Immunotherapy for Head and Neck Squamous Cell Carcinoma: A Review of Current and Emerging Therapeutic Options

The use of combination immunotherapies has shown profound synergy and may lead to further treatment advances compared with use as a monotherapy or with current cytotoxic regimens. This review focuses on recently completed checkpoint receptor inhibitor immunotherapy trials, emerging trials in head and neck squamous cell carcinoma, and the background for the development of these trials.

Abstract

Advances in the field of cancer immunotherapy have occurred rapidly over the past decade. Exciting results from clinical trials have led to new treatment options and improved survival for patients with a myriad of solid tumor pathologies. However, questions remain unanswered regarding duration and timing of therapy, combination regimens, appropriate biomarkers of disease, and optimal monitoring of therapeutic response. This article reviews emerging immunotherapeutic agents and significant clinical trials that have led to advancements in the field of immuno‐oncology for patients with head and neck squamous cell carcinoma.

Implications for Practice.

This review article summarizes recently developed agents that harness the immune system to fight head and neck squamous cell carcinoma. A brief review of the immune system and its role in cancer development is included. Recently completed and emerging therapeutic trials centering on the immune system and head and neck cancer are reviewed.

Introduction

Recognition of the immune system's role in contributing to cancer development was an important advancement in our original understanding of cancer immunology from the early 20th century [1], [2]. Premalignant cells are normally destroyed by the immune system before tumor formation can occur, termed “immune surveillance” (Fig. 1) [1], [3], [4]. After attack by the innate immune system, tumor antigens (TAs) are released and captured by antigen presenting cells (APCs), then processed and loaded onto major histocompatibility (MHC) complex, activating effector T cells [5], [6]. However, cell alteration and/or derangements in the immune system can lead to “immune escape.” Tumor cells circumvent the activated T cells and survive by developing resistance mechanisms and inducing T cell tolerance. The complex milieu of signaling between tumor cells and cells surrounding the tumor ultimately creates an immunosuppressive environment [3], [4], [6].

Figure 1.

The immune synapse and immune escape. Immune escape from each signal is required for development of antitumor immunity. Signal 1 is the TCR/human leukocyte antigen peptide/antigen interaction. Signal 2 denotes costimulatory (or coinhibitory) signals. Signal 3 represents cytokine secretion, either a proinflammatory type 1 antitumor signal, or a tumor permissive cytokine signal. These signals amplify or suppress antitumor response. Adapted with permission. © 2015 American Society of Clinical Oncology. All rights reserved. Ferris, RL. Immunology and immunotherapy of head and neck cancer. J Clin Oncol 2015;33:3293–3304.

Abbreviations: APC, antigen presenting cell; CCR, cytokine/chemokine receptor; CD, cluster of differentiation; IFN interferon; MHC, major histocompatibility; TCR, T cell receptor.

Several immune system alterations occur in head and neck squamous cell carcinoma (HNSCC) patients, suggesting that this cancer is an overall immunosuppressive process. In the peripheral bloodstream, HNSCC patients have less overall number of white blood cells, which are comprised of a greater proportion of suppressive regulatory T cells (Treg) [7]. Additionally, tumor‐infiltrating lymphocytes (TIL) in HNSCC tumors are composed of an even more suppressive population of Treg cells than in the periphery, suggesting an expansion of Treg in the tumor microenvironment (TME) [7], [8], [9], [10], [11], [12], [13]. The alterations noted in the immune cell populations of patients with HNSCC highlight the myriad of potential therapeutic targets. Preclinical and clinical data from TA‐targeted therapy in HNSCC provided significant evidence of the interplay between the immune system and tumor development [3], [14]. An additional layer of complexity regarding progression to malignancy in HNSCC stems from varied tumor etiology, from environmental factors to virally induced cancers. However, recent studies have indicated that despite the varied etiology, both tumor types display an inflammatory phenotype that is amenable to treatment with immunotherapy [15].

The HNSCC population endures significant treatment morbidity with standard therapy. Immunotherapy, on the otherhand, is generally well tolerated and may provide more efficacious therapy with long‐term patient benefit. Diverse treatment approaches have been developed in the realm of immunotherapy, including tumor‐specific monoclonal antibodies, cancer vaccines, immune‐modulating antibodies, oncolytic viruses, and adoptive cell transfer [11], [16]. The use of combination immunotherapies has shown profound synergy and may lead to further treatment advances compared with use as a monotherapy or with current cytotoxic regimens. This review will focus on recently completed checkpoint receptor inhibitor (ICR) immunotherapy trials, emerging trials in HNSCC, and background for the development of these trials.

Immunotherapy Background

T lymphocytes have been a major focus in the development of therapeutic agents as a means to manipulate an antitumor response. T cells selectively recognize non‐self peptides from cellular compartments and orchestrate diverse immune responses that lead to T cell‐mediated killing of the tumor cell [17]. Activation of the T cell requires two signals to occur in the context of a third signal (Fig. 1). Signal 1 occurs at the “immune synapse,” in which TAs bound to the MHC molecule on the surface of APCs are presented to a T cell receptor. Signal 2 consists of either a confirmatory costimulatory signal, such as the cluster of differentiation (CD)28/B7 interaction, or an inhibitory signal. The final signal, from immune‐activating cytokines such as interleukin 12 or type I interferon (IFN), confirms signal 2, directing the cell towards inhibition or stimulation [3], [18]. Stimulatory receptors allow for the development of a more robust immune response to an antigen. Inhibitory checkpoint receptors are present normally as a means to prevent autoimmunity against self antigens as well as an overabundant response to non‐self antigens. However, tumor cells develop many mechanisms to thwart immune recognition and response, a dynamic process termed “immunoediting” [14]. This includes upregulation of inhibitory receptors, recruitment of suppressive cells into the TME, and ineffective antigen presentation to T cells. The ultimate balance of suppressive cells, such as Treg and myeloid‐derived suppressor cells (MDSC), and effector cells, such as cytotoxic CD8‐positive (CD8⁺) T lymphocytes (CTL), determines if malignant cells can overcome an activated antitumor T cell response [12], [14].

