Cancer Immunotherapy: the Beginning of the End

The idea of exploiting the host’s immune system to treat cancer dates back decades and relies on the insight that the immune system can eliminate malignant cells during initial transformation in a process termed immune surveillance. Individual human tumors arise through a combination of genetic and epigenetic changes that facilitate immortality, but at the same time create foreign antigens, the so-called neo-antigens, which should render neoplastic cells detectable by the immune system and target them for destruction. Nevertheless, although the immune system is capable of noticing differences in protein structure at the atomic level, cancer cells manage to escape immune recognition and subsequent destruction. To achieve this, tumors develop multiple resistance mechanisms, including local immune evasion, induction of tolerance, and systemic disruption of T cell signaling. Moreover, in a process termed immune editing, immune recognition of malignant cells imposes a selective pressure on developing neoplasms, resulting in the outgrowth of less immunogenic and more apoptosis-resistant neoplastic cells.

Angiogenesis is a physiological process in which new blood vessels are formed from pre-existing vessels. It occurs during normal growth and development, as well as during wound healing. Under physiological conditions, angiogenesis is tightly regulated by the complex and coordinated actions of pro-angiogenic and anti-angiogenic factors according to the spatiotemporal requirements of cells or tissues. To date, many pro-angiogenic factors and their cognate receptors have been identified, including vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF), angiopoietin (Ang), hepatocyte growth factor (HGF), and epidermal growth factor (EGF).

Cancer immunotherapy is now considered a pillar of cancer treatment, alongside surgery, chemotherapy, and radiation. Ipilimumab and nivolumab/pembrolizumab are among the earliest immune checkpoint inhibitors (targeting CTLA-4 and PD-1, respectively) and are now moving from second-line to become first-line therapies of choice in advanced non-small cell lung cancer and melanoma. Treatment with these agents can induce resistance through upregulation of additional immune checkpoints, highlighting a need for new antitumor immune activating agents. Emerging drugs target not only lymphocytes associated with adaptive immunity − via blockade of immune-inhibitory checkpoints or as agonists of immunostimulatory pathways − but also innate immune processes mediated by macrophages and natural killer (NK) cells, pathways of broad relevance across many types of solid and hematopoietic cancers. The following emerging immune targets in cancer immunotherapy were selected based on their advanced stage of development in preclinical/clinical studies and on the limited number of review articles available describing some of these targets.

Cancer immunotherapy has become the fourth pillar of cancer care, complementing surgery, cytotoxic therapy, and radiotherapy. The field has a long history that started with Coley's toxins and Erlich's hypothesis that the immune system suppresses cancer development. Thomas and Burnet expanded Erlich's idea by proposing the immune surveillance hypothesis, and Prehn and Main subsequently demonstrated that carcinogen-induced tumours elicit tumour-specific immune responses. Intra-vesicular Bacillus Calmette Guerin was approved for superficial bladder cancer in 1990, and the cytokines interferon-α and interleukin-2 were approved for melanoma and renal cell carcinoma in 1986 and 1992, respectively. Schreiber more recently described immune-editing as a process that enables escape from immune surveillance to establish overt malignancy. More specific cancer immunotherapies were approved in recent years, including preventive and therapeutic cancer vaccines, the first immune checkpoint inhibitors, a bi-specific T-cell engager, and an oncolytic virus. Of these, immune checkpoint antagonists that target the PD-1 pathway have generated the most interest, with response rates across tumour types that average 20–30%. With many more immunotherapy drugs under development, a major challenge for the field is how to prioritise the most promising of these many immuno-oncology agents alone or in combination immunotherapies designed to achieve therapeutic synergy.

The PD-1 (CD279) (see Glossary) receptor can be detected at the cell surface of T cells during thymic development and in the periphery of several types of hematopoietic cell following T cell receptor (TCR) signaling and cytokine stimulation. PD-1 is expressed on CD4−CD8− thymocytes and inducibly expressed on peripheral CD4+ and CD8+ T cells, B cells, monocytes, natural killer (NK) T cells, and some dendritic cells (DCs) 1, 2. Persistent expression of PD-1 on T cells induces T cell exhaustion [3]. Exhausted CD8 T cells lose their effector function, evidenced by their inability to secrete cytolytic molecules such as perforin and their failure to secrete proinflammatory cytokines, such as IL-2, interferon gamma (IFN-γ), and tumor necrosis factor alpha (TNF-α) 4, 5.

The search for the “magic bullet” to selectively deliver a cytotoxic agent to the site of a cancerous cell has been the goal of clinical oncology for more than 100 years. Monoclonal antibody (mAb) therapy is arguably one of the most successful treatment strategies for patients with haematological and solid tumour malignancies. Antibody-based therapies have therapeutic effects through a range of mechanisms, including altering antigen or receptor function and signalling, inducing complement-dependent cytotoxicity (CDC) or antibody-dependent cytotoxicity (ADCC). Even with more than 20 monoclonal antibodies approved for therapeutic use in cancer patients, further development is required to increase their effectiveness and reduce their toxicity. Antibody–drug conjugates combine the ability to link a cytotoxic payload to a monoclonal antibody which specifically recognises a cellular surface antigen and deliver a toxic payload directly into the target cell. The development of an effective ADC therapy is influenced by the interplay of the three main structure elements of the molecule: the antibody, the linker, and the covalently attached cytotoxic agent. Depending of the target antigen in question, the use of monoclonal antibodies reduces the off target effects by limiting the exposure of normal tissues to the payload compared with conventional systemic therapies. The ability to uniquely deliver a cytotoxic payload to a tumour cell provides a tenable excitement to the cancer biology field.

Simultaneously with the development of these “next generation ADCs”, a paradigm shift in oncology drug development occurred, with immuno-oncology drugs becoming increasingly prominent due to their enhancement in the durability of anti-tumor responses. Recent treatment successes with antibodies that regulate immune activation such as CTLA-4 and PD-1 improved the fraction of patients with complete and partial responses relative to standard of care (SOC) treatment (reviewed). The first approval by the U.S. Food and Drug Administration (FDA) of an immune checkpoint inhibitor targeting CTLA-4 (ipilimumab) for the treatment of advanced melanoma occurred in 2013, coincident with the approval of the first ADC targeting solid tumor indications, T-DM1 in Her2 positive breast cancer. Subsequently, clinical trials with blocking antibodies targeting the immune checkpoint mediator programmed cell death 1 (PD-1) and its ligand (PD-L1) have resulted in objective and durable responses in cancer patients with treatment-refractory solid tumors, including melanoma and cancers of the lung and kidney.

Immunotoxins are recombinant proteins consisting of an antibody or antibody fragment targeting the tumour antigen, linked to protein toxins such as diphtheria toxin or pseudomonas exotoxin A.1 Up to now, the only immunotoxin approved by the US Food and Drug Administration (FDA) is denileukin diftitox for treatment of CD25-positive cutaneous T-cell lymphoma.2 Another immunotoxin, moxetumomab pasudotox, targeting CD22 has shown substantial activity in patients with hairy cell leukaemia and is now being assessed in a multicentre trial in patients with relapsed or refractory hairy cell leukaemia ( number, NCT01829711).3 In the case of solid tumours, immunotoxins have been less effective mainly because they induce an immune response restricting their activity. However, major tumour regressions were reported with an anti-mesothelin immunotoxin, SS1P, in patients with treatment-refractory mesothelioma when it was given in combination with pentostatin and cyclophosphamide.4 Advances in developing immunotoxins that are inherently less immunogenic show promise in preclinical studies and are now being evaluated in the clinic,5 but are outside the scope of this Review.

Scientists have known for decades that cancer cells are particularly efficient at suppressing the body’s natural immune response, which is why most treatments exploit other means, such as surgery, radiation therapy and chemotherapy, to eliminate neoplastic cells. It is now established that various components of the immune system play pivotal roles in protecting humans from cancer. Following numerous disappointing efforts and unequivocal clinical failures, the field of cancer immunotherapy has recently received a significant boost, encouraged primarily by the approval of the autologous cellular immunotherapy, sipuleucel-T, for the treatment of prostate cancer in 2010 and the approval of the anti-cytotoxic T lymphocyte-associated protein 4 (CTLA-4) antibody, ipilimumab, and of anti-programmed cell death protein 1 (PD1) antibodies for the treatment of melanoma in 2011 and 2014, respectively. These successes have revitalized the field and brought attention to the opportunities that immunotherapeutic approaches can offer.

