An integrated research approach for preventing oral cavity and oropharyngeal cancers: two etiologies with distinct and shared mechanisms of carcinogenesis

Abstract

Head and neck squamous cell carcinoma (HNSCC) was the 7^th most common malignancy worldwide in 2018 and despite therapeutic advances, the overall survival rate for oral squamous cell carcinoma (OSCC) (~50%) has remained unchanged for decades. The most common types are OSCC and oropharyngeal squamous cell carcinoma (OPSCC, survival rate ~85%). Tobacco smoking is a major risk factor of HNSCC. In the developed world, the incidence of OSCC is declining as a result of tobacco cessation programs. However, OPSCC, which is also linked to HPV infection, is on the rise and now ranks as the most common HPV related cancer. The current state of knowledge indicates that HPV-associated disease differs substantially from other types of HNSCC and distinct biological differences between HPV-positive and HPV-negative HNSCC have been identified. Although risk factors have been extensively discussed in the literature, there are multiple clinically-relevant questions that remain unanswered and even unexplored. Moreover, existing approaches (e.g. tobacco cessation, vaccination, and chemoprevention) to manage and control this disease remain a challenge. Thus, in this review, we discuss potential future basic research that can assist in a better understanding of disease pathogenesis which may lead to novel and more effective preventive strategies for OSCC and OPSCC.

Introduction

Head and neck squamous cell carcinoma (HNSCC), including oral cavity, oropharynx, nasopharynx, hypopharynx and larynx malignancies, was the 7^th most common cancer worldwide in 2018 accounting for 890,000 new cases and 450,000 deaths (1). Together, oral squamous cell carcinoma (OSCC) and oropharyngeal squamous cell carcinoma (OPSCC) are the most common HNSCC types accounting for >260,000 cases and >128,000 deaths worldwide, respectively, and, an estimated 53,260 new cases and 10,750 deaths in in the US in 2020 (2). As recently reported in an excellent clinical review, it is clear that the prognosis and multimodal therapeutic options for patients with HNSCC can vary widely depending on epidemiologic factors, anatomical locations and stage of the disease (3). Even if treated successfully, patients may be left with critical functional deficiencies including impaired abilities to eat, drink and vocalize effectively. The 5-year survival rate has remained unchanged for decades and varies by race (48% in blacks, 66% in whites) and stage of disease (2). While OSCC rates in developed countries have been decreasing overall due to reductions in smoking rates, rates of OPSCC (base of tongue and tonsil) have been increasing and it is the most common HPV-associated cancer in the US and worldwide sparking concern regarding its future control (4–8).

The goal of this review is to discuss the current state of knowledge of OSCC and OPSCC, challenges for prevention, and potential future basic research required for a better understanding of the mechanisms that account for the development of these diseases. Furthermore, this review focuses specifically on scientific and clinically-relevant questions that remain unanswered including, but not limited to: 1) Why is there a specific tropism for oropharyngeal tissues (e.g. tonsil, base of tongue) for HPV associated cancer? 2) Why do men have higher OPSCC incidence than women? 3) What is the role of local resident and recruited immune cells in OPSCC development? 4) What are the roles of tobacco smoke oxidants in disease development? 5) Is there a synergy in oncogenesis between genetic factors, HPV and specific tobacco smoke constituents? 6) What is the role of the oral microbiome in disease development? To answer these questions, there is an urgent need for interdisciplinary research approaches targeting the many varied factors which could impact disease development and susceptibility. We strongly believe that future basic research should rely on the development of well-defined animal models that reflect human exposure to relevant etiological agents (tobacco, alcohol, HPV), individually and in combination to provide key mechanistic insights into the major factors driving this disease. These data can form the basis for developing novel and more effective strategies to intercept and prevent this disease.

