Immune Checkpoint Inhibitors in Oral Cavity Squamous Cell Carcinoma and Oral Potentially Malignant Disorders: A Systematic Review

Cancers of the oral cavity cause significant cancer-related death worldwide. While survival rates have improved in recent years, new methods of treatment are being investigated to limit disease progression and to improve outcomes, particularly in oral cavity squamous cell carcinoma (OSCC) and oral potentially malignant disorders (OPMD). The emerging treatment modality of immunotherapy targets immune checkpoint molecules including PD-1 and its ligand PD-L1, CTLA-4, LAG-3, and TIM-3 to enhance the host immune response against tumours, and to limit the growth and progression of cancer cells. In this systematic review, we searched five databases for keywords pertaining to oral cancers and OPMDs, along with immune checkpoint inhibitors, in order to summarize the current status of their use and efficacy in these diseases. A total of 644 different articles were identified between 2004 and 2019, with 76 deemed suitable for inclusion in the study, providing a total of 8826 samples. Combined results show expression of PD-1 and PD-L1 in the majority of OPMD and OSCC samples, with expression correlating with increased progression and decreased survival rates. Immunotherapy agents pembrolizumab and nivolumab target PD-1 and have been shown to prolong survival rates and improve disease outcomes, especially in combination with chemotherapy or radiotherapy. Despite the equivocal nature of current evidence, there is support for the prognostic and predictive value of immune checkpoint molecules, especially PD-L1, and many studies provide support for the effective use of immune checkpoint inhibitors in the management of OSCC. Limited data is available for OPMD, therefore this should be the focus of future research.

CTLA-4 Tregs exhibited a higher prevalence in the peripheral circulation compared to normal controls CTLA-4 is an important functional marker of Tregs. 2 Ahn et al. 2017 [2] High PD-L1 expression was a favourable prognostic factor for overall survival only in the miR-197 high subgroup PD-L1 expression on IHC is associated with increased TILs and favourable prognosis in miR-197 high subgroup 3 Balermpas et al. 2017 [3] PD-1 expression was defined as being low in 88 patients and high in 73 patients PD-L1 in CD8 cells represents a promising prognostic marker and could be used to guide treatment with PD-1/PD-L1 inhibitors 4 Bauml et al. 2017 [4] 17% of patients with PD-L1 expression of >1% of tumour cells responded to treatment with nivolumab pembrolizumab exhibited clinically significant antitumor activity and an acceptable safety profile in heavily pretreated OSCC 5 Bharti et al. 2013 [5] High frequency of 1661G allele and AG genotype may be related to increased risk while low frequency of A allele suggests it to be protective in patients with tobacco related OSCC Association between several polymorphisms in CTLA-4 gene and OSCC supports the role of immune/inflammatory molecules in the susceptibility to tobacco-related OSCC 6 Bhosale et al. 2017 [6] CD274 expression was amplified in OSCC compared to normal tissue. CD274 was dysregulated in OSCC compared to leukoplakia CD274 and its ligand PD-1 are important targets of immunotherapy in OSCC 7 Cai et al. 2019 [7] PD-L1 was positive in 58% of tumour samples, compared to 0% in normal mucosa. In the tumour samples, PD-L1 was found to localize to cell membrane and cytoplasm and was highly expressed in poorly differentiated tumours. However, PD-L1 was weakly expressed in well and moderately differentiated tumours anti-PD-1 mAb may have better tumour permeability due to its smaller molecular weight. It may be more efficacious in poorly differentiated OSCCs with PD-L1 higher expression. 8 Chen et al. 2015 [8] Patients with both necrosis and positive PD-L1 expression in OSCC surrounding necrosis had worse outcome and disease control The aggressive behaviour of advanced OSCC could be related to PD-L1 immune escape, therefore patients with positive tumour PD-L1 expression may be good candidates for anti-PD-L1 immunotherapy. 9 Chen et al. 2018 [9] 72 samples were PD-L1 positive and 34 were PD-L1 negative. Positive p16INK4A expression was significantly higher and the mean age of patients was significantly higher in the group exhibiting positive expression of PD-L1 compared with the negative group.
