Studies Evaluating the Utility of N -Methyl- N -Nitrosourea as a Positive Control in Carcinogenicity Studies in the p53 + /– Mouse

Abstract

Studies conducted under the auspices of International Life Sciences Institute (ILSI) have suggested that an alternative mouse carcinogenicity study may be substituted for the traditional 2-year mouse bioassay typically conducted to support the development of drug candidates. The purpose of this study was to characterize the carcinogenic potential of N-methyl-N-nitrosourea (MNU), a DNA alkylating agent, in p53^{+ /–} knockout mice to determine its suitability as a positive control agent in an alternative carcinogenicity model. p53^{+ /–} knockout mice were administered a single oral dose of 90 mg/kg and maintained for up to 13 weeks prior to evaluation of neoplasms. Treatment was generally well tolerated; however, 4 of 30 mice died between the days of 75 and 92 due to neoplasms. MNU-related macroscopic observations included enlargement of the thymus, spleen, mandibular and mesenteric lymph nodes; and pale liver, heart, kidney, and bone marrow, which correlated with the diagnosis of lymphoma of the hematopoietic system, noted in the thymus of all affected animals and in the spleen, liver, lungs, and kidneys of some animals. Other treatment-related single neoplasms included a squamous-cell carcinoma in the nonglandular stomach and leiomyosarcoma in the glandular stomach. Non-neoplastic proliferative lesions included acanthosis and hyperkeratosis in the nonglandular stomach, focal papillary hyperplasia of the nonglandular stomach, glandular hyperplasia of the stomach, and adenomatous hyperplasia of the duodenum or ileum. The increased incidence of neoplastic and proliferative changes in MNU-treated mice suggests MNU could serve as a positive control in alternative carcinogenicity studies conducted in p53^{+ /–} knockout mice.

The 2-year bioassays in rodents evaluate carcinogenic risk associated with exposure to chemicals and drugs. These bioassays have been used successfully by the National Toxicology Program (NTP) and others to create a large database of carcinogen testing information and have gained a high degree of credibility with regulatory agencies worldwide (National Toxicology Program 1979). These conventional assays are typically carried out using B6C3F₁ or CD-1 mice and F344 or Wistar rats. In the past several years, transgenic mice that are genetically predisposed to cancer have become available, and efforts have been made to exploit these strains in alternative carcinogenicity studies (Tennant et al. 1999). In 1996, the International Conference on Harmonization (ICH) recognized that the conventional 2-year rodent bioassays had little or no relevance in predicting human cancer risk assessment based on the lack of correlation among positive findings in these bioassays to human cancer risk (MacDonald et al. 2004). In addition, this group acknowledged the utility of alternative carcinogenicity models for use in human cancer risk assessment by proposing that one of the 2-year rodent bioassays be complemented with a short-term (i.e., 6-month) study in one of the recently developed alternative mouse carcinogenicity models. One major incentive for use of these models in carcinogenicity studies is the shorter time to tumor formation in transgenic mice, typically 6 months for p53^{+ / –}, rasH2, and Tg.AC mice, and 9 months for XPA^{– / –} mice, and 1 year in the neonatal mouse. Compared to that of conventional 2-year bioassays, this shortened testing period has obvious advantages in reduced resources, earlier results, and a negligible background rate for spontaneous tumors in these strains. However, a comprehensive set of data demonstrating the predictability of these animal models for carcinogenicity is not available; hence, there is some hesitancy in going forth with the use of these models to characterize carcinogenic risk of new chemical entities. To further understand the predictability of these animal models, the International Life Sciences Institute (ILSI) Health and Environmental Science Institute (HESI) coordinated a large-scale collaborative research program involving worldwide industrial, government, and academic laboratories to evaluate several models proposed for use in carcinogenicity assessment (Robinson and MacDonald 2001). The ILSI HESI results confirmed that a few of the alternative carcinogenicity models had better concordance with human cancer incidence than the lifetime rodent bioassays (Alden, Smith, and Morton 2002). Specifically, the p53^{+ / –}transgenic mouse model could identify genotoxic carcinogens in a more accelerated manner than the standard 2-year rodent bioasays.

