Induction of Invasive Mouse Skin Carcinomas in Transgenic Mice with Mutations in Both H-ras and p53

Introduction

Nonmelanoma skin cancer is the most common cancer in the United States. Eighty percent of these skin cancers are basal cell carcinomas and 20% are squamous cell carcinomas (SCC; ref. 1). SCCs of the skin develop through a multistep process that involves activation of proto-oncogenes and/or inactivation of tumor suppressor genes in keratinocytes. The p53 tumor suppressor gene is mutated in numerous human cancers, including ∼50% of human skin cancers (2). Forty-six percent and 31% of human primary SCCs and basal cell carcinomas, respectively, contain H-ras mutations (3). Similar to human skin cancer, chemically induced mouse skin cancers are caused by the presence of multiple genetic alterations, including activation of proto-oncogenes and loss of tumor suppressor genes (4). Activated H-ras proto-oncogene has been found in >50% of chemically induced mouse skin papillomas and SCCs (5). Inactivation of several tumor suppressor genes, including p53, p15, and p16, occurs principally in late-stage tumors, such as SCCs (p53) or spindle carcinomas (p15 and p16). p53 mutations have been detected in both chemically induced and UV-induced mouse skin tumors (6, 7). Loss of heterozygosity studies have been conducted to identify the regions of frequent allele loss in skin tumors of various F₁ hybrid mouse strains to localize important tumor suppressor genes (8). Loci were most commonly detected on chromosomes 4, 6, 7, and 11 (8). Chromosome 4 was shown to harbor p15 and p16 genes, whereas chromosome 11 contains the p53 gene. These losses of heterozygosity were only detected in SCCs (chromosome 11) and spindle cell carcinomas (chromosomes 4, 6, and 7), suggesting a role for resident tumor suppressor loci in the malignant conversion of mouse skin tumors. Alterations in the expression of several genes have been reported in mouse skin carcinogenesis (9-12). For example, overexpression of cyclin D1 was observed in chemically induced papillomas and SCCs (11, 12). In UV-treated mice, c-fos, c-myc, and H-ras genes were overexpressed in both exposed skin and skin tumors (9).

Several transgenic mouse models for skin carcinogenesis have been developed (4). One of these models is the v-H-ras transgenic TG·AC mouse line. TG·AC mice were created by the introduction of an activated v-H-ras transgene linked to a ζ-globin promoter into FVB/N mice (13, 14). Consequently, the TG·AC mouse is a genetically initiated model for skin carcinogenesis that has a short latency period and a high susceptibility for skin papilloma induction by tumor promoters or carcinogens (15). Interestingly, this mouse model has been used to screen for agents with carcinogenic or tumor-promoting properties. Another potential tumor model for mouse skin carcinogenesis is the p53 transgenic mouse. Inactivation of the p53 gene in the germ line of mice by a dominant-negative transgene has made possible research with a model analogous in many respects to the human inherited cancer predisposition, Li-Fraumeni syndrome. Germ line p53 mutations were found to significantly increase the number of skin tumors and the propensity for multiple tumor development when treated with UV radiation (16). p53 transgenic mice are also a potential model for human Li-Fraumeni syndrome, because patients with this disease usually carry a germ line p53 mutation that predisposes them to multiple primary tumors, including breast and lung cancer, as well as diverse types of soft-tissue sarcomas and carcinomas (17).

In the present study, we investigated the role of germ line H-ras and p53 mutations in skin tumorigenesis in vivo. We hypothesized that the compound transgenic mice carrying activated H-ras and dominant-negative p53 could greatly accelerate mouse skin tumorigenesis because the mutant p53 acts synergistically with mutant ras to cause malignant transformation of primary rat embryo cells (18). In addition, the altered gene expression associated with H-ras and p53 genes was evaluated in normal skin and skin SCC following carcinogen exposure.