Risk Factors

Risk factors (Figure 1), as classified by the International Agency for Research on Cancer (IARC), for carcinogenicity in humans include smoking, alcohol consumption, and poor oral hygiene for OSCC and, to a lesser extent, OPSCC (9). Other factors such as, betel quid, smokeless tobacco, marijuana and certain dietary components have also been implicated for these sites. Diet has long been known to be an important factor impacting the risk for oral cancers with diets low in fruits and vegetables consistently being associated with increased risk for these diseases (10–13). Intake of specific foods such as whole grains have been inversely related to risk of OSCC and OPSCC, while for other foods and beverages, such as tea, meat and dairy products, results have been inconsistent (13). With respect to tea drinking, which is the second most popular beverage worldwide after water, the results of a more recent comprehensive and dose-response meta-analysis based on 14 case-control studies (MOOSE compliant) suggest that tea intake provided protection against oral carcinogenesis (14). Increased oral cancer risk has been linked to reduced intake a variety of specific nutrients including iron (15), retinols (16,17), glutathione (15,18), β-carotene, vitamin C, and selected flavonoids, although results from observational studies have not always been consistent (13). In part based on reported associations between dietary intake and oral cancer risk, a variety of chemopreventive agents have been developed and tested with varying degrees of success including retinoids, β-carotene, vitamin C and selenium (19,20).

Figure 1:

Risk Factors for Oral and Oropharyngeal Cancers and Promising Approaches for Prevention. Included are multiple molecular targets involved in cancer development following exposure to etiological factors and plausible strategies for cancer prevention.

Human papillomavirus (predominantly HPV16) infection is an important risk factor specifically for OPSCC (9). According to a recent CDC report, OPSCC is the most common HPV-associated cancer in men and has exceeded cervical cancer rates in women (5); sex hormones, oral sex and levels of infections could be cofactors that can account for the higher rate of OPSCC in men than women (5,21).

While HPV16 infection is an established cause of OPSCC, its etiologic role in OSCC remains unclear as HPV infection rates in OSCC remain lower than in OPSCC. Nevertheless, OSCC and OPSCC are likely to have shared environmental risk factors such as tobacco smoking (22,23) that deserve further investigation as proposed in this review. Over the past three decades, increasing OPSCC incidence, despite decreasing smoking rates, has been attributed to HPV16 infection. HPV-associated tumors most frequently occur in the tonsil and base of the tongue (21,24) and test positive for HPV DNA and E6 and E7 antibodies (25). Overall, these properties of HPV-associated OPSCC distinguishes this disease from the other categories of HNSCC, namely OSCC (26). Table 1 briefly provides a comparison of several parameters that distinguish OSCC from OPSCC (27–30).

Table 1:

Comparison of etiology, response to treatment, prognosis, molecular pathogenesis and clinical sites involved in OSCC and OPSCC

Parameter	OSCC (non-HPV)	OPSCC (HPV-associated)
Etiology	• Exposure to carcinogens (tobacco, alcohol) • DNA repair deficits • Genetic mutations to growth regulatory proteins	Integration of oncogenic HPV (primarily HPV16) into host keratinocytes’ DNA.
Response to treatment	Surgery only; wide surgical excision is necessary.	Tumor cells that respond to radiation and chemotherapy can therefore use multi-modal treatment (surgery, radiation, chemotherapy).
Prognosis	Contingent upon tumor site (e.g. lip has much better prognosis), tumor size and clinical stage, 49% disease free survival.	Due to tumor cell responsiveness, 85% disease free survival.
Molecular Pathogenesis	Heterogeneous; while often involves p53, FHIT and p16 InK4a, variety of other parameters including signaling pathways affected.	HPV E6 and E7 proteins silence p53 and Rb tumor suppressor genes, respectively.
Clinical sites involved	Can affect all surface oral epithelium. Typically the “pooling” area: floor of the mouth, ventral and lateral tongue, retromolar trigone and anterior tonsillar pillar.	Fenestrated surface epithelium overlying lymphoid tissue, i.e. palatine tonsils and lingual tonsils at the base of the tongue.