This study identified an association between PD-L1 and p16INK4A expression in non-OPHNSCC Chen et al. 2019 [10] PD-L1 was significantly associated with the OSCC pathological grade, and higher PD-L1 staining was observed in OSCC compared to OLK PD-L1 expression in OSCC and OLK was closely associated with disease progress and CD8+ TILs 11 Cho et al. 2011 [11] PD-L1 expression on OSCC cells was observed in 87% of cases. The staining patterns were membranous and/or cytoplasmic. The density of intratumoral CD8+ T lymphocytes showed a significant inverse correlation with the PD-L1 expression of tumour cells The association between PD-L1 and CD8 T cells+ may be relevant to the PD-L1/PD-1 interactions resulting in the negative regulation of activated T-cells. PD-L1 expression was also correlated with the histological grade of the tumours 12 Dong et al. 2017 [12] Tim-3+ monocytes suppressed IFN-γ secretion in CD8+ T cells Blocking Tim-3 and/or Gal-9 could inhibit the Tim-3/Gal-9mediated suppression of monocytes in vitro 13 Du et al. 2011 [13] PD-L1 was abundantly expressed on keratinocytes and slightly on the infiltrate T cells in the subepithelium PD-L1 was abundantly expressed in most OLP cases. PD-L1 mRNA levels in OLP patients' mucosa and blood were 1.43and 0.62-fold compared to the control 14 Falco et al. 2019 [14] Not Specified 22 patients achieved clinical benefit. Partial response seen in 10, stable disease 9, Complete response in 3 patients. For those with high PD-L1 expression, single-agent pembrolizumab also improves overall survival compared with Cetuximab plus chemotherapy. 15 Fayette et al. 2017 [15] Not Specified Cabazitaxel gave a signal of activity OSCC but was toxic. In future studies, Cabazitaxel could be used at 20 mg/m2 every 3 weeks or weekly at a lower dose like Paclitaxel or Docetaxel. 16 Feldman et al. 2015 [16] PD-1: 125/182 altered expression. PD-L1: 32/183 altered expression. PD-1-positive TILs were detected in a range of 65% to 72%, across disease stages. Both PD-1-positive TILs and PD-L1-positivity in tumour cells was found across primary disease sites The data supports the use of agents in clinical trials (PIK3CA, PD-1/PDL1), combination strategies (PIK3CA1/EGFR), or agents approved for other solid tumours (MGMT, HER2). We propose a comprehensive molecular profiling approach to enhance personalized therapy options for OSCC. 17 Ferris et al. 2017 [17] Tumour PD-L1 expression status could be evaluated in 260 of 361 patients (72.0%). 57.3% of evaluated patients had a tumour membrane PD-L1 expression level of 1% of cells per field of view (minimum 100 cells) or more.
Exploratory biomarker analysis indicated that patients who were treated with nivolumab had an average of 4.83 months longer overall survival than those treated with standard therapy, regardless of tumour PD-L1 expression, and nivolumab-treated patients had a risk of death that was 30% lower than the risk among patients assigned to standard therapy 18 Fiedler et al. 2017 [18] 43 Results show a nearly significant association between PD-L1 expression in tumour cells and complete tumour response upon irradiation (77.4% high PD-L1 showed complete response vs 54.0% of low PD-L1 expression). PD-L1 was expressed in 40% of patients indicating radiosensitivity and favourable survival in this group of patients 19 Foy et al. 2017 [19] PD-L1 was overexpressed in never smokers/drinkers compared to smokers/drinkers and in never smokers/drinkers at protein level. PD-L1-positive immune cells were mostly adjacent to tumour cells expressing PD-L1.
The results confirm that CD4+CD25+ T cells, but not CD4+CD25-T cells, from OSCC patients have potent inhibitory consequences on T effector function and promote the generation of IL-10 21 Ghapanchi et al. 2019 [21] At position PD-1.3, the genotype of GG was present in 80.8%, while the genotypes of GA and AA were found in 13.7% and 5.5% patients. The most prevalent genotype among patients was CT heterozygote.