Results

To examine the interactions between p53 and ras genes in the development of skin SCCs, we crossed p53 transgenic mice with v-H-ras transgenic TG·AC mice. All F₁ mice were treated with a topical application of benzo(a)pyrene (BaP) alone twice weekly (100 nmol/mouse) for 5 months. As shown in Fig. 1A, treatment of (TG·AC × p53^Val135/wt) F₁ wild-type (p53^wt/wtHras^wt/wt) mice with BaP produced a 26% skin tumor incidence, whereas treatment of p53^wt/wtHras^TG·AC/wt, p53^Val135/wtHras^wt/wt, and p53^Val135/wtHras^TG·AC/wt mice yielded tumor incidences of 75%, 77%, and 100%, respectively. Interestingly, p53 transgenic mice, TG·AC transgenic mice, and compound transgenic mice developed early-onset SCC. This was particularly striking for double transgenic mice, with the first carcinoma appearing after 8 weeks of treatment (Fig. 1A). Furthermore, an average of 0.33 tumor per mouse was observed in wild-type (p53^wt/wtHras^wt/wt) mice, whereas ∼1.54, 1.96, and 3.08 tumors per mouse were observed in p53^wt/wtHras^TG·AC/wt, p53^Val135/wtHras^wt/wt, and p53^Val135/wtHras^TG·AC/wt mice, respectively (Fig. 1B). There were even more significant effects on total tumor volume. We found ∼7-, 48-, and 588-fold increases in tumor volume in p53^wt/wtHras^TG·AC/wt, p53^Val135/wtHras^wt/wt, and p53^Val135/wtHras^TG·AC/wt mice, respectively, compared with wild-type (p53^wt/wtHras^wt/wt; Fig. 1B). Grossly (Fig. 2A-D), most carcinomas appeared as solid masses forming shallow encrusted ulcers or massive necrosis within the tumors. Some of them exhibited a cauliflower-like appearance. In some advanced cases in the p53^Val135/wtHras^TG·AC/wt, carcinomas exhibited an extremely rapid growth rate. For example, carcinomas in Fig. 2D developed to the sizes shown in 1 to 2 weeks without any apparent precursor lesion. The surfaces of these tumors were irregular and they had areas of necrosis and hemorrhage.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Skin tumor induction in p53^Val135/wtHras^TG·AC/wt mice by topical application of BaP. A. Cumulative percentage of mice with skin tumors in relation to time (weeks) after the onset of treatment. B. Effects of genotypes on skin tumor multiplicity and tumor load. Six- to 8-week-old females were used. There were 27 p53^wt/wtHras^wt/wt, 28 p53^wt/wtHras^TG·AC/wt, 26 p53^Val135/wtHras^wt/wt, and 26 p53^Val135/wtHras^TG·AC/wt mice. Bars, SD. *, P < 0.05, Student's t test.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Pathology of skin tumors found in p53^Val135/wtHras^TG·AC/wt mice. A-D. Gross photomicrographs of BaP-induced skin tumors from p53^wt/wtHras^wt/wt, p53^wt/wtHras^TG·AC/wt, p53^Val135/wtHras^wt/wt, and p53^Val135/wtHras^TG·AC/wt mice, respectively. E and G. Light photomicrographs of BaP-induced skin tumors in p53^wt/wtHras^wt/wt mouse at ×10 and ×40 magnifications, respectively. F and H. Light photomicrographs of BaP-induced skin tumors from p53^Val135/wtHras^TG·AC/wt mice at ×10 and ×40 magnifications, respectively.