Effective antigen presentation leading to T cell activation is enhanced and sustained by the induction of costimulatory cell surface molecules. OX40 is a costimulatory molecule expressed on the CD4‐positive (CD4⁺) T cell surface that promotes T cell proliferation, cytokine secretion, and development of memory [19], [20]. CD40L expressed on activated CD4⁺ T cells activates APCs that prime CD8⁺ T cells. CD137, a protein expressed on activated T cells, natural killer (NK) cells, and dendritic cells (DC), binds to its ligand (CD137L) on APCs, leading to enhanced proliferation, cytotoxic capacity, and survival of T cells [21]. APCs can express B7‐H3, a molecule induced by inflammatory cytokines, to provide a costimulatory signal to T cells [22]. Interestingly, B7‐H3 also has shown an inhibitory function depending on the tumor milieu [23]. Expression of B7‐H3 inversely correlated with intratumoral TIL in patients with hypopharyngeal squamous cell carcinoma, and higher B7‐H3 expression in this patient population was associated with development of distant metastases and poor prognosis [24]. Other sources for antigen stimulation, such as neoantigens that occur due to genomic instability from p53 suppressor loss, recruit infiltrates of T lymphocytes and NK cells into the TME [25], [26]. Tumor cells have devised mechanisms to alter the human leukocyte antigen (HLA) complex, leading to impaired recognition of neoantigens from malignant cells. While complete loss of HLA expression removes a key inhibitory signal for NK cells, HNSCC, for example, circumvents NK cell activation by maintaining altered HLA components through chromosomal and regulatory expression defects in HLA and antigen‐processing machinery encoding genes [27], [28], [29]. Lymphocyte activation gene 3 (LAG3) also functions as a negative regulator of T cells [30]. By impeding the interaction between Treg and DC, LAG3 expression leads to hampered maturation of DC and thus impaired antigen presentation and T cell activation [31].

Inhibitory checkpoint receptors expressed on activated immune cells, such as cytotoxic T lymphocyte antigen 4 (CTLA4) and program death 1 (PD‐1), play an important role in the TME (Fig. 2) [3], [32]. Transient expression of CTLA4 occurs on activated T cells, acting as a competitive inhibitor of CD28 through its higher affinity with the B7 ligands CD80 and CD86 [20], [33]. Ligand binding provides an inhibitory signal, preventing further immune system activation [20], [34]. Activated CD8⁺ T cells, NK cells, B cells, monocytes, and DC express PD‐1, a cell surface protein that, when bound by its ligand programmed death‐ligand 1 (PD‐L1) or programmed death‐ligand 2 (PD‐L2), serves as an inhibitory signal. Additionally, T cell exhaustion and tolerance have been observed with chronic antigen exposure and subsequent PD‐1 upregulation [20], [35]. Another inhibitory receptor involved in T cell tolerance is T cell immunoglobulin domain and mucin 3 (TIM3), a molecule selectively expressed on activated interferon gamma (IFN‐γ) producing CD4⁺ and CD8⁺ T cells, marking the most exhausted T cells [36]. Tolerance produces ineffective elimination of tumor cells as the T cell becomes exhausted and unresponsive to antigen stimulation.

Figure 2.

Costimulatory and coinhibitory signals. Costimulatory and coinhibitory signals modify antigen specific stimulation via the T cell receptor. Intrinsic suppressive signals on tumor antigen specific T cells or from extrinsic suppressive Tregs both can lead to impaired antitumor immunity. Receptors serve as potential targets for immunotherapeutic agents. The role of PD‐L2 and its location is under investigation. Adapted with permission. © 2015 American Society of Clinical Oncology. All rights reserved. Ferris, RL. Immunology and immunotherapy of head and neck cancer. J Clin Oncol 2015;33:3293–3304.

Abbreviations: CD, cluster of differentiation; CTLA4, cytotoxic T lymphocyte antigen 4; MHC, major histocompatibility; PD‐L1, programmed death‐ligand 1; PD‐L2, programmed death‐ligand 2; TCR, T cell receptor; Treg, regulatory T cell.

HNSCC tumor cells and the cells in the surrounding microenvironment have been noted to acquire many of the above described immune evasion mechanisms, and these alterations can be used as targets for immunotherapy. Nearly half of PD‐1‐positive (PD‐1⁺) CD8⁺ T cells are also TIM3‐positive in HNSCC patient TIL [37], and approximately 50%–60% of HNSCC tumors express PD‐L1, stimulated by an increase of IFN‐γ in the TME [38]. CTLA4‐positive (CTLA4⁺) Tregs in the HNSCC TME suppress the function of effector T cells [34], [35], [39]. However, an attempt at generating an effective T cell response with costimulatory protein expression seems to occur in HNSCC. In some studies, OX40 appears to be upregulated within the tumor and tumor‐draining lymph nodes of HNSCC patients, where 30% of T cells expressed OX40, compared with none of the peripheral blood lymphocytes [40]. Other studies have also shown a decrease in OX40 expression on CD4⁺ T cells in the peripheral blood of patients with advanced HNSCC [19]. Although the TME in HNSCC has a good immune infiltrate, the environment must change from a suppressive one to a stimulatory one so that this infiltrate can extinguish the tumor cells.