Polymorphism of PD-1.3 and PD-1.5 genes did not have any significant correlation with OLP susceptibility 22 Goltz et al. 2017 [22] 80 patients had high mPD-1 expression, 40 had low mPD-1 expression. mPDCD1high was associated with shorter overall survival mPDCD1 might potentially serve as a predictive biomarker for the response to immunotherapies targeting the PD-1/PD-L1 axis 23 Groeger et al. 2016 [23] All 15 oral squamous cell carcinomas investigated showed a positive expression of the PD-L1 receptor in the cancerous areas The 5-year survival rate of the patients whose tissues were positive for PD-L1 expression, was 73.33 % (11 of 15). Expression of PD-L1 may be a prognostic marker in oral squamous cell carcinomas. Intraepithelial CD8+ TILs express PD-1 at a high rate, and the stromal CD8+ TILs, as well as CD8+ T cells in OLP, express PD-1 at low rates. Expression of PD-1 in OLP ~100x10^-5, tongue cancer ~500x10^-5, compared to normal ~10x10^-5.
High expression rate of PD-1 by the intraepithelial CD8+ TILs suggests their exhausted functions 29 Larkins et al. 2017 [29] Not Specified Accelerated approval expands the FDA-approved indications for pembrolizumab for the treatment of recurrent or metastatic OSCC with disease progression on or after platinum-containing chemotherapy 30 Lecerf et al. 2019 [30] 85 tumours overexpressed at least one of TIGIT, CTLA4, PD-1/PD-L1 and OX40/OX40L and 33 tumours simultaneously overexpressed them all. No tumour exclusively overexpressed PD-1 or OX40. PD-L1 was exclusively overexpressed in only one sample PD-1 overexpression was associated with good prognosis, whereas low mRNA levels of PD-1 correlated with poor prognosis and high risk of recurrence 31 Lechner et al. 2017 [31] Significantly higher rate of PD-1 and CTLA-4 expression on TILs. The percentage of PD-1, PD-L1 and CTLA-4 expressing circulating T cells was increased in OSCC patients compared to healthy donors, indicating elevated proportions of regulatory or exhausted T-cell phenotypes Results demonstrate elevated proportions of regulatory or exhausted T-cell phenotypes in OSCC patients compared to healthy controls, with 3.7% of total cells compared to 1.6%, respectively 32 Leduc et al. 2017 [32] A significant increase in PD-L1 expression was observed in both tumour and immune cells post-TPF induction chemotherapy, with increases from 9.5% positivity prior to treatment compared to 38% positivity at the 5% cut-off level Results suggest combination strategies of concomitant administration of cytotoxic therapies and anti-PD-1/PD-L1 therapies might be relevant for OSCC 33 Lin et al. 2015 [33] High PD-L1 cytoplasm intensity was more likely in tumours from female than from male patients. High expression levels of PD-L1 were also more likely to occur in tumours from female High PD-L1-expression was significantly associated with distant metastasis and poor prognosis in male patients and smoking patients Results suggest patients with high PD-L1 than from male patients. High PD-L1 expression is an independent risk factor in males (hazard ratio = 1.556) and smokers (hazard ratio = 2.058) expression had poor clinical outcome (hazard ratio = 2. 74) and might require PD-L1-targeted immunotherapy to improve their prognosis 34 Linedale et al. 2017 [34] Amongst the CD8 infiltrate, the frequency of PD-1, CTLA-4 and Tim-3 expressing cells was significantly elevated in tumour relative to the blood across the patient Results suggests an enrichment of PD-1+ CD8 T cells in perineural tumour tissue relative to blood and that Tim-3 expressing T cells are enriched in perineural tumour and might be good candidates for targeted antibody therapy 35 Malaspina et al. 2011 [35] Numbers of CD8 T cells expressing PD-1 were significantly higher in OSCC patients than in control subjects. OSCC tumour samples contained elevated expression of PD-1 and present higher number of CD4+ PD-1+ T cells when compared with tissue from actinic chelitis patients PD-1 and PD-L1 molecules are present in different phenotypes of lymphocytes in blood actinic chelitis and OSCC. High PD-1 expression in CD4+ (43%) and CD8+ (68%) T cells may be used as a potential prognostic marker in oral tumours or in pre-malignant lesions 36 Malm et al. 2015 [36] Abundant PD-1 expression on CD4 and CD8 T cells at all sites, significantly higher than Lag-3. PD-1 was expressed at some level on both CD4 and CD8 T cells in all samples. 