Immunotherapy in Oral Cancer: A Fourth Dimension of Cancer Treatment

Although many components of the immune system are involved in the antitumor immune response, the most critical immune cell is the T cell. The Cancer Immunity Cycle, as described by Chen and Mellman, explains the steps required for the adaptive immune system to target tumors, including uptake and presentation of antigens by dendritic cells (DC), priming and activation of T cells in the lymph nodes, homing of T cells to the site of the tumor, and T cell recognition and destruction of tumor cells [1] (Fig. 10.1). Cancer cell death then exposes more potential antigens to immune recognition, which can perpetuate the cycle. Each step in this sequence has biochemical stimulators and inhibitors, and appreciation of these steps is critical in developing strategies to overcome immune evasion by tumors.

Open image in new window

The central role of the immune system in both the prevention and evolution of cancer has been described by Schreiber et al. in the concept of immunoediting [2]. Immunoediting starts with the surveillance of the body for abnormal cells by the immune system. In some cases, the immune system is successful in targeting and eliminating a tumor and preventing the occurrence of clinically detectable cancer. However, sometimes the immune system cannot eliminate a subset of the tumor cells, as those cells have acquired phenotypes that subvert immune targeting. After a period of equilibrium, where these resistant tumor cells exist in quiescence, they may begin to proliferate and spread. This is called immune escape and is now recognized as one of the hallmarks of cancer [3].

Advances in whole-exome sequencing have provided great insights into the molecular landscape of head and neck cancer. Remarkable efforts by several groups, including The Cancer Genome Atlas (TCGA), have identified the most commonly altered molecular pathways involved in head and neck squamous cell carcinoma (HNSCC). These data have shown that HNSCCs have a moderate-to-high mutational burden and are heterogeneous tumors. The most common involved pathways are tumor-suppressor genes such as p53, which are either inactivated or mutated. Successful restoration of functional tumor-suppressor pathways in HNSCC through targeted therapy has so far remained elusive. Oncogenic pathways which have been identified in a significant percentage of tumors, such as PI3K, are theoretically more targetable than tumor suppressors. However, results of therapies such as PI3K and mTOR inhibitors have been mostly disappointing [4]. The heterogeneity of cells within a tumor, even those tumors that test positive for targetable mutations, may play an important role in the difficulties encountered with targeted therapies [5].

While the relatively high mutational rate of HNSCC creates difficulty in applying targeted therapies, it may be advantageous for the application of immunotherapy through the presence of more mutations that may be recognized as non-self by the immune system. However, there are numerous escape mechanisms that tumors employ to suppress the natural antitumor capabilities of immune cells. Impaired antitumor responses in HNSCC may be caused by alterations in the generation, processing, and/or presentation of T cell epitopes derived from tumor-associated antigens (TAA) by human leukocyte antigen (HLA) class I and/or class II molecule [6]. Other mechanisms for evading immune targeting by tumors include upregulation of immune-suppressive “checkpoint” ligands (e.g., PD-L1) [7], release of inhibitory cytokines by suppressive immune cells including regulatory T cells (Treg), myeloid-derived suppressor cells (MDSC) and tumor-associated macrophages (TAM) in the tumor microenvironment [8, 9, 10], or the secretion of immune-suppressive mediators (e.g., TGF-beta) [11]. Knowledge of the specific mechanisms of immune suppression in HNSCC has led to numerous approaches to overcome immune suppression in the tumor microenvironment (Fig. 10.2).

Open image in new window

Treatment and prognosis of HNSCC vary depending on anatomical location. For oropharyngeal SCC, carcinogenesis is often driven by infection with high-risk strains of HPV, usually HPV-16. It has been shown that the molecular pathways and immunobiology of HPV-positive tumors are distinct from HPV-negative disease [12]. In addition, HPV-positive HNSCCs respond better to standard therapies and most immunotherapies [13]. The significance of HPV in oral cancer specifically remains uncertain, and there is currently no standard method for identifying HPV as the etiologic agent in oral cavity SCC that is both accepted and available universally [14]. This chapter is meant to focus on cancers arising in the oral cavity and therefore will limit the discussion to HPV-negative HNSCC.