Data from GLOBOCAN and others show a higher attributable fraction of HPV-associated HNSCC in high income countries and, from 1983 to 2002, OPSCC incidence rates have increased in men from predominately high income countries and in people age <60 (4,31,32). Based on these increasing rates of infection, OPSCC represents a potential epidemic and, to manage and control this disease, future studies should be aimed at elucidating molecular mechanisms that can identify targets which drive HPV-positive vs. HPV-negative disease.

Interactions of HPV Proteins with Molecular Targets of Carcinogenesis

In contrast to OPSCC in which HPV incidence is high (70-90%), the prevalence of HPV in other HNSCC subsites including OSCC and laryngeal cancer is much lower (<20%) (33–37). While molecular mechanisms responsible for this relationship remain unclear, data are accumulating regarding potential molecular targets of carcinogenesis impacted by HPV proteins. HPV E6 interacts with p53 to promote its degradation via the ubiquitin pathway, whereas HPV E7 forms a complex with retinoblastoma (Rb) protein leading to its functional inactivation and dysregulation of the cell cycle (38,39). Eckhardt et al. (40) described the first systematic study of HPV-host protein interactions combined with an integrative genomics approach to elucidate interactions that drive viral carcinogenesis. Interactions linking HPV to cancer included E6-p53 and E7-Rb as well as previously unknown connections involving E7-YAP1, E2 and proteins of the WNT/β Catenin signaling pathway, E5-p16 and E1-KEAP1 interactions (40). Clearly, there are many HPV proteins that could be playing critical roles in disease development and additional research is needed to further identify the range of proteins involved. In addition, there is a distinct need for establishing the specific molecular mechanisms by which these proteins, either individually or in combination with other etiologic agents, can impact the disease process, for which the development of well-defined and validated animal models will be essential.

Animal Models for HNSCC

The lack of appropriate animal models has hindered progress in HNSCC prevention research. Several animal models have been described, including xenograft, transgenic and chemically-induced models; the advantages and disadvantages have been previously reviewed by our team (41). More recently (42), Li et al. comprehensively provided a broader survey and summarized the benefits and disadvantages of the various mouse models that have been developed by numerous investigators. Therefore, in the present review we focused on our novel mouse papillomavirus model that can be employed to address the role of viral infection in combination with etiological factors such as tobacco smoking in the development of HNSCC. Some of the important characteristics of a biologically and clinically-relevant HNSCC model include the utilization of carcinogens found in the human environment (e.g. tobacco smoke) and a tumor progression profile that accurately mimics the heterogeneity and known cellular and molecular changes associated with initiation and progression of the human disease. Additionally, models are needed which address the role of HPV infection and specifically target oropharyngeal tissues as does HPV infection in humans. In this regard, our novel mouse papillomavirus model (described below) may address this important need as it mimics HPV associated infection in humans and prefers the base of the tongue. While it is clear that no single animal model can completely recapitulate human disease, careful selection of appropriate models based on the specific questions being raised is critical (41,42).

A Novel Mouse Papillomavirus Model for OSCC and OPSCC.

Our team has pioneered the development of a novel and unique mouse papillomavirus (MmuPV1) model and studies are actively being pursued toward a better understanding of the mechanisms of HPV-associated OPSCC (43–46). This model mimics many features of HPV-associated oral infections and diseases, despite differences in genome structure and molecular interactions of their respective oncogenes (47). In 2013, we successfully achieved MmuPv1 infections at several mucosal sites including the oral mucosae in both immunocompromised and immunocompetent mice (43–46). We established non-invasive lavage and swab techniques to quantitatively track viral DNA and RNA in the oral cavity (44) . Furthermore, we have demonstrated that MmuPV1 preferentially infects the base of the tongue, a specific region that HPVs target in the human oropharyngeal mucosa, and showed significantly more viral DNA in MmuPV1-infected male mice compared to female mice, consistent with sex differences observed in humans (5,9,21). Collectively, this model should be useful to understand the basis for specific tropism and to determine the impact of infection on immune regulation at different sites in the oral cavity including the base of the tongue. In addition, it can be utilized to investigate how the interaction between infection and immune response cooperates to promote HPV persistence leading to HPV-associated OPSCC. HPV-associated diseases including OPSCC are more prevalent in immune-compromised populations (48). Several human retrospective studies demonstrated the importance of immune cell infiltration and neutrophil-to-lymphocyte ratio (NLR) in disease outcome in OPSCC (49,50). High NLR was associated with a lower 5-year survival rate (51) and low leukocytosis was associated with better distant control in HPV positive and HPV negative patients (52). However, these studies did not define immune cell populations. Clearly evaluating the immune cell landscape in the tonsils and base of the tongue samples from male and female OPSCC patients with and without HPV infection can lead to the discovery of sex-specific immunological targets. To complement human tissue analysis mechanistic studies, using our MmuPV1 model, on the role of different immune cells on virus persistence and associated cancer development should be performed.