The p53 tumor suppressor gene is critical for cell cycle control and DNA repair. It acts to suppress DNA damage by inhibiting progression through the cell cycle and initiating apoptosis in response to damage (Tennant, French, and Spalding 1995; Tennant, Spalding, and French 1996). This gene is often found to be mutated or absent in human and rodent tumors (Donehower et al. 1992). The p53^{+ / –} knockout mouse was genetically engineered to have one allele of the p53 tumor suppressor gene deleted to mimic some heritable forms of cancer in humans. These mice are viable and show a low background tumor incidence (Donehower et al. 1992). These animals are susceptible to genotoxic carcinogens and exhibit accelerated tumorigenesis (French, Storer, and Donehower 2001). As of 2002, the p53^{+ / –}transgenic mouse is the predominant model of choice for carcinogenicity study protocols received by the Food and Drug Administration (FDA) that have contained an alternative animal model (MacDonald et al. 2004). This model is generally accepted for use in evaluating the carcinogenic potential of genotoxic agents (MacDonald et al. 2004). Although the NTP and ILSI HESI have demonstrated a strong concordance between positive responses for mutagenic carcinogens in the alternative models for carcinogenicity (Storer et al. 2001; Tennant, French, and Spalding 1995; Tennant, Spalding, and French 1996); there is still some uncertainty regarding the limited amount of data available in transgenic models and their replacement of the 2-year rodent bioassays (Blain 2003). To further evaluate the merit of some of the alternative carcinogenic models as a mode of evaluating carcinogenic potential, Pritchard et al. (2003) conducted a comprehensive analysis of the data from some of these models and compared to the traditional 2-year rodent bioassay. The authors note that “overall the alternative models performed well, but important issues of validation and standardization need further attention to permit their regulatory acceptance and use in human risk assessment.” Thus, until larger toxicological databases are available in these alternative models, the inclusion of a positive control in studies evaluating the carcinogenic potential of a novel agent should be considered. Studies were undertaken to determine the utility of N-methyl-N-nitrosourea (MNU) as a positive control in carcinogenicity studies in the p53^{+ / –} knockout mouse. Based on the increased incidence of neoplasms and other proliferative lesions in multiple sites in MNU-treated mice, the data suggest that a single dose of MNU could serve as a positive control in alternative carcinogenicity studies conducted in p53^{+ / –} knockout mice.

MATERIALS AND METHODS

Animals

Male and female (n = 15/sex/group) heterozygous p53^{+ / –}knockout mice (C57BL/6TacfBR-[KO]p53 N5), approximately 8 to 9 weeks old and weighing approximately 20 g, were obtained from Taconic (Germantown, NY, USA). Mice were housed in suspended, stainlesssteel, wire-bottom cages. Environmental controls for the animal rooms were set to maintain a temperature of 64°C to 79°C, a relative humidity of 30% to 70%, and a 12-h light/12-h dark cycle. Certified Purina 5002 pelleted diet (PMI Feeds, Richmond, IN, USA) and municipal tap water treated by reverse osmosis were supplied ad libitum throughout the duration of the study. Animals were randomized to treatment groups by random number generation. To reduce the number of animals used on the study, wild type controls were not included in the study design.

Treatments and Data Collection

MNU was obtained from Sigma (St. Louis, MO, USA) and prepared in 0.04 M citrate-buffered saline (pH 4.5) at 9 mg/ml on the day of use. Mice were administered a single dose of 90 mg/kg of MNU (10 ml/kg) in citrate-buffered saline (pH 4.5) by oral gavage (specified as day 1) within 2.25 h after the preparation of the formulation. Stability analysis was not conducted; however, given the robust response observed, one can presume a sufficient amount of intact MNU was present in the solution. Heterozygous p53^{+ / –} knockout control mice received a single oral dose of citrate-buffered saline (pH 4.5) only. In-life data collected for 13 weeks included clinical observations, body weight, food consumption, and mass palpation. Mice were euthanized under sodium pentobarbital by exsanguination via abdominal vasculature. At termination (days 92 and 93), observations included clinical pathology, macroscopic and microscopic observations, organ weights, and clinical pathology parameters.

Hematology parameters evaluated included hemoglobin concentration, hematocrit, red and white blood cells, mean cell volume, mean cell hemoglobin, mean cell hemoglobin concentration, platelets, red cell distribution width, reticulocytes (absolute and relative), total leukocyte number, and differential leukocyte (absolute and relative). Clinical chemistry measurements included aspartate aminotransferase (AST), glucose, total protein, globulin (calculated), creatinine, alanine aminotransferase (ALT), total bilirubin, potassium, hemolysis, blood urea nitrogen, sodium, albumin/globulin ratio (calculated), albumin, chloride, calcium, and 5′-nucleotidase. At termination, the following organs were weighed: brain, heart, liver, kidneys, testes, and ovaries (paired organs weighed together).