A total of 9 tumors from p53^wt/wtHras^wt/wt wild-type mice, 43 tumors from p53^wt/wtHras^TG·AC/wt mice, 50 tumors from p53^Val135/wtHras^wt/wt mice, and 76 tumors from p53^Val135/wtHras^TG·AC/wt mice were examined pathologically. The majority of the skin tumors found in wild-type (p53^wt/wtHras^wt/wt) mice were papillomas or well-differentiated SCCs, consistent with numerous previous studies (6, 19). As shown in Fig. 2E and G, these tumors are characterized by well-differentiated cells and monomorphic growth pattern. In contrast, ∼60% of tumors from p53^wt/wtHras^TG·AC/wt mice, 70% of tumors from p53^Val135/wtHras^wt/wt mice, and 100% from p53^Val135/wtHras^TG·AC/wt mice are SCCs composed of relatively poorly differentiated or undifferentiated cells with extensive local invasion (Fig. 2F and H). These data indicate that p53 deficiency caused by a dominant-negative p53 transgene can synergize with activated H-ras to enhance both the formation (increased tumor number) and the progression (pathologic grade) of the resulting tumors.

Gene expression profiles associated with altered p53 and H-ras status in skin tumorigenesis were analyzed to determine the role of specific genes or pathways in skin tumorigenesis in these transgenic mice (Fig. 3A and B). Tables 1, 2, and 3 illustrate the representative alterations of gene expression in clusters associated with p53, H-ras, or p53/H-ras status in skin tumorigenesis (e.g., growth factor and growth factor receptors, cell cycle control and apoptosis, signaling pathways, and transcription regulators). Using microarray analysis, the mRNA expression of tumor necrosis factor receptor superfamily, member 10b (also called KILLER/DR5) was found to be significantly higher in skin from p53^wt/wtHras^wt/wt mouse skin compared with that from p53^Val135/wtHras^wt/wt mice after exposure to BaP and compared with those tumors found in mice with p53 mutation (Tables 1 and 3). We used quantitative reverse transcription-PCR to confirm the differential expression of KILLER/DR5 found in mouse skin tissues after treatment with BaP (Fig. 3C, a). We also determined whether expression of v-H-ras would affect extracellular signal-regulated kinase (ERK) activity in the skin of p53^wt/wtHras^wt/wt wild-type versus p53^wt/wtHras^TG·AC/wt mice following treatment with BaP. ERK activity was determined by immunoblotting with anti-phospho-ERK antibody. As shown in Fig. 3C, b, the skin of p53^wt/wtHras^TG·AC/wt mice has higher ERK activity than that of p53^wt/wtHras^wt/wt wild-type following carcinogen treatment. Consistent with microarray data, these results provide a possible mechanism(s) for the enhancing effect of the v-H-ras transgene on skin tumorigenesis.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Differentially expressed genes in normal skin tissues and SCCs among p53^wt/wtHras^wt/wt, p53^Val135/wtHras^wt/wt, p53^wt/wtHras^TG·AC/wt, and p53^Val135/wtHras^TG·AC/wt mice after BaP treatment (P < 0.05, Student's t test). A. Differentially expressed genes in BaP-treated normal skins. B. Differentially expressed genes in BaP-induced skin SCC. Black, the value is near the mean; red, the value is above the mean; and green, the value is below the mean. C. (a) Quantitative reverse transcription-PCR analysis of KILLER/DR5 in BaP-induced (TG·AC× p53^Val135/wt) F₁ mouse skins. Lines 1 and 2, BaP-treated normal skin from p53^wt/wtHras^wt/wt mice; lines 3 and 4, BaP-treated normal skin from p53^Val135/wtHras^wt/wt mice; lines 5 and 6, BaP-treated normal skin from p53^wt/wtHras^TG·AC/wt mice; lines 7 and 8, BaP-treated normal skin from p53^Val135/wtHras^TG·AC/wt mice. (b) Detection of ERK activity in BaP-induced p53^wt/wtHras^wt/wt and p53^wt/wtHras^TG·AC/wt mouse skins. Western blot was probed with anti-phospho-ERK (top) and anti-ERK (bottom) antibodies. Lines 1 and 2, BaP-treated normal skin from p53^wt/wtHras^wt/wt mice; lines 3 and 4, BaP-treated normal skin from p53^wt/wtHras^TG·AC/wt mice.

