Targeting the PD-1/PD-L1 axis in multiple myeloma: a dream or a reality?

Malignant cells evade host immunity through the activation of biologic systems that suppress antigen presentation and effector cell function and create an immunosuppressive milieu in the tumor microenvironment. Immunotherapeutic strategies seek to activate native innate and adaptive anti-tumor immunity by reversing critical components of tumor-mediated immune suppression.¹  Antigen-presenting and immune-effector cells interact via a complex series of inhibitory and stimulatory signals that maintain the equilibrium between activation and tolerance. This system helps mount responses that target foreign pathogens while avoiding the expansion of autoreactive clones with ensuing tissue damage. A series of positive and negative co-stimulation signals have been identified that modulate the T-cell response between activation and anergy, respectively.²  The presence of danger signals induced by viral-mediated cytotoxic injury favors the increased expression of positive co-stimulatory signals and the induction of immunologic response.³  In contrast, a panel of negative co-stimulatory factors, including the programmed cell death ligand 1 (PD-L1)/programmed cell death protein 1 (PD-1), cytotoxic T-lymphocyte–associated protein 4 (CTLA-4)/CD28, TIM-3/galectin-9, and LAG3, mediate tolerance such that genetic deletion in animal models is associated with autoimmunity.^4-9  Tumor cells exploit these biologic mechanisms of tolerance to evade host immunity.¹⁰ 

The PD-1/PD-L1 pathway is a critical inhibitor of immune activation and plays an important role in mediating tolerance.⁷  The PD-1 receptor is expressed on T cells, B cells, monocytes, and natural killer (NK) T cells after activation.¹¹  PD-L1 and PD-L2 are expressed on antigen-presenting cells, including dendritic cells (DCs) and macrophages.¹²  In addition, PD-L1 is expressed on nonhematopoietic cells, including pancreatic islet cells, endothelial cells, and epithelial cells, thus playing a role in protecting tissue from immune-mediated injury.^7,8,13  Binding of PD-1 to PD-L1 or PD-L2 decreases secretion of Th1 cytokines, inhibits T-cell proliferation, results in T-cell apoptosis, and inhibits CTL-mediated killing. In the physiologic setting, this pathway plays a critical role in maintaining immunologic equilibrium after initial T-cell response, which prevents overactivation, collateral tissue damage, and the inappropriate expansion of autoreactive T-cell populations.^14,15  In pathologic settings such as chronic viral infection, signaling via the PD-1/PD-L1 pathway results in the induction of an exhausted T-cell phenotype characterized by the inability to mount protective immunologic response.^16-18  Similarly, in the context of malignancy, upregulation of this pathway serves to prevent the activation and function of tumor-reactive T-cell populations, which contributes to immune escape and tumor growth.^19-22  PD-L1 expression has also been noted in immunoregulatory cells of the tumor microenvironment, such as myeloid-derived suppressor cells that may work in concert with malignant cells to promote tolerance.

Antibody blockade of the PD-1/PD-L1 pathway has emerged as a highly effective therapeutic strategy for a subset of patients with solid tumors and hematologic malignancies.^23-27  Durable responses have been noted in melanoma, renal cancer, and non–small-cell lung cancer, which has resulted in US Food and Drug Administration approval of nivolumab for patients with advanced disease in these settings.^28-32  In addition, combination therapy with CTLA-4 blockade in melanoma demonstrates that concurrent blockade of several negative costimulatory pathways may be associated with a marked enhancement in efficacy.^33,34  A major area of investigation is defining biomarkers that predict sensitivity to these agents to identify cancer settings and biologic categories of disease that are likely to benefit from therapy with checkpoint blockade.²³  The predictive value of tumor expression of PD-L1 has been unclear and is complicated by varying patterns of expression within the tumor bed, inconsistency between in vitro and in vivo expression, and a lack of uniform techniques for assessment. An association between mutational burden and disease response has been noted in solid tumors, suggesting that the presence of neo-antigens generated from mutational events and for which high-affinity T cells remain in the repertoire may be predictive of efficacy of PD-1 blockade. The potency of this strategy in hematologic malignancies has been most pronounced in Hodgkin disease in which associated mutations in chromosome 9 drive PD-L1 expression and in which the tumor bed is characterized by a dense infiltrate of T cells.²⁴  These findings suggest that checkpoint blockade is most likely to be effective in malignancies for which immune regulation is important for disease progression, that increased expression of negative co-stimulation by tumor or accessory cells in the microenvironment is an important mediator of tolerance, and that the presence of tumor-reactive effector cells subject to expansion are present in the tumor bed and circulation.