Tumor angiogenesis is a hallmark of cancer and plays a crucial role in providing oxygen and nutrients to tumor cells during cancer progression and metastasis. Under pathological conditions, many pro-angiogenic factors and their receptors are upregulated; among these factors, VEGF is generally regarded as a key regulator of tumor angiogenesis. Bevacizumab, an anti-VEGF antibody, was recently developed as a cancer therapy to suppress tumor angiogenesis. Although bevacizumab is clinically effective for treating patients with a variety of cancers, it has frequent complications due to its inhibition of VEGF signaling in normal endothelial cells, which express high levels of VEGF receptors (VEGFRs). Importantly, bevacizumab treatment is associated with severe side effects, including bleeding, proteinuria, hypertension, gastrointestinal perforation, and stroke. Furthermore, in glioblastoma multiforme patients, bevacizumab treatment is associated with the presence of tumor cells that have an infiltrative phenotype, and high levels of matrix metalloproteinase (MMP)-2 and membrane-type 1 MMP. In addition, long-term bevacizumab treatment can lead to the development of drug resistance, due to the upregulation of other redundant tumor-derived angiogenic factors, including Ang, EGF, HGF, and PDGF.

Lymphocyte Activation Gene 3 (LAG-3) is a surface receptor expressed on activated T cells, an exhaustion marker with immunosuppressive activity. Major histocompatibility complex class II (MHC-II) is a ligand for LAG-3; additional ligands (e.g., L-selectin and galectin-3) have also been identified. Regulatory T cells (Tregs) expressing LAG-3 have enhanced suppressive activity, whereas cytotoxic CD8+ T cells expressing LAG-3 have reduced proliferation rates and effector cytokine production in cancer and autoimmune diabetes. A splice variant of LAG-3 cleaved by metalloproteinases and secreted in the cellular microenvironment has immune-activating properties when bound to MHC-II on antigen presenting cells.

A pressing challenge is transforming the majority of patients from immunotherapy non-responders to responders. This will likely require potent combination immunotherapies that effectively harness the cancer-immunity cycle described by Chen and Mellman. Cancer therapies result in tumour cell death and release of tumour antigens, which are presented by dendritic cells in the tumour-draining lymph nodes to prime and activate tumour immunity. Tumour-specific T cells then gain access to the circulation and traffic to tumours, where they infiltrate the tumour mass. T cell-mediated lysis of cancer cells releases more tumour antigens, thus perpetuating the cycle. Multiple opportunities for therapeutic intervention that enhance tumour immunity are possible at each step of this cycle. Developing novel immunotherapeutic agents, defining synergistic drug combinations, and understanding the tumour microenvironment—defects in antigen processing and presentation, and the number, type, quality and distribution of immune cells in a tumour, and the pathways that regulate them—are critical for ongoing clinical success.

CD4+Foxp3+ regulatory T cells (Tregs), a highly immunosuppressive subset of CD4+ T cells that is critical in maintaining tolerance and attenuating immune responses, express cell-surface PD-1, which contributes to their development, maintenance, and functional response. Ligand binding to the PD-1 receptor on Tregs in the presence of CD3 and transforming growth factor beta (TGF-β) leads to an increase in the de novo conversion of naive CD4+ T cells to Tregs. This induction generates heightened suppressive function and maintenance of Foxp3 expression through inhibition of Akt–mammalian target of rapamycin (mTOR) signaling and increased phosphatase and tensin homolog (PTEN) activity 7, 8. This indicates that PD-1 pathway stimulation results not only in a reduction in effector T cell function, but also an increase in immunosuppressive Treg function. This allows proper control of immune homeostasis and creates a high threshold for T cell activation.

The appropriate selection of the antigen-binding site is a critical developmental step for the eventual success of an antibody–drug conjugate. As such, the most effective antigens share certain characteristics. The monoclonal antibody selected, as the basic structural element of the ADC, binds to a target antigen present on the surface of the cell that is accessible via the bloodstream. Following binding, the complex must be rapidly internalised, allowing the release the cytotoxic agent within the tumour cell. Ideally, the antigen should be well-characterised, proportionally abundant, and accessible on tumour cells compared to surrounding normal cell populations. This is to allow the preferential binding and delivery of the ADC to malignant populations, reducing the potential for toxicity to normal cells. The antibody selected should have a high affinity for its target, increasing the potential for internalisation of the cytotoxic agent. Heterogeneity throughout the tumour population should be avoided, and antigens that shed and are abundant in the circulation should be avoided. The uptake of the ADC is limited by the rate of this target antigen-ADC complex internalisation. The rapid uptake of the ADC also reduces extracellular payload release. The ability to identify an appropriate antigen target that contains all these characteristics is, in practice, difficult. The selection of antigen in further complicated by the constant evolutionary pressure placed on cancer cell populations during treatment. Ideally, the target antigen should not be downregulated post-treatment to maintain cellular sensitivity to the therapy.

The key differentiating attributes of IO compounds are the increased recognition of tumor antigens by CD8+ T-cells and induction of tumor-specific immunological memory in cancer patients. These attributes are considered the biological drivers behind the long lasting responses seen in subsets of patients treated with IO compounds, frequently manifested during early treatment cycles. Given these clinical successes, cancer immunotherapy is likely to become a key part of the clinical management of cancers. Despite these early clinical successes, only a subset of cancer patients responds to single agent immunotherapies, and combination treatments with other immune checkpoint inhibitors or different therapeutic modalities are needed to increase the fraction of patients benefitting from IO treatment. As a consequence, combination studies with IO compounds have become a central focus of the current preclinical- and clinical development activities in oncology. Better understanding of the molecular and cellular mechanisms limiting the anti-tumor activities of current IO compounds is critical to inform the selection of optimal combination treatment regimens between IO compounds and other anti-cancer therapeutics for clinical development in oncology.

Antibody–drug conjugates make use of antibodies that are specific to tumour cell-surface proteins6 and have tumour specificity and potency not achievable with traditional drugs. Although the idea of linking drugs to tumour-targeted antibodies was clear, development of therapeutic antibody–drug conjugates needed several technological advancements. Early antibody–drug conjugates were mouse monoclonal antibodies covalently linked to anticancer drugs such as doxorubicin, vinblastine, and methotrexate. These conjugates had little success in clinical trials because of immunogenicity, scant potency, suboptimum target selection, and insufficient selectivity for tumour versus normal tissue. The lessons from these early efforts led to improvements in technology and renewed interest in antibody–drug conjugates. Replacing murine antibodies with humanised or fully human antibodies prevented immunogenicity. Potency was improved by using drugs that were 100–1000 times more potent. Careful target and antibody selection improved selectivity and efficiency of internalisation.

Immunotherapies against existing cancers include various approaches, ranging from stimulating effector mechanisms to counteracting inhibitory and suppressive mechanisms. Strategies to activate effector immune cells include vaccination with tumor antigens or augmentation of antigen presentations to increase the ability of the patient’s own immune system to mount an immune response against neoplastic cells. Additional stimulatory strategies encompass adoptive cellular therapy (ACT) in an attempt to administer immune cells directly to patients, the administration of oncolytic viruses (OVs) for the initiation of systemic antitumor immunity, and the use of antibodies targeting members of the tumor necrosis factor receptor superfamily so as to supply co-stimulatory signals to enhance T cell activity. Strategies to neutralize immunosuppressor mechanisms include chemotherapy (cyclophosphamide), the use of antibodies as a means to diminish regulatory T cells (CD25-targeted antibodies), and the use of antibodies against immune-checkpoint molecules such as CTLA-4 and PD1. This review summarizes the main strategies in cancer immunotherapy and discusses recent advances in the design of synergistic combination strategies.

