Ras genes are the most frequently mutated oncogenes in human cancer. However, the contribution of ras to tumor initiation still is unclear because ras expression in primary cells can cause cell cycle arrest and even cell death by apoptosis. Furthermore, when expressed in the epidermis of mice, mutant ras promotes the formation of benign papillomas, only few of which will progress into carcinomas. However, in these cases, ras-transgene expression often is restricted to suprabasal or follicular epithelial cells that may lack self-renewal capacity. Thus, it still is conceivable that expression of active ras in other epithelial compartments may exert a distinct ability to promote malignant progression. To address this possibility, transgenic mice carrying the tetracycline-inducible system (tet-on receptor) targeted to the basal layer of stratified epithelium, which includes the epithelial stem cells, were engineered and crossed with mice expressing the K-rasG12D oncogene under the control of tet-regulated responsive elements. On doxycycline administration, proliferative lesions ranging from hyperplasias, papillomas, and dysplasias to metastatic carcinomas developed in squamous epithelia of the skin, oral mucosa, salivary glands, tongue, esophagus, forestomach, and uterine cervix within just 10 to 20 days. The most noticeable lesions were invasive squamous carcinomas of the skin and oral mucosa. These findings suggest that the expression of oncogenes in an epithelial compartment that includes the stem cells may be sufficient to promote squamous carcinogenesis. They also provide a molecularly defined conditional animal model system in which the mechanisms responsible for cancer initiation, maintenance, and metastatic spread can be readily investigated.
Mutations in the three human ras genes, H-ras, K-ras, and N-ras, have been detected with high frequency in human cancers, including those from the lung, colon, urinary bladder, gall bladder, pancreas, breast, and ovary, with frequencies ranging from 10 to 90% (1) . Activating mutations in K-ras and H-ras also have been reported in human squamous cell carcinomas (SCCs), primarily in those caused by exposure to carcinogens present in betel quid (2) . Similarly, H-ras mutations frequently are found in experimental mouse models of chemical carcinogenesis (3) . However, the role of ras in tumor initiation still is controversial because ras expression in primary cells can cause cell cycle arrest or even apoptosis (4 , 5) . The prevailing view is that human and experimental carcinogenesis involves the sequential accumulation of activating mutations in several oncogenes along with inactivation of a number of tumor suppressor genes (6) . In line with this view, expression of mutant ras in the skin of mice, when caused by chemical carcinogens or by transgenic strategies, promotes benign papilloma formation with a variable rate of delayed carcinoma conversion (3) . Because in these cases ras expression often is restricted to suprabasal or follicular cells that may lack self-renewal capacity, the possibility still exists that activation of ras in other epithelial compartment may display a distinct ability to promote tumor initiation and progression (3 , 7) . We show here that the conditional expression of activated K-ras in the basal layer of the epidermis, which includes the stem cell compartment (8) , is sufficient to induce the rapid (10 to 30 days) formation of malignant SCC in the skin and other squamous epithelia.
Materials and Methods
Generation of Transgenic Mice.
The K5-rTA transgene, referred herein as K5-tet-on, was constructed by adding a nuclear localization signal to the reverse tetracycline transactivator (rTA; ref. 9 ) and cloned into a unique NheI site in a pBSKII-derived vector, downstream from the bovine keratin 5 (K5) promoter and upstream from a polyadenylation signal (10) . K5-tet-on mice were produced by injecting the 9.452 kb Asp718-excised fragment into FVB/N blastocytes. Founders were screened for transgene insertion by Southern blot analysis, and subsequent generations were screened by PCR using specific primers. tet-K-ras transgenic mice, expressing active K-ras4bG12D oncogene, referred herein as K-rasG12D, under the control of seven tet-responsive elements, have been described previously (11) . Doxycycline (RPI) was administered via the drinking water at a concentration of 500 mg/L.
Tissue Preparation, Histology, and Immunohistochemistry.