It is becoming clear that ICR alone may not induce a de novo immune response, but may enhance preexisting suppressed immune responses, and is the rationale for combination immunotherapy trials. Therefore, agents that stimulate an effective signal 1 through antigen‐specific immunity appear to be excellent partners with a signal 2 coinhibitory blockade. There are multiple potential combinations that can be used with immunotherapy agents together and/or with standard cytotoxic regimens.

Immunotherapy in HNSCC

Defining the appropriate regimen with the least toxicity and with durable responses is the goal of immunotherapy clinical trials today. Targeting redundant pathway mechanisms that lead to cancer progression will likely provide the best chance for curative therapy. Traditionally, systemic chemotherapy was thought to impart its effects solely through direct tumor killing. However, recent studies have shown significant immune stimulation not only with chemotherapy and radiation therapy (RT) [41] but also at lower doses of systemic cytotoxic therapy [42]. RT has been shown to induce upregulation of PD‐L1 on both tumor cells and MDSC. These changes could alter the TME, making it more responsive to PD‐1 pathway‐blocking agents. Ionizing radiation stimulates the adaptive immune response through several other mechanisms, any of which may be synergistic with immunotherapy [43]. Understanding these mechanisms provides evidence to support a trimodality approach with immunotherapy as a component of the regimen [44].

Ionizing radiation stimulates the adaptive immune response through several other mechanisms, any of which may be synergistic with immunotherapy. Understanding these mechanisms provides evidence to support a trimodality approach with immunotherapy as a component of the regimen.

Adding immunotherapy to current standard of care therapies is the immediate frontier for cancer treatment. The optimal dose (low vs. high) and timing (neoadjuvant, concurrent, or adjuvant) of cytotoxic therapy in combination with immunotherapy remains to be determined. Furthermore, studies are needed to identify biomarkers and other predictive factors for response to therapy. Emerging trial results, to be reviewed, will hopefully provide the answer on whether immunotherapy will replace current standard of care regimens.

Several phase II and III immunotherapy trials have been designed for three groups based on biologic subsets of HNSCC and include previously untreated and locally advanced (PULA) human papillomavirus (HPV)‐positive disease, PULA HPV‐negative disease, and lastly, recurrent/metastatic (R/M) disease. Trials are investigating immunotherapy alone, immunotherapy in combination with standard of care regimens, and combination immunotherapy trials.

R/M HNSCC

HNSCC patients with R/M disease have a very poor prognosis, with a median overall survival (OS) of just 10 months [45]. The “EXTREME” trial showed that cetuximab in combination with platinum and 5‐Fluorouracil (5‐FU) led to improved progression‐free survival (PFS) and OS in R/M disease. Despite this advance, these therapies are not curative, and there exists a large void for treatment options in this patient population. Until this year, cetuximab has been the only agent introduced in the past decade for treatment of R/M HNSCC.

PD‐1 pathway targeting has been the most extensively studied immunotherapy regimen in HNSCC patients. The early results from the Keynote 012 trial led to the accelerated U.S. Food and Drug Administration (FDA) approval of pembrolizumab, an immunoglobulin G4 (IgG4) antagonistic anti‐PD‐1 monoclonal antibody (mAb), in August 2016 for the treatment of platinum refractory R/M HNSCC. This phase I multicohort clinical trial included participants with various PD‐L1‐positive (PD‐L1⁺) solid tumor types. For the HNSCC cohort, response rates were similar in both HPV‐positive and HPV‐negative groups, but HPV‐positive patients did fare better in terms of median PFS and OS compared with HPV‐negative patients (17.2 vs. 8.1 weeks and median OS not reached vs. 9.5 months, respectively) [46]. A pooled analysis was performed on patients from the HNSCC cohort with PD‐L1⁺ tumors and HNSCC patients treated irrespective of PD‐L1 status (a total of 173 patients). This analysis revealed a statistically significant median PFS and OS for the whole population of 2.2 and 9.6 months, respectively [47]. Biomarker analysis revealed that expression of PD‐L1 correlated with response rate, although pembrolizumab did show clinical efficacy in some PD‐L1‐negative cases as well [46], [47]. The second cohort included patients treated regardless of PD‐L1 expression, HPV status, or prior systemic therapy [46], [47], [48], [49]. As a follow‐up, pembrolizumab as a monotherapy in platinum refractory patients is being compared with standard of care in the Keynote 040 phase III randomized trial (NCT02252042). Keynote 048 will also evaluate pembrolizumab as a first‐line therapy for patients with R/M disease (NCT 02358031). Patients will be randomized to receive pembrolizumab, pembrolizumab with 5‐FU/cisplatin, or investigator's choice [50].