11% of analysed tumours were PD-L1 negative, 56% had regional expression, and 44% had diffuse expression with infiltrating lymphocytes PD-1 is expressed on T cells from HPV-negative patients with OSCC and the abundance of its ligand in tumour tissue between 44-56%. 37 Maruse et al. 2018 [37] PD-L1 and PD-1 expression was not observed in the adjacent non-malignant oral epithelium. Increased expression of PD-L1 and PD-1 was significantly associated with 3.67 times greater incidence of cervical lymph node metastasis Patients with PD-L1-positive expression had a significantly more unfavourable outcome than those with PD-L1negative expression (69.2% vs 91.0% 5 year survival). The co-expression of PD-L1 and PD-1 is predictive of a poor prognosis in OSCC patients. PD-L1 expression in cancer cells is more critical than PD-1 expression in the infiltrating inflammatory cells in the prediction of the prognosis of OSCC patients. 38 Mattox et al. 2017 [38] 42/42 samples expressed PD-L1 at >1% of membranous expression by tumour and/or immune cells. PD-L1 and CD4 and CD8 TIL expressions were not significantly associated with clinical outcomes Our study suggests that tumour-specific CD4+ T cells may be a key regulator of PD-L1 expression within the Tumour microenvironment. 39 Moratin et al. 2019 [39] Expression levels of PD-L1 were significantly higher in female patients than male patients (9% compared to 7.2%). Mean expression of PD-L1 was 7.9% of cells per section The expression scores of PD-L1 correlated significantly with the tumour size, the presence and the stadium of neck node metastases, clinical stage, and age. There was a trend toward worse overall survival for patients with higher PD-L1 expression AntiPD-1/PD-L1 therapy may be of therapeutic use even in early stage OSCC to prevent further disease progression 40 Moreira et al. 2010 [40] A lower percentage of CTLA-4+ cells was observed in OSCC compared with Lip SCC (3.39% vs 13.12%).
Higher CTLA-4 expression was significantly associated with a low tumour proliferative index No association between CTLA-4 expression and survival 41 Muller et al. 2017 [41] PD-L1 expression was at the membrane and cytoplasm of tumour cells. A subset of peritumoral and tumour infiltrating lymphocytes also revealed strong immunoreactivity for PD-L1.
There was no correlation between PD-L1 with tumour stage, lymph node involvement, lymphatic invasion, vascular invasion, tumour grade or extracapsular expansion PD-L1 expression is a suitable prognostic biomarker, independently of other well-known prognostic factors such as tumour stage and tumour grade (hazard ratio: 4.269 and 2.845 for cohort 1 and 2). PD-L1 expression was a strong predictor for poor outcome 42 Naruse et al. 2019 [42] Among 121 OSCC, 54.5% were positive for PD-1 and 57.9% were positive for PD-L1 by IHC. PD-L1 was expressed primarily in the cytoplasm and nuclei of tumour cells, with particularly strong expression observed at the invasive front.
A significant decrease in 5-year disease specific survival rate for patients with combined PD-1+/PD-L1+ expressions. Inhibition of the PD-1/PD-L1 axis may be a good strategy to improve prognosis in OTSCC patients with local recurrence after NAC 43 Ngamphaiboo n et al. 2019 [43] 79 of 94 OSCC samples (84%) were classed as positive for PD-L1 expression.
Patient sex, smoking status, site of primary cancer, stage at diagnosis, and p16 status were not associated with PD-L1 expression Highly expressed PD-L1 (≥50%) was an independent prognostic factor for poor overall survival in anti-PD-1/PD-L1 untreated OSCC patients 44 Okada et al. 2018 [44] PD-L1 expression was negative in 11 cases, weakly positive in 9 cases, and strongly positive in 6 cases.