VEGF, first identified as a factor that promotes vascular permeability and vascular endothelial cell mitosis in the 1980s, is a key player in angiogenesis, as well as in endothelial proliferation, migration, and nitric oxide (NO) release [26,27,28]. The mammalian VEGF proteins are dimeric glycoproteins with a molecular weight of approximately 40 kDa. Although VEGF-A is the most well characterized VEGF isoform, the VEGF family consists of five distinct isoforms: VEGF-A, VEGF-B, VEGF-C, VEGF-D, and placenta growth factor (PlGF). Structurally, VEGF proteins have eight regularly spaced cysteine residues, which form intramolecular disulfide bonds that generate three loops; two intermolecular disulfide bonds allow for a homodimer structure between two VEGF molecules. In addition, VEGF has various alternative splice variants, which exhibit different binding affinity for VEGFR co-receptors, including neuropilins and heparin sulfate proteoglycans. For instance, VEGF-A, VEGF-B, and PlGF can be divided into five (VEGF-A121, VEGF-A145, VEGF-A165, VEGF-A189, and VEGF-A206), two (VEGF-B167 and VEGF-B186), and four (PlGF-1, PlGF-2, PlGF-3, and PlGF-4) variants. Among the VEGF-A variants, VEGF-A165 and VEGF-A189 bind neuropilins and heparin sulfate proteoglycans, whereas VEGF-A121 does not bind either. Importantly, this molecular diversity alters bioavailability and activity, which in turn, allows for the initiation of various cellular responses.

Multiple early phase clinical trials are testing antagonistic LAG-3 agents in combination with anti-PD-1 and/or anti-CTLA-4 therapy. In view of the activating properties of soluble secreted LAG-3, a soluble agonist LAG-3 antibody (IMP321) was tested in advanced solid malignancies as a single agent, and demonstrated sufficient tolerability and efficacy to warrant advancement to phase II. T-cell Immunoglobulin- and Mucin-domain-containing molecule 3 (TIM-3) is an immune-inhibitory molecule first identified on CD4+ Th1 (helper) T-cells and CD8+ Tc1 (cytotoxic) T-cells, then later on Th17 T-cells, regulatory T-cells, and innate immune cells. TIM-3 is activated primarily by its widely-expressed ligand, galectin-9, leading to effector T-cell death through calcium influx, cellular aggregation, and apoptosis. When TIM-3 signalling is active, interferon-producing T-cells become exhausted, resulting in Th1 suppression and immune tolerance. TIM-3 expression is commonly observed during chronic infection, as a characteristic marker of exhausted T cells.

Experience with ipilimumab (a CTLA-4 antagonist), nivolumab and pembrolizumab (PD-1 antagonists) and atezolizumab (a PD-L1 antagonist) in treating cancer has defined several key principles of cancer immunotherapy. First, checkpoint inhibitors engage T cells with inherent capacity for adaptability and memory. This mechanism of action underlies the durable responses and long-term survival observed with these agents. Second, immunotherapy treats the immune system. It can work regardless of either tumour histology or the presence of driver mutations. Third, the side effects of checkpoint inhibitors are distinct from those of chemotherapy and targeted agents. Finally, the efficacy of immunotherapy can be improved by combining it with other treatment strategies. Although PD-1 has best been characterized in T cells, its function in other cell subsets have also become apparent. The regulation of PD-1 expression is tightly controlled during B cell differentiation. Levels are undetectable in pro-B cells, an early precursor in B cell development, and increase as B cell differentiation. Additionally, surface levels of PD-1 can be greatly enhanced in mature B cells following stimulation with Toll-like receptor 9 (TLR9) agonists. Blockade of PD-1 on B cells has been shown to increase antigen-specific antibody responses, suggesting that PD-1 plays a role in inhibiting B cell clonal responses.

The chemical conjugation of the antibody and the cytotoxic payload has a major influence on the pharmacokinetics, selectivity, and the therapeutic index of ADC-based therapies. A linker is the required covalent connection between the cytotoxic compound and the antibody. The linker influences the stability and drug to antibody ratio (DAR) of the therapeutic agent. These parameters are critical for the overall success of an ADC-based design. With the high potency of the payloads chosen, linkers must be stable within the bloodstream to limit early payload release and allow tumour site targeting to occur. However, this stability must be context-dependent: once inside the cell, these linkers must be efficient in releasing the payload to induced the cytotoxic effect.

In response to antineoplastic agents, the composition of the tumor immune infiltrates can be predictive for outcomes of therapy. An increased number of CD3+ T-lymphocytes as well as an increased ratio of cytotoxic CD8+ T-lymphocytes (CTLs) over FOXP3+ regulatory T-cells (Tregs) within tumors following chemotherapy treatment was predictive of favorable therapeutic responses in human breast and colorectal cancer patients treated with anthracyclines and oxaliplatin, respectively. Therefore, combination treatments of IO compounds and SOC regimens, in tumor indications showing responses to single agent IO compounds including melanomas, lung and renal cancers, represents an attractive clinical development strategy. However, there are several concerns when combining IOs with certain SOCs including chemotherapy. One being the notion that the dose-limiting toxicities of SOC cytotoxic regimens, in particular lymphopenia and neutropenia, may interfere negatively with the mechanism of action of IO compounds, and to impair clonal expansion of effector lymphocytes and/or disturb the homeostasis of immune cells. In support of these concerns, meta-analysis of multiple clinical trials indicated that severe lymphopenia (<1000 lymphocytes/μL) correlates negatively with the response to chemotherapy in multiple solid tumor indications. The drop in peripheral lymphocytes induced by many standard chemotherapeutic regimens may thus limit the response to IO compounds, as the activities of the latter depend on the presence of tumor infiltrating leukocytes. Additionally, the tumor microenvironment has been shown to actively impede effector cell functions, thereby limiting the efficacy of TILs activated and recruited to tumors by immune-based therapies. A potential way to circumvent such negative interference between cytotoxic and IO compounds is by staging the two modalities, and by providing sufficient time after cytotoxic treatment for the lymphoid cell population to recover prior to initiating IO treatment. In support of this concept, combination of SOC chemotherapeutics with IO compounds administered concomitantly failed to improve clinical outcome. In contrast, when chemotherapy was given prior to IO treatment (sequentially), an increase in progression free survival was observed.

As a result of this work, gemtuzumab ozogamicin was granted accelerated FDA approval for the treatment of acute myeloid leukaemia in 2000; however, this conjugate was withdrawn from the market in 2010 because it failed to meet efficacy targets in post-marketing clinical trials. Two antibody–drug conjugates have since achieved FDA approval: trastuzumab emtansine was approved in 2013 for the treatment of metastatic breast cancer, and brentuximab vedotin was approved in 2011 for the treatment of refractory Hodgkin's lymphoma and systemic anaplastic large-cell lymphoma. More than 40 antibody–drug conjugates are now in or nearing clinical trials.

ACT is a promising form of immunotherapy which exploits the antitumor properties of lymphocytes to eradicate primary and metastatic tumor cells. Lymphocytes are firstly isolated from patients’ peripheral blood, tumor-draining lymph nodes or tumor tissue, expanded ex vivo, and reinfused back into the patient. This strategy would, in theory, circumvent the baffling duty of breaking tolerance to tumor antigens and produce a large amount of high avidity effector T cells. Indeed, over the last two to three decades, autologous T cell therapies have demonstrated their potential to induce dramatic clinical responses (and have become a viable therapeutic option). VEGF-mediated cellular functions occur via the activation of three receptors: VEGFR-1 (Flt-1), VEGFR-2 (KDR/Flk-1), and VEGFR-3 (Flt-4). VEGFRs are members of the receptor tyrosine kinase (RTK) superfamily and are composed of an extracellular domain with seven immunoglobulin (Ig)-like domains, a transmembrane domain, a juxtamembrane domain, and an intracellular domain with a split tyrosine kinase domain and C-terminal tail. Ig-like domains are involved in ligand binding; in particular, Ig-like domains 2 in VEGFR-1 and Ig-like domains 3 in VEGFR-2 are associated with ligand-binding site and ligand-binding specificity, respectively. Upon ligand binding, VEGFRs form homo- and heterodimers, which induce tyrosine transphosphorylation of the intracellular kinase domains and activates signal transduction.

