|Home | About | Journals | Submit | Contact Us | Français|
Cancer immunotherapy comprises a variety of treatment approaches, incorporating the tremendous specificity of the adaptive immune system (T cells and antibodies) as well as the diverse and potent cytotoxic weaponry of both adaptive and innate immunity. Immunotherapy strategies include antitumor monoclonal antibodies, cancer vaccines, adoptive transfer of ex vivo activated T and natural killer cells, and administration of antibodies or recombinant proteins that either costimulate immune cells or block immune inhibitory pathways (so-called immune checkpoints). Although clear clinical efficacy has been demonstrated with antitumor antibodies since the late 1990s, other immunotherapies had not been shown to be effective until recently, when a spate of successes established the broad potential of this therapeutic modality. These successes are based on fundamental scientific advances demonstrating the toleragenic nature of cancer and the pivotal role of the tumor immune microenvironment in suppressing antitumor immunity. New therapies based on a sophisticated knowledge of immune-suppressive cells, soluble factors, and signaling pathways are designed to break tolerance and reactivate antitumor immunity to induce potent, long-lasting responses. Preclinical models indicate the importance of a complex integrated immune response in eliminating established tumors and validate the exploration of combinatorial treatment regimens, which are anticipated to be far more effective than monotherapies. Unlike conventional cancer therapies, most immunotherapies are active and dynamic, capable of inducing immune memory to propagate a successful rebalancing of the equilibrium between tumor and host.
The relationship between the immune system and human cancer is dynamic and complex. Individual human tumors harbor a multitude of somatic gene mutations and epigenetically dysregulated genes, the products of which are potentially recognizable as foreign antigens.1 However, the overriding relationship between the immune system and growing cancers is one of tolerance, in which, paradoxically, foreign molecules expressed by tumor cells are viewed as self.2
Growing cancers contain tumor-infiltrating lymphocytes (TILs), which are ineffective at tumor elimination in vivo but can exert specific functions (eg, proliferation, cytokine secretion, cytolysis) outside the immunosuppressive and toleragenic tumor microenvironment. This is because the tumor milieu contains suppressive elements including regulatory T cells and myeloid-derived suppressor cells; soluble factors such as interleukin 6 (IL-6), IL-10, vascular endothelial growth factor, and transforming growth factor beta; and ligands for coinhibitory receptors that downmodulate TIL activity.3 The clinical responsiveness of melanoma and renal cell carcinoma (RCC) to systemically administered pro-inflammatory cytokines such as IL-2 demonstrates the antitumor potential of an activated immune system; however, this nonphysiologic method of reversing immunologic tolerance exerts global rather than localized effects, resulting in serious systemic toxicities.4,5 The recent molecular characterization of toleragenic mechanisms mediated by human tumors has sharpened the focus of cancer immunotherapy on more specifically targeted methods for overcoming tolerance, revealing new therapeutic opportunities. Promising immunotherapies based on recombinant and cellular agents that harness innate as well as adaptive immune responses are the subject of this review. They illustrate the diversity of antitumor immunity and highlight the need to incorporate multiple approaches into synergistic combinatorial treatment strategies.
Monoclonal antibodies (mAbs) have had a major impact on the practice of clinical oncology. Indeed, the three top-selling cancer drugs (ie, rituximab, trastuzumab, and bevacizumab) are mAbs. Most preclinical models used to test mAbs are not designed to assess the active role of the host immune response in mediating mAb-induced anticancer responses, resulting in an underestimation of the importance of this phenomenon. In vitro assays exploring the immune effects of anticancer mAbs involve extensively manipulated lymphocytes and span a few hours. In contrast, therapeutic levels of mAbs are present for months in treated patients, allowing for more extensive lymphocyte trafficking and activation and lysis of cancer cells. In vivo studies of mAbs often involve animal tumor models with limited heterogeneity, extremely rapid growth, and limited infiltration with immune effector cells; many are performed with xenografts in immunodeficient mice.