The 5-year overall survival rates in the low and high PD-L1 groups were 72.5% and 16.7%, respectively. High PD-L1 expression is an independent poor prognostic factor in OSCC patients with surgically resected pulmonary metastasis 45 Oliveira-Costa et al. 2015 [45] Cytoplasmic PD-L1 expression in 47/96 cases while membrane expression was in 7/96. PD-L1 cytoplasmic expression was histologically found in areas with undifferentiated cells Cytoplasmic PD-L1 expression was significantly associated with tumour size (hazard ratio: 7.618). PD-L1 was expressed translationally and at the protein level in OSCC-derived circulating tumour cells. 46 Pekiner et al. 2012 [46] CTLA-4 was expressed an average of 1.01 times greater than CD8 in OLP, but an average of 0.74 times lower in controls No significant difference in the percentage of CTLA-4 in OLP and controls (P>0.05) 47 Poropatich et al. 2017 [47] PD-1+ CD8+ T-cell levels were positively correlated with increased size of the primary site tumour (r = 0.63). CTLA-4 and PD-1 expression on CD8+ T cells were significantly higher in stage IV HPV-negative than in stage IV HPV-positive patients. Treg TIM3 expression was 51.94% in HPV-negative compared to 14.88% in HPV-positive samples PD-1 and TIM-3 T-cell expression were specifically elevated in OSCC patients as compared with healthy controls 48 Quan et al. 2016 [48] Compared with the matched PBMCs in both CD41 and CD81 TIL-Ts we found a significant increase in the expression of PD-1 and Tim-3 T-cells expressing PD-1 and Tim-3 indicate T cell exhaustion and contribute to an immunosuppressive environment 49 Rasmussen et al. 2019 [49] Using a 1% cut-off value to define positivity, 36% of the specimens were concordant with tumour proportion score and 52% were concordant with combined positive score. With a 50% cut off, the concordance was higher at 70% with tumour proportion score and 54% with combined positive score.
The assessed PD-L1 positivity varies markedly within the tumour in this patient series which limits the utility of this biomarker 50 Ryu et al. 2017 [50] When comparing immune cell compositions among p16 patterns, both the MOSAIC and STRONG patterns similarly revealed high infiltration of CD3+, CD8+, and PD-1+ T cells. PD-1+ T cells were higher in NUCLEAR than in the ABSENT pattern s of staining The association of the nuclear pattern with worse prognosis may be related to a relatively high proportion of exhausted and dysfunctional PD-1+ T cells, although with fewer infiltrating CD8+ cytotoxic T cells PD-L1 expression in more than 5% of the SCC cells was present in 40 cases. Cell membrane and cytoplasmic staining were observed in this cohort and there was no nuclear staining. Weak 1+ staining intensity was observed in 26, moderate 2+ staining PD-L1 expression was observed in 18% of the cases in the current cohort intensity was present in 7 and strong 3+ staining intensity was seen in 7 cases 54 Scognamiglio et al. 2017 [54] Fifty-eight of 96 cases (60%) showed PD-L1 expression in the immunocytes in the tumour microenvironments, whereas 38 (40%) of the cases were negative for PD-L1 Frequent discordance of PD-L1 expression between primary and metastatic tumours would indicate that the tissue typing for PD-L1 expression should ideally be performed on the metastatic tumour if PD-L1 positivity is included as an eligibility criterion in anti-PD-1/PD-L1 clinical trials 55 Seiwert et al. 2016 [55] 81 out of 104 (84%) screened patients were positive for PD-L1 at the 1% cut-off level PD-L1 expression by IHC was predictive of best overall response and improved progression-free survival. Greater anti-tumour activity was recorded in OSCC that expressed higher levels of PD-L1 56 Shayan et al. 2017 [56] Expression of macrophage PD-L1 was increased by approximately one-fold change when incubated with cetuximab plus motolimod Addition of a PD-1 inhibitor to the combination of cetuximab and motolimod would further augment the antitumor response. Otolimod plus cetuximab decreased induction of Treg and reduced markers of suppression, including CTLA-4 57 Singh et al. 2017 [57] PD-L1 was significantly up-regulated from 1 to 2 fold changes in most of the tumours but downregulated in two tumours Expression of PD-L1 was upregulated in 10 of the 12 studied tumours suggesting these subsets of tumours may be susceptible for checkpoint blockade therapy 58 Sridharan et al. 2016 [58] All checkpoint receptor bearing cell populations increased following treatment. Low levels of soluble PD-L1 in all patients was found Radiation-induced effects on the local tumour microenvironment in OSCC patients may translate into quantifiable immune effects in circulating immune mediators, T cell receptor repertoires, and potential antitumour antibody responses. Higher baseline levels of soluble PD-L1 correlated with nodal status, higher levels in patients with node-positive disease 59 Strati et al. 2017 [59] Median fold change of PD-L1 expression in the EpCAM-PBMC fraction was 1.03 in controls and 0. 