Ipilimumab prolongs overall survival (OS) with no impact on overall response rate (10%) or progression-free survival (PFS) in melanoma. A recent phase III study comparing nivolumab to everolimus in metastatic kidney cancer showed no difference in median PFS (about 4.5 months), but the median OS for patients treated with nivolumab was longer (25 versus 19.6 months, HR 0.73, P ≤ 0.0148). This clinical benefit pattern may be related to the immune response evolution over time, or to pseudo-progression, defined as an increase in size of the tumour mass on imaging from immune infiltration rather than tumour growth. Atypical/non-conventional response patterns led to the development of immune-related response criteria (irRC), which preserve the potential of benefiting from immunotherapy despite apparent disease progression on imaging. Thus, immunotherapy should be continued in the face of apparent disease progression by imaging if the patient is doing well, and until progressive disease is confirmed by a second scan 4 weeks later.

PD-1 has two binding ligands, PD-L1 (B7-H1, CD274) 11, 12 and PD-L2 (B7-DC, CD273) 13, 14, with PD-L1 being the most prominent in regulation. PD-L1 is inducibly expressed on both hematopoietic cells and non-hematopoietic cells following cell-specific stimulation. Cytokines such as IFN-γ and TNF-α upregulate the expression of PD-L1 on T cells, B cells, endothelial cells, and epithelial cells, furthering its role in the maintenance of peripheral tolerance. Data also link genetic changes seen in cancer cells to the induction of PD-L1, although this can vary by cancer type. PTEN dysfunction in human glioma cells induces Akt activation and subsequently PD-L1 expression, while human melanoma cells show no association between PTEN or Akt and PD-L1 induction 15, 16. Recent data show that PD-L1 binds to B7-1 (CD80) in addition to PD-1. While PD-L1 expression is induced on a wide array of both hematopoietic and non-hematopoietic cells, PD-L2 expression is restricted to inducible expression on DCs, macrophages, mast cells, and some B cells in response to IL-4 and IFN. The affinity of PD-L2 for PD-1 is three times greater than that of PD-L1, which indicates competition between the two ligands. Recent data confirm a second cognate receptor for PD-L2, repulsive guidance molecule B (RGMb). Despite recent research efforts surrounding PD-L2, little is known regarding the transcriptional regulation of the ligand.

Conventional ADC conjugation involves methods of alkylation of reduced interchain disulphides and alkylation or acylation of lysine residues. These conjugation methods generate heterogeneous mixtures of ADCs with variable drug per antibody ratios (DARs). This varies between zero and eight conjugated payloads per antibody. T-DM1, for example, has on average a DAR of 3.5. The position and number of payloads bound to the antibody can have profound effects on the binding to the antigen, the aggregation of the ADC, the pharmacokinetic characteristics of the antibody construct, and even the safety profile of the ADC. One mechanism by which cytotoxic compounds induce anti-tumor immunity is via direct activation and maturation of dendritic cells (DCs). The second mechanism is tumor cell intrinsic and is known as immunogenic cell death (ICD), preceding tumor cell death (reviewed in next section). Importantly, both mechanisms have been shown to engage the adaptive immune response through improved cross presentation of tumor derived antigens and priming of specific CD8+ effector T-cells, thereby triggering an immune response towards the tumor. Given the potential of both mechanisms to address some of the current limitations of single agent IO treatments, combination of IO compounds with ADCs represent a promising area of future ADC research, both pre-clinically and clinically.

There are important considerations to be made in antibody–drug conjugate design. A major consideration is the selection of the target. Arguably, the target is the most important contributor to antitumour activity and tolerability of an antibody–drug conjugate. Targets for antibody–drug conjugates can be present either on tumour cells, tumour-associated cells (eg, tumour endothelial cells), or in the tumour microenvironment. The target antigen should be expressed preferentially on the surface of a tumour compared with normal cells.12, 13, 14, 15, 16 For example, the antigens targeted by antibody–drug conjugates approved for use by regulatory agencies are expressed strongly on tumour cells, but have low expression elsewhere. Uniformly high expression of CD30 is seen in anaplastic large-cell lymphoma, whereas amplification and overexpression of HER2 happens in about 15–20% of breast cancers. In addition to differential expression on cancer cells, antibody–drug conjugate targets must have an extracellular epitope amenable to specific antibody binding and be capable of internalisation into target cells where the drug can be released.

The inclusion of a lymphodepleting conditioning regimen for patients prior to TIL infusion has resulted in durable, complete regression of melanoma. Host lymphodepletion is speculated to improve TIL functionality not only by eliminating immunosuppressive cells, such as Treg and myeloid-derived suppressor cells (MDSCs), in the tumor microenvironment but also by increasing levels of homeostatic cytokines IL-7 and IL-15. In a series of recent clinical trials, 93 patients with metastatic melanoma refractory to standard therapies were infused with autologous TILs in conjunction with IL-2 administration following three different lymphoconditioning regimens. The objective response rates ranged from 49 % to 72 % and the rate increased with a greater degree of lymph depletion. A complete tumor regression was observed in 20 of 93 patients (22 %) and this response was durable, continuing for 37 to 82 months in 19 (95 %) of those 20 patients. Other centers involved in large scale trials (such as the MD Anderson Cancer Centre and the Sheba Medical Centre) have reported consistently high response rates and long-lasting tumor regression following TIL therapy.

VEGF/VEGFR signaling is critical for vessel development during embryogenesis and tumorigenesis. In particular, VEGF-A and its receptors have been well studied in physiological and pathological neovascularization processes, such as tumor angiogenesis. Several studies have demonstrated that VEGF-A is induced by exposure to hypoxic conditions and secreted by tumor cells and tumor-associated stroma. VEGFRs are also present in both liquid and solid tumors, such as leukemia, non-small cell lung cancer (NSCLC), gastric cancer, and breast cancer. In NSCLC, the five-year survival rates of patients expressing low and high VEGF mRNA were 77.9% and 16.7%, respectively. Xenograft mouse models of VEGF-A-deficient lung carcinoma display delayed tumor growth and reduced tumor weight, as well as the inhibition of angiogenesis and induction of apoptosis. In breast cancer, VEGF-A has been reported to promote the proliferation, survival, and metastasis of breast cancer cells in vitro and in vivo. VEGF-A expression in breast cancer cells, including MCF-7, BT-474, and T47-D cells, induces proliferation and survival in vitro through Bcl-2, an anti-apoptotic protein. Additionally, inhibition of VEGF-A in a xenograft mouse model of murine breast carcinoma 4T1 cells suppresses primary tumor growth and prevents pulmonary metastases.

B7-H3 is highly expressed in normal tissues. and has been shown to be overexpressed in melanoma and numerous carcinomas in most cases, expression is associated with worse outcomes. Enoblituzumab (MGA271), a monoclonal antibody targeting B7-H3, inhibits tumor growth in renal and bladder carcinoma xenografts and is currently being investigated in at least four phase 1 clinical trials, including in combinations with pembrolizumab or ipilimumab. Preliminary single agent results (NCT01391143) report good tolerability and tumor shrinkage (2–69% at 12 weeks) across several tumor types. A monoclonal antibody against B7-H3 labeled with iodine-131 for intratumoral delivery of radiation has shown promise in preclinical studies and is being investigated in phase 1 trials. MGD009, a dual-affinity re-targeting protein bispecific for B7-H3 and CD3 is at a similar stage of development.

Patients who stop immune checkpoint blockade for reasons other than progression (primarily toxicity) may continue to benefit. Eighty-five percent of the patients who discontinued single-agent nivolumab due to toxicity had a complete or partial response, and 70% continued to respond despite stopping treatment. Similarly, although grade 3–4 toxicity occurred in 55% of the patients treated with ipilimumab/nivolumab and led to treatment discontinuation in 30%, the median OS was not reached at ≥18 months follow-up for the population as a whole, or patients who discontinued treatment due to toxicity. The optimal duration of immune checkpoint blockade thus remains unknown, and future studies should investigate this question.