Despite these limitations, data supporting the role of the immune response in general, and antibody-dependent cell-mediated cytotoxicity (ADCC) in particular, as a major mechanism of mAb activity are convincing. These data are strongest for rituximab.6 Studies in vitro, animal models, and correlative clinical investigations indicate that the interaction between mAb and Fc receptor (FcR) contributes to the clinical antitumor activity of rituximab. Patients with lymphoma and a polymorphism encoding high-affinity FcR (more specifically, FcγRIII) have a better response rate to single-agent rituximab than do patients with low-affinity FcR.7–9 Cancers growing in mice lacking activating FcR fail to respond to anticancer mAbs, including rituximab and trastuzumab.10 Trastuzumab can alter human epidermal growth factor receptor 2 signaling; its ability to mediate ADCC likely also contributes significantly to its antitumor activity.11 This also applies to other mAbs that target antigens on the surface of cancer cells such as other epidermal growth factor receptor family members.
There is growing evidence of extensive interactions—both synergistic and antagonistic—among various mechanisms of action that modulate the immune effects of mAbs (Fig 1). mAb-induced tumor cell lysis mediated by a number of different mechanisms may lead to enhanced uptake and cross presentation of the targeted antigen, thereby generating an adaptive cellular anticancer immune response, at least in preclinical models.12 Complement has complex effects. There is little evidence that complement-mediated cytotoxicity (CMC) contributes to antitumor activity in solid tumors, but it may contribute in hematologic malignancies, in which target cells are exposed to complement in the circulation.13 The anti-CD20 mAbs rituximab and ofatumumab, and the anti-CD52 mAb alemtuzumab, kill target cells rapidly in vitro via CMC. On the other hand, complement can block interactions between mAb and activating FcR on natural killer (NK) cells and may reduce ADCC. Components of complement clear apoptotic bodies, thereby limiting development of an active immune response.14 Thus, in some circumstances, complement may mediate CMC and enhance response to mAb, whereas in others, it could inhibit ADCC-dependant anticancer immune responses and so blunt responses to mAb.
The majority of mAbs approved for clinical use contain a human immunoglobulin (Ig) G1 heavy chain. For many years, investigators have tried to enhance the effects of mAb therapy by producing modified mAbs and conjugates. Persistence has paid off, and these approaches are finally showing promise. Changing the sequence or glycosylation of Fc regions can enhance interactions with FcR on immune effector cells.15,16 Bifunctional antibody-like molecules that bind to an antigen on a cancer cell with one arm and activating molecules such as CD3 with the other are being used to retarget T cells toward cancer cells.17,18 Drug-antibody conjugates are also showing considerable promise, particularly in hematologic malignancies.19,20 Although much of the antitumor effect of such conjugates results from the cytotoxic effects of the drugs, it is likely that immune response also plays a role.
Moving forward, how can knowledge about the immune effects of anticancer mAbs influence our ability to use them therapeutically? First, such information informs the identification of biomarkers that can enhance our ability to select those patients who are most likely to respond to mAb therapy. The most obvious biomarker is tumor expression of the target antigen. Genetic polymorphisms, including polymorphisms in FcγR and complement, can also affect the likelihood of response to mAb therapy. Second, information on mechanisms of action supports the development of combinatorial therapies of mAbs with other forms of cancer therapy, including agents that activate NK cells or other immune effector cells. Finally, we now have the ability to modify mAb structure and thereby modulate effector mechanisms such as the ability to bind to FcR or fix complement. Understanding the mechanisms of action responsible for the immune effects of antitumor mAbs, and using this information to apply and modify mAbs in a way that enhances these effects, will allow us to optimize this already effective class of immunotherapeutic agents.