Multiple myeloma (MM) is associated with progressive immune dysregulation characterized by decreased antigen-presenting and effector cell function, loss of myeloma-reactive effector T-cell populations, and a bone marrow microenvironment that promotes immune escape.^35-37  The role of the PD-1/PD-L1 pathway in mediating immune escape in MM and the corresponding therapeutic efficacy of PD-1/PD-L1 blockade has emerged as an area of great interest.³⁸  PD-L1 is highly expressed on plasma cells isolated from patients with MM but not on normal plasma cells. Notably, PD-L1 is not expressed on plasma cells isolated from patients with monoclonal gammopathy of undetermined significance.^39-42  PD-1 is expressed on circulating T cells isolated from patients with advanced MM, whereas expression of PD-1 on circulating T cells is reduced in patients who achieve a minimal disease state following high-dose chemotherapy.⁴¹  Ligation of PD-1 on potentially reactive T cells induces anergy and apoptosis. PD-L1 expression is associated with increased proliferation and increased resistance to anti-myeloma therapy.⁴³  PD-L1 expression on plasma cells is upregulated in the setting of relapsed and refractory disease, which suggests that it might have a role in the development of clonal resistance.⁴⁴  The mechanisms by which PD-L1 expression is regulated are currently an area of investigation. Our laboratory has demonstrated that microRNA plays a role in regulating the expression of PD-L1, and in MM, MUC1 expression on plasma cells upregulates the expression of PD-L1 via miR-200c. In a study of 81 patients with newly diagnosed myeloma, higher serum PD-L1 levels were associated with poorer responses and shorter progression-free survival.⁴⁵  Accessory cells in the bone marrow microenvironment, including plasmacytoid DCs and myeloid-derived suppressor cells (MDSCs), express PD-L1, consistent with their immunoregulatory phenotype.^44,46,47  PD-1 blockade restores the capacity of plasmacytoid DCs to evoke cytotoxic T-lymphocyte killing of myeloma targets in vitro.⁴⁴  Of note, PD-L1 expression on malignant plasma cells is upregulated in the presence of interferon γ or Toll-like receptor ligands consistent with the presence of a counterregulatory mechanism that blunts killing of myeloma cells in the setting of immune activation.³⁹  PD-L1 is upregulated in the presence of stromal cells in an interleukin-6–dependent manner,⁴⁴  whereas PD-L1 blockade inhibits stromal cell–mediated plasma cell growth.⁴⁸  Increased PD-1 expression is observed on NK cells derived from patients with myeloma and is associated with loss of effector cell function⁴⁹  that is restored via PD-1 blockade. These findings support the role of the PD-1/PD-L1 pathway in contributing to immune escape in MM and suggest that blockade may be an effective therapeutic strategy.

In contrast, several findings suggest that PD-1 blockade alone will be insufficient to induce clinically meaningful anti-myeloma immunity. A recent report suggested that downregulation of effector cell function in myeloma may be attributed in part to senescence rather than PD-1–mediated exhaustion.^50,51  Clonal expansion within the T-cell repertoire was observed in 75% of myeloma patients who were characterized by low PD-1 expression, and it was thought to represent a population of tumor-reactive cells with a senescent phenotype in contrast to the nonclonal T cells that expressed high levels of PD-1. The presence of senescence as an alternative mechanism for T-cell inactivation points to the potential importance of combining checkpoint blockade with strategies that evoke the expansion of activated, myeloma-reactive T cells.