Persistent expression of PD-1 by T cells is highly indicative of an exhausted phenotype, noted by a decrease in effector function This phenotype has been observed in various types of tumor-infiltrating lymphocyte (TIL) and linked to poor prognosis and tumor recurrence, highlighting PD-1 as an important molecule in regulating antitumor activity. Similar to PD-1, PD-L1 and PD-L2 also possess prognostic capacities in some human malignancies. Some clinical studies associate high expression of PD-L1 in tumors to tumor size, lymph node (LN) involvement, grade, and overall survival, while PD-L2 has generally been tied only to a trend in decreased survival that is not of statistical significance. PD-L1 generally has a much broader expression pattern compared with PD-L2. This indicates that the regulation of PD-L2 depends much more on environmental stimuli than that of PD-L1. Data from these studies provide a solid rationale for investigating the immunological mechanisms behind the clinical associations. The poor prognosis indicated by the expression of PD-1 on TILs and of PD-L1/2 on tumor cells supports the targeting of the pathway therapeutically.

The majority of ADCs in clinical trials attach the linker to a cysteine of the antibody. Cysteines are present within the intra and interchain disulphide bridges within an antibody. The presence of 4 interchain disulphide bridges within in an IgG1 antibody typically results in only eight possible conjugation sites. This reduces potential ADC heterogeneity resulting from these reactions. The linking to antibody-derived cysteine is commonly mediated through maleimide-type linkers. The two most common are maleimidocaproyl (mc) and maleimidomethyl cyclohexane-1-carboxylate (mcc). An example is the vc-MMAE linker-payload, which contains a mc spacer attached to the valine-citrulline (vc) dipeptide, which in turn is linked to the self-immolative spacer, para-amino benzyloxycarbonyl (PABC), which is attached to a monomethylauristatin E (MMAE) payload (mc-vc-PABC-MMAE). The mc spacer allows for the lysosomal vc processing via cathepsin B, liberating an unstable PABC-MMAE product, from which PABC disintegrates leaving a chemically unmodified MMAE. There are four lysine-based linkers currently used for ADCs in the clinic: N-succinimidyl-4-(2-pyridyldithio) butanoate (SPDB), N-succinimidyl-4-(2-pyridyldithio)-2-sulfo butanoate (sulfo-SPDB), maleimidomethyl cyclohexane-1-carboxylate (MCC), and hydrazone. In comparison to the limited number of cysteines, an IgG1 contains approximately 90 lysines, 30 of which are available for conjugation. This is advantageous, as no antibody structural changes are required, but they increase the potential for ADC heterogeneity during production.

ADCs are targeting the tumor cells by means of specific binding to the tumor antigen by the antibody, followed by internalization and intracellular release of the payload. The first mechanism of action of ADCs is related to the direct induction of tumor cell death by the cytotoxic payload employed. The second mechanism is triggered by certain cytotoxic compounds inducing ICD, resulting in the release of “danger signals” by the tumor cell. Such damage-associated molecular patterns (DAMPs) include calreticulin (CRT), high-mobility group protein B1 (HMGB1), and adenosine triphosphate (ATP). These molecules released by the dying tumor cell stimulate dendritic cell maturation and activation. DAMPs are recognized by specific cell surface receptors expressed on dendritic cells, known as DAMP receptors. The third mechanism of action of ADCs is the direct activation and maturation of dendritic cells by the cytotoxic payloads released in the tumor microenvironment. The molecular mechanisms mediating direct dendritic cell maturation and activation by cytotoxic compounds, including tubulin inhibitors, remain to be determined. Combined, these three mechanisms of action of ADC payloads can lead to increased cross priming of CD8+ T-cells in the lymph nodes by activated dendritic cells, CD8+ T-cell infiltration into the tumor core followed by cytolytic killing of tumor cells by granzyme B and perforin released by effector T-cells.

Targets of antibody–drug conjugates are mostly unmutated proteins; however, as mutations in cancer cell-surface proteins are discovered, development of antibody–drug conjugates with greater selectivity might be possible. Technologies for antibody discovery, including phage display libraries and humanised mice, can produce fully human antibodies, and mouse antibody humanisation can result in highly specific non-immunogenic antibodies. The antibody–drug conjugate–target complex must also internalise into the target cells where the drug can be released.

Ongoing efforts aim not only at improving TIL therapy but also on broadening TIL to battle malignancies. Advances in T cell culturing methods and genetic T cell engineering ensure that clinically relevant numbers of tumor-specific T cells can be generated and delivered as therapy in a timely manner. There are two basic strategies that are being explored in clinical testing of engineered T cells. The first strategy involves the expression of T cell receptor (TCR) α and β chains that confer the engineered T cell with antigen-specificity of the transferred TCR. This therapy is potentially accessible to any patient whose tumor carries the cognate human leukocyte antigen allele and expresses the target antigen recognized by the TCR. However, the clinical use of highly avid TCRs has been associated with significant secondary destruction of healthy tissues expressing the same target antigen. Ongoing efforts are focused on improving gene transfer efficiencies, designing TCR structural modifications, and identifying target antigens that are highly selective for tumor cells rather than normal cells. Chimeric antigen receptors (CARs) constitute the second approach and consist of an Ig variable domain fused to a TCR constant domain. The advent of CARs omits the need for tumor cells to carry a functional antigen processing machinery or to express antigens through MHC class molecules since the engineered T cells obtain the antigen-recognition properties of antibodies and are thus potentially targeted against any cell surface target antigen.

PDGFs are assembled in the endoplasmic reticulum as inactive precursor disulfide-linked homodimers composed of AA-, BB-, CC-, and DD-polypeptide chains, and the heterodimer PDGF-AB. Generally, the structure of PDGF family members consists of a growth factor domain of about 100 amino acid residues that is involved in receptor-binding and dimerization and a prodomains of various length that is an attached amino acid sequence found in N-terminal extension. These isoforms are activated by proteolytic processing: PDGF-AA, -AB, and -BB have short N-terminal extensions that are cleaved by intracellular proteolysis in secretory vesicles, whereas PDGF-CC and -DD have a distinct protein domain called the CUB-domain, which is cleaved by extracellular proteolysis. Active isoforms of PDGF bind to the α- and β-RTKs, PDGFRα and PDGFRβ, and induce the activation of downstream signaling pathways following tyrosine autophosphorylation.

In mouse models, blocking VISTA increases immune infiltration in tumors while preferentially decreasing myeloid-derived suppressor cells. Combining anti-VISTA and anti-PDL1 agents decreases tumor size and increases survival. In humans, VISTA+ TILs were reported in 46% of gastric cancer patients, with a small percentage of tumor cells also expressing VISTA. Recently, Oliveira P et al. showed that epigenetic factors can regulate VISTA expression in gastric cancer cell lines and that VISTA is associated with the epithelial-mesenchymal transition phenotype. In oral squamous cell carcinoma patients, VISTA expression associates with poor overall survival in patients with low CD8+ TILS. VISTA+ TILs and macrophages are upregulated in prostate cancer and melanoma patients following ipilimumab (anti-CTLA-4) treatment, with a greater percentage of VISTA+ macrophages being of the immunosuppressive M2 phenotype, suggesting that VISTA may represent a compensatory resistance mechanism. As of this writing, only phase 1 clinical trials of anti-VISTA agents are open, with combination strategies anticipated once safety is established.

The safety profile of checkpoint inhibition differs from chemotherapy or targeted therapy. Immune-related adverse events (irAEs) result from immune activation once inhibition by CTLA-4 and/or PD-1 is released. The most frequent side effects of ipilimumab involve the skin (pruritus and cutaneous rash), gastrointestinal tract (colitis and diarrhoea), liver (autoimmune hepatitis) and endocrine system (thyroid dysfunction and hypophysitis). Immune-related neuropathy, myositis, arthritis and uveitis also occur uncommonly. The safety profile of PD-1 blockade is similar and also includes pneumonitis. Pneumonitis is typically grade 1–2 and does not result in cessation of treatment. The incidence of grade 3–4 (severe) irAEs is 10–15% with ipilimumab, 5% with nivolumab or pembrolizumab, and 55% with ipilimumab/nivolumab. Combination therapy did not cause new toxicities or deaths. Early diagnosis and prompt treatment with steroids are critical for the effective management of irAEs. Algorithms have been developed to guide treating physicians.