Many of the current immunotherapy trials in the R/M setting seek to answer which treatment regimen will provide a more durable and robust response rate with the least side effects for patients with platinum refractory disease: immunotherapy alone or immunotherapy in combination with current standard of care regimens (Tables 1 and 2). Another PD‐1‐blocking mAb, nivolumab, is expected to obtain FDA approval in the near future based off the results from the Checkmate 141 trial. Patients treated in the phase III trial received nivolumab at a dose of 3 mg/kg intravenously (IV) every 2 weeks versus weekly IV single‐agent chemotherapy (methotrexate 40 mg/m², docetaxel 30 mg/m²) or cetuximab (400 mg/m² once, then 250 mg/m²) in a 2:1 ratio (NCT02105636) [51], [52]. Two hundred forty of the 361 mostly male patient cohort received nivolumab, with 121 receiving investigator's choice (methotrexate, docetaxel, or cetuximab). Fifty five percent of patients had undergone therapy with two or more prior agents. The trial was closed early because the statistical boundary for OS was crossed. Patients treated with nivolumab resulted in a statistically significant 30% reduction in risk of death (regardless of the p16 status or PD‐L1 expression) [51], [52], [53] and doubling of the one‐year OS compared with the control arm (36% vs. 16.6%). All subgroups, regardless of HPV status or PD‐L1 expression, showed a statistically significant survival benefit with nivolumab. Grade 3 and 4 adverse events (AEs) were much lower in the nivolumab arm compared with the cytotoxic arm (13% vs. 35%) [51], [52], [53].

Table 1. Immunotherapy with cytotoxic therapy.

Abbreviations: CD, cluster of differentiation; CSF1R, colony stimulating factor 1 receptor; CTLA4, cytotoxic T lymphocyte antigen 4; CXCR2, C‐X‐C motif chemokine receptor 2; FLT3, Fms‐related tyrosine kinase 3; HNSCC, head and neck squamous cell carcinoma; IDO, indoleamine 2,3 dioxygenase; KIR, killer inhibitory receptor; KIT, stem cell factor receptor; LA, locally advanced; LAG3, lymphocyte activation gene 3; M, metastatic; mAb, monoclonal antibody; MEK 1 and 2, mitogen activated protein kinase extracellular signal regulated kinase; MGA 271, humanized anti‐B7‐H3 monoclonal antibody; NCI, National Cancer Institute; NKG2A, killer lectin‐like receptor subfamily C member 1; NPC, nasopharyngeal carcinoma; PD‐1, program death 1; PD‐L1, programmed death‐ligand 1; poly ICLC, carboxymethylcellulose stabilized polyriboinosinic/polyribocytidylic acid; PI3K, phosphoinositide 3‐kinase; R/M, recurrent/metastatic; SOC, standard of care; TAT3, signal transducer and activator of transcription 3; SYK, spleen tyrosine kinase; TIM3, T cell immunoglobulin domain and mucin 3; TLR, toll‐like receptor

Table 2. Immunotherapy as monotherapy.

Abbreviations: HNSCC, head and neck squamous cell carcinoma; LA, locally advanced; MGA271, humanized anti‐B7‐H3 monoclonal antibody; NCI, National Cancer Institute; NPC, nasopharyngeal carcinoma; PD‐L1, programmed death‐ligand 1; poly ICLC, carboxymethylcellulose stabilized polyriboinosinic/polyribocytidylic acid; R, recurrent; R/M, recurrent/metastatic; SOC, standard of care; TLR, toll‐like receptor

Despite these exciting results, it is important to recognize that over half of the patients in the Checkmate 141 trial had progression of disease at or within 6 months of platinum‐based therapy. The results of the Checkmate 141 trial do not suggest that nivolumab be used as a monotherapy in the first‐line treatment setting, but it is a new standard of care option for patients with progressive disease after cisplatin therapy. Nivolumab was recently approved by the FDA for treatment of platinum refractory R/M HNSCC disease in November 2016. The current first‐line EXTREME regimen provides an early, yet usually unsustained response; therefore, investigators are hopeful that immunotherapy may provide this long‐lasting response that other regimens have not provided.

Nivolumab and pembrolizumab are quickly entering the clinical repertoire, but additional agents targeting the PD‐1 pathway have also been developing rapidly. Durvalumab, a human IgG1 mAb that selectively targets PD‐L1, is currently being evaluated in several phase I trials in patients with multiple tumor types, including HNSCC. Patient populations include those with advanced disease resistant to standard therapy (NCT01938612, NCT02586987, NCT01693562), treatment‐naïve patients with R/M disease (NCT02262741), and treatment‐naïve advanced or metastatic disease (NCT02658214) [50]. This selective targeting of PD‐L1 may confer advantages over other ICR agents, including, but not limited to, a decrease in PD‐L2 mediated toxicities. An additional benefit of this agent is conferred by alterations made to the fragment crystallizable domain during its development that will reduce IgG‐mediated side effects [54]. Preliminary results from the R/M HNSCC trial showed that PD‐L1⁺ tumors have a trend towards improved OS (15% vs. 11% in PD‐L1‐negative tumors). However, this agent did not show a significant reduction in side effects, as was postulated to occur with selective ligand targeting [55]. HAWK is a phase II open‐label trial of durvalumab in platinum refractory patients with PD‐L1⁺ HNSCC (NCT02207530) [50].

Trials evaluating the role of costimulatory agonist mAb provide an alternative approach to immune stimulation. Treatment with an agonist OX40 mAb demonstrated benefit after surgery or radiation in mice [56]. Clinical phase I trials demonstrated 9B12, a murine‐derived OX40 agonist mAb, to be a potent immunostimulator in patients with late‐stage solid tumor malignancies [57]. Patients had only moderate toxicities, and a maximum tolerated dose was not reached. These results led to the development of a humanized/mouse chimeric antibody, MEDI6469 [57]. MEDI6469 is being tested in several phase I trials for patients with platinum refractory R/M HNSCC and other advanced solid malignancies (NCT02205333, NCT02318394, NCT02315066, NCT02221960) [50], [56], [57]. Additional trials with this agent will assess its use in the preoperative setting in a window of opportunity trial (NCT02274155). MEDI6469 will be given to patients at a standard dose of 0.4 mg/kg for 3 doses at various time intervals prior to surgery (e.g., 3 weeks prior, 2 weeks prior, or 1 week prior). This trial may provide information about the safety of neoadjuvant immunotherapy and determine if these patients are at increased risk for surgical complications.