The clinical efficacy of PD-1 blockade is most pronounced in malignancies such as melanoma and Hodgkin disease, which are characterized by the presence of infiltrating effector cells in the tumor bed. In addition, therapeutic efficacy has been correlated with mutational burden and is thought to be associated with the presence of neo-antigens derived from mutational events that produce non–self-epitopes targeted by high-affinity T cells.⁵²  In contrast, myeloma is characterized by low levels of infiltrating effector cells and a relatively modest mutational burden compared with solid tumors, which suggests a more restricted neo-antigen profile. Thus, it is likely that checkpoint blockade will be more potent when coupled with therapies that stimulate myeloma-reactive T cells. Such approaches, including combining checkpoint blockade with immunomodulatory drugs, transplantation, and cellular therapies such as tumor vaccines, are currently being studied in the context of clinical trials.

A phase 1b study of PD-1 blockade (nivolumab) was recently completed in patients with relapsed or refractory hematologic malignancies.⁵³  Of the 81 patients treated on that study, 27 had MM. The median age of the myeloma patients was 63 years, 96% had been treated with 2 or more prior regimens, 56% had undergone prior autologous transplantation, and 24 of 27 patients had experienced disease progression after being treated with an immunomodulatory drug and proteasome inhibitor. The median follow-up duration for patients with MM was 65.6 weeks (range, 1.6 to 126 weeks). Stabilization of disease was observed in 17 MM patients (63%), which lasted a median of 11.4 weeks (range, 3.1 to 46.1 weeks). No significant evidence of disease regression was observed.

Lenalidomide reduces PD-1 expression on NK cells, helper cells, and cytotoxic T cells and downregulates PD-L1 expression on tumor cells and MDSCs in patients with MM.^49,54  Importantly, in preclinical studies, lenalidomide enhances the effect of PD-1/PD-L1 blockade on T-cell– and NK-cell–mediated cytotoxicity.^49,54  The combination of lenalidomide and PD-1 or PDL-1 blockade increased interferon γ expression by bone marrow–derived effector cells in myeloma and were associated with increased apoptosis of MM cells.⁴⁸  These in vitro effects strongly support the potential for synergy between lenalidomide and PD-1 blockade. Several clinical trials are evaluating the therapeutic efficacy of combining PD-1 or PD-L1 antibodies with lenalidomide or pomalidomide.^55-57  Preliminary results from a phase 1 study evaluating pembrolizumab (anti-PD1 antibody) in combination with lenalidomide and low-dose dexamethasone in patients with relapsed/refractory MM (NCT02036502) demonstrated disease response in 13 (76%) of 17 patients.⁵⁸  A phase 3 randomized trial evaluating pembrolizumab in combination with pomalidomide and low-dose dexamethasone in patients with relapsed/refractory MM is being initiated (NCT02576977).⁵⁹  In newly diagnosed patients, a phase 3 clinical trial evaluating pembrolizumab in combination with lenalidomide and low-dose dexamethasone compared with lenalidomide and low-dose dexamethasone alone (KEYNOTE-185) is ongoing, with a target enrollment of approximately 640 patients (NCT02579863).⁶⁰  In addition, clinical trials are ongoing to evaluate the combination of antibodies targeting PD-L1. In patients with newly diagnosed MM, durvalumab is being evaluated in combination with lenalidomide (NCT02685826). In patients with relapsed/refractory disease, durvalumab is being studied alone and in combination with pomalidomide (NCT02616640). Durvalumab in combination with daratumumab is also being studied in patients with refractory MM and in combination with pomalidomide, dexamethasone, and daratumumab (NCT02807454). Atezolizumab is being evaluated in combination with daratumumab in patients with refractory MM (NCT02431208). In patients with asymptomatic MM, atezolizumab is being administered in a clinical trial with the goal of assessing the biological and clinical effects of therapy (NCT02784483).

