Programmed Cell Death Ligand 1 as a Biomarker in Head and Neck Cancer

INTRODUCTION

The role of immune checkpoint pathways in various cancers, including those of the head and neck, has recently become a major focus of interest in the field of medicine because of their demonstrated clinical benefit for a subset of patients. The study of immune checkpoint pathways is already affecting the practice of pathology, in which the results of immunohistochemical testing from companion diagnostics are being applied to formulate customized therapies (personalized medicine). Given the use of fine-needle aspiration, often the first way and sometimes the only way in which a tumor sample is obtained, cytopathologists may soon have a role to play in the application of special tests related to immune pathway detection in formalin-fixed, paraffin-embedded tissues, in direct smear slides, or in other cytological preparations.

The immune system is our first line of defense against cancer cells and is often able to detect and recognize a transformed cell as foreign, and a number of stimulatory and inhibitory immune pathways are involved in this antitumor response.^1,2 One way in which tumors can evade attack by the immune system is the exploitation of inhibitory checkpoint pathways that suppress the activity of antitumor T cells.¹ Many recent studies have focused on blocking these immune checkpoint molecules, and one of the most popular approaches has been the blockade of the programmed cell death 1 (PD-1)/programmed cell death ligand 1 (PD-L1) pathway.³ The anti–PD-1 agents nivolumab and pembrolizumab have been demonstrated to be clinically useful and are approved for select indications, including melanoma, non–small cell lung cancer, squamous cell carcinoma of the head and neck, renal cell carcinoma, and Hodgkin lymphoma.^4–15 The purpose of this commentary is to introduce the reader to a prominent immune marker, PD-L1, and its role in the application of immune checkpoint inhibitors for select cancers of the head and neck.

PD-L1, also known as B7-H1 and CD274, belongs to the B7 superfamily of molecules, which also includes B7-1 (CD80), B7-2 (CD86), B7-H2, B7-H3, B7-H4, and B7-H6.¹⁶ PD-L1 binds to PD-1^17–19 and can be expressed by T cells, B cells, myeloid-derived dendritic cells, and macrophages located in various tissues.^20,21 As described previously, the PD-1:PD-L1 axis has gained much attention over the past several years because of reports demonstrating that the blockade of PD-1 or PD-L1 by specific monoclonal antibodies (mAbs) can enhance antitumor immunity.^20,22

HEAD AND NECK SQUAMOUS CELL CARCINOMA (HNSCC)

PD-L1 (or B7-H1) expression was first assessed in HNSCCs by Strome et al²² with the 5H1 mAb. PD-L1 expression was found in 66% of HNSCCs at a variety of primary HNSCC sites, which included the lip and oral cavity, hypopharynx, larynx, skin, and paranasal sinuses as well as cervical lymph nodes (for metastatic HNSCC). More recently, PD-L1 expression was evaluated in human papillomavirus–positive head and neck squamous cell carcinomas (HPV-HNSCCs) with the same PD-L1 mAb.²³ Seventy percent of HPV-HNSCCs expressed PD-L1, with greater than 90% displaying membranous staining in the periphery of the tumors, and PD-L1 was associated with the presence of PD-1–expressing tumor-infiltrating lymphocytes, which demonstrated impaired T cell function. This finding suggested that the PD-1:PD-L1 pathway was indeed activated in these tumors and was a mechanism of immune evasion used by HPV-HNSCCs. On the basis of these data and others, an open-label, multicenter, phase 1b clinical trial (Keynote-012) was launched in June 2013, and it evaluated the safety and clinical activity of an mAb targeting the PD-1 receptor, pembrolizumab (anti–PD-1 or MK-3475), in a variety of solid tumors, including a cohort of recurrent and/or metastatic (R/M) squamous cell carcinomas of the head and neck, with the goal of abrogating the PD-1:PD-L1 axis.⁸ Eligibility criteria for the study included the expression of PD-L1 within the tumor microenvironment, which was defined as more than 1% of tumor cells and/or immune cells within the stroma staining positive with the anti–PD-L1 mAb 22C3. Pembrolizumab was well tolerated, and HPV-HNSCC patients had the highest overall response rate at 25%, whereas the rate was 14% for human papillomavirus–negative HNSCC patients.⁸ Subsequently, on August 5, 2016, the US Food and Drug Administration (FDA) granted accelerated approval of pembrolizumab (Keytruda; Merck Sharp & Dohme Corp) for the treatment of patients with R/M HNSCC with disease progression on or after platinum-containing chemotherapy. This was followed on November 10, 2016, by FDA approval of nivolumab (Opdivo; Bristol-Myers Squibb Company) for the same indications on the basis of the results of a multicenter, open-label, randomized phase 3 clinical trial (CheckMate 141).⁹ The study demonstrated a statistically significant improvement in overall survival with a median overall survival of 7.5 months for the nivolumab-treated arm versus 5.1 months for patients who received single agent chemotherapy. p16-positive HNSCCs (a surrogate marker for HPV-HNSCCs) demonstrated a greater magnitude of effect from nivolumab therapy in comparison with the p16-negative tumors.