39  Marked PD-L1 expression was observed in 26/88 (29%) OSCC specimens and 73/88 (83%) showed considerable presence of PD-1-positive TILs. Relevant PD-L1 expression was noted significantly more often in cancers from mandibular and oral tongue than maxilla or soft palate. OSCC specimens with high PD-L1 expression displayed infiltration with PD-1+ TILs PD-L1 expression in OSCC was associated with an increased metastasis and greater levels of infiltrating PD-1+ TILs. PD-L1 expression in OSCC might differ depending on its anatomic origin 65 Wang et al. 2019 [65] Nimotuzumab therapy significantly increased the expression of TIM-3, LAG-3, IDO, PD-L1, and CTLA-4 in the tumour microenvironment of OSCC patients compared with baseline Aberrant expressed LAG-3 and PD-L1 were more likely to be associated with a worse survival (hazard ratio = 1.504) and could be considered as indicators of poor prognosis in OSCC 66 Weber er al. 2018 [66] PD-L1_4 and PD-L1_2 showed significantly increased expression, up to 3-fold, in OSCC compared to normal oral mucosa PD-L1 expression in peripheral blood might be an indicator of the existence of metastatic disease via nodal involvement in OSCC. There is also an association between a systemic state of immune tolerance and a more aggressive tumour 67 Wirsing et al. 2018 [67] PD-L1 expression in OSCC cells correlates with increased infiltration of CD4+ cells and small tumour size PD-L1-expressing tumour cells correlated positively to a tumour microenvironment rich in CD4+ cells but had no prognostic significance. PD-1/PD-L1-targeted immunotherapy might be successful in tumours rich in CD4+ cells 68 Wong et al. 2006 [68] No significant difference between samples and controls for all CTLA-4 phenotypes (Phenotype A frequency: 59.9% in OSCC vs CTLA-4 A/A genotype polymorphism is associated with younger age of onset and poorer survival in OSCC patients 60.5% in controls, and phenotype G frequency: 89.9% in OSCC vs 83.0% in controls) 69 Wu et al. 2017 [69] IHC staining showed that CTLA-4 was highly expressed on the TILs in the tumour microenvironment. PD-L1 and CTLA-4 were overexpressed in OSCC and VISTA was positively correlated with PD-L1 (r=0.342) and CTLA-4 (0.294) VISTA is overexpressed in OSCC and is correlated with PD-L1 and CTLA-4 indicating that it may regulate antitumor immunity 70 Wu et al. 2019 [70] PD-1 was co-expressed with TIGIT on human CD4+ and CD8+ T cells and TILs, and that this co-expression was higher on TILs than PBMCs (62.61% total cells vs 9.22% total cells) The data indicated that blocking PD-1 and TIGIT corporately may elicit better antitumor effects 71 Xu et al. 2019 [71] PD-L1 immunopositivity was more frequent in the salivary duct carcinoma ex-pleiomorphic adenoma group than in the salivary duct carcinoma de-novo group. PD-1 expression in immune cells was seen in 35 salivary duct carcinomas. PD-L1 and PD-1 expression did not predict the risk of lymph node metastases Expression of tumour-specific antigens and over-expression of PD-1 and PD-L1 indicate that salivary duct carcinoma patients may benefit from novel immune therapy approaches 72 Yagyuu et al. 2017 [72] Oral pre-cancerous lesion patients with subepithelial PD-L1+ cell count below median showed a 5-year malignant-free survival rate of 96.8%, whereas the survival rate for those with above median counts showed a decreased survival rate of 86.1%. For every additional increase in the number of subepithelial PD-L1+ cells and PD-L1 positivity score of epithelium, the risk of malignant transformation increased 1.07 and 3.38 times, respectively Subepithelial PD-L1-positive cell count and epithelial PD-L1 positivity were significantly associated with malignant transformation, whereas low subepithelial PD-L1 expression was associated with greater 5-year malignant free survival 73 Yang et al. 2019 [73] Plasmacytoid dendritic cell infiltration was positively correlated with PD-1 (r = 0.3628), LAG3 (r = 0.3241), and TIM3 (r = 0.4834) staining Plasmacytoid dendritic cell high infiltration correlated with an adverse outcome in human primary OSCC patients