The p53 heterozygous knockout mouse as a model for chemical carcinogenesis in vascular tissue.

Heterozygous p53 knockout mice were investigated as a potential model for vascular tumor carcinogenesis. Groups of 20 male mice were exposed by gavage for 6 months to the vascular carcinogen urethane at 1, 10, or 100 mg/kg body weight/day. Wild-type and heterozygous p53 knockout control groups were exposed by gavage to the vehicle alone. Another group of 20 male mice received d-limonene by gavage (d-limonene is noncarcinogenic in mice). The high dose of urethane caused early mortality in the majority of mice associated with histopathologic evidence of toxicity and tumors, including a high incidence of benign and malignant vascular tumors, in all animals. At the intermediate dose, toxicity was less marked and 3 of 20 mice had tumors; mice that received the low dose did not have signs of toxicity or neoplasia. The two control groups had no tumors and the d-limonene group had one tumor of the prostate, which was considered spontaneous. We conclude that the p53 knockout mouse is a useful tool for investigating vascular tumorogenesis.

knockout control groups were exposed by gavage to the vehide alone. Another group of 20 male mice received etlimonene by gavage (dtlimonene is noncarcinogenic in mice). The high dose of urethane caused early mortality in the majority of mice associated with histopathologic evidence oftoxicity and tumors, including a high incidence of benign and malignant vascular tumors, in all animals. At the intermediate dose, toxicity was less marked and 3 of 20 mice had tumors; mice that received the low dose did not have signs of toxicity or neoplasia. The two control groups had no tumors and the ilimonene group had one tumor of the prostate, which was considered spontaneous. We conclude that the p53 knockout mouse is a useful tool for investigating vascular tumorogenesis. Key words: carcinogenicity models, p53 knockout mouse, vascular tumors. Environ Healt Persect 108:61-65 (2000). [ Online 14 December 1999] http://ehpnetl.niehs.nih.gov/docs/2000/108p61-65carmichaellabstracthtml The p53 knockout mouse is the subject of great interest for its potential as a tool to study chemical carcinogenesis. The heterozygote could be used in carcinogenicity testing strategies that would potentially reduce the number of full-scale lifetime studies in conventional mouse models (1,2). Because heterozygous p53 knockout mice should be more sensitive to carcinogens than normal mice, such rodent bioassays would use fewer animals and take less time. Furthermore, because heterozygous p53 knockout mice have a low incidence of tumors until 9-12 months of age, it should be easier to identify the treatment-related effects of chemicals in a mouse strain with a low background of spontaneous cancers (3,4). After 12 months of age, the rate of tumor development accelerates and by 18 months of age, 50% of the heterozygous mice have succumbed to tumors, mainly osteosarcomas, malignant lymphomas, and hemangiosarcomas. Therefore, 6 months is a suitable test duration; the expected yield of spontaneous tumors should be low.
Mice, as with other experimental species, have a strain-dependent profile of tumor types. Hepatocellular tumors are common in many strains of mice and are the subject of much discussion with respect to their relevance in carcinogen evaluation (5,6). Lung tumors and lymphoreticular tumors are also relatively common. Although vascular tumors are also found, they usually occur at a relatively low incidence (7,8). The most commonly found tumors in p53 homozygous and heterozygous knockout mice are sarcomas, malignant lymphomas, and hemangiosarcomas (4). Vascular tumors are increased in the mouse by relatively potent chemical carcinogens such as vinyl chloride (9). Because of the nature of vascular tissue, which does not lend itself easily to mechanistic studies, little is known about the susceptibility of this tissue.
To explore the potential of the p53 knockout mouse for studying vascular tumors we selected the genotoxic compound urethane, which produces these tumors in lifetime studies in mice. Urethane, which is genotoxic via the formation of a reactive electrophilic metabolite, vinyl carbamate epoxide (10), has well-documented carcinogenic activity in rodent models; studies date back more than 20 years (11). Urethane was previously used as an anesthetic, but is also a natural compound found in low concentrations in many fermentation products (12).