To further examine the interaction between dominant-negative p53 and activated H-ras during tumorigenesis, we compared the expression profiles of normal skin and tumors from animals with different p53 and/or H-ras genotypes using the GenMAPP. The altered genes were found to be involved in apoptosis, cell cycle, G-protein, and transforming growth factor-β pathways. Figure 4A and B represent gene expression alteration associated with apoptosis and cell cycle. H-ras and p53 double transgenic mice had significantly more and unique genes that were differentially expressed. For example, several key genes, such as Apaf1, BID, and PARP, were altered only in double transgenic mice (Fig. 4A). Similarly, key cell cycle genes, such as Ywhag, Bub1, Mps1, cyclin D1, cyclin E1, cyclin A1, cyclin-dependent kinase 4, Cdc61, and p107, were also only altered in p53^Val135/wtHras^TG·AC/wt mice (Fig. 4B). Thus, GenMAPP revealed that the concurrent expression of dominant-negative p53 and activated H-ras genes seem to influence the expression of a different set of genes involved in signal transduction pathways in skin tumorigenesis (Fig. 4A and B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

A. GenMAPP apoptosis. B. Cell cycle pathways integrate the expression data (cutoff: P < 0.05, fold change > 1.5). Yellow, genes changed in tumors from p53^wt/wtHras^TG·AC/wt mice; green, genes changed in those from p53^Val135/wtHras^wt/wt mice; red, genes changed in those from p53^Val135/wtHras^TG·AC/wt mice; gray, the selection criteria were not met, but the gene is represented on the array; white, the gene was not present on the chip.

Discussion

Mutations in the p53 and ras genes are the most frequent molecular events in human cancer (20, 21). Germ line mutations in p53 are associated with the Li-Fraumeni family cancer syndrome, which predisposes individuals to multiple, unpredictable, aggressive, and often lethal tumors (22-25). Our previous studies showed that p53 transgenic mice with a germ line missense mutation (Ala¹³⁵Val) were more susceptible than wild-type mice to lung adenocarcinoma, uterine sarcomas, and colon carcinomas following treatment with several chemical carcinogens (26-29). However, the possibility of a synergistic interaction between H-ras and p53 in skin carcinogenesis was not explored previously. In this study, we have developed and systematically characterized a novel mouse model to study of skin tumor progression. We have shown clearly that mice with a dominant-negative p53 and a mutant H-ras are extremely susceptible to chemical induction of skin SCC. This unique mouse model can be not only used for further detailed mechanistic studies but also applied directly in prevention or therapeutic studies against in situ tumors with mutations in both p53 and ras. Thus, the results from this study are highly significant.

TG·AC mice carry the coding sequence of v-H-ras linked to a ζ-globin promoter and a SV40 polyadenylation signal sequence (30). The transgene confers the property of genetically initiated skin (30). These mice have been used to screen for both a variety of chemical carcinogens and tumor promoters (13, 15). Consistent with previous studies (13, 15), skin tumor incidence and multiplicity were significantly increased in these mice when treated with BaP. Furthermore, we observed that high levels of ERK activity correlated with a high degree of susceptibility to skin tumor development in TG·AC mice. These results suggest a possible mechanism for promoting the effect of v-H-ras transgene on tumor development through increased ERK activity. Similarly, p53 transgenic mice exhibited an increased susceptibility to skin SCC in response to carcinogen treatment. This implies that the mutant p53 allele increases skin tumor susceptibility through a promotion function of the missense protein (26, 27). Introduction of a p53 transgene that expresses the mutant p53 protein will inactivate endogenous wild-type p53 (26, 27). The p53 protein helps to regulate genomic stability by preventing cell cycle entry and initiates apoptotic cell death in response to DNA damage (21), which is reflected in a decrease in tumor latency, as shown Fig. 1, and routinely an increase in tumor multiplicity. Following BaP treatment in the p53 mutant mice, we found that tumor latency was significantly decreased and the size of the individual tumors was strikingly increased. We presume that the increase in tumor multiplicity is associated with a decrease in repair of DNA damage. This parallels our prior findings in lung that mice with a p53 mutation showed an increase in tumor multiplicity following treatment with various carcinogens, including BaP (26). Thus, loss of normal p53 function leads to deregulation of the cell cycle, allowing accumulation of genetic mutations and continued cell cycle progression after DNA damage (21). Altered p53 protein expression has been associated with increased proliferative rates in other tumors, such as human breast cancer (31). The imbalance produced by an enhanced proliferative activity and a decreased apoptotic rate may make tumor cells harboring p53 alterations more aggressive. The observation in this study is the profound effect of the mutant p53 transgene on development of skin tumors in p53^Val135/wt mice, strongly indicating that skin tumor formation was highly p53 dependent.