Appreciating the consequences of the upregulation of the PD-1/PD-L1/2 axis aids our progress in manipulating an immunosuppressive cancer microenvironment. The cytoplasmic tail of PD-1 contains two signaling motifs. One is an immunoreceptor tyrosine-based inhibitory motif (ITIM) and the other is an immunoreceptor tyrosine-based switch motif (ITSM). Binding of PD-L1 or PD-L2 to PD-1 on activated T cells, along with TCR signaling, leads to phosphorylation of the cytoplasmic domain tyrosines and recruitment of a Src homology 2-containing tyrosine phosphatase (SHP-2) to the ITSM. Consequently, SHP-2 dephosphorylates TCR-associated CD-3ζ and zeta chain-associated protein kinase 70 (ZAP70), resulting in inhibition of downstream signaling including blocking phosphoinositide 3-kinase (PI3K) and Akt activity, disrupting glucose metabolism and IL-2 secretion.

Peptide linkers of the protease sensitive class are more stable than an acid-sensitive-based linker and has been successfully used in the clinically approved brentuximab vedotin linking the payload monomethylauristatin E (MMAE) to the anti-CD30 antibody. Peptide linkers, such as the valine-citrulline (vc) dipeptide, contain an engineered lysosomal-specific protease (i.e., cathepsin B) cleavage site, allowing for the chemically unmodified release of cytotoxic payload from the lysosome compartment. Taking advantage of the increased glutathione concentration within tumour cells, the reducible, disulphide bond-based glutathione sensitive linkers are a third major sub-category of cleavable linkers. However, low levels of glutathione is also present within the circulation (2.8 ± 0.9 µM). Increasing levels of disulphide steric hindrance through the insertion of methyl groups has proportionally reduced their untimely release within the circulation. Lorvotuzumab mertansine (IMGN901) is the most clinically advanced ADC employing a hindered disulphide linker. This anti-CD56 targeting ADC with a maytanisinoid (DM1) payload has shown efficacy against Merkel cell carcinoma in early phase trials and has been granted orphan status by the U.S. Food and Drug Administration (FDA).

Immuno-histochemical analysis of tumor samples from patients treated with an ADC, SGN-35, comprised of an anti-CD30 antibody conjugated to vcMMAE revealed significant changes in the population of inflammatory cells. In SGN-35 treated Hodgkin’s lymphoma tumors, a 2–3 fold increase in the number of intra-tumoral CD8+ effector T-cells was noticed. These studies provided evidence for tubulin destabilizing poisons to stimulate the cancer immunity cycle. However, additional studies are required to identify the molecular pathways engaged by tubulin inhibitors leading to activation and maturation of tumor resident DCs. A deeper understanding of the molecular events triggered by tubulin inhibitors leading to DC activation will help to inform rational combination treatments of IO compounds. While increased numbers of CD8+ effector T-cells in preclinical and clinical tumors following treatment with cytotoxic compounds were reported, additional preclinical experiments combining ADCs with IO compounds are needed to understand how broadly these observations apply for ADC development. In conclusion, identification of optimal combination regimens between ADCs and IO compounds holds strong promise to overcome the limitations of current immune checkpoint inhibitors by increasing the recruitment and infiltration of antigen specific CD8+ effector T-cells to the tumor.

Currently, two antibody–drug conjugates have been approved for use in the USA and Europe: brentuximab vedotin and ado-trastuzumab emtansine. In general, these antibody–drug conjugates are well tolerated, with toxic effects consistent with the known mechanism of action of the cytotoxic payloads—eg, neutropenia and neuropathy with brentuximab vedotin, and rises in hepatic aminotransferase levels with ado-trastuzumab emtansine. However, the exact mechanisms of toxic effects that are attributable to the antibody–drug conjugates are complex, with contributions from every component of the conjugate—ie, the monoclonal antibody, the linker, and the cytotoxic payload. Mechanisms include non-specific systemic release of the cytotoxic drug because of premature lysis of the linker, and internalisation of the antibody–drug conjugate by cells not expressing the target. For example, thrombocytopenia, a common toxic effect of ado-trastuzumab emtansine, is thought to be due to the internalisation of the antibody–drug conjugate by megakaryocyte precursor cells via an Fc receptor-mediated process, after which the cytotoxic DM1 component of ado-trastuzumab emtansine impedes differentiation to mature megakaryocytes and subsequent platelet formation.36 Toxic effects can also result from a bystander effect, which refers to non-selective cytotoxicity of target-negative cells that are located in close proximity to target-positive cells.

Human cancers carry a multitude of somatic gene mutations and epigenetically altered genes, the products of which are potentially recognizable as foreign antigens. Although an endogenous immune response to cancer is observed in preclinical models and patients, this response is not efficient because tumors induce tolerance among tumor-specific T cells and by expressing ligands that bind inhibitory receptors and dampen T cell functions within the tumor microenvironment. One approach to trigger antitumor immune responses has been termed “checkpoint blockade”, referring to the blockade of immune-inhibitory pathways activated by cancer cells.

Signal transduction via PDGF and PDGFR has been implicated in several physiological and pathological processes, including wound healing, embryogenesis, bone formation, and tumor growth. PDGF and PDGFR signaling is also associated with the normal development of the kidney, brain, and respiratory systems. Under physiological conditions, endothelial-derived PDGF-BB activates pericytes and smooth muscle cells through paracrine signals, which in turn, affects vascular remodeling, maturation, and stability. PDGF-BB also modulates endothelial proliferation, migration, and tube formation, and contributes to angiogenesis in vitro; in contrast, PDGF-AA has no such effect or induces a weak angiogenic response. PDGF-CC promotes the angiogenesis of the mouse cornea through PDGFRαα and -αβ dimers. In addition, PDGF-CC is abundantly expressed in angiogenic tissues, such as the placenta, ovary, and embryo. Finally, PDGF-DD plays an important role in increasing interstitial fluid pressure, macrophage recruitment, and blood vessel maturation during angiogenesis in the skin and skeletal muscles. Interestingly, inhibition of PDGF-DD also suppresses ocular neovascularization.

Many preclinical reports on GITR anti-tumor in vivo activity have used a murine GITR agonist IgG1 monoclonal antibody (DTA-1) in solid cancer mouse models. GITR agonism, in addition to depleting and inhibiting Tregs (similar to CTLA-4 antagonists and OX40 agonists), suppresses myeloid derived suppressor cells and IL-10 production. Combining of GITR agonists with other immune modulating agents leads to additive antitumor effects. One of the most interesting finding from these studies is that GITR agonists suppress tumor growth and increase survival not only in immunogenic tumor models (colon, bladder, lung, melanoma) but also in poorly immunogenic tumors (breast, B16 melanoma mouse model, ovarian), putting GITR agonists in a unique position in comparison to other immune checkpoint inhibitors for which pre-existing immunity appears a prerequisite for the agents to work. By immunohistochemistry or flow cytometry, GITR expression has been reported in many human solid cancers. In breast and endometrial cancer patients, GITR expression on Tregs is higher in TILs than in the peripheral blood. The prognostic value of GITR+ TILs has yet to be investigated in detail. At least four GITR agonists are being investigated alone and in combinations with other checkpoint inhibitors in early phase clinical trials.