Preclinical studies have shown promising immune stimulation with both agonistic CD40 mAb and recombinant CD40L, which increases the ability of APCs to cross‐prime naïve T cells to TAs [58], [59], [60]. Current trials with CD40 agonists, dacetuzumab and lucatumumab, are underway for hematologic and solid malignancies and have shown exciting results thus far [61], [62], although CD40 expression on APCs is increased in HNSCC after surgical resection [63], and therefore may be a great target as an adjuvant therapy. There are no trials currently evaluating these agents in HNSCC [50].

PULA HNSCC

Developing trials for locally advanced disease presents challenges in malignancies with established standard of care regimens. There is a need for targeted therapy for PULA HPV‐positive patients, as this patient population frequently responds to treatment more effectively and has better outcomes than HPV‐negative patients. Current treatment modalities can render these patients with needless lifelong post‐treatment morbidity. Although the majority of patients with HPV‐induced HNSCC do well, a subset of patients develop recurrent and/or metastatic disease. Impactful studies are needed to address this uncommon yet lethal condition of distant metastatic disease in HPV‐induced malignancy.

Preclinical data suggest that HPV‐associated oncogenes contribute to immunogenicity in the setting of HNSCC. In addition, targeting unique viral TAs present in HPV‐positive tumors may provide improved response rate to immunotherapy and OS in this patient population. Current open trials aim to eliminate systemic chemotherapy by combining intensity‐modulated radiation therapy (IMRT) with cetuximab and the anti‐CTLA4 mAb ipilimumab (NCT01935921). This trial uses a staggered therapy approach with ipilimumab exposure beginning at week 5 of cetuximab‐RT. The Radiation Therapy Oncology Group (RTOG) Foundation trial #3504 (NCT02764593) will also be evaluating patients with “intermediate‐risk” HPV‐positive and “high‐risk” HPV‐negative disease (tumor node metastasis [TNM] classification T2N2b‐N3 with >10 pack‐years smoking status or T3/4N0‐N3 or <10 pack‐years stage T4N0‐3 or T2‐3N2c‐N3) with concurrent, weekly cisplatin chemoradiation therapy (CRT) or cetuximab CRT and the antagonistic anti‐PD‐1 mAb nivolumab. An additional arm evaluating nivolumab with IMRT includes patients either with age over 70 years, with a history of neuropathy or hearing loss, or evidence of renal insufficiency. This randomized phase III trial with a phase I lead‐in will investigate the effect of this combination treatment on OS for this patient population [50]. For patients with locally advanced disease unfit for cisplatin therapy, the NCT02938273 trial offers a treatment regimen consisting of a PD‐L1 agent, avelumab, in conjunction with cetuximab and radiation therapy [50].

The majority of HNSCC worldwide results not from HPV but from environmental carcinogens, and OS of tobacco‐ and alcohol‐induced HNSCC has only marginally improved over the last 2 decades. Patients with surgically resected, locally advanced disease have an unacceptable relapse rate of 30%–50% within 3 years, suggesting that current regimens are inadequate [64], [65]. The current adjuvant therapy recommendations for patients with surgically resected HNSCC arose from trials demonstrating that patients with one or more high‐risk features, such as extracapsular nodal extension (ECE) or positive surgical margin, had a clinical benefit when treated with concurrent cisplatin‐RT compared with RT alone [66], [67]. Subsequent meta‐analyses revealed local and distant failure rates of 50% and 15%, respectively, despite improved locoregional control initially [68]. These results emphasize the current need for new intensification approaches with the integration of immunotherapy. Due to the high mutational load in HPV‐negative disease, there is considerable interest in ICRs in this population. Mutational load serves a potential stimulus for the immune system, as these peptides are loaded onto self HLA and presented to T cells. Reversal of suppression and/or exhaustion of these tumor‐specific T cells with immunotherapy could be a potential therapeutic benefit in this patient population, as has been realized with other malignancies with high mutational burden such as melanoma [69]. An upcoming trial for patients with surgically resected, high‐risk, HPV‐negative HNSCC will evaluate adjuvant pembrolizumab with cisplatin CRT (RTOG HN‐003). A similar trial currently recruiting patients, NCT02641093, is also investigating the combination of pembrolizumab with cisplatin CRT for this patient population. A trial on pembrolizumab in the neoadjuvant and adjuvant settings for high‐risk patients with surgically resectable, HPV‐negative disease is recruiting patients (NCT02296684). This trial evaluates neoadjuvant pembrolizumab with cisplatin CRT adjuvant therapy as dictated by surgical pathology (ECE or positive margin). Patients with high‐risk pathologic features will also be treated with pembrolizumab after resolution of toxicities from CRT [50].

Future Directions in Immunotherapy

Combination Immunotherapy

Despite the enthusiasm regarding ICR, the majority of patients do not benefit from anti‐PD‐1 therapy. Interest has turned to combining ICR agents with the hope of overcoming multiple layers of resistance to enhance efficacy in a synergistic manner while maintaining an acceptable toxicity profile. It remains unclear how blockade of one immune checkpoint receptor affects other checkpoint receptors and if blockade leads to cross talk downstream with other pathways. Nevertheless, combining costimulatory agonists with one or more ICR agents may provide the key to harnessing the immune system to fight malignancy in an effective and durable manner.