Tissues were fixed in 4% paraformaldehyde overnight and transferred to 95% EtOH and embedded in paraffin. Five-micrometer sections were cut and stained with hematoxylin and eosin or analyzed for expression of proliferating nuclear antigen (PCNA; Zymed, San Francisco, CA) and active phosphorylated forms of Akt and extracellular signal-regulated kinase (ERK) (Cell Signaling, Beverly, MA). Cryostat sections were obtained from snap-frozen, OCT-embedded tissues samples, fixed in 4% paraformaldehyde in 1× PBS, washed, and processed for β-galactosidase staining (10) .
Results and Discussion
Transgenic expression of mutant ras genes in suprabasal epithelial cells or proliferating cells within the hair follicle causes hyperplasias and papillomas, <15% of which would slowly progress into carcinomas after several months (12 , 13) . Thus, expression of ras in differentiating or committed cells may not be sufficient to promote cancerous growth. Because available evidence suggests that squamous carcinomas may instead arise from genetic alterations in more primitive epithelial cells (3 , 7) , we decided to explore the consequences of expressing mutant ras from the K5 promoter, which targets transgene expression to a basal epithelial cell compartment that includes the stem cells (8) . However, numerous attempts to express activated ras using this promoter failed to generate viable transgenic animals (data not shown), suggesting that ras expression may compromise mouse development and/or viability. Thus, as an alternative approach, we set out to express ras in the epithelium of adult animals using a tetracycline-inducible system. Mice were engineered to express the rTA (tet-on) transgene under the control of the K5 promoter and were first crossed with mice expressing the β-galactosidase gene under the control of tet-responsive elements to test the effectiveness of this in vivo gene expression system (Fig. 1) ⇓ . β-Galactosidase activity was readily detected in the epidermis of double transgenic animals but not in single transgenic mice and that was strictly dependent on the administration of doxycycline, a tetracycline analog, thus confirming the specificity of this system. β-Galactosidase activity was observed in follicular and interfollicular epithelium, as well as in other stratified squamous epithelium, such as in the tongue and forestomach, mirroring the pattern of expression of K5 (10) .
We next crossed the K5-tet-on animals with mice expressing an activated allele of K-ras, K-rasG12D, under the control of tet-responsive elements (tet-o-K-ras). Double transgenic animals were viable and healthy. Administration of doxycycline caused skin keratinocyte hyperproliferation, as judged by the immunodetection of proliferating nuclear antigen in nearly all of the basal epithelial cells, even after only 2 days of treatment (Fig. 1) ⇓ . Furthermore, these animals began to exhibit a wide range of visible lesions within 2 weeks of doxycycline administration, particularly in the skin and the mouth (Fig. 2) ⇓ . Animals sacrificed 3 to 4 weeks after initiating doxycycline treatment displayed overt lesions of the skin and squamous epithelium from other organs, such as the forestomach and esophagus (Table 1) ⇓ . Histologically, a wide variety of epithelial alterations were observed, ranging from benign hyperplasia to dysplasia and SCCs. The latter were detected in ∼40% of the animals (6 of 17 mice), some of which had already metastasized to lymph nodes (Fig. 3) ⇓ , indicating that expression of activated alleles of K-ras in this particular epithelial compartment promotes malignant transformation. Furthermore, tumor progression correlated with the persistent activation of Ras-initiated biochemical pathways, such as the activation of extracellular signal-regulated kinase and Akt, as judged by the immunodetection of the active, phosphorylated form of these signal-transducing molecules (data not shown).