In terms of the effect of p53 or H-ras individually on tumorigenesis, our findings are consistent with some of the previously published studies (3, 18, 32). For example, our finding on the effect of p53 is consistent with the results from Halevy et al. (18) in which the mutant p53 gene was found to act synergistically with the mutant ras gene to cause malignant transformation of primary rat embryo cells. Pierceall et al. (3) reported that inactivation of the p53 tumor suppressor gene as well as activation of ras oncogenes may be involved in the pathogenesis of some human skin SCCs. Fisher et al. (32) reported that, in the absence of p53 gene, doxycycline-mediated induction of K-ras4b^G12D initiated lung tumorigenesis much more rapidly than in a wild-type background. Finally, employing the same mutation in p53 in a lung tumor model that invariably develops K-ras mutations, we have observed a synergy between these two genes (p53 and K-ras) in that tumors grew faster and were more likely to progress to adenocarcinomas in p53 mutant animals as contrasted with wild-type mice (26, 29). These results suggest that p53 plays a major role in tumorigenesis and can cooperate with mutant ras to induce tumor progression. These data indicate that concurrent expression of a dominant-negative p53 with an activated H-ras can accelerate skin tumor formation.

Perhaps the most interesting finding of this study is the observation of synergistic induction of invasive mouse skin carcinomas by mutations in both H-ras and p53. In this study, a compound transgenic mouse with both a dominant-negative p53 transgene and a mutant v-Ha-ras gene were created to determine the possible interaction between the two genes in promoting tumor progression. We not only observed growth-promoting synergies in these mice carrying both p53 and v-H-ras transgenes but also were surprised with the striking distinction between bitransgenics and the other littermates in terms of the rapidity and aggressiveness of tumor development. Both tumor multiplicity and volume were profoundly increased in bitransgenics compared with wild-type, p53 transgenic, or TG·AC mice. Microarray and GenMAPP together with reverse transcription-PCR analysis in skin tumors and normal tissues also suggest that germ line H-ras and p53 mutation seem to affect several cellular processes involved in the tumorigenesis, including apoptosis, cell cycle arrest, and Ras-mitogen-activated protein kinase (MAPK) pathway. Furthermore, we also found a significant number of genes whose expression was only altered in mice with both a dominant-negative p53 and an activated H-ras, which may provide a clue for the observed effect on tumor progression. For example, several key genes involved in apoptosis (TNFR1, FAS, TRAF1, IAP3, C-JUN, and NF-κBp105) and those involved in cell cycle regulation (SMAD4, ATM, Bub1, Mps1, cyclin D2, cyclin E1, cyclin A2, and p107) were altered only in double transgenic mice (Fig. 4A and B). However, the exact mechanism for the observed synergistic induction of tumor progression remains to be elucidated.

Materials and Methods

Reagents

BaP was purchased from Sigma Chemical Co. (St. Louis, MO). Acetone was purchased from Fisher Scientific (Fair Lawn, NJ). BaP was prepared immediately before use in bioassays.