There are now several trials evaluating various combinations of PD‐1 and CTLA4 blockade for many malignancy types. Melanoma patients with tumors expressing high levels of PD‐L1 did not respond to anti‐CTLA4 mAb and radiation until the addition of PD‐L1‐blocking agents. These results seem to implicate PD‐1/PD‐L1 signaling in resistance to CTLA4 blockade [70]. Combination immunotherapy trials in melanoma and non‐small cell lung cancer (NSCLC) using ipilimumab (CTLA4‐blocking mAb) and nivolumab (PD‐1‐blocking mAb) showed improved overall response (40% and 57%, respectively) and similar safety profiles compared with either monotherapy alone [71], [72]. A phase I trial combining durvalumab (PD‐L1‐blocking mAb) and tremelimumab, a CTLA4‐blocking mAb, in NSCLC patients showed an overall response rate of 25%, and, interestingly, PD‐L1 expression was insignificant. Results from this trial allowed for selection of the specific dose and schedule that provided clinical response with acceptable toxicity profile for future phase III trials (durvalumab 20 mg every 4 weeks and tremelimumab 1 mg/kg every 4 weeks) [73].

The combination of durvalumab with tremelimumab is being investigated in several trial designs for patients with platinum refractory HNSCC (NCT02319044, NCT02369874), treatment‐naïve HNSCC patients in the R/M setting (NCT02262741), and HNSCC patients with locally advanced disease who are not candidates for curative chemoradiotherapy (NCT02658214) [50]. The phase III EAGLE trial is testing this combination regimen for patients with R/M platinum refractory disease, stratifying patients by PD‐L1, HPV, and smoking status with co‐primary endpoints of PFS and OS (NCT02369874). Patients are randomized (1:1:1) to receive durvalumab, tremelimumab plus durvalumab or investigator's choice (cetuximab, taxane, methotrexate, or fluorpyrimidine) [50]. The CONDOR trial is another randomized open phase II trial assessing this combination regimen in immunotherapy‐naïve, platinum refractory patients with PD‐L1‐negative R/M HNSCC (NCT02319044). Patients are randomized to receive durvalumab monotherapy, tremelimumab monotherapy, or durvalumab plus tremelimumab. The patients are stratified according to HPV and smoking status, with a primary endpoint being objective response rate with Response Evaluation Criteria In Solid Tumors (RECIST) v1.1 criteria [50]. The KESTREL study is a phase III study evaluating patients with no prior systemic therapy (unless part of multimodal treatment for locally advanced disease) or immunotherapy (NCT02551159). This study is using a 2:1:1 randomization comparing durvalumab, tremelimumab plus durvalumab, or standard of care EXTREME regimen (carboplatin or cisplatin plus 5‐FU and cetuximab), with co‐primary endpoints of PFS and OS [50].

Other trials are using nivolumab and ipilimumab to study the effects of combining PD‐1 and CTLA4 blockade. Checkmate 651 is a phase III trial evaluating patients with R/M disease that are treatment naïve (unless prior treatment is part of multimodal therapy and completed 6 months prior to enrollment; NCT02741570). Patients will be randomized to receive nivolumab with ipilimumab versus the EXTREME regimen, with co‐primary endpoints of PFS and OS [50]. The Checkmate 714 trial is looking at combinatorial immunotherapy with nivolumab and ipilimumab compared with nivolumab with an ipilimumab‐placebo in R/M disease (NCT02823574). This is one of the only trials that includes a placebo arm.

In addition to CTLA4‐blocking agents, emerging trials are beginning to evaluate PD‐1 blockade in combination with other checkpoint receptors. MGA271, a humanized IgG1 mAb targeting B7‐H3, is currently undergoing investigation as a second‐line therapy in combination with ipilimumab (NCT02381314) and pembrolizumab (NCT02475213) [50]. The NCT02381314 trial is a phase I study evaluating patients with B7‐H3‐expressing HNSCC, melanoma, and NSCLC. This is a dose escalation trial of weekly IV MGA271 for up to 1 year in combination with ipilimumab IV every 3 weeks for 4 doses. Preclinical models in mice treated with anti‐PD‐1 mAbs and anti‐LAG3 mAbs resulted in curative treatment in the majority of tumors that were resistant to either mAb alone [74]. TIM3 upregulation has also been shown to occur after blockade of PD‐1 in HNSCC TIL [75], [76]. Early phase trials looking at PDR001, a human anti‐PD‐1 mAb, in combination with LAG525, an anti‐LAG3 mAb, (NCT02460224), and MBG453, a human anti‐TIM3 mAb (NCT02608268), are both recruiting patients [50]. Correlating TIM3 upregulation in these patients treated with anti‐PD‐1 therapy, an event noted in preclinical models, may be of clinical benefit in determining patients who could achieve a clinical response with this combination regimen. Patients with HNSCC are eligible for two novel trials with the PD‐L1‐targeting mAb atezolizumab. Both of these trials evaluate the safety and efficacy in combination with CPI‐444, an A2 receptor antagonist, or varlilumab, a CD27 costimulatory agonist (NCT02543645) [50].