Current therapies and those in development are focused on strategic targeting of the various steps in the Cancer Immunity Cycle. The great promise of immunotherapy lies with the potential for lasting, or durable, responses. This is referred to as the “tail at the end of the survival curve” (Fig. 10.5) and is perhaps the most exciting prospect of immunotherapy.

There is abundant preclinical evidence supporting a combinatorial approach to cancer immunotherapy. For example, CTLA-4 acts relatively early in the cancer immunity cycle during T cell priming and activation, while PD-1 comes into play later in the cycle by modulating immune effector cell function within tumors. Additionally, blockade of one immune checkpoint can lead to the increased expression of other checkpoints by tumor cells leading to immune escape [29]. Therefore, it has been hypothesized that adding one or more additional CPI to PD-1/PD-L1 therapy may improve response rates, particularly in PD-L1-negative patients, and prevent resistance to single-agent therapy (Fig. 10.6).

Open image in new window

Two large studies are currently testing nivolumab (anti-PD-1) alone or in combination with ipilimumab (anti-CTLA-4) in R/M HNSCC. CheckMate714 is a double-blinded, randomized, phase II study in both the platinum-refractory and first-line settings, and CheckMate651 is an open-label, randomized phase III study of the same combination as first-line therapy compared to chemotherapy (NCT02823574, NCT02741570). This combination has been approved for the treatment of patients with advanced melanoma and advanced renal cell carcinoma. Combination nivolumab + ipilimumab has also shown significant promise in the treatment of patients with advanced non-small-cell lung cancer [33]. Data from these trials are not yet mature and will be presented at future meetings.

Tremelimumab (anti-CTLA-4) is now being tested in numerous clinical trials for HNSCC in combination with durvalumab (anti-PD-L1). The CONDOR trial was a phase II, randomized trial of durvalumab and tremelimumab alone and in combination in patients with platinum-refractory R/M HNSCC who had low PD-L1 levels. Median overall survival was 7.6 months in the combination arm, 6.0 months for durvalumab alone, and 5.5 months for tremelimumab alone [34]. EAGLE (NCT02369874) and KESTREL (NCT02551159) are randomized phase III trials studying this combination in platinum-refractory disease and as first-line treatment of R/M HNSCC, respectively.

There is a large body of preclinical evidence for the antitumor activity of other combinations of C. In fact, many of the newer CPI are being developed as combinations from the start, most commonly in combination with PD-1/PD-L1 pathway antagonists. For example, all the current anti-TIM-3, anti-TIGIT, and anti-LAG-3 trials evaluating tolerability and efficacy in advanced solid tumors are testing these agents in combination with anti-PD-1 or anti-PD-L1 antibodies. While clinical trials for these other combinations are still in early phases for the most part, there are early reports of success such as the phase I/IIa trial of relatlimab (anti-LAG-3) with nivolumab (anti-PD-1) in advanced solid tumors. In an expansion cohort of melanoma patients who were refractory to PD-1 therapy alone, 11.5% had objective responses and 37.7% had stable disease with the combination. Patients with ≥1% LAG-3 expression had an ORR of 18%, and patients who also had prior exposure to CTLA-4 had an ORR of 24% [35]. This trial includes head and neck patients as well, and study completion is expected in late 2019.

In addition to T cell receptor (TCR) recognition of MHC-presented antigens, the activation of T cells requires specific costimulatory signals (Fig. 10.5). TCR ligation without the second costimulation signal leads to T cell anergy and immune tolerance [36]. The development of agonist antibodies that activate costimulatory receptors has added a new dimension to cancer immunotherapy. While the activity of checkpoint inhibitors has been described as “releasing the breaks” on the immune system, costimulatory agonists have been described as “stepping on the gas.” Preclinical studies have shown that costimulatory agonists have synergistic activity with CPI, and there are currently many such agents being tested in clinical trials alone and in combination with other immunotherapies.