Adoptive cell transfer (ACT) is a form of immunotherapy in which antitumor T cells are manipulated ex vivo and then infused into the patient. The earliest form of effective ACT was bone marrow transplantation (BMT) for hematologic malignancies. Enhanced graft versus leukemia/lymphoma effects from allogeneic versus autologous and T cell–replete versus depleted BMT were associated with improved clinical outcomes, although at the cost of increased graft versus host disease.21,22 In patients with relapsing chronic myeloid leukemia or indolent lymphoma after allogeneic BMT, the efficacy of donor lymphocyte infusions directly demonstrated the antitumor activity of transferred lymphocytes.23,24 Although T cells were long presumed to be responsible for graft versus leukemia/lymphoma effects, recent studies have elucidated an important role for alloreactive NK cells when donor killer inhibitory receptors are incompatible with host ligands.25
With the discovery in the 1980s that human T cells isolated from peripheral blood, tumor-draining lymph nodes, or tumor tissue could manifest selective antitumor reactivity in vitro, the cancer immunotherapy field undertook to develop specifically targeted ACT protocols. Melanoma TILs are a rich source of tumor-specific CD4+ and CD8+ T cells relative to other malignancies.26 The heightened immunogenicity of melanoma compared with other human cancers has provided a model system in which to define and characterize tumor antigens, comprising shared nonmutated tumor-associated proteins as well as uniquely mutant molecules.27 Autologous unfractionated TILs expanded in vitro and infused into patients with metastatic melanoma, in conjunction with systemic IL-2, have mediated objective responses in 34% to 50% of patients.28,29 Biomarker studies correlating clinical responses with the in vitro tumor specificity of TIL30 have led to the development of more complex methods to subculture tumor-reactive cells. Combined with more intense chemoradiotherapy preconditioning regimens, objective clinical response rates of 49% to 72% were observed in patients with melanoma receiving highly selected TILs.31 Because clinical TIL studies have been performed sequentially and have not been randomized, these different treatment regimens cannot be directly compared.
In contrast to the substantial therapeutic impact of ACT with polyclonal TIL cultures, reduced clinical efficacy has been encountered with CD4+ or CD8+ clones specific for a single melanoma antigen (MART-1/Melan-A, gp100, NY-ESO-1; Fig 2).32–34 Clinical trials with T-cell clones have provided a platform for proof-of-principle studies, including precise monitoring of proliferation, trafficking, and persistence of transferred cells. The outgrowth of antigen-loss tumor variants in treated patients has reflected successful antigen targeting while underscoring the capacity of rapidly adaptable tumor cells to evade narrowly focused therapies. Reduced persistence and/or trafficking of T-cell clones may also underlie reduced clinical activity.
Individualized and cumbersome microculture techniques for cloning tumor-reactive T cells are being superseded by genetically engineered T-cell receptor (TCR) –transduced T cells for ACT, in which genes encoding TCRs with defined antitumor properties are transduced into short-term cultured peripheral blood lymphocytes (CD4+ and CD8+), conferring tumor recognition. This therapy is potentially accessible to any patient whose tumor possesses the cognate human leukocyte antigen (HLA) allele and expresses the target antigen recognized by the TCR. Objective response rates after ACT with HLA-A2–restricted TCRs against MART-1/Melan-A or gp100 have ranged from 13% to 19% using native TCRs to 30% with modified TCRs designed to enhance normally low TCR avidities for tumor major histocompatibility complex (MHC) –antigen complexes.35 However, the clinical use of highly avid TCRs has been associated with significant collateral destruction of normal tissues sharing the target antigen. Current efforts aim to optimize gene transfer efficiencies, design TCR structural modifications, and identify target antigens, the expression of which is highly selective in tumor rather than nontransformed cells.36 ACT with an HLA-A2–restricted TCR specific for NY-ESO-1, a cancer-testis antigen expressed by a variety of human cancers and testis but not other normal adult tissues, resulted in objective tumor regressions in five of 11 patients with melanoma without incurring serious autoimmune toxicities. In the same study, four of six patients with treatment-refractory synovial cell sarcoma demonstrated objective responses,37 showing that principles of immunotherapy established in melanoma can now be successfully extended to other forms of cancer.