Animals

p53 transgenic mice (p53^Val135/wt) were obtained from the National Institute of Environmental Health Sciences (Research Triangle Park, NC). TG·AC mice were purchased from Taconic (Germantown, NY). Animals were housed in plastic cages with hardwood bedding and dust covers in a HEPA-filtered, environmentally controlled room (24 ± 1°C, 12/12-hour light/dark cycle). Animals were given Rodent Lab Chow (Purina, St. Louis, MO) and water ad libitum. At the age of 6 weeks, the animals were paired to produce (TG·AC × UL53-3) F₁ mice.

H-ras and p53 Genotype

p53^Val135/wt mice were produced by microinjection of FVB/J mouse oocytes with a BALB/c mouse genomic clone of the p53 gene containing a point mutation in codon 135 (Ala¹³⁵Val) at exon 5. The mutation, a C-to-T transition, created a RFLP with a new HphI restriction enzyme cleavage site (recognition site: GGTGA). This mutation was used to genotype p53^Val135/wt mice using the PCR-RFLP method as described previously (26). The F₁ mice used in this study were genotyped by Southern blot analysis for identification of the H-ras genotype. The 250-bp ζ-globin promoter probe was amplified by the PCR method. The primer sequences were as follows: TG·AC-forward (5′-CCCTCAGTGCTAAGTGAGAGG-3′) and TG·AC-reverse (5′-ACACATGGTCAGGGACCTGTC-3′). The methods for TG·AC mouse genomic DNA isolation and Southern blot protocol were supplied by the manufacturer, with the exception of the DNA probe.

Skin Tumorigenesis Studies

Using (TG·AC× p53^Val135/wt) F₁ female mice, we did a standard skin carcinogenesis study. Seven-week-old (TG·AC× p53^Val135/wt) F₁ female mice were randomized into four groups according to the p53 and H-ras genotypes. All mice had their backs shaved weekly. Forty-eight hours before initial treatment, the dorsal skin of mice was shaved. All mice were treated topically twice weekly with BaP (100 nmol/mouse) dissolved in 200 μL acetone. Skin tumors were monitored and recorded twice weekly for the duration of the studies. This bioassay was terminated at 5 months after BaP exposure. For each mouse, the skin tumors were enumerated and three diameters were measured. Portions of the dorsal normal skin tissue and tumors were isolated and placed in individual tubes and immediately frozen in liquid nitrogen. The rest of the tumors were fixed in 10% neutral-buffered formalin overnight followed by 70% ethanol and paraffin embedding. Tissue sections (5 μm) were stained with H&E for histopathologic evaluation. Mean monthly body weights were compared by Student's t test. A gross necropsy was also done. Based on the tumor shape, the total tumor volumes of skin tumors were calculated by V = (4/3)πr³ or V = 2rπh (r is radius).

