Evaluation of the Carcinogenic Potential of Clofibrate in the rasH2 Mouse

Abstract

The purpose of the study was to support of the International Life Sciences Institute (ILSI) alternative carcinogenicity models initiative to evaluate the carcinogenic potential of the nongenotoxic carcinogen, clofibrate, a peroxisome proliferator-activated receptor (PPAR) α agonist, following oral administration to rasH2 mice. Peroxisome proliferators are one of the most widely studied of the nongenotoxic carcinogens and have diverse industrial and therapeutic uses (Gonzalez et al. J. Nat. Cancer Inst. 90: 1702–1709, 1998); however, the nongenotoxic mechanism of carcinogenicity is currently unknown. Male mice were administered doses of clofibrate at 50, 100, or 200 mg/kg/day and female mice were administered doses of 50, 150, or 250 mg/kg/day by oral gavage at 10 ml/kg for 27 weeks. In addition, rasH2 male and female mice were treated with N-nitroso-N-methylurea (NMU). Nontransgenic male and female mice were treated with 200 and 250 mg/kg/day, respectively, of clofibrate. The NMU-treated mice were given a single intraperitoneal dose of 75 mg/kg, which was followed by a 90-day observation period; all others were sacrificed after 6 months of daily dosing. Hepatocellular neoplasms were observed in clofibrate-treated rasH2 male mice after 6 months of treatment but not in nontransgenic males or females. Clofibrate treatment (250 mg/kg/day) of female rasH2 mice was associated with a slight increase in the incidence of various neoplasms (harderian gland, lungs, skin, spleen, tail, thymus, and uterus) compared with untreated transgenic mice and with similarly treated nontransgenic mice. Non-neoplastic changes were found in the liver of transgenic and nontransgenic mice of both sexes and in the kidneys of male mice. NMU produced findings are consistent with previous studies. The data suggest that the rasH2 mice are a good model for testing epigenetic carcinogens in a shorter timeframe than conventional mouse carcinogenicity bioassays.

Alternative carcinogenicity models have been proposed for use in human cancer risk assessment as a replacement for one of the rodent 2-year bioassays (Robinson and MacDonald 2001). These models include the rasH2 transgenic mouse model, the Tg.AC transgenic mouse model, the p53^{+ / –} knockout mouse model, the XPA^{– / –} knockout mouse model, the XPA^{– / –}/p53^{+ / –} double-knockout mouse model, and a neonatal mouse model (Robinson and MacDonald 2001). The p53^{+ / –}knockout mouse and the rasH2 transgenic mouse were two of the first genetically modified models developed for potential alternative carcinogen bioassays (Dass et al. 1999; Spalding et al. 1999; Tennant, French, and Spalding 1995). One major incentive for use of these models in carcinogenicity studies is the decreased time to onset of tumor induction in transgenic mice, typically 6 months for p53^{+ / –}, rasH2, and Tg.AC mice, 9 months for XPA^{– / –} mice, and 1 year in neonatal mice. Compared to that of conventional 2-year bioassays, this model has obvious advantages in reduced labor costs, as well as increased sensitivity due to negligible background rate for spontaneous tumors in these strains.

The rasH2 mouse is a hemizygous transgenic alternative carcinogenicity model carrying the human c-Ha-ras gene (5 or 6 copies) within the promoter region which encodes the prototype c-Ha ras gene product (i.e., p21) (Tamaoki 2001; Yamamoto et al. 1997; Yamamoto, Urano, and Nomura 1998; Tennant et al. 2001). The p21 protein is a key member of the signal transduction pathway for cell proliferation, differentiation, and death. The GTP-bound active form of p21 modulates cell proliferation through intracellular gene expression of mitogen-activated protein (MAP) kinase and c-fos (Tamaoki 2001). The transgene is expressed in the tumors and in normal tissues in man, but not in normal mouse tissues, with an increased level of p21 protein in transgenic mice compared with nontransgenic mice. Point mutations in the Ha-ras gene are common occurrences contributing to human cancers (Tamaoki 2001). The transgenic mouse rarely develops spontaneous tumors until 6 months of age. However, N-nitroso-N-methylurea (NMU), a methylating agent, induces various neoplasms in this particular transgenic mouse including thymic lymphomas, mammary tumors, and gastrointestinal tumors (Qin et al. 1999) and induces forestomach tumors within 6 months with an increase in gene expression of the human transgene (Tamaoki 2001). These data suggest that a rapid onset and a higher incidence of malignant tumors can be expected after treatment with various genotoxic carcinogens in rasH2 mice due to increased p21 gene expression.