Chimeric antigen receptors (CARs) were developed to overcome limitations of MHC restriction and intracellular antigen processing imposed by ACT with conventional T cells. CARs are single-chain constructs composed of an Ig variable domain (extracellular) fused to a TCR constant domain (extracellular/intracellular); when introduced into T cells, they combine the antigen-recognition properties of antibodies with T-cell lytic functions, broadening the spectrum of tumor antigen recognition.38 To optimize CAR signaling and activation potential, first-generation CARs incorporating the cytoplasmic domain of CD3-zeta were followed by second- and third-generation CARs, the cytoplasmic domains of which incorporate multiple signaling modules from costimulatory receptors.39 Encouraging early clinical results with second-generation anti-CD19 CARs have been observed in patients with lymphoma.40,41 However, the high affinity for target cells conferred by the Ig component of CARs, combined with amplified nonphysiologic T-cell signaling in second- and third-generation constructs, has been associated with serious adverse events.42 Reducing on-target toxicities while maintaining antitumor efficacy is an important goal of current investigations. Possible solutions include use of lower-affinity Ig variable domains, more judicious selection of target antigens, infusion of defined T cell subsets, methodic dose escalation, alternative gene transfer technologies,43 and co-opting of the more physiologic T-cell signaling mechanisms of virus-specific T cells.44
Historically, the primary approach to specifically activate host T cells against tumor antigens (ie, active immunotherapy) has been therapeutic cancer vaccination. Long-standing interest in cancer vaccines comes from the tremendous successes of prophylactic vaccines for infectious diseases and is based on immunobiology demonstrating the capacity of T cells to recognize target antigens in the form of peptides complexed to surface MHC molecules. Because immunogenic peptides can be derived from proteins in every cellular compartment, essentially any protein has the potential to be recognized by T cells as a tumor-specific or tumor-selective antigen. Successful vaccination marshals multiple immune effector arms including CD4+ and CD8+ T cells to generate a potent antitumor response.45
Despite anecdotal reports and promising phase I and II clinical trial results with cancer vaccines evaluated since the 1960s, a string of failures in randomized clinical trials has bred significant skepticism as to the ultimate clinical value of therapeutic cancer vaccines.46–48 However, in the past few years, a number of important successes with cancer vaccines have dramatically altered the perception of their potential value.
The first successful randomized phase III cancer vaccine trial used a putative dendritic cell (DC) vaccine—sipuleucel-T—to treat patients with advanced hormone-resistant prostate cancer.49 This vaccine is based on the concept that optimal T-cell activation requires antigen processing and presentation by a specialized cell—the DC—with the capacity to concomitantly deliver strong costimulatory signals in the form of membrane ligands and secreted cytokines. Sipileucel-T is a patient-specific vaccine produced by transiently incubating the patient's own peripheral blood mononuclear cells with a fusion protein consisting of prostatic acid phosphatase (prostate/prostate cancer–specific antigen) linked to the DC growth and differentiation factor granulocyte macrophage colony-stimulating factor (GM-CSF). A 4-month overall survival (OS) benefit relative to the control arm (uncultured peripheral blood mononuclear cells without prostatic acid phosphatase–GM-CSF fusion protein) in the absence of objective tumor regressions or effect on time to progression emphasizes a developing paradigm: immunotherapy can potentially provide OS benefits that are not reflected in progression-free survival (PFS) or objective response rate (ORR). The survival benefit of sipuleucel-T ultimately led to US Food and Drug Administration approval in 2010.