One of the assumptions with the knockout model is that a mutation at the intact p53 allele is necessary for development of the carcinogenic process. In principle, nongenotoxic compounds that induce tumors by other mechanisms should be negative in this system. Therefore, d-limonene, which is neither genotoxic nor carcinogenic in mice but is carcinogenic in the male rat by a nongenotoxic mechanism (13)(14)(15), was included as a negative control substance.
The criteria established for assessing the usefulness of this model were zero incidence of vascular tumors in untreated p53 knockout mice, untreated wild-type mice, and in d-limonene-treated p53 knockout mice (at 6 months); a high incidence of vascular tumors in p53 knockout mice that receive a toxic dose of urethane; and a dose-related decrease in vascular tumors in p53 knockout mice with lower doses of urethane.
The animals were checked daily for clinical signs, moribundity, and mortality.
Detailed physical examinations were performed weekly during the treatment period. Body weight and food consumption were measured weekly during the first 14 weeks and monthly thereafter. All animals were necropsied; adrenal gland, brain, heart, kidney, liver, spleen, testis, and thymus were weighed fresh at final sacrifice only. The tissues (adrenal glands, aorta, articular surface, bone, bone marrow, brain, cecum, colon, duodenum, epididymides, esophagus, eyes, gallbladder, Harderian glands, heart, ileum, jejunum, kidneys, larynx, liver, lung, mammary gland, mesenteric lymph nodes, ovaries, pancreas, Davidson's fixative (20). All of the standard protocol tissues (with the exception of the larynx) were embedded in paraffin wax, sectioned at 5 pm, and stained with hematoxylin and eosin for routine light microscopic histopathologic examinations.
During the study, the care and use of animals were in accordance with regulations of the Guide for the Care and Use of Laboratory Animals (21)

Wild-Type Mice
The body weight of the p53 knockout mice was higher than the body weight of wildtype animals from the end of the first week of the study ( Figure 1). This difference, which ranged between 3 and 8%, was statistically significant during most of the study blefor. time of..death, significant rous,finding at oecropey. and-factors that contribute."to the. death of animalsi eitheJound Timhe ofdet Trwtmentgreap~~Animal no.
(day on study) Major-gross findings and was related, in part, to di food consumption between the I p53 knockout versus wild-type n In this study, there were no organ weight or gross findings between the untreated p53 knm and their wild-type counterparts. examinations revealed midzona fatty change characterized by th tion of intracellular large vacuol the p53 knockouts as compared wild type. In the kidney, 200/c knockout mice showed minor changes (cortical basophilic inflammatory changes (peripelh clear cell infiltrates) as compare 5%, respectively, in the wild Focal epithelial hyperplasia of was noted in 20% of the p53 kn as compared to none in the Mononuclear cell infiltrate of was noted in 25% of the p53 kr as compared to 5% in the wild t plastic changes were noted in t: p53 knockouts or in the wild-typ Urethane Treatment: p53 B Mice Clinical observations, mortality, The highest dose of urethane mg/kg bw/day) was toxic fron month in the study, and it ha effects on body weight (Figure mg/kg bw/day, the mean body 5.5% lower than the vehicle cot from day 85 of the study; thi reached up to 15% at the end By the end of the 6-month exp 17 of the 20 mice at this dose group was internal hemorrhage related to sisting of hemangiomatous-like endothelial differences in and most likely secondary to vascular tumors hyperplasia were present in the liver of 2 of at necropsy (Table 1). Although a few sporadic deaths 20 mice and in the heart of a single mouse ockout mice were noted in other groups, these were con-from the 100 mg/kg/day urethane group.