Many combination trials with ICR antagonist and costimulatory agonist combinations are underway for patients with HNSCC (Table 3). A multicenter phase I trial evaluating durvalumab alone or in combination with an OX40‐stimulating mAb, MEDI1638, in R/M solid tumors, including platinum refractory HNSCC, has completed patient recruitment (NCT02221960). Subjects receive therapy until disease progression [50]. Nivolumab in combination with varlilumab is another example (NCT02335918) [50]. Preclinical models have shown efficacy of agonistic CD137 mAb. Although similar results have not been seen with this agent as monotherapy in HNSCC models [77], anti‐CD137 mAb has shown synergism with chemoradiation in HPV‐positive HNSCC models [78], [79]. Currently, there are two humanized mAb against CD137: urelumab (IgG4) and PF‐05082566 (IgG2). These agents are being evaluated in early phase trials for melanoma, NSCLC, and lymphoma [50]. For HNSCC, trials with these agents are underway in combination with other immune‐modulating agents [20]. Several groups have demonstrated a cetuximab‐mediated upregulation of CD137 in HNSCC and colorectal cancer, which has led to a phase I trial combining urelumab and cetuximab for patients with colorectal cancer (NCT 02110082) [80], [81]. Urelumab is also being tested in combination with nivolumab in solid tumors, including HNSCC (NCT02253992) [50]. The Keynote 036 trial evaluating PF‐05082566 in combination with pembrolizumab has completed accrual (NCT02179918). Patients receive pembrolizumab IV at 2 mg/kg every 3 weeks and PF‐05082566 with a starting dose of 0.45 mg/kg IV every 3 weeks, with subsequent dose escalation. PF‐05082566 in combination with avelumab, a PD‐L1‐blocking mAb, is being evaluated in another phase I trial (NCT02554812). This multicohort study includes patients with several types of cancer, including HNSCC, melanoma, and NSCLC. In the phase II cohort expansion arm, patients with HNSCC undergo treatment with avelumab until disease progression. PF‐05082566 is given at one of three dose levels to determine optimal dosage of this agent in combination with avelumab [80].

Table 3. Immunotherapy with chemoradiation.

Abbreviations: 5‐FU, 5‐Fluorouracil; HNSCC, head and neck squamous cell carcinoma; IMRT, intensity‐modulated radiation therapy; LA, locally advanced; M, metastatic; NCI, National Cancer Institute; PULA, previously untreated locally advanced; R, recurrent; R/M, recurrent/metastatic; SBRT, stereotactic body radiation therapy; TLR, toll‐like receptor; XRT, radiation therapy.

Targeting other components in the TME in conjunction with T cell targeting is also under investigation. A phase I/II study is investigating multiple solid tumor response to tremelimumab with durvalumab in combination with TME modulator polyICLC (a toll‐like receptor [TLR] 3 agonist; NCT02643303). Patients can either be HPV‐negative with prior therapy history or HPV‐positive with no prior therapy history. Preclinical data showed that the TLR8‐targeted mAb motolimod (previously named VTX‐2337) enhanced cetuximab‐mediated antibody‐dependent cellular cytotoxicity [82]. This led to a clinical trial combining motolimod with cetuximab for platinum refractory patients with locally advanced and R/M HNSCC (NCT01334177). A trial comparing combination of motolimod and cetuximab with or without nivolumab in the neoadjuvant setting in patients with locally advanced HNSCC is currently underway (NCT02124850). Motolimod in combination with standard cytotoxic therapy (platinum agent, 5‐FU, and cetuximab) is also being tested as first‐line therapy in patients with R/M HNSCC (NCT01836029).

Future studies from combination clinical trials will continue to advance our understanding of the complex interactions of these pathways. It appears that combination of immunotherapy agents will yield better response rates than immunotherapy as a monotherapy.

Monitoring Therapeutic Response

Identifying patients who would benefit from immunotherapy prior to starting treatment would eliminate subjecting patients to autoimmune side effects who will not benefit from ICR. The results of studies evaluating PD‐L1 as a biomarker for disease response thus far, however, have been mixed [83], [84]. Potential biomarkers of disease response, such as PD‐L1, are actively being evaluated. Different companies use different immunohistochemistry thresholds for the ligand, ranging from >1% to >50% of cells staining for PD‐L1, resulting in an unclear definition of positive staining [83], [85]. Also, the concentration of PD‐L1 on tumor cells versus peritumoral tissue may play a role in these varied results [84]. In normal tissue, PD‐L1 expression is induced by IFN‐γ as a mechanism for tissue protection, and evaluation of PD‐L1 and IFN‐γ may represent a way to determine the presence of TIL [38]. HNSCC tumor cells express PD‐L1 but also demonstrate infiltration of PD‐1⁺ Treg cells, a finding that may be more common in HPV‐positive tumors [3], [55]. Therefore, IFN‐γ, p16 (a biomarker for HPV positivity), and PD‐L1 together may give information to predict response to anti‐PD‐1 treatment. Emerging data from patients with PD‐L1⁺ HNSCC tumors from the Keynote 012 trial were recently released regarding an IFN‐γ signature as a predictive biomarker for HNSCC [86]. An analysis was performed of six IFN‐γ regulated genes with a gene signature panel (consisting of chemokine ligand 9, chemokine ligand 10, indoleamine, IFN‐γ, HLA‐DR alpha chain, and signal transducer and activator of transcription 1) that had been previously developed for melanoma in the Keynote 001 trial [86], [87]. This IFN‐γ gene signature showed a statistically significant association with PFS such that signature scores and PFS readily segregated patients into two groups (>5 months vs. <5 months PFS). With an optimal cutoff for the top‐performing IFN‐γ signature, the positive predictive value for response was 40.0%, and the negative predictive value was 95.0%. Results from this gene signature compare well with PD‐L1 gene expression (CD27) [48], [86]. Saloura et al. identified molecular correlates of an inflamed TME in HNSCC. Analysis of the cancer genome atlas revealed that one third to one half of HNSCC patients showed an inflamed phenotype based on gene expression signature. The inflamed phenotype may be able to serve as a biomarker to predict patients who will be more responsive to immunotherapy, thus allowing for personalization of this treatment approach [15]. Therefore, it appears that the “inflamed phenotype” signatures may allow for prediction of patients who will have clinical benefit from anti‐PD‐1 treatment in PD‐L1⁺ preselected HNSCC patients. However, further study is needed to evaluate this gene signature in PD‐L1‐negative patients [86].