Recently, two positive randomized cancer vaccine trials were reported. A melanoma vaccine consisting of a modified gp100 peptide plus systemic IL-2 was compared with systemic IL-2 alone in patients with advanced melanoma,50 yielding a statistically higher ORR in the vaccine plus IL-2 arm, improved PFS, and improved OS (P = .06). Of note, the same peptide vaccine, when combined with anti–cytotoxic T-lymphocyte antigen 4 (CTLA-4), demonstrated no improvement in patients with advanced melanomas relative to anti–CTLA-4 alone,51 underscoring the importance of context when evaluating vaccines as components of combinatorial therapies. Another trial comparing a poxvirus–prostate specific antigen prime/boost vaccine regimen plus GM-CSF versus nonantigen expressing viruses in patients with advanced prostate cancer demonstrated a significant (8 months) OS benefit for the vaccine arm but no effect on PFS or ORR.52 Finally, a single-arm clinical trial using long peptides (selectively processed and presented by DCs) derived from human papillomavirus 16 (HPV-16) E6 and E7 antigens induced complete regressions in nine of 19 patients with vulvar intraepithelial neoplasia, an HPV-associated preneoplastic condition, with a spontaneous regression rate of less than 2%.53
Why has there been a recent spate of successful cancer vaccine trials after such a long drought? Basic immunology advances in the past decade have directly affected vaccine design and clinical trial design. We now know that as cancer grows, it induces tolerance among T cells specific for its antigens,54 usurping normal mechanisms of self tolerance and dampening immune responses within the microenvironment through a variety of mechanisms.55,56 Thus, in contrast to conventional prophylactic vaccines for infectious diseases, therapeutic cancer vaccines must break tolerance to reactivate antitumor immune cells. Critical means for achieving this goal include targeting high quantities of antigen to DCs, expanding DC numbers, and providing DCs with appropriate activation signals (Fig 3).57 Activated DCs traffic to draining lymph nodes, where they present antigen to T cells. Although locally elaborated GM-CSF dramatically expands DC numbers at the vaccine site, additional signals are required for DC activation. These are mediated by pattern-recognition receptors (PRRs; toll-like receptors), which bind common microbial molecules, and a set of cytosolic sensors (eg, melanoma differentiation-associated antigens, retinoic acid-inducible gene 1, and nucleotide-binding oligomerization domain–leucine-rich repeat proteins).58 Most early-generation cancer vaccines did not incorporate agonists for these PRRs and thus would not have been expected to break tolerance; however, newer vaccine formulations are indeed incorporating both synthetic and natural PRR ligands.
Another component of vaccines in which scientific advances can be leveraged is the choice of target antigen. Many earlier vaccine formulations used whole cells or cell lysates as polyvalent sources of tumor antigens because relevant tumor-specific or tumor-selective antigens had not yet been defined. However, cell-based vaccines contain thousands of self antigens that provide no tumor specificity. For T cells, specific tumor antigens were first defined by the pioneering work of Boon et al59 and Kawakami et al60 in melanoma. Since then, tumor antigens recognized by T cells from patients with cancer have been defined for many common cancer types. The ideal tumor antigen is: first, expressed in a significant proportion of patients with a particular cancer type; second, not expressed (or expressed at low levels) in normal tissues; and third, vital to the cancer's growth and/or survival (avoiding outgrowth of resistant antigen-loss tumor variants). Relatively few antigens in current cancer vaccines fit all these criteria, although viral oncogene products in virus-associated cancers (ie, HPV E6 and E761) as well as certain self antigens, such as Wilms tumor 1 (leukemia62) and mesothelin (pancreas, ovarian, and lung cancers63) do. Thus, these findings have enabled the development of antigen-specific vaccines that can be engineered to codeliver tumor antigens with DC activation signals, optimally promoting effective antitumor immunity.
In addition to shared antigens selectively expressed by tumors, investigations into the nature of the human antitumor immune response have revealed a vast array of unique antigenic targets derived from mutated genes found in individual tumors. These are not approachable with generic methods of immunization, ACT, or tumor-specific mAbs. With the relatively recent realization that cancer exerts an immune-tolerizing influence in the host, new trends in immunotherapy have focused on methods to interrupt tolerogenic pathways and reactivate endogenous immunity against unique as well as shared tumor antigens.