Previous studies demonstrated that sex hormones such as progesterone increased the susceptibility of viral infections at the lower genital tract by suppressing local anti-viral activity (53–55). Whether male hormones play a similar role in the oral cavity remains unclear. Therefore, this model can be used to further examine the impact of modulators including sex hormones and receptors on disease development which may explain why males have a higher incidence of OPSCC (9).

The MmuPV1 mouse model can also be employed to assess the effects of co-factors such as tobacco carcinogens on infection and disease progression. Carcinogens including polycyclic aromatic hydrocarbons (PAHs) such as dibenzo[a,l]pyrene (DB[a,l]P) and tobacco-specific nitrosamines (TSNA) such as N’-Nitrosonornicotine (NNN), play a critical role in OSCC development(41); the metabolic activation of DB[a,l]P and NNN leading to metabolites that can damage DNA, a prerequisite step in the multi-step carcinogenesis, is shown in Figures 2 and 3, respectively. Accordingly, we have pioneered the development of an OSCC mouse model using DB[a,l]P (41). The induction of DNA damage by DB[a,l]P and its ultimate carcinogenic diol-epoxide, DB[a,l]PDE stimulated a follow-up study directed to determine their effects on in vivo mutagenesis in the oral cavity of Big Blue C57BL/6 mice (56,57). Both carcinogens were powerful mutagens and induced mutation profiles in the lacI reporter gene similar to those observed in P53 gene in human HNSCC; specifically 50% of the mutations-induced by both carcinogens are G:C→T:A and G:C→A:T substitution and about 30% of the mutations AT base pairs and these percentages are similar to those found in the P53 gene in human HNSCC (41,58). We have recently reported, for the first time, that the mutation profile induced by a mixture of DB[a,l]P + NNN in the oral cavity of lacI mice resembled that of the P53 gene in human HNSCC more than either agent alone, suggesting that they may be acting synergistically in disease development (59). We also showed that DB[a,l]P and DB[a,l]PDE upregulated the expression of p53 and COX-2 proteins (56,57). We further demonstrate that DB[a,l]P resulted in upregulation of several inflammatory-related genes in the mouse oral tissue (60). In addition, we showed that hypomethylation of Fgf3 is a potential biomarker for early detection of OSCC in mice treated with DB[a,l]P (61). Fgf3 is among a large fibroblast growth factor superfamily genes, which are involved in numerous biological activities, including cell survival and regulation of epithelial-mesenchymal transition (EMT) pathway (62–64). Frequent amplification of Fgf3 gene is observed in HNSCC (65–67). Taken together, it would be useful to examine the effects of select tobacco carcinogens (DB[a,l]P, NNN) as co-factors to promote viral persistence leading to increased papillomavirus associated OPSCC in both immune deficient and immune competent mice (68).

Figure 2:

Metabolic Activation of DB[a,l]P

Figure 3:

Metabolic Activation of NNN

Genetic Mouse Models for OSCC and OPSCC

Although mechanisms of oral carcinogenesis remain to be fully defined (69,70), several molecular targets have been identified (Figure 1) as important factors in malignant transformation including cell cycle regulators CDK4, CDK6, and cyclin D1 and Notch 1; the latter can exert both oncogenic and tumor suppression activities (71). Over-expression of COX-2 is well-established in OSCC (72) and is also apparent in our DB[a,l]P-induced mouse model (41).

