Chemical Carcinogenesis Models for Evaluating Molecular-Targeted Prevention and Treatment of Oral Cancer

Head-and-neck squamous cell carcinomas (HNSCC) mostly arise in the oral cavity, and represent the sixth most common cancer worldwide (1), resulting in nearly 11,000 deaths each year in the United States alone (2). Although alcohol and tobacco use, betel nut chewing, and human papillomavirus infection have long been recognized as the most prevalent risk factors (3, 4), most patients with HNSCCs are still diagnosed at advanced disease stages that often fail to respond to available therapies (3, 4). Indeed, <50% of patients with HNSCC survive >5 years after diagnosis (2). Advancing our understanding of the molecular mechanisms underlying HNSCC progression, however, may afford the opportunity to develop novel molecular-targeted strategies for HNSCC prevention and treatment.

HNSCC progression involves the acquisition of genetic and epigenetic alterations in oncogenes and tumor suppressors and involves the aberrant activity of molecular networks controlling cell growth, migration, and survival (3–5). Although some of the key molecular events contributing to HNSCC initiation and progression have been identified recently, translating these findings into the development of novel mechanism-based preventive and therapeutic approaches for HNSCC has been hampered by a paucity of appropriate animal models of oral carcinogenesis. To address this need, investigators have recently used genetically engineered mice to begin modeling HNSCC progression (6–8) and have initiated parallel efforts to recapitulate the tumor heterogeneity and complexity of the clinical setting by optimizing carcinogen-induced oral-specific carcinogenesis models in immunocompetent mice. As previously reviewed (9), the latter effort includes exposing experimental animals to chemical carcinogens such as coal tar, cigarette smoke, 20-methylcholanthrene, 9,10-dimethyl-1,2-benzanthracene, and 3,4-benzpyrene. Although 9,10-dimethyl-1,2-benzanthracene remains one of the most widely used chemical carcinogens, it is a potent irritant and may have limited use in studying early HNSCC lesions because it produces an inflammatory response, necrosis, and the appearance of granulation tissue when applied locally in the oral cavity (see ref. 9 and references therein). Addressing this limitation, early work revealed that the synthetic water-soluble chemical carcinogen 4-nitroquinoline-1 oxide (4NQO) is more effective for inducing oral carcinogenesis in rats compared with 9,10-dimethyl-1,2-benzanthracene and does not cause an extensive inflammatory response (see ref. 9 and references therein). Therefore, 4NQO was adapted as early as in the mid-1960s for the development of oral squamous cell carcinoma (SCC) in mice, the experimental animal most frequently used for carcinogenesis studies. Table 1 includes information on specific mouse strains, dose and duration of 4NQO exposure, route of administration, and length of observation, which should be useful in judging the best experimental system for investigating the molecular events contributing to oral cancer progression and for testing novel preventive and treatment strategies. We recommend that the reader also consult the original study reports (10–37), many of which represent seminal contributions to this field.

4NQO forms DNA adducts, causes adenosine for guanosine substitutions, and induces intracellular oxidative stress resulting in mutations and DNA strand breaks, all similar to the genetic alterations provoked by tobacco carcinogens, and so 4NQO serves as a surrogate for tobacco exposure (9, 38, 39). Oral-specific carcinogenesis can be achieved by the local delivery of this carcinogen, most often by labor-intensive procedures including brushing the hard palate, tongue, or gingiva with concentrated water solutions of 4NQO, which leads to the progressive acquisition of oral premalignant lesions and SCC (Table 1). Early analysis of the ability of low doses of 4NQO in drinking water to induce the appearance of dysplastic oral lesions and DNA mutations (20, 21, 25, 32, 33) led to the seminal demonstration by Tang et al. that prolonged exposure to higher doses of 4NQO in drinking water was sufficient to cause dysplastic lesions and frank SCCs in the tongue, oral mucosa, and esophagus of both C57BL/6 and CBA mice (31). The tested doses were 20, 50, and 100 μg/mL for 8 and 16 weeks in CBA mice, and 100 μg/mL for 16 weeks in C57BL/6 mice. This approach was found to be better than brushing in terms of ease of application and reproducibility, and 100 μg/mL of 4NQO in drinking water led to oral and esophageal lesions including SCCs in 100% of CBA or C57BL/6 mice (31), which is one of the most frequently used mouse strains for the analysis of genetically engineered alleles (40). 4NQO at the lower dose of 50 μg/mL in drinking water, however, caused minimal esophageal carcinogenesis (microcancer or papilloma) in C57BL/6 mice, whereas also producing the progressive appearance of tumoral lesions in the tongue and oral mucosa that were preceded by clearly identifiable premalignant events in 100% of these mice (10). This model helped reveal that a clinically relevant inhibitor of the mammalian target of rapamycin signaling pathway (rapamycin) represents a potential chemopreventive approach for halting progression in the oral cavity to HNSCC (10). Hasina et al. (17) administered 4NQO at a dose of 100 μg/mL in drinking water to CBA mice for 8 weeks, which caused elevated incidences of hyperkeratoses, dysplasias, and SCCs in the oral cavity and minimal morbidity and mortality over a period of 32 weeks. In these mice, the use of an antiangiogenic agent, ABT-510, reduced the overall incidence of SCC in the oral cavity, supporting the further evaluation of angiogenesis inhibitors in HNSCC prevention. These and prior 4NQO studies provide the basis for the future extended use of the 4NQO oral carcinogenesis model to explore novel chemopreventive approaches for HNSCC. Furthermore, the study of 4NQO in C57BL/6 mice (frequently used for genetic engineering) may facilitate future investigations of the molecular mechanisms underlying HNSCC initiation and progression in mice both treated with 4NQO and engineered to express modified versions of relevant genes.

An emerging body of data suggests that the progressive changes occurring in oral tumoral lesions of the 4NQO mouse model reflect changes that occur in human HNSCC. These shared changes include altered expression of differentiation markers such as keratins, decreased expression or acquisition of potential inactivating mutations in tumor suppressor genes such as p16 and p53, increased expression of the epidermal growth factor receptor and cyclooxygenase-2, enhanced angiogenesis, increased levels of vascular endothelial growth factor, and increased activity of the phosphoinositide-3-kinase–Akt–mammalian target of rapamycin pathway (9, 10, 17, 31). Therefore, the in-depth analysis of 4NQO-induced premalignant and malignant lesions may provide a framework for future explorations of the earliest molecular events in HNSCC progression. On the other hand, we expect that the information summarized in Table 1 may guide the selection of appropriate animal strains and 4NQO dose and exposure duration in developing mouse models that incorporate both genetic engineering and 4NQO-induced carcinogenesis for future studies of specific molecular events that contribute to HNSCC initiation and progression. For example, giving 4NQO to mice defective in a tumor suppressor gene relevant to HNSCC would be expected to lead to the development of either more lesions or earlier tumors; such a model might also better reflect the complexity and heterogeneity of clinical tumors (41). We expect that the use and further development of the 4NQO oral chemical carcinogenesis model in mice may ultimately expedite the evaluation of novel molecular-targeted chemopreventive strategies to halt HNSCC progression. These efforts may also facilitate the identification of new mechanism-based treatment options for patients with HNSCCs.