PD-L1 AS A BIOMARKER OF RESPONSE TO PD-1:PD-L1 BLOCKADE

Understanding tumor and/or patient characteristics that might predict who will or will not respond to a particular therapy can facilitate personalized patient care and allow the most efficient use of health care resources. PD-L1 expression within the tumor microenvironment is currently the single best predictive biomarker of response to PD-1:PD-L1 blockade. In the first reported phase 1 study of anti–PD-1 therapy (nivolumab), the group reported a correlation between membranous PD-L1 expression on 2 melanoma cancers and 1 renal cell carcinoma and response to treatment.²⁴ Subsequently, intratumoral PD-L1 expression was evaluated in a larger cohort of melanoma, non–small cell carcinoma, colorectal cancer, castration-resistant prostate cancer, and renal cell carcinoma patients treated with nivolumab.²⁵ Pretreatment biopsies obtained from 42 patients were available to assess the role of intratumoral PD-L1 expression in the modulation of the PD-1:PD-L1 axis. PD-L1 expression was correlated with a higher treatment response to blocking anti–PD-1 antibodies. However, because clinical responses have also been observed in a subset of PD-L1–negative patients, efforts are ongoing to identify additional biomarkers of response to these immunotherapeutic agents.

PD-L1 AS A PROGNOSTIC BIOMARKER IN HNSCCs

Although PD-L1 may serve as the best single biomarker of response to this class of immune checkpoint inhibitors, several groups have evaluated its prognostic potential in a limited number of HNSCCs. One study of oral squamous cell carcinoma demonstrated that PD-L1 positivity was an independent prognostic factor.²⁶ However, in another study that evaluated PD-L1 expression in oropharyngeal cancers, PD-L1 expression was not associated with any differences in overall survival.²⁷ There is an ongoing international, multicenter study correlating the PD-L1 status with clinical outcomes in R/M HNSCC patients (ClinicalTrials.gov identifier NCT02543476). The results of this study will be informative and will be able to definitively determine whether PD-L1 can serve as a prognostic biomarker in HNSCC independent of the treatment regimen.

WELL-DIFFERENTIATED THYROID CARCINOMA

PD-L1 expression has also been explored in thyroid cancers. Cunha et al²⁸ was the first to report on PD-L1 expression in well-differentiated thyroid carcinomas. Using immunohistochemistry, they found that 82.5% of papillary thyroid cancers (PTCs) and 87.5% of follicular thyroid cancers were positive for PD-L1 protein expression. They did not find PD-L1 expression to be associated with aggressive histopathological features such as multifocality, extrathyroidal invasion, tumor size, and tumor stage. Furthermore, a log-rank test failed to demonstrate PD-L1 expression as a prognostic marker. Another study reported that PD-L1 expression was present in 52.3% of PTCs.²⁹ However, in contrast, positive PD-L1 staining was linked to multifocality and extrathyroidal extension. Furthermore, PD-L1 expression was associated with worse recurrence-free survival.²⁹ These studies highlight some of the challenges that pathologists and cytopathologists could face with the interpretation of PD-L1 immunohistochemical stains. The Cunha study used a rabbit anti-B7H1 polyclonal antibody (ab82059; Abcam, Cambridge, United Kingdom), and PD-L1 positivity was determined on the basis of the cytoplasmic reactivity of tumor cells. Shi et al²⁹ used 2 PD-L1 antibodies: the same Abcam antibody and a mouse monoclonal anti–PD-L1 (MABC290; Millipore, Darmstadt, Germany), but they did not report whether PD-L1 positivity was scored on the basis of cytoplasmic or membranous staining. More recently, Chowdhury et al³⁰ reported on subcellular PD-L1 expression in thyroid tumors. They found that 66.5% of PTCs demonstrated cytoplasmic PD-L1 expression, whereas 40% demonstrated membrane localization. They concluded that PD-L1–positive expression in PTCs did correlate with a greater risk of recurrence and shortened disease-free survival.