Transgenic mice have been employed to study oral carcinogenesis; however, due to low physiological levels of transgene products, the use of two or more genes (e.g. K-ras or AKT) to drive tumor formation may be necessary to reflect tumor development in humans (41). We have generated and mechanistically characterized a genetic mouse model in which conditional loss of the tumor suppressor p120ctn in the squamous oral cavity resulted in OSCC that precisely phenocopies the histologic features of human OSCC (73,74). P120ctn protein expression is decreased or lost in approximately 75% of HNSCC and is associated with poor prognosis (75). Tumors in this model develop over 9-12 months in the background of dysplastic lesions containing a chronically inflamed microenvironment dominated by myeloid lineage cells. This immune response is likely induced by NF-κB activation in p120ctn-null cells with commensurate cytokine secretion. Indeed GM-CSF, M-CSF, MCP-1 and TNF-alpha are secreted from these cells. The immune system plays a pivotal role in control of tumor growth and HNSCC is characterized by profound immunosuppression (76). Increased density of tumor infiltrating lymphocytes (TILS) was observed in HPV-positive as compared to HPV-negative OSCC (26). Investigations aimed at determining the functional roles of immunological biomarkers and the complex tumor-immune cell interactions could define subgroups of patients that are more likely to respond to immunotherapy.

To represent the complex environment that mimics human exposure it would be important to explore how p120ctn inactivation cooperates with other genetic and environmental factors (tobacco carcinogens) and HPV infection to induce oral cancers. An estimated 30%-40% of HNSCC have mutations in the PI3K pathway (77–79) and we observed that p120ctn down regulation worked cooperatively with PIK3CA mutations to drive invasion in oral keratinocytes in vitro through increased MMP1 expression (74). Analysis of human HNSCC confirmed that p120ctn loss and P13K pathway activation were associated with increased MMP1 expression (80). Thus, building on these established in vitro and in vivo models in combination with the MmuPV1 mouse model, investigations of how HPV infection may be enhanced by p120ctn loss can be conducted. In addition to establishing p120ctn’s relevance in HNSCC (75), p120ctn has been implicated in cellular responses to tobacco carcinogens. Carcinogens from tobacco exposure are known to induce DNA damage and also induce decreases in p120ctn protein levels (81). Over time, exposure to these carcinogens can result in chronic inflammation and tumor development which is analogous to the effects induced by p120ctn loss in the mouse oral cavity. It is possible that p120ctn loss may be an important event in the pathogenesis of HNSCC resulting from tobacco carcinogens. It is also possible that p120ctn may be acting on pathways that are independent from tobacco carcinogens in the development of HNSCC. Identifying which of these two relationships exist between p120ctn loss and tobacco carcinogens would be of great importance in understanding the pathobiology of HNSCC. Thus, studies examining the mechanisms by which p120ctn potentiates disease development induced by tobacco carcinogens are urgently needed.

A variety of epigenetic alterations have been described in the progression of tobacco/alcohol related OSCC (70). These include promoter hypermethylation of tumor suppression genes, global hypomethylation, and changes in gene-specific methylation patterns, non-coding RNAs, and histone modifications. The biological functions of these genes (cell cycle regulation, cell growth, survival, apoptosis, and epithelial-mesenchymal transition [EMT]) have been described. We showed that DB[a,l]P induced hypomethylation of Fgf3, which is involved in EMT pathways, can occur early in the mouse oral cavity (61). Genetic and epigenetic differences between OSCC and OPSCC have been noted including a higher frequency of hypermethylation of the P16 gene promoter in human tongue cancers compared to pharyngeal and laryngeal cancers (82). Further, hypermethylation frequency of p16 was also greater in tumors of females compared to males (82). Collectively, pathways identified in preclinical studies should also be analyzed in human HNSCC samples using a multi-omics approach to assess their clinical relevance.