Microscopic sidered unrelated to treatment and did not Neoplasms of the vasculature, benign J hepatocyte exceed 2 of 20 in any group. At necropsy, all hemangioma ( Figure 2), and/or malignant ie accumula-of the high-dose urethane p53 knockout hemangiosarcoma ( Figure 3) were noted in es in 25% of mice presented with masses or red colored the liver of a total of 18 (90%) mice treated to 0% in the spots on the liver. Dark fluid, presumably at 100 mg/kg/day urethane (  Table 2). Vascular sys-type were detected at microscopic examinan the fourth tem. We noted nonproliferative vascular tion of any of the organs examined. ,d significant changes in a high proportion of mice treated Liver. At 100 mg/kg/day urethane, we c 1). At 100 with the two higher dose levels of urethane. noted large areas of hepatocyte zonal necroy weight was Angiectasis, consisting of dilated vascular sis affecting essentially the centrilobular or ntrol animals spaces lined by endothelial cells and filled midzonal areas (functional zones 1 and 2) of is difference with erythrocytes, was present in the liver of the hepatic lobules in 70% of the heterozyof the study. 8 of 20 and 9 of 20 p53 knockout mice at gous p53 knockout treated mice. These osure period, 10 and 100 mg/kg/day, respectively. A single degenerative changes, which were located in had died or focus of angiectasis was also noted in the the liver parenchyma not affected by the presence of tumors, were sometimes associate for the number of animals affected by selected microscopically ed with signs of regeneration of hepatocytes such as marked cytomegaly or increased other tissues (liver, kidney, lung, spleen, heart, ng/kg/day. cDose level 10 mg/kg/day. dDose level 100 mg/kg/day. 0ose level 250 adrenal gland, sternum, subcutis, aorta, and ble for the eye in the 100 mg/kg/day group. 9All types of tumors. spinal cord). Six (30%) of the 100 mg/kg/day mice had generalized atrophy/involution of the thymus as did one mouse at 10 mg/kg/day. Cortical lymphocytolysis/apoptosis was noted in 5 (25%) of the mice that received 100 mg/kg/day. No changes were noted at 1 mg/kg/day. Lung. We found bronchioloalveolar adenomas ( Figure 5) in 5 mice (25%) treated at 100 mglkg/day. Interstitial mononuclear cell infiltrate of the pulmonary parenchyma was noted in 15 of 20 (75%) of the mice at the same dose level. No significant changes were found at 10 and 1 mg/kg/day. Subcutis. A large subcutaneous sarcoma was present in a single p53 knockout mouse treated at 10 mg/kg/day and in one treated at 100 mg/kg/day.
Testis. Twenty-five percent of the mice from the 10 and 100 mg/kg/day group had diffuse atrophy of the seminiferous tubules.
Eye. Twelve of 13 (92%) of the mice treated with urethane at 100 mg/kg/day had bilateral retinal atrophy consisting of the complete loss of the bacillary, outer nuclear, and outer plexiform layers of the eye. d-Limonene Treatment: p53 Knockout Mice Clinical observations, mortality, and necropsy. No significant changes were found in the dlimonene-treated p53 knockout mice as compared to the untreated control group.
Histopathology. We observed hyperplasia of the nonglandular stomach in 85% of the d-limonene-treated p53 knockout mice. With the exception of a single adenoma of the prostate, no neoplasms were observed in this dose group.

Discussion
The p53 knockout mouse model is used in fundamental studies of the mechanisms of carcinogenesis (3,(23)(24)(25)(26). To better understand the role of the p53 gene in this process, the model is also under validation for its usefulness as a more sensitive rodent model for identifying potential carcinogens (1,3).