Whereas ras readily transforms cell lines in culture, this oncogene does not stimulate cell proliferation but instead causes cell cycle arrest or even apoptosis when expressed in primary rodent and human fibroblasts (4 , 5) . In our animal model, however, conditional expression of K-ras potently stimulated cell proliferation, an observation in line with recent reports that activated ras promotes cell growth when conditionally expressed in other epithelial tissues, such as the lung, pancreas, and colon (14 , 15) . Collectively, the available data support the emerging concept that ras can display a potent proproliferative activity in vivo, which is distinct from its biological effects in primary cells in vitro. However, the observation that the sole expression of ras can result in the rapid malignant conversion of squamous epithelium is intriguing because mutations in the ras oncogene usually are considered insufficient to directly promote carcinogenesis (16) . Because mice have long telomeres, unlike humans, one explanation could be that murine cells may be more prone to cell transformation because they do not need to reactivate telomerase activity to initiate tumorigenesis (17) . We also have observed that ras-transformed cells display abnormal nuclear morphology consistent with changes in ploidy, whose nature is under investigation. Thus, chromosomal instability secondary to ras expression may promote the rapid genetic and epigenetic inactivation of tumor suppressor genes. Alternatively, functional inactivation of key tumor suppressor proteins may be achieved by the ability of Ras to stimulate signal-transducing pathways and cell cycle–regulating molecules. These, as well as additional possibilities, are under current investigation.
In this regard, it is tempting to speculate that the remarkable effects caused by ras in our animal model could be because of the expression of this oncogene in a particularly susceptible cell population, such as the epidermal stem cells, which might exhibit distinct requirements for malignant transformation. In contrast to the fast onset of malignant conversion observed here, when ras is expressed in other epithelial cell compartments, SCCs arise only after an extended period in which additional stochastic genetic lesions are likely to occur (3) . Classical chemical carcinogenesis studies further support this explanation because activation of endogenous ras by carcinogens appears to occur primarily in cells that have left the stem cell niche (3 , 7) , thus also causing benign tumors, few of which can progress into frank malignancies after prolonged treatment with tumor promoters. The primitive stem cells, few in number and buried deep into the epidermis, may be protected from mutagens because they are less accessible to environmental carcinogens than are the more superficial epithelial transit-amplifying cells (3 , 7) . Stem cells also divide infrequently; therefore, they have fewer chances to incorporate mutations or can repair them more effectively (18) and are susceptible to apoptosis secondary to DNA damage, often choosing to self-eliminate rather than undergo extensive error-prone DNA repair (18) . Conversely, recently developed mathematical models suggest that exponential growth during development may result in the accumulation of mutations in a small population of stem cells, which, although not tumorigenic in nature, may nonetheless predispose to certain late-life cancers (19) . Thus, it is possible that in our animal model the seemingly explosive growth of cancerous cells may result from the transformation of this limited pool of already predisposed cells that do not require the accumulation of additional mutations subsequent to ras activation for malignant conversion. This exciting possibility may be relevant to human cancer. Individuals may harbor few already predisposed epithelial stem cells or cells that have regained self-renewal capacity, which may accumulate additional mutations over the years by the exposure to carcinogens, but without presenting any clinically identifiable lesion. In this scenario, aberrant activation of proliferative pathways (e.g., by mutations in ras or other oncogenes) may promote the rapid cancerous growth of this particular population of susceptible cells. Additional work in this novel experimental animal model may help address this possible stem cell origin of ras-induced tumors, as well as the nature of the preexisting or subsequent genetic and epigenetic events that enable tumor progression. We envision that the future use of this molecularly defined conditional animal model system may help in current efforts to unravel the mechanisms responsible of cancer initiation, maintenance, and metastatic spread, thus aiding in the search for novel chemopreventive strategies and molecular-targeted human cancer therapies.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Requests for reprints: J. Silvio Gutkind, Oral and Pharyngeal Cancer Branch, NIDCR, NIH, 30 Convent Drive, Building 30, Room 212, Bethesda, MD 20892-4340. Phone: 301-496-6259; Fax: 301-402-0823; E-mail:
- Received July 22, 2004.
- Revision received October 2, 2004.
- Accepted October 22, 2004.
- ©2004 American Association for Cancer Research.