Impact of Oxidative Stress on OSCC and OPSCC

Most preclinical studies on HNSCC have focused on individual carcinogens such as DB[a,l]P and NNN as etiological agents (41). However, tobacco smoke contains over 7000 chemicals including high levels of reactive oxidants, including free radicals and aldehydes known to impact various pathways involved with initiation and post-initiation phases of carcinogenesis (83). Epidemiologic evidence indicates that higher levels of oxidative stress can act cooperatively with other tobacco smoke exposures to increase susceptibility to lung cancer (84). Furthermore, we have demonstrated that individuals with decreased capacity to protect against oxidative stress are at increased risk for oral as well as lung cancer (85). However, the impact of these tobacco smoke oxidants in the different phases of HNSCC has received little attention in the laboratory. We demonstrated that like conventional cigarettes, e-cigarettes also emit highly reactive free radicals and aldehydes (86,87), all of which can induce oxidative stress and inflammation. Recent findings suggest that e-cigarette aerosol exposure may have co-mutagenic activities through induction of phase-I carcinogen-bioactivating enzymes and enhanced oxidative DNA damage (88,89). Using our mouse models (DB[a,l]P, MmuPV1, p120ctn), it would be of great interest to elucidate the role of oxidants contained in whole tobacco smoke and e-cigarette aerosols in OSCC and OPSCC development.

Interactions of Known Risk Factors with the Oral Microbiome

The human oral cavity contains a diverse community of ~700 bacterial species, many of which may contribute to carcinogenesis. Several studies using different experimental approaches have examined and characterized the interactions of oral cancer risk factors with the microbiome (90–92). Alterations in the structure, diversity and function of the oral microbiome that occur in association with established risk factors (tobacco smoking, betel nut chewing, alcohol) may contribute to oral cancer development (91,93,94). A nested case-control study was carried out using data from 2 large prospective cohorts: the American Cancer Society Cancer Prevention Study II (CPS-II) and the Prostate and Lung, Colorectal and Ovarian Cancer Screen Trial (PLCO) (92). Overall microbiome composition was not associated with risk of developing HNSCC. However, greater abundance of genera Corynebacterium and Kingella was associated with decreased risk of HNSCC with potential implications for cancer prevention (92). The results of this study (92) appear to be in line with findings observed in an experimental mouse model that utilize the synthetic oral carcinogen 4-nitroquinoline-1-oxide (95) where a decrease in Corynebacterium was observed in the OSCC group as compared to control mice. Recently, Abusleme et al. (96) developed an experimental protocol for characterizing the oral microbiome in murine models which should prove useful for future animal model investigations. After the gut, the oral cavity houses the second largest microbial community in humans (97), including a wide range of microorganisms that can be modified by environmental exposures linked to carcinogenesis (smoking, diet, alcohol). For example, oral bacteria can metabolize alcohol to acetaldehyde that can lead to mutagenesis and carcinogenesis. In a pilot study, Stewart et al. (98) tested the impact of e-cigarettes or tobacco smoking on the oral and gut microbiota in comparison to non-smoking controls. In summary, Stewart et al. (98) found that tobacco smoking significantly altered the bacterial profiles in feces, buccal and saliva samples. Changes in the gut microbiota of tobacco smokers were associated with increased relative abundance of Prevotella while decreased relative abundance of Bacteroides. However, exposure to E-cigarettes, compared to controls, had no effect on the oral or gut communities. Based on the microbial ecology perspective, the authors stated that E-cigarettes may represent a safer alternative to tobacco smoking; however, they noted further research aimed at assessing other endpoints besides the microbiota is needed to determine the impact of E-cigarettes on human health and disease. HPV infection was found to coincide with the presence of certain dysbiosis-associated bacteria which correlate with oral carcinogenesis (99,100). Also poor oral health increases the risk of HPV infection and may contribute to HPV-related OPSCC (101). Such relationships may account, in part, for tissue tropism. Overall, studies aimed at determining the relationships between etiological factors (smoking, HPV, alcohol), oral microbiome diversity, and HNSCC development as proposed in this review are needed.