Histopathological Evaluation

Tissues collected for histopathological evaluation were fixed in 10% phosphate-buffered formalin, except for the eyes and optic nerves and the testes and epididymides, which were fixed in Bouin’s solution. Tissues were subsequently embedded in paraffin, sectioned at 5 μm, and stained with hematoxylin and eosin for histopathological examination. The following tissues were examined: adrenal glands, aorta, brain (forebrain, midbrain, and hindbrain), cecum, colon, duodenum, epididymides, esophagus, eyes and optic nerves, femur/joint/bone marrow, harderian glands, heart, ileum, jejunum, gallbladder, kidneys, liver, lungs with bronchi, lymph nodes (mesenteric and mandibular), nasal cavity, ovaries, pancreas, peripheral nerve, pituitary, prostate, rectum, salivary glands, seminal vesicles, skeletal muscle, skin with mammary glands, spinal cord (cervical, thoracic, and lumbar), spleen, sternum/bone marrow, stomach (glandular and nonglandular), testes, thymus (thymic area), thyroid and parathyroid glands, tongue, trachea, tumors/masses, urinary bladder, uterus with cervix, and vagina. The animal identification site and head were collected but not examined. In addition, gross lesions were similarly collected, processed, and examined by light microscopy.

Statistical Methods

Data are expressed as mean ± standard deviation of the mean.

RESULTS AND DISCUSSION

Rationale for Dose Administered and Route of Exposure

Oral administration was chosen for these studies because it was considered to the most appropriate route of administration for comparison to studies evaluating novel agents that would have an oral route for exposure. The dose selected was based on previous studies where MNU was given by the intraperitoneal route. Administration of a single intraperitoneal dose of 75 mg MNU to (Tg)rasH2/CB6F1^{+ / –} mice resulted in an increased incidence of skin and forstomach hyperplasia and papillomas, forestomach squamous cell carcinoma, lymphoma, granulocytic leukemia, angiosacoma, and duodenum adenomas in male and female mice 14 weeks after administration (Yamamoto et al. 1996). The administration of 90 mg MNU by oral gavage produced approximately equivalent DNA damage in the liver of mice to that produced by a single intraperitoneal dose of MNU (ILSI HESI Study 1997). DNA alkylation by MNU is considered to be a critical factor in NMU-induced tumorigenesis (Frei et al. 1978). Therefore, an oral dose of 90 mg was considered to be comparable to an intraperitoneal dose of approximately 75 mg, which is similar to doses used in multiple studies evaluating the effects of MNU.

In-Life Findings

MNU treatment was generally well tolerated; however, 4 (one male and three females) of 30 mice died between days 75 and 92 due to neoplasms. Specifically, the cause of death for the male mouse was a leiomyosarcoma noted in the nonglandular stomach; the 3 female mice were diagnosed as having lymphoma in multiple tissues, including the liver, heart, kidney, lung, heart, and bone marrow.

Treatment-related clinical observations consisted of decreased activity, alterations in breathing patterns (labored, rapid, or noisy), dehydration, hunched posture, and coolness to the touch (data not shown). Dehydration occurred in one male mouse on days 6 through 9 only and in a second mouse beginning on day 84. Other clinical signs were evident beginning on day 63 or later. Total numbers of animals exhibiting clinical signs at any time were 3 of 15 males and 4 of 15 females. There were no palpable masses noted by clinical observation.

Body weight data are shown in Table 1. For the mice treated with MNU, both male and female mice had lost weight in the first week of the study compared to the first day of dosing. Both control and MNU-treated heterozygous p53^{+ / –} knockout mice generally gained weight thereafter throughout the study. However, the week 1 weight loss effect persisted in the MNU-treated males, such that there was a slight decrease in body weights at the end of the study for males (11.5%); terminal female body weights were similar to the controls. In contrast, no treatment-related effects on food consumption in males or females were noted (data not shown).