RNA Isolation and Quantitative Reverse Transcription-PCR

Total RNAs were isolated from mouse normal tissues using the TRIzol reagent (Life Technologies, Gaithersburg, MD). Total RNA (2 μg) was used to synthesize cDNA in a total reaction volume of 40 μL. After incubating the RNA in an adjusted amount of DEPC-treated water at 70°C for 10 minutes, the following components were added: 1.5 μg (dT), 8 μL of 5× reaction buffer [250 mmol/L Tris-HCl (pH 8.3), 375 mmol/L KCl, 15 mmol/L MgCl₂], 10 mmol/L DTT, 40 units RNasin, and 300 units Moloney murine leukemia virus reverse transcriptase (Life Technologies). The reaction mixture was incubated at 37°C for 1 hour followed by terminating the reaction at 95°C for 10 minutes. A 2-μL aliquot added into 25-μL total reaction mixture was used to perform a quantitative reverse transcription-PCR analysis. Sequences of PCR primers for p53 downstream target genes were listed previously (27): 25 amplified cycles at 94°C for 1 minute, 55°C for 1 minute, and 72°C for 1 minute in 0.2 mmol/L deoxynucleotide triphosphates, 1.5 mmol/L MgCl₂, 1 μmol/L downstream gene-specific primer, 1 μmol/L [γ-³²P]ATP end-labeled upstream gene-specific primer, and 0.025 unit Taq DNA polymerase (Promega, Madison, WI). A pair of primers specifically flanking a fragment of the glyceraldehyde-3-phosphate dehydrogenase gene was also coamplified as an internal control. The forward primer of glyceraldehyde-3-phosphate dehydrogenase was also end-labeled with [γ-³²P]ATP by T4 polynucleotide kinase. The PCR products were resolved in an 8% denaturing polyacrylamide gel and the results were quantitated by densitometry using a Shimadzu Dual-Wavelength TLC scanner CS-930 (Lenexa, KS) and by densitometry analysis using ImageQuant software.

Western Blot Analysis of ERK Activity

Mouse skin tissues were lysed directly in SDS sample treatment buffer. A total of 80 μg protein/lane was loaded on a 15% SDS-PAGE gel and analyzed by Western blotting with anti-ERK antibody or anti-phospho-ERK antibody (Sigma Chemical) as described previously (33).

RNA Amplification

In vitro transcription-based RNA amplification was done on each sample. cDNA for each sample was synthesized using a SuperScript cDNA Synthesis kit (Invitrogen, Carsbad, CA) and a T7-(dT)₂₄ primer: 5′-GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGG-(dT)₂₄-3′. The cDNA was cleaned using phase-lock gel (Fisher Scientific) phenol/chloroform extraction. Then, the biotin-labeled cRNA was transcribed in vitro from cDNA using a BioArray High-Yield RNA Transcript Labeling kit (Enzo Biochem, New York, NY) and purified, again using the RNeasy Mini kit.

Affymetrix GeneChip Probe Array and Clustering Analysis

The labeled cRNA was applied to the Affymetrix Mu74Av2 GeneChip (Affymetrix, Santa Clara, CA), which contains >12,000 genes and expressed sequence tags on one array according to the manufacturer's specifications. Every gene or expressed sequence tag is represented by a probe set consisting of ∼16 probe pairs (oligonucleotides) of 25-mer oligonucleotides. One sequence of a probe pair represents the complementary strand of the target sequence, whereas the other has a 1-bp mismatch at the central base pair position. This mismatch sequence serves as an internal control for specificity of hybridization. Four or five independent samples were collected for each group. Array normalization and gene expression estimates were obtained using Affymetrix Microarray Suite 5.0 software. The array mean intensities were scaled to 1,500. These estimates formed the basis for statistical testing. Differential expression was determined on the combined basis of statistical testing using Student's t test and based on ratio with a cutoff of P < 0.05 and fold change ≥ 2 being called positive for differential expression. For the selected genes, expression indices were transformed across samples to a N(0,1) distribution using a standard statistical Z-transform. These values were input to the GeneCluster program of Eisen et al. (34) and genes were clustered using average linkage and correlation dissimilarity.

GenMAPP

Signal transduction pathways, metabolic pathways, and other functional groupings of genes were evaluated for differential regulation using the visualization tool GenMAPP (University of California at San Francisco; http://www.genmapp.org). GenMAPP is a recently reported tool for visualizing expression data in the context of biological pathways (35). We imported the statistical results of our data set into the program and used GenMAPP to illustrate pathways containing differentially expressed genes. Differential gene expression was based on p53 genotype status (P < 0.05, fold change > 1.5).

Statistical Analysis

Student's t test was used to determine the difference in the number and size of skin tumors per mouse between transgenic mice and wild-type mice. Fisher exact test was used to determine the difference in the incidence of skin tumor development between transgenic and wild-type mice.