Challenges to Disease Prevention

Treating cancer at late stages, even with recent advances in targeted therapies, continues to be a major challenge (41). Thus, prevention clearly represents the most effective strategy to manage and control this disease. Avoidance of risk factors (e.g. tobacco cessation) has only been partially successful and survival rates have not improved despite advances in therapeutic approaches (41).

In contrast to cervical cancer, currently there is no US Preventive Services Task Force recommended screening for other HPV-associated cancers (102) and less is known about the carcinogenic progression of HPV- infections at other sites including oropharynx (103). Although smoking is a risk factor for OPSCC and smoking rates in the US have declined, increases in OPSCC have been attributed to HPV infection (104). Despite the availability of vaccines against HPV infection (105), their utilization are not yet widespread, especially in boys and male adolescents. Furthermore, they do not address the large number of currently infected patients. For these patients and those with HPV-negative cancers, there is an urgent need to understand the molecular basis for how HPV infections progress to OPSCC development. Several studies demonstrated that OPSCC is more sensitive to radiotherapy and chemotherapy compared with other types of HNSCC, cf. Table 1 (106–108); why HPV-positive patients are more responsive remains to be explored. Nevertheless, the impact of smoking on HPV-positive OPSCC emphasizes the continuous need for smoking cessation programs for primary prevention (109–112).

A previous study provided proof-of-principle that the HPV vaccine may prevent HPV-induced OSCC and OPSCC; i.e. primary prevention (113) and three prophylactic HPV vaccines have been approved by the FDA. However, individuals >30 years of age may not benefit from vaccination because prophylactic vaccines do not clear existing HPV infections. In 2017, uptake of a single HPV vaccine dose and HPV vaccine series completion was 66% and 49%, respectively (114). Further, HPV vaccination rate is lower in males than females (115,116). We would expect a reduction in OPSCC incidence with increased vaccination rates in boys, girls and adolescents (117).

Clinical diagnosis of HPV-associated HNSCC includes the detection of HPV DNA and RNA and the surrogate marker, p16 in tumor tissues and HPV-specific antibodies in serum; however, these do not allow for early HNSCC detection (118). At present, there are no universally-accepted, FDA-approved oral HPV detection methods. The use of clinically relevant oral fluids for HPV detection and assessment of cellular alterations in infected cells will likely assist in early detection and novel intervention. Oral rinse HPV testing has had moderate to poor sensitivity, but results with HPV16E6 serum antibody have been promising (119). HPV testing must not be considered in isolation since there are important interactions with other factors such as tobacco exposure (120).