The National Toxicology Program (NTP) and the International Life Sciences Institute (ILSI) 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). However, limited published data sets are available in transgenic models, and their use as replacements for the 2-year rodent bioassays is not fully accepted for chemicals other than pharmaceuticals (Blain 2003). To further support the evaluation of alternative models to the mouse carcinogenicity study, this study investigated the carcinogenic potential of clofibrate, a nongenotoxic peroxisome proliferator-activated receptor (PPAR) α agonist (NDA 1993), following oral administration to the rasH2 mouse for 27 weeks. Chronic administration of PPARα ligands typically leads to hepatocarcinogenesis in rats and mice (Kluwe et al. 1982; Reddy, Rao, and Moody 1976). For example, the PPARα agonists fenofibrate and clofibrate induce liver tumors in mice at doses ≥60 mg/kg/day and ≥200 mg/kg/day, respectively (NDA 1993). Recent data from our laboratory indicate that clofibrate produced papillomas in Tg.AC mice after dermal application (Torrey et al. 2005b) but is noncarcinogenic in p53^+/– mice after 6 months of exposure (Torrey et al. 2005a), or neonatal mice up to 1-year after treatment on litter days 9 and 16 (Nesfield et al. 2005) or Tg.AC after 6 months of oral exposure (Torrey et al. 2005c). Nontransgenic rasH2 mice, not containing the human c-Ha-ras gene, were included in the study to confirm that these mice are less susceptible to genotoxic carcinogens. The positive-control dose of NMU was the standard used for this type of study (Yamamoto, Urano, and Nomura 1998) as recommended by ILSI guidelines.

MATERIALS AND METHODS

Animals

Male and female (n = 15/sex/group in the toxicology groups and 4 to 20/sex/group in the toxicokinetic groups; Table 1) CB6F1-Tg rasH2 mice and nontransgenic rasH2 mice were purchased from the Central Institute for Experimental Animals (Kawasaki, Japan). Mice were housed individually in plastic solid-bottomed cages. Environmental controls for the animal rooms were set to maintain a temperature of 18°C to 24°C, a relative humidity of 31% to 69%, and a 12-h light/12-h dark cycle. Either Rat and Mouse No. 1 Expanded Diet (Special Diets Services, Witham, Essex, United Kingdom) and domestic tap water meeting United Kingdom Water Supply (Water Quality) Regulations, were supplied ad libitum throughout the duration of the study.

Treatments and Data Collection

Clofibrate (2-(4-Chlorophenoxy)-2-methylpropanoic acid, ethyl ester; batch number 514) was obtained from Zeneca Pharmaceuticals (Cheshire, UK), prepared in 0.5% hydroxypropylmethylcellulose in sterile water at concentrations ranging from 5 to 25 mg/ml, and stored refrigerated (2–8°C) for 24 h. NMU was obtained from Sigma Chemical (Dorset, UK), prepared in citrate-buffered saline (pH 4.5) at 15 mg/ml, and stored refrigerated (2–8°C) for 24 h. Formulation analysis demonstrated that the actual clofibrate dosing suspension concentrations were within 10% of the target concentrations (data not shown).

Table 1 illustrates the treatment groups. Transgenic male mice were administered daily doses of clofibrate (groups 2 through 4 and 9 through 11) at 50, 100, or 200 mg/kg/day (males) or 50, 150, or 250 mg/kg/day (females) by oral gavage at 10 ml/kg for 27 weeks. Transgenic male and female mice (group 7) were administered a single intraperitoneal dose of the positive control (NMU) at 90 mg/kg at 5 ml/kg and observed for 90 days. Vehicle-control animals (groups 1, 5, 8, and 12) received 0.5% hydroxymethylcellulose in water only. Nontransgenic male and female mice (groups 6 and 13) also received the high-dose of clofibrate. The high doses of clofibrate (200 [males] or 250 [females] mg/kg/day) were based on the findings of an earlier range-finding study in wild-type B6C3F1 mice in which dosages between 400 and 500 mg/kg/day were lethal (data not shown). Lethality in the range-finding study showed that B6C3F1 males are more susceptible than females to toxicity induced by clofibrate. The low dose (50 mg/kg/day) was the lowest dose level used in the range-finding study. The intermediate doses (100 [males] or 150 [females] mg/kg/day) were the approximate geometric means of the low and high doses. Based on lethality noted in dose range–finding study, males are more susceptible than females to toxicity induced by clofibrate. Therefore, males and females received different dosages.