Immunotherapy can be combined effectively with chemotherapy, targeted therapy, and radiotherapy. The efficacy of chemotherapy combined with immunotherapy depends on the drug, and the relative timing of immunotherapy and chemotherapy. Dacarbazine combined with ipilimumab was more effective than chemotherapy alone in melanoma. In NSCLC, the combination of ipilimumab with chemotherapy (paclitaxel and carboplatin) was evaluated on a concurrent schedule (four doses of ipilimumab plus chemotherapy followed by two doses of placebo plus chemotherapy) and a phased schedule (two doses of placebo plus chemotherapy followed by four doses of ipilimumab plus chemotherapy) [46]. Median PFS (5.1 and 4.1 months) and OS (12.2 and 9.7 months) were longer with the phased than with the concurrent regimen. There are similar data for SCLC [47]. More recently, a randomised phase 2 study evaluated the addition of pembrolizumab to carboplatin and pemetrexed for advanced non-squamous NSCLC, with a ORR of 55% for pembrolizumab with chemotherapy compared with 29% for chemotherapy alone. Importantly, there was no apparent relationship between PD-L1 expression and response in this study. Combining targeted therapy with immunotherapy is another interesting approach. The addition of BRAF or BRAF and MEK inhibitors to ipilimumab is limited by a high incidence of side effects, and combinations involving anti-PD-1/PD-L1 agents might be more feasible. Increased understanding of the immunological effects of conventional and targeted cancer therapy is essential to guide the design of clinical studies testing them in combination with immunotherapy.

mAbs have been developed for cancer immunotherapy by enhancing T cell function via blockade of the binding between PD-1 and PD-L1 or PD-L2. Many of these studies have shown that blockade of PD-1 alone or PD-L1 leads to an increase in T cells and IFN-γ at the tumor site, along with decreases in the percentages of the highly immunosuppressive myeloid-derived suppressor cell (MDSC) population. Increase in the effector-to-suppressor cell ratio usually supports an antitumor microenvironment. These results demonstrate that the neutralization of PD-1, PD-L1, or PD-L2 can be effective in controlling tumor growth by changing the dynamic of the tumor microenvironment.

The first generation of ADC payloads were developed around clinically approved cytotoxic agents. The well-known safety and efficacy of agents such as doxorubicin, 5-fluorouracil, and methotrexate was seen as advantageous; however, these agents did not achieve clinical benefit as ADCs due to their moderate cytotoxic potential, lack of selectivity, and low intracellular drug concentration. Subsequent approaches have involved identifying drugs that were suitable for antibody conjugation and would deliver an effective cytotoxic dose, even with small efficiencies in intracellular delivery. The two most commonly used cytotoxic payloads focus on either the tubulin-targeting anti-mitotic agents or DNA-damaging drugs. The majority of payloads in current clinical development reside in the broad class of tubulin-targeting anti-mitotic agents, the maytansinoids and auristatins. These agents inhibit spindle and microtubule dynamics during interphase, resulting in G2/M mitotic arrest. Payloads that target the microtubule network may not be limited to inhibiting cellular division and may also inhibit intracellular trafficking of proteins essential for cancer cell function. This is particularly important for ADC-based targeting of tumour initiating cells (TIC) or quiescent cell populations.

In conclusion, ICD is a mechanism whereby standard chemotherapy can stimulate the immune system and induces strong inflammatory responses in tumors. Certain cytotoxic agents induce the sequential generation of secreted markers which activate dendritic cell maturation, leading to MHC mediated tumor antigen cross presentation, and ultimately antigen specific CD8+ T-cell recruitment to the tumors. In particular, treatment regimens including anthracyclines (Doxorubicin) and oxaliplatin appear attractive for combination with IO compounds, as both compounds were associated with a strong induction of immunogenic tumor cell death, thereby potentially enhancing anti-tumor immune responses. Some of the compounds are currently being developed in clinical programs as ADC payloads and were shown to induce ICD. However the relative contributions of immunogenic cell death to the overall pharmacodynamic effects of ADCs in cancer patients remain poorly understood and additional studies are needed to fully understand and harness the ICD induced by ADCs.

In 2011, brentuximab vedotin became the first antibody–drug conjugate to be approved by the FDA under its accelerated approval regulations.37 This agent is a CD30-directed antibody–drug conjugate consisting of the chimeric anti-CD30 IgG1 antibody, the microtubule-disrupting agent monomethyl auristatin E, and a protease-cleavable linker that attaches the cytotoxic agent covalently to the antibody. After binding CD30, the antibody–drug conjugate is internalised rapidly and transported to lysosomes, where the peptide linker is cleaved selectively. Monomethyl auristatin E is then released into the cell, binds tubulin, and prompts arrest of the cell cycle between the gap 2 phase and mitosis, causing cell apoptosis.38 Brentuximab vedotin is approved for patients with relapsed or refractory CD30-positive Hodgkin's lymphoma after autologous stem-cell transplantation (ASCT) or after at least two previous therapies when ASCT or multiagent chemotherapy is not a treatment option. Moreover, brentuximab vedotin is approved for use in patients with relapsed or refractory systemic anaplastic large-cell lymphoma. It is given intravenously at a dose of 1·8 mg/kg over 30 min every 3 weeks. Approval of brentuximab vedotin for treatment of these two disorders was based on data from two single-arm phase 2 trials.39, 40 In 77 patients with Hodgkin's lymphoma who relapsed after ASCT and were subsequently treated with brentuximab vedotin,40 the number of patients who achieved an objective response was 58 (75%; 95% CI 65–83) and the number of patients with a complete response was 26 (34%; 25–44). The median duration of the objective response was 6·7 months and 20·5 months for those 58 patients who achieved a complete response. For 58 patients with relapsed or refractory anaplastic large-cell lymphoma treated with brentuximab vedotin,39 the number of patients achieving an objective response was 50 (86%; 95% CI 77–95), and 33 achieved a complete response (57%; 44–70). The median duration of the objective response was 12·6 months (95% CI 5·7–not estimable [NE]) and 13·2 months (10·8–NE) for those 50 patients who achieved a complete response. Among all patients treated in both trials, the most common adverse events were neutropenia, peripheral sensory neuropathy, fatigue, nausea, anaemia, upper respiratory infection, diarrhoea, pyrexia, rash, thrombocytopenia, cough, and vomiting.41 Important serious adverse reactions reported were Stevens–Johnson syndrome, tumour lysis syndrome, and progressive multifocal leukoencephalopathy. Phase 3 trials of brentuximab vedotin as monotherapy and in combination with chemotherapeutic agents are ongoing in patients with CD30-positive cutaneous T-cell lymphoma (NCT01578499), Hodgkin's lymphoma (NCT02166463), and CD30-positive mature T-cell lymphoma (NCT01777152).

A phase I clinical trial with ipilimumab (anti-CTLA-4) combined with nivolumab (anti-PD1) reported tumor regression in 50 % of treated patients with advanced melanoma. A more recent randomized, placebo-controlled phase II study comparing ipilimumab combined with nivolumab versus ipilimumab alone reported even better responses. Patients with previously untreated metastatic melanoma who received the combination treatment showed an objective response rate of 61 % while, of the patients assigned to the ipilimumab monotherapy, only 11 % demonstrated an objective response. According to a recent, randomized, three-arm phase III clinical trial which compared monotherapy with either ipilimumab or nivolumab to their combination in patients with melanoma, nivolumab alone was less toxic and showed greater clinical benefit than ipilimumab alone. Nivolumab as monotherapy and in combination with ipilimumab demonstrated better objective response rates compared to ipilimumab. From this study the overall survival results are anticipated to shed light on the full effect of combination immunotherapy. On account of these promising efficacy results, there are ongoing clinical trials with anti-CTLA-4 (ipilimumab, Bristol-Myers Squibb or tremelimumab) plus anti-PD1 or anti-PD-L1 in other tumor types such as renal cell carcinoma, NSCLC, small-cell lung, triple-negative breast, pancreatic, gastric, and bladder cancer.

Nesvacumab is a fully human monoclonal antibody developed by Regeneron Pharmaceuticals (Tarrytown, NY, USA). It selectively binds Ang-2 with high affinity (Kd = 24 pM) and blocks Ang-2 binding to Tie-2 . In preclinical studies, nesvacumab significantly inhibited tumor growth in xenograft models of prostate adenocarcinoma, colorectal adenocarcinoma, and epidermoid carcinoma. Nesvacumab also had no apparent adverse toxic effects in Sprague Dawley rats and cynomolgus monkeys. On the basis of its safety and anti-tumor activity in preclinical studies, nesvacumab entered phase I trials for advanced solid tumors, which reported that nesvacumab had anti-tumor activity within an acceptable safety profile [153]. In addition, the combination trials of nesvacumab with aflibercept have completed enrollment (Clinical trials information: NCT01688960).

