cancer, and renal cell carcinoma. Preliminary results of a large, single-agent, dose-escalation, phase 1 study of nivolumab that included 27 patients with relapsed/refractory myeloma led to a best response of stable disease (SD) in 17 (63%) in myeloma, with 15% of patients progression free at 24 weeks.⁴⁰  One patient achieved complete recovery with radiotherapy to a single bony lesion.⁴⁰  Because nivolumab appears to have prolonged receptor binding (>100 days), a follow-up study of a small number of these patients (n = 8) was conducted to evaluate response to the next line of therapy following nivolumab. In this cohort, 6 of 8 patients (4 with SD and 4 with progressive disease on nivolumab) obtained clinical benefit in the postnivolumab line of therapy, including 1 patient who underwent radiation of an isolated plasmacytoma followed by 97 additional weeks of nivolumab therapy. At the time of the report, this patient was alive and without evidence of further relapse 48 weeks after completion of nivolumab.⁴¹  At least 4 trials are ongoing or planned with nivolumab in myeloma, either alone or in combination with other immune agents (NCT01592370 and NCT02726581) or after autologous (NCT02681302) or allogeneic (NCT01822509) transplantation.

Pembrolizumab (Keytruda, MK-3475), a humanized, IgG₄ PD-1 mAb approved for use in melanoma and non–small cell lung cancer, is currently being studied in more than 2 dozen tumor types and settings. Results presented from a dose-escalation, phase 1 trial of pembrolizumab in combination with lenalidomide and dexamethasone (KEYNOTE-023, NCT02036502) demonstrated a 50% overall response rate.⁴²  Thirty-eight percent of patients with myeloma refractory to lenalidomide achieved objective responses, and responses were seen in patients with myeloma in whom no response was observed to multiple, prior lines of therapy. A phase 3 study of lenalidomide/dexamethasone with or without pembrolizumab (KEYNOTE-185, NCT02579863) in newly diagnosed patients ineligible for stem cell transplantation is now recruiting.

Interim results from a phase 2 study of pembrolizumab in combination with pomalidomide and dexamethasone have also been presented in which an overall response rate of 60% was observed in 27 patients with heavily pretreated, relapsed/refractory myeloma.⁴³  In this trial, 75% of patients were classified as having myeloma refractory to both immunomodulatory agents (IMiDs) and proteosome inhibitors; in this subgroup, the overall response rate was 55% (2 very good partial responses, 9 partial responses).⁴⁴  A phase 3 study of pomalidomide/dexamethasone with or without pembrolizumab (KEYNOTE-183, NCT02576977) is current accruing patients in the relapsed/refractory setting.

In addition to the agents targeting PD-1, durvalumab (MEDI-4736) and atezolizumab (MPDL3280A) are in clinical development for myeloma targeting PD-L1. At this time, no data are available on studies of these agents in myeloma; however, there are at least 3 studies of durvalumab ongoing: a phase 1/2 trial with lenalidomide with or without dexamethasone in newly diagnosed myeloma (NCT02685826), a phase 1b study with pomalidomide with or without dexamethasone in relapsed/refractory myeloma (NCT02616640), and in the setting of autologous stem cell transplant (NCT02716805). In addition, a currently open, dose-escalation, phase 1 study of atezolizumab as a single agent includes patients with myeloma (NCT01375842), and a myeloma-specific, phase 1b study is also accruing to atezolizumab either alone following autologous stem cell transplant or in combination with lenalidomide in the relapsed/refractory setting (NCT02431208).

In summary, including the trials summarized here, there are at least 18 studies planned or currently ongoing evaluating the PD-1/PD-L1 in myeloma from the smoldering setting (NCT02603887) to combinatorial strategies for active myeloma and in the peritransplant and relapsed/refractory settings.

Three trials with IPH2101, a fully human IgG₄ mAb that targets KIR2D receptors, have been published in myeloma. Thirty-two patients were enrolled in a single-agent, dose-escalation, phase 1 trial in relapsed/refractory myeloma in which the planned, biologic end point of full KIR blockade over the dosing interval was achieved without dose-limiting toxicity.²⁷  No objective responses were observed and a best response of SD was recorded in 34% of patients. A phase 2 trial of IPH2101 was published in patients with smoldering myeloma in which 1 minor response and 6 SD responses were seen in 9 patients.⁴⁴  Finally, a dose-escalation, phase 1 trial of IPH2101 in combination with lenalidomide (without dexamethasone) in relapsed/refractory myeloma demonstrated an acceptable safety profile with no apparent effect of lenalidomide on IPH2101 pharmacokinetics and pharmacodynamics.⁴⁵  In this trial, 5 of 15 enrolled patients achieved objective responses (2 very good partial responses, 3 partial responses), 1 additional patient achieved minor response, and 6 patients had SD. The median progression-free survival was 24 months. Lirilumab (BMS-986015, formerly IPH2102) is a second-generation, human, IgG₄ anti-KIR mAb now in ongoing development in myeloma with 1 study in combination with elotuzumab (NCT02252263) and a second trial in combination with nivolumab (NCT01592370).