Cytokine therapy with IFN-α or IL-2 has been utilized in the treatment of cancer for over 30 years. IFN-α use has greatly diminished due to marginal effectiveness and concerns for acute toxicities [51]. High-dose IL-2 (HD-IL2) is also associated with severe, although generally reversible, acute toxicities including hypotension, renal failure, and thrombocytopenia. However, HD-IL2 is capable of producing durable responses in a minority of patients with metastatic melanoma and renal cell carcinoma (RCC). Patients must be selected for good functional status and organ function, and the therapy must be administered in specialized centers with hospitalization required for the duration of treatment [52]. The antitumor activity of intravenous HD-IL2, however, has not been demonstrated in cancers outside of melanoma and RCC.

The well-established immune-stimulating effects of various cytokines have sustained interest in researching clinical applications for cancers outside of melanoma and RCC, including HNSCC. In 2002, results were published from a randomized phase III trial of perioperatively administered perilymphatic IL-2 in Stage II-IVb oral or oropharyngeal SCC. This therapy appeared to be safe and efficacious, significantly improving disease-free survival and OS. The 5-year survival rate in the perilymphatic IL-2 group was 73%, compared with 55% in the control group, and disease-free survival (DFS) rates were 64% and 51%, respectively [53]. Despite the impressive improvement in outcomes from this randomized, phase III study, perilymphatic IL-2 has not received much attention in the field. One possible reason for this includes the intensive dosing regimen used in the study, including daily injections into the neck for 10 days prior to surgery, and injections five times a month for up to a year after surgery.

Building on this work, another group is investigating in HNSCC, a blend of cytokines termed IRX-2, which primarily consists of IL-2, IL-1β, IFN-γ, and TNF-α. The formulation is administered perilymphatically and is combined with systemic low-dose cyclophosphamide for Treg depletion, along with indomethacin and zinc to inhibit immunosuppressive elements within the systemic circulation and TME. In 2011, nonrandomized phase II data were presented from 27 patients with Stage II-IVa HNSCC receiving perilymphatic IRX-2. The regimen was found to be well-tolerated, with no grade 4 or higher toxicities. 3-year OS and DFS rates were 69% and 62%, respectively; median DFS and OS were not reached after follow-up of at least 3 years [54]. A larger, phase II trial in HNSCC is currently underway (NCT02609386) and a phase Ib trial in combination with durvalumab (anti-PD-1L) and tremelimumab (anti-CTLA-4) is planned to commence in October 2018 (NCT03381183).

Another cytokine mixture-dubbed Multikine has been tested in HNSCC. This formulation contains 14 interleukins, interferons (IFN), chemokines, and colony-stimulating factors. In 2005, published data from a phase II clinical trial in T2-3, N0-2, M0 HNSCC showed an ORR of 41% in 17 patients treated with neoadjuvant perilymphatic Multikine injections [55]. A large phase III trial of Multikine in HNSCC completed enrollment in late 2016, and results are pending (NCT01265849).

IL-15 is cytokine of particular interest currently in cancer immunotherapy. IL-15 shares the immunostimulatory characteristics of IL-2; however, IL-15 does not promote Treg expansion or activation-induced cell death (AICD) of effector T cells, which are characteristics of IL-2 [56]. Two variants of IL-15, modified to improve the pharmacokinetic and pharmacodynamic properties when administered intravenously, are currently in clinical development. Recombinant human IL-15 (rhIL-15) was found to be well-tolerated and produced substantial increases in circulating NK and CD8⁺ T cells [57]. Additional phase I trials are combining CPI with rhIL-15 or hetIL-15, another modified IL-15, for metastatic or refractory cancer (NCT03388632, NCT02452268).

Traditionally, the immune system has been divided into “innate” and “adaptive” components. The term “innate” refers to sensors that do not require rearrangement of genes. Receptors present on innate immune cells, such as dendritic cells and macrophages, are highly conserved between individuals and species. “Adaptive” immune components require rearrangement of genes, leading to great diversity in receptors but are therefore specific to individuals. Adaptive immune components include T cell receptors, B-cell receptors, and antibodies.

The innate and adaptive immune systems are not separate systems but are in fact intimately related. Innate immune cells are involved in the activation of the adaptive immune system through cytokine signaling and antigen presentation [58]. Therefore, the innate immune system is capable of targeting threats directly through “innate” pathways and also of “priming” adaptive immune cells including T cells. Appreciation of this relationship has made targeting the innate immune system attractive for cancer immunotherapy, as well as autoimmune diseases.