Prevention approaches should be directed toward molecular mechanisms responsible for cancer induction by environmental carcinogens using targeted intervention/chemoprevention strategies. We have previously reviewed (41) the role of chemopreventive agents employed in preclinical and clinical studies against HNSCC development. Recently, Tsai and Dillon (20) published an overview of the current state of several chemopreventive agents that have been examined in oral cancer in clinical trials and provided commentary on potential future research directions in this field. Agents discussed in this review included N-acetyl-L-cysteine, the antibiotic bleomycin, tea polyphenols, curcumin, retinoids, carotenoids, non-steroidal anti-inflammatory drugs and ligands for PPARɣ activators. Early studies used retinoids for oral cancer chemoprevention. Administration of 13-cis-retinoic acid for 3 months resulted in a 67% regression of oral leukoplakia vs. 10% for placebo; dysplasia was reversed in 54% of the drug groups and in 10% of the placebo group. The toxic effects of the drug were acceptable in all but 2 patients; however, relapse occurred in 9 of 16 patients, 2-3 months after treatment ended (121). In a follow-up study (122), daily treatment with high doses of 13-cis-retinoic acid was effective in preventing second primary tumors in patients who have been treated for HNSCC, although it did not prevent recurrences of the original tumors. Other safe and effective agents (vitamin E, Bowman-Bink inhibitor concentrate derived from soy beans, curcumin, and green tea polyphenol) have been tested with varying levels of success in oral leukoplakia patients (41). Furthermore, the presence of numerous potential chemopreventive agents in black raspberry (BRB) (vitamin E, C, A, folic acid, calcium, selenium, B-sitosterol, elegiac acid, ferulic acid, quercetin) make it an attractive chemopreventive agent based on several preclinical and clinical studies (41). For example, several studies explored the tolerability and feasibility of BRB; Mallery and her team and Ugalde et al. showed the feasibility of delivering a “muco-adhesive gel” into the oral cavity of healthy volunteers (123–125). When dysplastic lesions were treated with 10% BRB gel (4 times/day for 6 weeks), there was histologic regression of ~60% of the lesions (126,127). Knobloch et al. showed that BRB phytochemical-rich troche suppressed proinflammatory and prosurvival markers in oral cancer patients (128). To understand the role of microbiome in cancer prevention by green tea and to provide insights into the observed inconsistency in clinical trials (129–131), Adami et al. (132) demonstrated that tea consumption can consistently change oral bacteria in humans, which may affect carcinogenesis, however, the authors argue that green tea effects on oral epithelial miRNA expression in humans vary between subjects. In addition to its antimicrobial activities, tea catechins have been shown to exert antiviral activities (133,134). Among different catechins, epigallocatechins-3-gallate is the most prevalent and believed to be the major constituent that can account for cancer prevention by tea (135). Collectively, several promising chemopreventive agents have been identified based on in vitro and in vivo studies; however, their clinical benefits remain limited. Thus, the search continues for safe and more effective chemopreventive agents to clinically manage this disease (41).

Tobacco use remains a significant risk factor resulting from exposure to potent tobacco carcinogens. Collection of oral mucosal cells is relatively simple and such cells are an excellent source of material for evaluating DNA adducts derived from structurally varied carcinogens identified in tobacco smoke and molecular alterations potentially related to cancer. In fact, the extent of DNA damage derived from NNN was quantified in buccal cells of smokers; such damage was shown to be an independent risk factor for HNSCC (136) and thus can provide a novel target for chemoprevention agents which enhance DNA repair capacity and possibly reverse epigenetic alterations induced by etiological agents that drive disease progression (Figure 1). Furthermore, to complement smoking cessation programs, increased HPV screening and prevention are essential.

Summary and Future Directions

Future basic research as proposed in this review are summarized below:

• Examine the role of etiological agents (tobacco smoking, HPV) individually and in combination on molecular targets (genetic and epigenetic) in well-defined animal models (MmuPV1-infected and genetically modified mouse models) during different stages of disease progression.

• Decipher molecular mechanisms that can fully account for development of papillomavirus-positive vs. negative disease.

• Understand the basis for HPV tropism (back of the tongue) and the impact of infection on immune regulation at different sites in the oral cavity.

• Determine the functional roles of the various immunological markers and the complex tumor-immune cell interaction.

• Examine the role of tobacco-smoking induced oxidative stress in infected and transgenic animal models.

• Determine the impact of etiological factors on the oral microbiome composition and function as related to disease development.

• Compare pathways identified in animals with those in human HNSCC tissues using a multi-omics approach to assess the relevance of these pathways in disease development.

• Corroborate pathways identified in the MmuPV1 mouse model with human tissue from HPV-associated OPSCC.

• Determine the efficacy and mechanisms of chemopreventive agents in the above-mentioned animal models.

These studies can lead to the development of sensitive and specific biomarkers for early detection that can help reduce oral cancer-related burden and improve overall patient survival rates. Considering the availability of different mouse models (chemically-induced, wild type, genetically engineered and MmuPV1-infected) combined with overwhelming evidence supporting the role of tobacco and viral infection in HNSCC development, we strongly believe the studies proposed above will provide key mechanistic insights required for answering the critical questions stated in this review (cf. Introduction). Of particular significance the outcome of these preclinical studies can also assist in the design of novel prevention strategies for HNSCC and avoid a potential epidemic in the USA resulting from OPSCC.