HNSCC tumor cells express PD‐L1 but also demonstrate infiltration of PD‐1⁺ Treg cells, a finding that may be more common in HPV‐positive tumors. Therefore, IFN‐γ, p16 (a biomarker for HPV positivity), and PD‐L1 together may give information to predict response to anti‐PD‐1 treatment.

Current monitoring strategies used for responders to cytotoxic chemotherapy may not be applicable for immunotherapy. RECIST criteria used for cytotoxic agents is based on the premise that an effective agent results in shrinkage of the tumor, and tumors resistant to an agent enlarge. In contrast to cytotoxic agents, immunotherapeutic agents have been shown to produce responses with an assortment of kinetic patterns, even including transient tumor swelling [88]. This “tumor flare” response, represented by increased tumor diameter radiographically, may be due to lymphocytic infiltration of tumor. Additionally, this tumor flare may be a result of delayed immune cell activation, during which time the tumor may grow while the immune system is preparing an antitumor response. The use of standard radiographic criteria may lead to erroneous cessation of the immunotherapeutic agent in this scenario [88]. Alternative endpoints of disease response are necessary for immunotherapeutic agents.

Optimizing Side Effects and Safety

The toxicities associated with immunotherapy differ from traditional systemic therapy. Endocrinopathies and other endocrine sequela such as pneumonitis have been observed with immunotherapy (seen in 2 of 240 patients in the Checkmate 141 trial). Although side effects such as anemia and nephrotoxicity may occur with immunotherapy, the frequency of these AEs in clinical trials such as Checkmate 141 has been much lower than traditional cytotoxic chemotherapy [53]. In general, any organ can be affected, as immunotherapy can lead to a systemic activation of the immune system. Autoimmune side effects are most frequently seen in the skin, endocrine organs, and gastrointestinal tract. Overall, grade 3 and 4 AEs with ipilimumab were reported in 24% of patients in a recent meta‐analysis; however, the overall incidence of immune‐related AEs was 72% [89]. The development of AEs with ipilimumab appears to be dose dependent, with 61% of patients sustaining AEs on the approved 3 mg/kg dose for melanoma compared with 79% on the 10 mg/kg dose [89], [90]. Interestingly, AEs with anti‐PD‐1 therapy are less common. The AE profile for anti‐PD‐1 agents differs from anti‐CTLA4 agents, with pneumonitis seen more frequently in the former and colitis more frequently in the latter [91]. Melanoma trials have confirmed a lower percentage of AE with pembrolizumab and nivolumab when compared with ipilimumab (10%–16% and 20%–27%, respectively) [92]. Although immune‐related AEs differ from those seen with standard systemic chemotherapy regimens, these events can be life‐threatening and require early recognition and implementation of treatment. It is important, however, to note that toxicities may arise after a prolonged period of agent cessation. Also, the field of immunotherapy is in its infancy, and thus the long‐term impact of immunotherapy on quality of life has not been fully elucidated [93]. Patients may gain long‐term benefit or cure from immune checkpoint blockade, and therefore it is prudent to carefully monitor patients for late‐onset immune‐related AEs [93].

Conclusion

Understanding the complex balance of immune cell interactions and cell signaling has advanced significantly and has led to renewed interest in immunotherapy as a potential cure for HNSCC. It is of paramount importance, however, that rational clinical trial designs are developed to identify potentially serious autoimmune reactions and other AEs. Current results from immunotherapy trials have shown for the first time an improved response rate and OS in HNSCC patients with R/M disease while maintaining a safe toxicity profile. Although data are still limited and long‐term follow‐up is needed for these therapies, patients appear to tolerate treatment reasonably well for longer periods of time than conventional chemotherapy. Future studies will be directed at biomarkers for therapeutic response as well to further advance personalized therapy and improve efficacy. Although studies are still needed to answer the question of ICR efficacy as a monotherapy, the available trial results are promising. It is prudent that every oncologist considers each case carefully, and the EXTREME regimen may be the preferred treatment modality for certain patients.

Author Contributions

Collection and/or assembly of data: Jessica M. Moskovitz

Manuscript writing: Jessica M. Moskovitz, Tanguy Y. Seiwert, Robert L. Ferris

Final approval of manuscript: Jessica M. Moskovitz, Robert L. Ferris

Disclosures

Tanguy Seiwert: Amgen, AstraZeneca, Bristol‐Myers Squibb, Celgene, Eisai, Eli Lilly, Innate, Jounce, Merck, Merck Serono (H); Robert Ferris: AstraZeneca/MedImmune, Bristol‐Myers Squibb, Eli Lilly, Merck, Pfizer (C/A), AstraZeneca/MedImmune; Bristol‐Myers Squibb, Merck, VentiRX (RF). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board