The fine specificity of T cells for their targets is mediated by the interaction of TCRs with antigenic peptide-MHC complexes displayed on the cell surface. However, the functional consequences of antigen recognition are mediated by coregulatory receptors expressed on T cells, which recognize cognate ligands displayed on target cells including antigen-presenting cells and tumor cells (Fig 4). These coreceptors can induce stimulatory or inhibitory signaling cascades, thereby modulating T-cell proliferation, cytokine secretion, and cytolysis. A dominance of coinhibitory receptor ligation induces tolerance. The best-studied group of coregulatory molecules is the CD28-B7 family,64 and a receptor for B7-1 and B7-2 termed CTLA-4 was the initial target for immune-modulatory antibodies. CTLA-4 is a coinhibitory TCR, the natural function of which is to downmodulate immunity at the appropriate time, avoiding collateral normal tissue damage. Although there is no tumor specificity in the expression of B7-1 or B7-2, potent antitumor properties of CTLA-4 blocking mAbs were nonetheless observed in preclinical models65 and then validated in the clinic. Two anti–CTLA-4 blocking mAbs—ipilimumab (Bristol-Myers Squibb, Princteon, NJ) and tremelimumab (Pfizer, New York, NY)—demonstrated similar properties in early-phase clinical trials in patients with advanced solid tumors, mediating objective response rates of 10% to 15% in patients with metastatic melanoma and RCC.66–68 Response characteristics included delayed onset, mixed regressions (ie, concomitant regressing/progressing lesions), and long-term complete remissions in a small percentage of patients. Ipilimumab (Yervoy; Bristol-Myers Squibb) was recently approved as first-line therapy for patients with melanoma with metastatic disease, based on phase III trials in which this drug, administered alone or in combination with a gp100 peptide vaccine or with dacarbazine, demonstrated superior OS and PFS compared with vaccine alone51 or dacarbazine alone,69 respectively. Approximately 20% of patients in both studies achieved long-term survival benefit; this exceeded the reported ORRs of 10% to 15%, suggesting that, as with other immunotherapies, ipilimumab may induce a state of equilibrium between the immune system and cancer, resulting in prolonged disease stabilization but not regression in some patients. As forecast by the lethal hyperimmune/autoimmune phenotype of CTLA-4 knockout mice,70 grades 3 to 5 immune-related adverse events have been observed in 10% to 35% of patients undergoing CTLA-4 blockade. The most frequently affected organs are colon, endocrine glands, and skin; the diverse spectrum of inflammation was unanticipated from patient medical histories.
The occurrence of immune-related adverse events validates the immunologic mechanism of action of CTLA-4 mAbs, but it also mandates the exploration of alternative immune checkpoint pathways with potentially improved benefit-to-toxicity ratios as targets for cancer therapy. Among mAbs targeting other members of the CD28-B7 family, anti–programmed cell death 1 (PD-1) and anti–B7-H1/PD ligand 1 (PD-L1) are farthest along in clinical development. This pathway is of particular interest because B7-H1/PD-L1, unlike B7-1/-2, is selectively upregulated by many human cancers.71 The PD-1 pathway normally plays a protective role in modulating immune-mediated tissue destruction but can be exploited by cancer to protect itself from tumor-specific T cells.56 Although CTLA-4 regulates de novo immune responses, the PD-1 pathway exerts its major influence on ongoing (effector) immune responses; this is supported by the distinct phenotypes of PD-1 genetic knockout mice, which develop delayed-onset organ-specific inflammation as opposed to the uncontrolled global T-cell proliferation seen in CTLA-4 knockouts. Of three anti–PD-1 mAbs currently in the clinic for cancer therapy—MDX-1106/BMS936558 (Medarex, Princeton, NJ; Bristol-Myers Squibb), CT-011 (CureTech, Yavne, Israel), and MK-3475 (Merck, Whitehouse Station, NJ)—most experience has involved MDX-1106.72 A first-in-human phase I trial of intermittent dosing showed durable objective responses in three of 39 patients with treatment-refractory metastatic solid tumors (melanoma, RCC, and colorectal cancer), and clinical responses correlated with pretreatment expression of B7-H1/PD-L1 in the tumor.73 An ongoing trial administering MDX-1106 biweekly has shown preliminary evidence of durable objective tumor responses in approximately one third of patients with advanced melanoma and RCC; grade 3 or greater adverse clinical events occurred in 12% of 126 patients and included the same kinds of immune-related phenomena encountered with anti–CTLA-4.74,75 Of interest, objective tumor responses to MDX-1106 have also occurred in patients with treatment-refractory non–small-cell lung cancer, highlighting activity against a nonimmunogenic tumor. A blocking antibody against the major ligand for PD-1—B7-H1/PD-L1 (MDX-1105/BMS936559)—is also in phase I clinical testing in patients with advanced solid tumors, and preliminary evidence of clinical activity against melanoma, RCC, and non–small-cell lung cancer has been shown. Although these results validate the PD-1 pathway as a target for immunotherapy, anti–B7-H1/PD-L1 might be expected to have a unique spectrum of clinical activity based on B7-H1/PD-L1 biology.