Our purpose was to test whether heterozygous p53-deficient mice are a good model to detect vascular tumors caused by genotoxic carcinogens using urethane as a positive control. Urethane is a well-known carcinogen that forms a reactive electrophilic metabolite (10). Mirvish et al. (11) showed that urethane induced malignant lymphoma, hepatoma (depending on the strain), mammary carcinoma, hemangioma, and lung adenoma in mice. In a later study, Schmihl et al. (2X) reported lung adenoma, mammary tumors, and hemangioendothelioma after 2 years of treatment. Although these are old studies, the findings are consistent from one study to another. In the present study, our findings are similar to those of Mirvish et al. (11) and Schmihl et al. (27): lung adenoma, hemangioma and/or hemangiosarcoma, hepatocellular tumors and malignant lymphoma were observed in urethane-treated mice. We did not observe mammary tumors because only male animals were used. The incidence of vascular tumors in p53-deficient mice administered with 100 mg/kg/day urethane was high (90%). The other tumor types (malignant lymphoma and sarcoma), which are usually detected in aging p53 Figure 2. Benign hemangioma in the liver of a heterozygous p53 knockout mouse treated with urethane at 100 mg/kg for 180 days. The lesion is characterized by dilated, blood-filled spaces lined by a single layer of prominent uniform endothelial cells without atypia. They are differentiated from angiectasis, which consists of cystic dilated normal blood vessels with no evidence of endothelial proliferation. Bar= 100 pm. *!11111111~.k . A_   Volume 108, Number 1, January 2000 * Environmental Health Perspectives knockout mice, were also present in the 100 mg/kg/day urethane group. We also obtained a dose-related effect with urethane treatment; no tumor was detected at 1 mg/kg/day and there was only a low incidence of tumors at 10 mglkg/day. Inai et al. (19) used a range of urethane doses that can be compared to ours for the purposes of making conclusions about the sensitivity of the model. In their study, dose levels equivalent to 1, 10, and 100 mg/kg bw/day were administered in drinking water to B6C3F1 mice for 70 weeks. Their study also found early death in the highest dose (600 ppm; approximately 100 mg/kg/day). In addition, the authors considered the early mortality to be attributable to the rupture of the vascular tumors, which occurred in approximately 80% of the high dose mice. Inai et al. (19) showed a few vascular tumors at 10 mg/kg/day; at lower doses the incidence was comparable to background. Their study showed a lower threshold for detecting lung tumors (significant increases were seen at 10 mg/kg/day), whereas we had none at this dose. However, Inai et al. (19) had an 18.4% incidence of this tumor in controls. Considering the differences in experimental design, the coherence of the results between these two studies appears remarkably good.
No tumors were seen in untreated heterozygous p53 knockout mice or in the wild-type mice. d-Limonene, which is not genotoxic or carcinogenic in mice, did not induce any treatment-related tumors in heterozygous p53 knockout mice. The single adenoma of the prostate was considered incidental. d-Limonene caused signs of chronic irritation of the nonglandular stomach that were associated with hyperplasia but not neoplasia. Field and Roe (28) reported similar lesions, associated with the irritant effect of the compound.
Thus, the p53 model proved efficient for the induction of vascular tumors. It was possible to induce a high incidence of hemangioma and/or hemangiosarcoma in the heterozygous p53 knockout mice treated with urethane in a short period of time as compared to a classical 2-year bioassay for carcinogenicity studies. The appearance of the tumors was dose related with neither neoplastic nor nonneoplastic proliferative findings at the low dose. There were no vascular tumors in the untreated animals or in animals treated with d-limonene. Therefore, we consider that our three criteria for the usefulness of the model were met, namely: zero incidence of vascular tumors in untreated p53 knockout mice, untreated wild-type mice, and in d-limonene-treated p53 knockout mice (at 6 months); a high incidence of vascular tumors in p53 knockout mice receiving toxic doses of urethane; and a doserelated decrease in vascular tumors in p53 knockout mice with lower doses of urethane.
In conclusion, the heterozygous p53 knockout mouse seems to be a good model for identifying vascular tumors. The essentially nonexistent level of vascular tumors in control p53 knockout mice might make this system particularly suitable for studying low potency compounds suspected of being vascular carcinogens.