In-life data collected included clinical observations, body weight (weekly), food consumption (weekly), and mass palpation (in general, once weekly from day 64 onwards). On day 182 at 0 (predose), 0.5, 1, 3, and 24 h post dosing, toxicokinetics samples at 1–2/sex/clofibrate group/time point were collected in tubes containing EDTA via cardiac puncture after adminstration of CO₂. At 3 h post dosing, samples were collected from mice treated with the vehicle control material. Samples were placed on wet ice immediately after collection, then centrifuged at approximately 3500 or 4500 rpm for 15 min at 4°C to 5°C. Plasma was transferred to labeled sample tubes and stored frozen at approximately −20°C. Samples were analyzed for the metabolite of clofibrate, clofibric acid (Cayen 1980). Briefly, 50 μl of mouse plasma, calibration standard or quality control was mixed with 150 μl of water and 50 μl of 10% perchloric acid. After the mixing by vortex and centrifugation 30 μl of the supernatant was injected into the high-performance liquid chromatography (HPLC) system. The mobile phase consisting of 23% acetonitrile, 23% methanol in 0.1% phosphoric acid was pumped at 1.5 ml/min. The column used for separation was BDS Hypersil C8, 150 × 4.6 mm, 5 μ particle size. Detection system used was UV at 220 nm. Each batch contained calibration standards at 0.5, 1.0, 10, 50, 200, and 500 μg/ml in duplicate, as well as quality control samples at 1.5, 75, and 350 μg/ml, also in duplicate.

Mice were euthanized under isoflurane anesthesia by exsanguination via abdominal vasculature at study termination. At termination (day 105 for group 7, and day 190/191 for groups 1 through 6; Table 1), observations included clinical pathology, macroscopic and microscopic parameters, and organ weights. Hematology parameters evaluated included hemoglobin concentration, hematocrit, red blood cells, mean cell volume, mean cell hemoglobin, mean cell hemoglobin concentration, platelets, reticulocytes (absolute and relative), total leukocyte number, and differential leukocytes (absolute and relative). Clinical chemistry measurements included alkaline phosphatase, aspartate aminotransferase (AST), glutamate dehydrogenase, glucose, total protein, globulin (calculated), creatinine, alanine aminotransferase (ALT), total bilirubin, potassium, sodium, albumin/globulin ratio (calculated), albumin, globulin, chloride, calcium, cholesterol, triglycerides, high-density lipoprotein (HDL), low-density lipoprotein (LDL), very-low-density lipoprotein (VLDL), and bile acids. At termination, the following organs were weighed: brain, heart, liver, lungs, spleen, kidneys, testes, and thymus (paired organs weighed together).

Histopathological Evaluation

Tissues collected for histopathological evaluation were fixed in 10% phosphate-buffered formalin. Tissues were subsequently embedded in paraffin, sectioned at 5 μm, and stained with hematoxylin and eosin for histopathological examination. The following tissues were examined from groups treated with the vehicle control (group 1), NMU (group 7), and the high-dose of clofibrate (group 4) only: adrenal glands, aorta, brain, cecum, colon, duodenum, epididymides, esophagus, eyes and optic nerves, femur/stifle joint/bone marrow, harderian gland, heart, ileum, jejunum, lymph nodes (cervical and mesenteric), lungs, nasal chambers, ovaries, pancreas, peripheral nerve, pituitary gland, prostate gland, rectum, salivary glands, seminal vesicles, skeletal muscle, skin with mammary glands, spinal cord, spleen, sternum/bone marrow, stomach, testes, thymus (thymic area), thyroid gland with parathyroid gland, tongue, trachea, tumors/masses, urinary bladder, uterus, and vagina. Only, target organs were examined, including the kidneys (males only) and the liver (with gallbladder), in all the toxicology treatment groups (groups 1 through 7). In addition, gross lesions were similarly collected, processed, and examined by light microscopy.

Statistical Methods

Data are expressed as mean of treatment group or percent change from control. Results of comparisons are indicated only when significance at p < .05 is attained. Analysis of covariance or analysis of variance was used for data for other occasions, where appropriate, after transforming by identity (i.e., none), log or rank. Pairwise comparisons between the dose groups and the principal control have been performed (Conover 1980; Snedecor and Cochran 1968). Differences of the means or medians or ratios of geometric means, statistically adjusted for covariance and imbalance in the data, are presented. In addition, the following comparisons were investigated using the pairwise comparisons procedure: group 5 and group 6. Statistical analysis was not performed for group 7, hematology, clinical chemistry, and organ weights due to the lack of concurrent controls.