Momentum gained from clinical results with mAbs blocking coinhibitory pathways has generated exploratory studies of agonistic or antagonistic mAbs against new targets, including T-cell costimulatory receptors in the tumor necrosis factor receptor family. Agonistic antibodies against 4-1BB (CD137),76 OX40 (CD134),77 glucocorticoid-induced tumor necrosis factor receptor family–related gene (GITR), and CD27 are currently or soon to be in the clinic. Although these agents may be effective as monotherapies, preclinical models indicate that maximum impact will be achieved in treatment combinations exploiting their unique roles in generating and maintaining antitumor immunity.
The complexity of successful immune responses with coordinate recruitment of innate and adaptive immunity, including soluble and cellular factors, has been established in preclinical and clinical models of infectious diseases and transplantation and applies to antitumor immunity as well. This review of immunotherapeutic agents in cancer barely scratches the surface of potential potency achievable by combinatorial strategies targeting distinct effector arms and both early (activation) and late (execution) stages of immune response. For example, combinations of vaccines with blocking mAbs against immune checkpoint receptors such as CTLA-4 and PD-1 demonstrate dramatic synergy in murine tumor models, in which the individual components offer little or no therapeutic efficacy.78,79 Another therapeutic opportunity involves the rational combination of distinct checkpoint inhibitors based on their biologic properties. For example, the CTLA-4 checkpoint plays a major role in dampening initial T-cell activation, whereas PD-1 inhibits effector T-cell responses within tissues. Thus, anti–CTLA-4 and anti–PD-1 have demonstrated synergy in animal tumor models, and this combination is in clinical testing.80 Although such combinations may significantly enhance antitumor immunity, they may also generate additive or synergistic immune toxicities, requiring careful dose titrations to define windows of clinical efficacy.81,82 Finally, preclinical models suggest that certain chemotherapies and targeted kinase inhibitors induce an immunologic cell death resulting from rapid release of both tumor antigens and self molecules from dying cancer cells, activating toll-like receptor pathways in DCs and promoting inflammation and heightened antitumor immunity.83,84 These and many other potentially synergistic treatment combinations are under active exploration and will be required to achieve the true potential of cancer immunotherapy.
We thank Jonathan Powell, MD (Johns Hopkins University School of Medicine), for helpful discussions.
Supported by National Institutes of Health (NIH) Grants No. P30 CA86862, P50 CA97274, and R01 CA137198 (G.J.W.); NIH Grant No. R01 CA142779 (S.L.T., D.M.P.); the Melanoma Research Alliance (S.L.T., D.M.P.); and the Barney Family Foundation (S.L.T.).
Terms in blue are defined in the glossary, found at the end of this article and online at www.jco.org.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: None Consultant or Advisory Role: Suzanne L. Topalian, Bristol-Myers Squibb (U), Amplimmune (C); George J. Weiner, Genentech (C); Drew M. Pardoll, Bristol-Myers Squibb (U), Amplimmune (C), Aduro Biotech (C), Celgene (C), GlaxoSmithKline (U) Stock Ownership: None Honoraria: Suzanne L. Topalian, Millennium Pharmaceuticals; Drew M. Pardoll, Abbott Laboratories, Immune Design Research Funding: Suzanne L. Topalian, Medarex/Bristol-Myers Squibb; George J. Weiner, Pfizer Expert Testimony: None Other Remuneration: None
Manuscript writing: All authors
Final approval of manuscript: All authors