Clinical Laboratory Measurements

Hematology data are summarized in Table 2. Mean white blood cell counts were increased for males treated with MNU. Mean absolute counts of all the leukocytes (granulocytes, monocytes, and lymphocytes) in MNU-treated males were increased compared to the controls. These high treatment group means, however, were attributed primarily to the high values exhibited by a single animal. This animal had systemic lymphoma, and the high white blood cell count (approximately 3.4-fold above the mean value of the control animals) was attributed to a neoplastic condition. In treated males, the red blood cell parameters were not remarkably different from control male values. In MNU-treated females, the total white blood cell counts were not considerably different from control female values. Mean absolute neutrophil counts were slightly higher (approximately 29%) and mean absolute lymphocyte counts were slightly lower (approximately 39%) than control values, but these differences were considered to reflect some individual biological variability and associated effects of lymphoma in some of the MNU-treated animals. Mean hematocrit, hemoglobin, and red blood cell counts were slightly reduced in treated females approximately 13%, 14%, and 14%, respectively. These red blood cell changes also were believed to be associated with lymphoma in the treated females. All other parameters were unaffected.

Clinical chemistry data are summarized in Table 3. Serum ALT and AST levels were increased in males and females to levels between 68% to 237% with no correlate to increased liver weight. These higher serum enzyme levels were attributed primarily to high values of several individual animals and could not be explained by morphologic appearance of hepatocyte degeneration; however, all these animals were diagnosed with lymphoma that included hepatic involvement.

Organ Weights

There were no treatment-related changes in organ weights for the brain, heart, liver, kidney, testes, or ovary in either males or females (data not shown). Enlargement of the thymus, spleen, and mandibular and mesenteric lymph nodes was noted on macroscopic examination but these tissues were not weighed.

Macroscopic Pathology and Proliferative Lesions

The historical incidence of background tumors in p53^{+ / –}knockout mice is relatively low, the highest being subcutaneous sarcomas (1% males; 3.37% females), lymphomas (1.7% males; 2.9% females), and osteosarcomas (0.5% males; 0.7% females) (Storer et al. 2001). In addition, tumors in multiple tissues, including the lung, thymus, forestomach, skin, bone marrow, and spleen, were noted 26 weeks after treatment of rasH2 mice with a single dose of MNU at 75 mg/kg (Adachi et al. 2002).

Treatment-related macroscopic observations were noted at necropsy in males and females (Table 4). Thymic masses and/or thymic enlargement were seen in 6 of 15 and 8 of 15 MNU-treated males and females, respectively. This finding was accompanied by splenic enlargement in four males and in one female, and one treated male had enlarged mandibular and mesenteric lymph nodes. These thymus, spleen, and lymph node lesions were diagnosed as lymphoma. Lymphomatous cells surrounded the thymus and heart in one treated male. The pallor noted macroscopically in the liver, heart, kidney, and bone marrow was due to a lymphomatous cell infiltration. There was no microscopic correlate for the pale appearance of the brain, lung, or pancreas noted in some of the mice.

MNU-treated mice had an increase in the incidence of neoplasms and nonneoplastic proliferative lesions compared to those treated with the vehicle control material (Table 5). Lymphoma of the hematopoietic system was diagnosed in 10 of 15 males and 14 of 15 females. The lymphoma involved the thymus in all affected animals and was limited to the thymus in three treated males and seven treated females. In other affected animals lymphoma was multicentric, most commonly involving the spleen, liver, lung, and kidney. The increased incidence in lymphoma in MNU treated mice (67% in males and 93% in females) is consistent with a previous study in p53 heterozygotes and wild-type mice where a dose of 50 mg/kg caused lymphoma in 68% of the heterozygotes and 30% of the wild-type mice (Reese, Allay, and Gerson 2001). The background incidence for lymphoma in p53 heterozygous mice is approximately 2% (Mahler et al. 1998; Storer et al. 2001). Similarly, MNU has been shown to increase the incidence of lymphoma in other transgenic mouse models, including LMO1 transgenic mice (Allay et al. 1997) and CD2–cyclin E transgenic mice (Karsunky et al. 1999). The response in the LMO1 transgenic mice was comparable to those obtained in this study (91%) 1 year after a single dose but was more robust than in the CD2–cyclin E transgenic mice where 54% of the transgenic mice had lymphoma 220 days after dosing. Differences in susceptibility to MNU-induced tumors is not limited to the transgenic models. Strain-related differences in susceptibility to MNU-induced effects have been demonstrated using multiple strains of mice (Angel, Morizot, and Richie 1993; Harris and Lawley 1994), which has been hypothesized to be due to various basal genetic alterations (Richie, Angel, and Rinchik 1996; Harris and Lawley 1994). Simple modifications of study design, such as time of administration during the day, can significantly influence the response to a simple chemical such as MNU (Beland et al. 1988). These variations in the response reflect the complexity of the carcinogenic process. Other treatment-related neoplasms included a squamous-cell carcinoma in the nonglandular stomach of a male and a leiomyosarcoma in the glandular stomach of a male. Multiple neoplasms were not diagnosed in any animals; however, several nonneoplastic proliferative lesions were seen in animals with neoplasms.