Numerous types of cancer vaccines have been tested in preclinical models and in clinical trials. These include peptide vaccines, tumor lysates, DNA or RNA vaccines, and cellular vaccines including dendritic cells that have been exposed to antigen and danger signals. Sipuleucel-T (Provenge), an autogenous cellular vaccine targeting prostatic acid phosphatase, was approved by the FDA for the treatment of metastatic prostate cancer in 2010. While Sipuleucel-T was shown to increase OS in a randomized phase III trial, there was no increase in PFS [71]. Sipuleucel-T remains the only FDA-approved therapeutic vaccine for the treatment of cancer.

The ability to identify neoantigens through next-generation sequencing (NGS) techniques is appealing, in that neoantigens would theoretically avoid self-reactivity seen with tissue differentiation antigens or overexpressed antigens. However, this approach is based on patient-specific antigens and relies on complex mathematical modeling to predict binding to the patient’s MHC types. Short-lived peptides are another potential source of tumor-associated antigens that have not been tolerized, due to the usual rapid degradation in autophagosomes. By “freezing” these vesicles in vitro, preventing fusion with lysosomes and subsequent destruction of the antigens, a source of tumor-associated antigens can be obtained which otherwise would have been “thrown away” by normal cell metabolism. This technique has been used to create a vaccine known as DRibbles, which has been shown to contain shared antigens capable of cross-recognition between different tumors and is currently in phase I clinical trials in multiple tumor types, with plans to expand to oral cancer [72].

Adoptive cell therapy (ACT) is an immunotherapeutic approach that involves the extraction of T cells from a patient, expansion of a T cell population ex vivo, and infusion of the T cells back into the patient, usually after chemotherapeutic lymphocyte depletion and followed by administration of cytokines such as IL-2 [78]. Several methods of ACT are in development, including extracting tumor-infiltrating lymphocytes (TIL), which are expected to include some T cells that have specificity for TAA. Alternatively, T cells may be extracted from the peripheral blood and either (1) selected for tumor reactivity ex vivo, (2) exposed to dendritic cells loaded with specific TAA, (3) transduced with specific T cell receptors (TCR) with affinity for known TAAs, or (4) transduced with a chimeric antigen receptor (CAR) that has the antigen-recognizing domain of an antibody and the signaling domain of a TCR [60]. These approaches essentially skip the majority of the cancer immunity cycle, including antigen recognition, T cell activation and proliferation, by directly introducing cancer-fighting T cells into the patient. ACT is therefore termed “passive immunity,” as it circumvents reliance on the patient’s immune response.

In 2015, the first FDA approval of an oncolytic virus was granted to a genetically engineered, GM-CSF-transduced Herpes Simplex Virus (HSV-1), Talimogene Laherparepvec or T-VEC, for use in locally advanced or nonresectable melanoma. In HNSCC, T-VEC was tested in a small phase I/II study in combination with chemoradiotherapy in advanced HNSCC, showing an OS of 70.5% at median follow-up of 29 months. Patients all received post-therapy neck dissection, which was not standard of care for many of these patients; however, a pathologic CR rate of 94% in neck dissection specimens was a promising finding [85]. T-VEC is being tested in many other types of cancers and is currently in a phase Ib/III study in HNSCC in combination with pembrolizumab [86]. HF10, a spontaneously occurring mutant HSV-1 virus, was shown to be well-tolerated in a phase I trial in HNSCC [87], and in a phase II trial in advanced melanoma, HF10 combined with nivolumab showed an ORR of 41% with a 16% rate of CR [88].

Another example is Cavatak™, a coxsackievirus developed by Viralytics, which seeks out and attaches itself to a protein that is highly expressed on the surface of many cancer cells, intercellular adhesion molecule-1 (ICAM-1). Since ICAM-1 is expressed in HNSCC [89], a phase 1 clinical trial studying Cavatak with pembrolizumab has been designed and is currently in its final stages of preparation before opening for recruitment. GL-ONC1, an attenuated vaccinia virus, was well-tolerated with concurrent chemoradiotherapy in HNSCC [90] and is in phase II development in recurrent ovarian cancer (NCT02759588). Oncorine (H101), an E1B-deleted adenovirus, was approved in China in 2005 for use in HNSCC after a phase III study showed an ORR of 78% when combined with cisplatin and 5-FU [91, 92]. The use of H101 so far remains limited to China.