RESULTS AND DISCUSSION

Toxicokinetics

C _max of clofibric acid, the metabolite of the parent compound, did not increase proportionally with dose (Table 2). This nonlinear kinetics observed in this study was different from other alternative carcinogenicity models treated with clofibrate (Torrey et al. 2005a) and may be related to structural differences rodent PPARα receptor and/or tissue/species-specific coactivators/corepressors (Cheung et al. 2004). There was no significant difference between transgenic and nontransgenic rasH2 mice. No statistical differences were noted between the sexes. No clofibric acid was detected in samples from control mice. Previous Tg.AC studies with clofibrate have shown mean plasma concentrations that were similar at 6 month to those concentrations obtained after 1 month of dosing (Torrey et al. 2005c).

Mortality

Unscheduled deaths are listed in Table 3. An increased incidence of unscheduled deaths was noted in clofibrate-treated transgenic mice compared to those treated with the vehicle control material.

Eight unscheduled deaths (three males and five females) occurred in the clofibrate-treated rasH2 mice. Microscopic findings in these animals included subcutaneous hemangiosarcoma or polysystemic arteritis, bronchoalveolar carcinoma, hepatocellular adenoma, and splenic hemangiosarcoma. In addition, one nontransgenic toxicokinetic animal treated with clofibrate (200 mg/kg/day) was found dead. In the vehicle-treated rasH2 control group, one female was diagnosed with thymic lymphoma with widespread metastases.

Two unscheduled deaths (one male and one female) occurred in the NMU-treated rasH2 transgenic mice. Microscopic findings included splenic hemangiosarcoma, histocytic sarcoma of the thymus, bilateral retinal atrophy of the eyes, duct hyperplasia of the mammary glands, and squamous papilloma of the stomach.

Clinical Observations

In males, a very low incidence of subdued behavior, tiptoe gait, piloerection, unsteady gait, reluctance to move, and hunched or low posture was noted in both transgenic and nontransgenic mice treated with 200 mg/kg/day of clofibrate. Masses were noted for one control transgenic mouse and two transgenic mice treated with 100 mg/kg/day of clofibrate from day 120 onwards. A nontransgenic mouse treated with vehicle control material was also noted as having a transient mass between days 127 and 129. Masses were first noted from day 77 for mice treated with NMU. There were no other clinical signs related to NMU treatment.

In females, no clinical signs related to clofibrate or NMU treatment were noted. Masses were observed for transgenic mice treated with clofibrate and also for one transgenic mouse treated with vehicle. These were first noted from day 56. Nontransgenic mice were not affected. Masses were noted beginning on day 70 in transgenic mice treated with NMU.

Body Weights and Food Consumption

No treatment-related effects on body weight or food consumption were noted in male transgenic or nontransgenic mice treated with clofibrate. In females, body weight gain was increased by 59% for transgenic mice treated with 250 mg/kg/day of clofibrate, compared to controls with no effect on food consumption (data not shown).

Male mice treated with NMU showed generally higher food consumption compared to mice treated with vehicle control with no effect on body weights (data not shown). This effect was not observed in females.

Hematology

Some white blood cell parameters were slightly increased for transgenic male and nontransgenic female mice treated with clofibrate (data not shown). However, the toxicological significance of these findings is unclear. There were no other hematological findings.

For mice treated with NMU, decreases in erythrocyte counts, hemoglobin concentration, hematocrit, mean cell hemoglobin and mean cell hemoglobin concentration, lymphocytes, monocytes, neutrophils, eosinophils, and platelets, and increases in leuckocytes and reticulocytes (data not shown) may be related to increased spleen weight and an inhibitory effect of NMU on proliferating bone marrow red and white blood cell precursor cells and apoptotic cell death (Shilkaitis et al. 2000). These data are consistent with studies showing that NMU treatment results in sever damage to the hematopoietic system (IARC 1978).

Clinical Chemistry

Clofibrate treatment-related clinical chemistry changes are illustrated in Table 4. Collectively, the clinical chemistry changes noted for clofibrate-treated animals were related to the changes in the liver and in lipid concentrations, which are a well-established effect of clofibrate (Yki-Jarvinen 2004). There were few significant differences between the transgenic and nontransgenic mice.

Alkaline phosphatase activity was increased up to 87% in clofibrate-treated transgenic animals. For nontransgenic male and female mice treated with 200/250 mg/kg/day, alkaline phosphatase activity was increased approximately 90%. At the high doses (200/250 mg/kg/day), ALT activity was slightly increased 37% and 13% for transgenic male and female mice, respectively, and 27% and 39% in male and female nontransgenic mice, respectively. In males only, AST activity was slightly increased (up to 22%) for transgenic mice treated, and urea concentration was decreased up to 19% for transgenic mice. Albumin:globulin ratio was decreased approximately 7% for transgenic male mice treated with doses ≥100 mg/kg/day. Calcium was increased (4% to 6%) for both male and female transgenic and male nontransgenic mice treated with 200/250 mg/kg/day.