Several possible precancerous non-neoplastic proliferative lesions were seen in the gastrointestinal tract of mice given MNU. Acanthosis and hyperkeratosis were present in the nonglandular stomach of all treated males and 12 of 15 treated females, and focal papillary hyperplasia of the nonglandular stomach was diagnosed in two treated females. The occurrence of a squamous-cell carcinoma and two focal papillary hyperplasias suggest the acanthosis/hyperkeratosis may be a precancerous effect of MNU administration. Glandular hyperplasia of the stomach occurred in five treated males and four treated females and adenomatous hyperplasia of the duodenum or ileum occurred in three treated males and seven treated females. Spindle-cell hyperplasia, also known as subcapsular hyperplasia of the adrenal cortex, is a common background finding in mice (Goodman 1996). The incidence of this finding in both treated males and females was slightly increased compared with control values. The incidence of reticular hyperplasia with small lymphocyte depletion of the mesenteric lymph node was also increased in treated animals compared with control values. This lymph node change, however, is more consistent with an immunologic response rather than a primary cellular proliferative change.

Several other non-neoplastic findings merit mention as occurring more commonly in mice administered MNU compared with control animals. Retinal atrophy occurred in 8 of 15 treated males and 6 of 15 treated females, but was not diagnosed in any control animals’ eyes. This finding was characterized by atrophy or loss of the outer sensory and granular layers of the retina. Increased extramedullary hematopoiesis (EMH) of the spleen was diagnosed in 9 of 15 treated females and 1 of 15 treated males. An increase in EMH occurs when there is a strong systemic demand for hematopoietic cell responses, such as infection, neoplasia, or loss of normal bone marrow. Lymphoma in the bone marrow probably accounted for the increased splenic EMH in several of the affected females. The affected male with increased EMH also had a gastric leiomyosarcoma.

Treated males exhibited a higher incidence of seminiferous tubule atrophy and degeneration in the testes (9 of 15 treated animals compared with 5 of 15 controls). This testicular finding was of minimal severity in all treated and control males, and is generally considered a normal background finding when present at this level. The slight increased incidence in treated males probably reflects normal biological variability and is considered to be incidental to MNU administration.

Hemosiderin accumulation in the heart valves was slightly increased in treated males and females, but this was considered incidental to treatment with MNU. This is a common background finding in mice, and its diagnosis is often dependent upon a fortuitous tissue section containing significant valvular tissue. It is generally considered a nonadverse and incidental microscopic finding.

Lymphohistocytic infiltrates were increased in the nonglandular stomach of treated males and females. This finding is interpreted to be associated with the acanthosis/hyperkeratosis of the nonglandular stomach seen in the same animals.

MNU has been used as a positive control in carcinogenic studies using alternative carcinogenicity models, including the rasH2 mouse (Urano et al. 2001). A single intraperitoneal dose of MNU at 75 mg/kg to mice induced malignant lymphomas of the thymus and/or bone marrow, papillomas of the forestomach and/or skin, and adenomas of the lungs. The high incidence of MNU-induced neoplasms and proliferative lesions in the current 13 week oral gavage study in p53^{+ / –} knockout mice is consistent with the findings observed previously in the rasH2 mouse (Urano et al. 2001), corroborating the proposition that MNU may be a useful positive control in studies conducted in alternative carcinogenicity models to evaluate the carcinogenic potential of genotoxic agents.

CONCLUSION

A single dose of MNU was generally well tolerated and resulted in an increased incidence of neoplasms and other proliferative lesions after 13 weeks after treatment. The increased incidence of neoplasms and other proliferative changes in MNU-treated p53^{+ / –} knockout mice was consistent with the findings observed previously in the rasH2 mouse (Urano et al. 2001), suggesting that MNU may be a useful positive control in alternative carcinogenicity models to evaluate the carcinogenic potential of genotoxic agents.