Animal models for the study of squamous cell carcinoma of the upper aerodigestive tract: A historical perspective with review of their utility and limitations part B. Transgenic mouse models

Experimental multistage carcinogenesis involves three broad sequential steps: tumor initiation, tumor promotion and tumor progression, and appear to be similar to human carcinogenesis. However, we expect that these processes in humans will be more complex and may not occur in a predictable order. The net result of a number of events such as activation of oncogenes, inactivation of tumor suppressor genes and expression of apoptosis-regulating genes may likely contribute to the accumulation of multiple genetic alterations, leading to tumor formation. In addition, humans are exposed to multiple mutagenic agents that may act at the same time on tumor initiation, promotion or progression. The classical animal models fall short of providing important information concerning carcinogen–host and tumor–host interactions. The development of animals with overexpression or inactivation of carcinogenesis-related genes has expedited studies on carcinogenesis and its prevention and treatment. These transgenic animals provide a powerful model to explore mechanisms of gene expression as well as the regulation of cellular and physiologic processes involved in initiation, promotion and progression of cancer. With this technology, targeted expression of a protein can be achieved in a controlled fashion by expressing the corresponding gene in vivo ectopically, ubiquitously or in a tissue-specific manner, using its own promoter. These newer models may provide more precise control over gene expression at various stages of development through activation of mutant alleles at precise points in the developmental process using conditional Cre-Lox and tetracycline-inducible technology. The Cre-LoxP and the tetracycline (Tet)-inducible systems are useful methods of conditional gene expression that allow spatial (cell-type-specific) and temporal (inducer-dependent) control. The Cre-LoxP system, which has been widely used for gene manipulation in animal models, uses P1 phage Cre-recombinase to catalyze the excision of DNA located between flanking loxP sites for conditional gene inactivation. Tetracycline-controlled gene expression relies on two components: (i) a tetracycline-responsive transactivator (tTA) that is, a fusion of E. Coli tet repressor to the transactivation domain of VP16 protein of herpes simplex virus, and (ii) a tTA-dependent promoter consisting of a minimal RNA polymerase II promoter fused to tet operator sequences. By using a reverse tTA, which becomes transcriptionally active only when bound to tetracycline or its derivatives, expression of a target gene can be induced by administration of tetracycline or doxycycline.1, 2, 3 In recent years, several transgenic mouse models with overexpression or inactivation of cancer-related genes using the Cre-LoxP technology have been used to study the initiation, development and metastasis of epidermal/cutaneous squamous cell carcinoma (SCC). However, the use of these animal model systems for the study of head and neck squamous cell carcinoma (HNSCC) is still in its infancy. Studies of several major cancer-related genes in HNSCC carcinogenesis using transgenic models are discussed below. Squamous epithelia of the upper airway as well as a variety of human tumors show abnormal overexpression of cyclin D1. Nakagawa et al. used an Epstein–Barr virus (EBV) ED-L2 promoter in a construct to target human cyclin D1 in transgenic mice.4 They observed the expression of the cyclin D1 transgene in oral and esophageal epithelia and a phenotype that showed mild dysplasia by 6–12 months, with progression to moderate–severe dysplasia by 15–16 months. Overexpression of PCNA and abnormalities in cyclin-dependent kinase-4, EGFR and p53 during early HNSCC carcinogenesis were also demonstrated,5 but no evidence of cancer was noted. Cross breeding the cyclin D1 expressing mice with p53 knockout mice, however, led to the development of mice with histologic evidence of invasive oral cancer,6 suggesting that cyclin D1 requires cofactors to cause malignant epithelial transformation. The TP53 tumor suppressor gene is mutated in over 50% of human cancers and is considered a central event in the progression of human malignancies. Mice with heterozygous knockout or mutations in the TP53 tumor suppressor gene show a high sensitivity to a variety of carcinogens and, thus, provide relevant tumor models to study carcinogenicity of pharmaceutical agents. However, there is a paucity of studies specific to HNSCC using this model. Nishikawa et al. showed a high susceptibility of p53 (+/−) knockout mice to esophageal carcinogenesis after administration of N-dibutylnitrosamine (DBN) in their drinking water for 8 weeks.7 This was likely attributed to early mutations of the residual TP53 allele and increased cellular proliferation in the target organs. Mutations in the K-ras and H-ras genes have been reported in HNSCC to be between 30 and 50% depending on geographical region.8 However, the role of –ras genes in HNSCC carcinogenesis is still debatable. Vitale-Cross et al. (2004) engineered transgenic mice carrying the tetracycline-inducible system (tet-on receptor) targeted to the basal layer of stratified epithelium, which includes the epithelial stem cells, by crossing with mice expressing the K-ras (G12D) oncogene under the control of tet-regulated responsive elements. Progeny animals developed lesions ranging from hyperplasias, papillomas and dysplasias to metastatic carcinomas in epithelia from skin, oral mucosa, salivary glands, tongue, esophagus and uterine cervix. The most noticeable lesions, however, were invasive squamous carcinomas of the skin and oral mucosa,9 suggesting that expression of K-ras oncogenes in epithelial compartments containing stem cells may be sufficient to promote squamous carcinogenesis. Importantly, Caulin et al. developed a novel system based on the generation of transgenic mice that express an inducible Cre recombinase under the control of a promoter specific for stratified epithelia. They fused Cre to a deletion mutant of the human progesterone receptor (PR), which