Because BRAF V600E mutations are associated with a worse prognosis for a subset of PTCs, Angell et al³¹ evaluated PD-L1 expression in the context of BRAF V600E mutations. They reported that tumors with BRAF (V600E) mutations more often expressed high levels of PD-L1 than BRAF wild-type tumors. This is an interesting observation because BRAF V600E mutations and PD-L1 positivity have been reported to be independent prognostic biomarkers in PTCs, and in this study, it seems that these 2 biomarkers were often associated.

FUTURE DIRECTIONS

Currently, a major dilemma in the field of immune checkpoint therapies is the lack of standardization by researchers and pathologists of what is considered PD-L1–positive. As an increasing number of groups and centers test for PD-L1 expression in various tumors to gain prognostic information and/or determine whether to treat a patient with a blocking PD-1:PD-L1 therapy, pathologists and cytopathologists will increasingly be requested to score the PD-L1 staining of the tumor. However, how does one determine whether a tumor is PD-L1–positive? Is it cytoplasmic or membranous staining? Is it staining of the tumor cells, the immune cells, or both that is clinically relevant? What qualitative or quantitative cutoff should be used to determine PD-L1 positivity? The answers to these questions differ according to the tumor type and the different therapeutic antibodies being used. There are 4 diagnostic anti–PD-L1 antibodies associated with the administration of a specific PD-1/PD-L1 inhibitor (Table 1). Three of the 4 antibodies (SP142, 22C3, and 28-8) are currently FDA-approved as a companion diagnostic for specific tumor types. Recently, Gaule et al³² performed a quantitative comparison of 6 antibodies against PD-L1 (SP142, E1L3N, 9A11, SP263, 22C3, and 28-8) and demonstrated that all of the antibodies had a high level of concordance (R² = 0.76–0.99). However, there is a strong need to develop a standardized approach to the interpretation of the PD-L1 immunohistochemical stains being performed across various centers and among different clinical trials so that a common language is being spoken when PD-L1 positivity is being discussed. Currently, PD-L1 protein expression is defined differently for the 4 diagnostic anti–PD-L1 antibodies according to the expression within the tumor cells and/or immune cells, and there are different thresholds placed on the basis of the tumor type being stained (Table 1). Head and neck cancer does not currently have a companion diagnostic for the application of PD-1/PD-L1 checkpoint inhibitors. However, the response to these drugs is reported to be higher in PD-L1–expressing head and neck cancer cells and/or immune cells, and this may bias clinicians to order PD-L1 staining to guide treatment options.^8,9

TABLE 1.

Comparison of PD-L1 Diagnostic Antibodies

Company	BMS	Merck	AstraZeneca	Roche
Clone	28-8	22C3	SP263	SP142
Instrument	Dako	Dako	Ventana	Ventana
Host species	Rabbit IgG4	Mouse IgG4	Rabbit IgG1	Rabbit IgG1
Binding site	Extracellular	Extracellular	Intracellular	Intracellular
Indication	NSCLC	NSCLC	Urothelial	NSCLC
	Melanoma			Urothelial
Drug trade name	Nivolumab	Pembrolizumab	Durvalumab	Atezolizumab
Cell type scored	TCsa	TCsb	TCsc	TCs + ICsd

Abbreviations: IC, immune cell; IgG, immunoglobulin G; NSCLC, non–small cell lung cancer; PD-L1, programmed cell death ligand 1; TC, tumor cell; TPS, tumor proportion score.

^aPD-L1 protein expression is defined as the percentage of tumor cells exhibiting positive membrane staining at any intensity.

^bPD-L1 protein expression is determined with the TPS, which is the percentage of viable tumor cells showing partial or complete membrane staining at any intensity. A specimen should be considered to have PD-L1 expression if the TPS is ≥1% and to have high PD-L1 expression if the TPS is ≥50%.

^cPD-L1 status is determined by ≥25% tumor membrane staining.

^dThe determination of the PD-L1 status is indication-specific, and an evaluation is based on either the proportion of the tumor area occupied by PD-L1–expressing tumor-infiltrating ICs of any intensity or the percentage of PD-L1–expressing TCs of any intensity. PD-L1 expression in 5% of ICs in urothelial carcinoma is associated with an increased overall response rate, and PD-L1 expression in 50% of TCs or 10% of ICs in NSCLC may be associated with enhanced overall survival.

In summary, surgical pathologists and cytopathologists are already being called upon to aid in the development and interpretation of laboratory tests for non–small cell lung cancer to identify patients who could potentially benefit from anti–PD-1/PD-L1 therapies. This trend will also undoubtedly be seen in head and neck cancers. The development, interpretation, and implementation of PD-L1 testing are the focus of ongoing research and will likely affect treatment options for those patients deemed to be PD-L1–positive until a better prognostic biomarker is identified that can predict a response to PD-1/PD-L1 blockade.