The observed clinical benefit obtained with therapeutic checkpoint inhibition across a variety of solid and hematologic malignancies has ushered in a burgeoning new arena of discovery, progress, and hope in cancer immunotherapy. The encouraging results observed particularly in combining immune checkpoint inhibitors with other agents in the relapsed/refractory setting have rapidly accelerated development into the smoldering and newly diagnosed settings of myeloma. Overall, it has been anticipated that, by 2023, the cancer immunotherapy market will involve sales of more than $13 billion annually, driven largely by immune checkpoint inhibitors.³⁸ 

Still, at least in myeloma, there are several outstanding, unresolved challenges to overcome in optimizing this approach to its greatest effectiveness. First, because myeloma has become increasingly understood as a complex and heterogeneous malignancy even with competing subclones present on a case-by-case basis,⁴⁶  it is similarly possible that the immune responses occurring in patients may be heterogeneous as well. Such a phenomenon might be related simply to the age of a patient, to the particular immunogenicity of the myeloma tumor cells associated with a given molecular, genomic subtype, the immunoediting stage in which the malignancy is encountered: for instance, in “equilibrium” (eg, MGUS/smoldering myeloma) vs “escape” (eg, progressive/relapsed or refractory disease),³³  or even the impact of prior therapies (eg, IMiDs, proteosome inhibitors, alkylating agents, corticosteroids, radiation) on the interface between immunity and myeloma. Second, it may be that multiple immune checkpoint systems are concurrently active in limiting host immunity against myeloma. As an example, NK cells may be inhibited through KIR-ligand interaction as described previously; however, the inhibitory receptor NKG2A may also independently restrict NK-cell function against myeloma.⁴⁷  This notion might explain the absence of objective responses observed in a number of studies examining single-agent activity with checkpoint inhibitors in myeloma.^27,40  Third, myeloma induces both innate and adaptive immune dysregulation through many complementary and independent mechanisms.^48,49  One important consideration in this regard is the exact nature of the clonal CD8⁺ cytotoxic T-cell subset observed in about one-half of patients with myeloma.⁴⁹  This subset has been associated with improved survival outcomes^4,5 ; however, they may display a “senescent” phenotype (rather than an “exhausted” phenotype) and thus, in fact, may not be readily responsive to immune checkpoint blockade.⁴⁹  This distinction might also explain the lack of objective responses with single-agent PD-1 blockade⁴⁹  and reinforce the notion of pursuing combination therapies (eg, lenalidomide appears to downregulate PD-L1 expression on myeloma cells, suggesting a possible complementary mechanism of action).^15,16 

At present then, the 2 exciting paths forward are being paved through ongoing development of rational combinatorial strategies^50,51  and by the discovery and translation of novel checkpoints for targeting into clinical trials.^38,52  Combinations of checkpoint inhibitors with IMiDs, for instance, have already shown early evidence of clinical activity even in patients with myeloma refractory to IMiDs (and proteosome inhibition)^42,43  and even without the concomitant use of corticosteroids.⁴⁵  Combining an immune checkpoint inhibitor with a tumor-targeting mAb also appears to enhance antimyeloma efficacy,³⁵  and studies are evaluating this approach currently (eg, NCT02252263, NCT02726581) as well. Studies combining currently available checkpoint inhibitors (NCT01592370 and NCT01592370) are also under way in myeloma. Additional research is ongoing toward combining immune checkpoint blockade with tumor vaccines, chimeric antigen receptor cellular therapy, oncolytic viruses, epigenetic agents, cytotoxic chemotherapies, and radiation, among other opportunities in myeloma. Finally, myriad immune checkpoint targets lie on the rapidly approaching horizon as well (Table 1).^39,53 

Don M. Benson Jr, 898 Biomedical Research Tower, The Ohio State University Comprehensive Cancer Center, 460 W 12th Ave, Columbus OH 43210-1240; e-mail: don.benson@osumc.edu.