fails to bind progesterone, but can be activated by progesterone antagonists, such as RU486. This fusion protein (CrePR1) is sequestered in the cytoplasm. After activation with RU486, CrePR1 translocates to the nucleus, where it mediates the excision of DNA sequences flanked by LoxP sites. In this matter, they generated mice that express CrePR1 under the control of the K5 or K14 promoter, which are expressed in the basal cells of stratified epithelia. These investigators demonstrated the formation of squamous papillomas that may represent early stages of tumor progression in oral cavity of mice in this inducible model with focal activation of K-rasG12D allele. This is the first study to use a novel inducible mouse model that allows oncogene activation or inactivation of tumor suppressor genes exclusively in stratified epithelia of the oral cavity.10 The HPV 16 E6 and E7 oncogenes can inactivate the tumor suppressor proteins p53 and Rb (the retinoblastoma susceptibility gene product), potentially resulting in carcinogenesis. Transgenic mice harboring the HPV-16 E7 oncogene showed histologic hyperplasia of stratified squamous epithelia at multiple sites including the skin, mouth palate, esophagus, forestomach and exocervix.11 Thus, HPV-16 E7 appears to alter epithelial cell growth sufficiently to promote tumorigenesis in mice, especially in the epidermis. The erbB family of receptor tyrosine kinases has been shown to be important in development of human SCC. Xie et al. (1999) developed a k14-rTA/TetRE-ErbB2 ‘Tet-on’ bitransgenic mouse system to investigate the significance of ErbB2 signaling in epithelial tissues. These mice were normal until the ErbB2 transgene was induced by exposure to doxycycline (Dox). ErbB2 transgene expression was observed after the administration of doxycycline.12 In adult transgenic mice, severe hyperplasias were noted in the skin, cornea, eyelids, tongue and esophagus. These were accompanied by increased expression of TGF-α, a ligand of the epidermal growth factor receptor. The results of this study suggested that ErbB2 may play an important role in the development of epithelial neoplasia. Mice expressing a human transgene can be used to help predict the broad principles of generating immunity to a self, tumor antigen. However, these systems possess several disadvantages. Transgenic mice are limited in availability and are expensive to use. For these reasons, they are not suitable for large scale drug testing and screening. As well, it remains unclear how a transgenic mouse model reflects human disease in terms of degree of gene expression during embryonic development and the expression levels of the antigen in normal tissues. Transgenic mice contain the transgene in every cell type and, under ideal conditions, express it in organ-specific subsets of cells. These conditions do not mimic those seen in the initiation of human cancer, where genetic alterations are sustained in only a few cells and not simultaneously altered in all cells of a given lineage. Transgenic strategy has other inherent limitations. If the genetic alteration is introduced in the germ line, the lack of its protein, if it has essential function during embryogenesis, can result in early lethality. This may preclude analysis of its possible functions at subsequent times in the adult animal. Germ line mutations may also be compensated during development, thus averting the appearance of an abnormal phenotype in the adult animal. Newer research tools, however, have made it possible to focally control gene expression so that effects of genetic activation or inactivation are restricted to a particular tissue type.9, 10 Spatially and temporally generated cell-type specific somatic mutations, however, will require careful evaluation of the pattern of expression and activity of the gene in question in transgenic animals, since mosaic expression in the target tissue can lead to incomplete gene inactivation or activation. In addition, expression of the transgene in nonrelevant tissue may lead to undesirable mutations that may complicate analysis of the phenotype. Preclinical studies are essential in translating basic mechanistic research from the bench to bedside. As such, animal models play a key role in contemporary cancer research for the study of carcinogenesis, chemoprevention and molecular/immunologic alterations leading to progression of carcinoma. Research on human HNSCC was for many years hampered because of the very limited availability of experimental model systems. Until the development of the transplanted xenogeneic animal model systems, it was difficult to study human tumor tissue in vivo. Undoubtedly, these animal models have supported drug development and have been useful for identifying genes involved in suppression and promotion of HNSCC. By providing novel targets for the development of new therapies, they have contributed significantly to the understanding and treatment of HNSCC. Although it is of critical importance to be able to extrapolate the findings obtained from animal models to human HNSCC, lack of true representation of the clinical disease (local aggressiveness and metastatic potential) must limit interpretation of studies involving these models. Transgenic mouse models have allowed advances in the understanding of specific genetic alterations in the development and progression of human cancer. For HNSCC, transgenic animals with inactivation of tumor suppressor genes or activation of oncogenes have yielded animals with severe hyperplasia (ErbB2, HPV E6 and E7, K-ras expression), dysplasia (cyclin D1, K-ras) or invasive HNSCC (cyclin D1 with p53, K-ras). However, newly developed transgenic conditional mouse models that control activation or inactivation of specific genes in a time and tissue-specific manner have the potential to overcome important drawbacks in conventional mouse model systems. The use of this sophisticated technology however, is still in its infancy in studies of HNSCC carcinogenesis. Nevertheless, with these novel reporter systems, it may be possible in the near future to follow the initiation and progression of HNSCC in vivo, thus, allowing identifying of novel genes involved in HNSCC carcinogenesis, progression, invasion and metastasis.