A trial called REO 018 trial was initially designed as a randomized phase III study of Reolysin, a reovirus, in combination with carboplatin and paclitaxel in platinum-refractory HNSCC. The study was reformatted after interim analysis found differential responses in patients with locoregional disease versus patients with metastatic disease alone. The company claimed a statistically significant increase in PFS and OS in patients with locoregional disease, but stated that there were too few patients to power a statistical analysis for patients with distal metastases alone [93]. From review of the company’s webpage, it appears that further development is currently focused on myeloma, breast, and pancreatic cancer [94].

While the efficacy of traditional cancer treatments, including chemotherapy, radiation, and targeted therapies, has historically been ascribed to direct cytotoxicity or inhibition of cellular activities, there is increasing appreciation for the immune-stimulating effects of these treatments. As described above, the Cancer Immunity Cycle begins with the release of cancer cell antigens. This is achieved by a process known as Immunogenic Cell Death (ICD), in which the killing of tumor cells can elicit an antitumor immune response. In addition to promoting recognition of tumor antigens through ICD, many chemotherapeutic agents have been shown to modulate immunosuppressive influences, e.g., through depletion of Tregs or MDSCs [95].

Additionally, there are mechanistic rationales for combining targeted therapies with immunotherapy. The antitumor activity of cetuximab in HNSCC is now appreciated to be primarily ADCC, as opposed to direct cytotoxicity. By combining cetuximab with CPI or other immunotherapies, immune-suppressive forces within the TME could potentially be counteracted leading to increased efficacy over either agent alone [102]. An interim safety analysis of a phase II trial of pembrolizumab and cetuximab in R/M HNSCC showed good tolerability with no DLTs [103]. There are several other efficacy studies underway which combine cetuximab with CPI with or without chemotherapy and radiation.

Immunotherapy is rapidly changing the standard of care in oncology. The appearance of a tail at the end of the survival curve with checkpoint inhibition in advanced cancers provides a graphic representation of the durable responses that can be achieved with this new group of therapies. This is the great promise of cancer immunotherapy, that is, the possibility of achieving lasting responses or even cures.

As evidence of the enthusiasm around immunotherapy, the number of new products in development and early-phase clinical trials has skyrocketed in the last decade. In 2017, it was estimated that there were 800–1000 cancer immunotherapy trials in the US involving over 100,000 patients [107, 108]. The same year, a report from the Pharmaceutical Research and Manufacturers of America found that there were 248 immuno-oncology agents in clinical trials, which only included “the most recognized classes of immunotherapy” [109]. In addition to the therapies described in this chapter, there are many other immunotherapies and immune adjuncts in development, including but not limited to agents that target tumor metabolism (e.g., IDO-1 and the adenosine pathway) [110, 111, 112], therapies to deplete or inhibit Tregs, MDSC, or TAM (e.g., anti-CCR4, PDE-5 inhibitors, anti-CSF1R) [113, 114, 115], and checkpoint inhibitors that target NK cells (e.g., anti-KIR, anti-NKG2A) [116, 117].

Moving forward in this new era of cancer immunotherapy will require continuing integration between the clinic and laboratory. Not only will preclinical science remain critical in developing new approaches in patient care, but laboratory evaluation of pathologic tumor responses, immune cell infiltrates, and circulating immune components will allow full-circle analysis and understanding of the physiologic effects of experimental treatments. The increasing number of neoadjuvant trials will facilitate this route of scientific discovery by providing postimmunotherapy tissue samples for comparison with pretreatment biopsies. Cutting-edge technologies for specimen analysis, such as NGS allowing whole-exome, RNA and T cell receptor sequencing, as well as advanced imaging techniques for multiplex immunohistochemistry, will allow for greater understanding of the in vivo effects of various immunotherapies on tumor biology. This work, along with clinical outcome correlations, will be critical in identifying predictive biomarkers and prognostic indicators and will provide evidence for future directions in cancer research.