The majority of transplantable tumour models assess the effects of immunotherapy early after tumour implantation in mice with small tumours. This is unlikely to recapitulate the natural development of cancer in patients, where tumours develop over time and induce immune tolerance. Overcoming tumour-induced tolerance is a major hurdle that must be overcome to generate potent anti-tumour immunity in patients. Genetically engineered mouse models (GEMMs) that develop tumours spontaneously more closely recapitulate both the natural tumour microenvironment and the tumour-specific immune tolerance that characterise human tumours. However, these models may be limited in that they typically do not have the high mutational burden that correlates with the response of some human tumours to immune checkpoint blockade.

Pidilizumab (CT-011) was the first mAb to PD-1 to reach clinical trials. It was initially identified as a mAb binding to the B lymphoblastoid cell line that stimulated murine lymphocytes and showed antitumor activity in mice [30]. It stimulated human peripheral blood lymphocytes and enhanced cytotoxicity toward human tumor cell lines. The first Phase I trial with pidilizumab recruited patients with hematologic malignancies, including acute myeloid leukemia (AML), chronic lymphocytic leukemia (CLL), non-Hodgkin's lymphoma (NHL), Hodgkin's lymphoma (HL), and multiple myeloma (MM). Dose levels ranged from 0.2 to 6 mg/kg. A maximum tolerated dose (MTD) was not reached and the drug was well tolerated. Of the 17 patients enrolled in the study, one patient experienced a complete response, four had stable disease, and one had a mixed response, amounting to a 33% clinical benefit rate. Durable responses of greater than 60 weeks were noted. This was followed by two Phase II clinical trials 31, 32. Patients with diffuse large B cell lymphoma (DLBCL) or primary mediastinal B cell lymphoma (PMBCL) who underwent autologous hematopoietic stem cell transplantation (ASCT) and who had chemosensitive disease were treated with pidilizumab at 1.5 mg/kg every 42 days for three cycles starting 30–90 days post-transplantation. The study enrolled 72 patients. Sixteen-month progression-free survival (PFS) for eligible patients was 72%, meeting the primary end point of the study. Intention-to-treat analysis revealed a 16-month PFS of 68%. The overall response rate for patients with measurable disease after ASCT was 51%. The most common grade 3 or 4 toxicities included neutropenia and thrombocytopenia. Correlative studies of select lymphocyte subsets revealed an increase in the number of activated CD25+PD-L1+CD4+ T cells, PD-L1+PD-L2+CD14+ monocytes, and circulating peripheral and central memory CD8 T cells as well as central memory CD4 T cells. These results suggest that pidilizumab may reverse PD-1-mediated inhibition of T cell survival and proliferation.

The US Food and Drug Administration (FDA) has approved several ADCs for clinical use in haematological malignancies and some solid tumour types. Many more ADCs are in late stage clinical development and have shown promising initial results. The first ADC to gain FDA approval (in 2000) was gemtuzumab ozogamicin (Mylotarg®) for patients over the age of 60 who suffered their first relapse of CD33 positive acute myeloid leukaemia (AML) and were ineligible for chemotherapy. This ADC comprised a humanised IgG4 CD33 antibody coupled with a DNA-binding calicheamicin derivitive. Although this agent sparked renewed clinical interest in the development and design of the ADC as a therapeutic class, this particular agent was voluntarily withdrawn by its manufacture following the post-approval Phase III study, which failed to identify an improved survival within the chemotherapy and gemtuzumab ozogamicin combinational groups when compared to chemotherapy alone with previously untreated AML. Given the heterogeneous nature of AML, it is still believed that a subpopulation of AML patients would benefit from the addition of a gemtuzumab ozogamicin regime. More recent randomised studies have displayed statistically significant overall survival with gemtuzumab ozogamicin with or without chemotherapy. The reason for this discrepancy still remains unclear, but it could in part be due to the lower doses of daunorubicin used in most of these trials or the fractionation of gemtuzumab ozogamicin treatment, which appeared to be better tolerated, allowing for a greater overall dose to be administered. These studies have reignited the clinical debate regarding the clinical use of gemtuzumab ozogamicin.

Despite the potential for synergies between ADCs and IO compounds, surprisingly few reports have studied combination treatments between both modalities. One reasons for this knowledge gap is the difficulty to assess the contributions of the host species immune-system to the anti-tumor activities of ADCs when tested in mice xenografted with human tumors. The use of immune-compromised mice, devoid of most lymphoid cell lineages, represented a key milestone in the field of preclinical oncology research, as these mice enabled growth of human tumors in the absence of the adaptive mouse immune system. These technical restrictions may however have prevented a more in depth analysis of the ability of ADCs to stimulate anti-tumor immunity in preclinical models, despite the well characterized pro-inflammatory properties of certain cytotoxic compounds employed as ADC payloads. The development of surrogate ADCs, which bind to the corresponding mouse antigen, may enable testing of ADCs in syngeneic mouse models to better understand the therapeutic potential of combining ADCs with IO compounds. Alternatively, syngeneic mouse tumor cell lines overexpressing the human antigen may have utility to study the combinatorial effects of current clinical ADCs with IO compounds. Ultimately, additional clinical studies will be needed to better assess whether ADCs can stimulate anti-tumor immunity significantly and whether combining ADCs with IO compounds will translate into improved survival or lead to negative interference with immune effector cells, which are required to mediate anti-tumor activity of IO compounds. Sequential treatment regimens starting with ADC followed by IO, and/or the reduction of neutropenia by second generation ADC compounds, may be required to avoid potential negative interference when combining both modalities in clinical trials.

Approval of ado-trastuzumab emtansine was based on results of a phase 3 trial in 991 patients with HER2-positive metastatic breast cancer.45 In that trial, patients were randomly allocated either ado-trastuzumab emtansine (n=495) or lapatinib plus capecitabine (n=496). The co-primary endpoints were progression-free survival and overall survival. Significant improvements in these endpoints were recorded in patients assigned ado-trastuzumab emtansine compared with those allocated lapatinib plus capecitabine. Median progression-free survival was 9·6 months in the ado-trastuzumab emtansine group, and 6·4 months for the lapatinib plus capecitabine group (hazard ratio [HR] 0·65, 95% CI 0·55–0·77; p<0·001). Median overall survival was 30·9 months and 25·1 months, in the ado-trastuzumab emtansine group and the lapatinib plus capecitabine group, respectively (0·68, 0·55–0·85; p<0·001). The most common adverse drug reactions were fatigue, nausea, musculoskeletal pain, thrombocytopenia, headache, increased aminotransferase levels, and constipation. Adverse events occurring more frequently in the ado-trastuzumab emtansine group than in the lapatinib plus capecitabine group included thrombocytopenia (31% vs 3%), constipation (27% vs 11%), increased aminotransferase levels (29% vs 14%), headache (28% vs 15%), epistaxis (23% vs 8%), arthralgia (19% vs 8%), pyrexia (19% vs 8%), dry mouth (17% vs 5%), and myalgia (14% vs 4%). Rare but serious adverse events included hepatotoxicity, which can potentially lead to liver failure, and reduced left-ventricular ejection fraction.

This field is rapidly advancing and extremely active for drug development and clinical trials. Toxicities still need to be determined, particularly of combination strategies, which risk enhanced autoimmune side effects. Given limited resources and patients available for clinical trials, emerging agents with acceptable toxicity will need to be prioritized based on factors including not only the strength of evidence implicating their role in cancer immuno-oncology, but also their frequency of expression in areas of clinical need not well-served by existing agents. While many of these agents may not ultimately find a place in the growing armamentarium of anticancer immuno-oncology drugs, the pathways under investigation are so many, and the early data so promising, that it is likely that several truly effective new treatment strategies will emerge.