In addition to the PD-1/PD-L1 pathway, other negative co-stimulatory receptors are expressed on T cells isolated from patients with MM, and they may play a role in mediating tolerance and provide a mechanism of immune escape in patients treated with PD-1 blockade alone. In a recent study, it was shown that CTLA-4, LAG3, and TIM-3 are expressed on T cells isolated from patients with MM 3 and 12 months after autologous transplant.⁶¹  Studies to assess the clinical effect of blocking these pathways alone and in combination with PD-1 blockade will be critical.

Disease evolution in myeloma is associated with loss of myeloma-reactive clones in the T-cell repertoire. The incorporation of strategies to expand myeloma-specific T cells and repair the effector cell repertoire will likely be critical to enhance the efficacy of checkpoint blockade. One strategy has been the use of cytotoxic therapy to deplete suppressor populations, which facilitates reconstitution of myeloma immunity. In a murine myeloma model, PD-L1 blockade administered after low-dose total body irradiation resulted in prolonged survival.⁶²  Lymphopoietic reconstitution after high-dose chemotherapy with stem cell rescue is associated with the depletion of regulatory T cells and concurrent expansion of myeloma-specific clones.^61,63  An alternative strategy for enhancing therapeutic efficacy of PD-1/PD-L1 blockade in myeloma is through the use of tumor vaccines to expand myeloma-reactive T-cell clones for potential further activation with checkpoint blockade. We have developed a myeloma vaccine in which patient-derived tumor cells are fused with autologous DCs such that a broad array of tumor antigens, including neo-antigens generated from mutational events, are presented in the context of DC-mediated co-stimulation.^63,64  A multicenter trial has been initiated through the Clinical Trials Network to assess the efficacy of vaccination in conjunction with lenalidomide. Of note, PD-1 blockade amplifies response to a DC/myeloma fusion vaccine in vitro.⁴¹  Similarly, in a murine model, PD-L1 blockade was shown to augment response and prolongation of survival when administered with a tumor vaccine after transplantation.⁶⁵ 

Alternatively, engineered T cells are being explored as immunotherapy for hematologic malignancies, including MM.^66,67  Chimeric antigen receptor (CAR) T cells involve the incorporation of an antibody variable chain and positive costimulatory molecule into the T-cell ζ receptor such that engagement with the antigenic target provokes T-cell–mediated killing. A potential concern is that ligation of PD-1 on the CAR T cell will induce anergy.⁶⁸  Preclinical models have demonstrated enhanced efficacy of a CAR T cell administered in conjunction with PD-1 antibody.⁶⁹  Concerns remain regarding potential toxicity as a result of overexcitation of immune effectors.

Preclinical studies support a critical but not exclusive role of the PD-1/PD-L1 pathway in mediating effector cell dysfunction and immune escape in patients with myeloma. The limited clinical results available to date have not suggested significant single-agent clinical activity.⁵³  This observation is likely a result of the complex nature of immune dysfunction present in the tumor microenvironment in myeloma. In contrast, checkpoint blockade is now being examined as a part of combination therapies that reverse tumor-mediated immune suppression and expand myeloma-reactive T cells. Although it demonstrates great potential, the therapeutic efficacy of PD-1/PD-L1 blockade specifically and immune-based strategies in general will likely depend on a sophisticated understanding of the immunologic milieu in a given disease setting and a coordinated effort to repair what is broken.

Contribution: J.R. and D.A. contributed equally to writing this review.

Conflict-of-interest disclosure: J.R. receives research funding from Bristol-Myers Squibb. D.A. serves on an immuno-oncology advisory board for Celgene.

Correspondence: Jacalyn Rosenblatt, Beth Israel Deaconess Medical Center, Harvard Medical School, 330 Brookline Ave, KS 134, Boston, MA 02215: e-mail: jrosenb1@bidmc.harvard.edu.