Cholesterol concentrations were increased during the study at all doses (11% to 38%) for transgenic and nontransgenic mice. Triglycerides were reduced between 16% and 48% for all mice treated with doses ≥100/150 mg/kg/day. Increases were seen in HDL (16% to 30%) and VLDL (33% to 79%) for all mice treated at ≥100/150 mg/kg/day. In females, LDL was decreased (19% to 41%) in all clofibrate-treated animals. In males, glucose concentration was increased (10% to 18%) for all mice treated with ≥100 mg/kg/day.

Decreases in plasma triglyceride concentrations for animals treated with ≥100 mg/kg/day are related to the therapeutic action of this class of compound (Yki-Jarvinen 2004). The mechanism of increases in cholesterol, HDL, and VLDL is unclear. It is known that rodents have limited usefulness for studying the effect of lipid-lowering compounds because of substantial differences in lipoprotein metabolism from humans (Overturf and Loose-Mitchell 1992). These increases in efficacy parameters (cholesterol, HDL, and VLDL) correlate with previous studies with PPARα agonists in normal, healthy animals. However, these data are in contrast to those observed in diseased animal and humans (Hoivik et al. 2004). Other changes considered to be related to hepatic effects include shifts in plasma proteins, glucose, and numerous enzymes. The liver is also intimately involved in glucose and lipid metabolism, protein/lipoprotein synthesis, and the enzymes associated with the various metabolic pathways (including alkaline phosphatase, ALT, and AST). Alteration of liver metabolism or effects on muscle are, therefore, almost certain to have a number of effects on the biosynthetic function of the liver and, therefore, the plasma levels of these endogenous molecules which is consistent with the lack of degenerative or necrotic effects in the liver.

Effects on NMU treatment in male and/or female mice included increases in glutamate dehydrogenase, ALT, AST, calcium, triglycercides, urea, potassium, and creatinine and decreases in bilirubin, bile acids, total protein, albumin, globulin, and albumin:globulin ratio (data not shown). Alterations in liver enzymes, plasma proteins, electrolytes, and triglycerides suggest that NMU has an effect on the metabolic profile of the liver; however, no degenerative changes in the liver were noted for NMU-treated mice (as noted below).

Organ Weights

Clofibrate-induced organ weight effects are illustrated in Table 5. Liver weight was increased at all doses; increases in the high-dose groups (200 [male] and 250 [female] mg/kg/day) were 28% in transgenic males, 24% in transgenic females, 21% in nontransgenic males, and 28% in nontransgenic females. These findings are consistent with activation of PPARα with subsequent proliferation of peroxisomes and smooth endoplasmic reticulum (Amacher et al. 1997; Milton, Elcombe, and Gibson 1990; Simpson 1997). In males, spleen weight was decreased up to 19% in transgenic mice treated with ≥100 mg/kg/day with no effect observed in nontransgenic mice. In females, kidney weight was increased up to 11% in all clofibrate-treated mice. In addition, thymus weight was increased up to 15% in female transgenic mice at 250 mg/kg/day. No thymus weight changes were observed in nontransgenic mice.

NMU treatment–related organ weight changes included increased liver, lungs, spleen, and thymus weights, and decreased kidney weight in one or both sexes (data not shown). Alterations in liver weight, and correlative liver enzymes, plasma proteins, electrolytes, and triglycerides suggest that NMU has an effect on the metabolic activity of the liver consistent with no microscopic degenerative change in the liver. The changes in spleen weights correlated with previous studies (da Silva Franchi et al. 2003). Changes in lung and thymus weights correlated with neoplastic findings. The significance of the decrease in kidney weight in females treated with NMU is unknown.

Macroscopic Pathology

In males, notable treatment-related macroscopic findings (variable-sized foci, pale areas, or nodules) were present in the liver of transgenic animals treated with ≥100 mg/kg/day (data not shown). Other findings included a splenic mass in a clofibrate-treated transgenic animal. In females, no macroscopic findings clearly related to clofibrate treatment were observed. Macroscopic findings associated with microscopic evidence of neoplasms (data not shown) included the following: at 50 mg/kg/day, one mouse had pallor and enlargement of the right intermediate lobe of the lungs (bronchoalveolar carcinoma) and another mouse had a cutaneous pale nodule (1 mm) on the head (squamous papilloma); at 150 mg/kg/day, one mouse had a mass (3 mm) on the tail (squamous-cell carcinoma) and one animal had a dark mass (6 to 10 mm) in the uterus (hemangiosarcoma); and at 250 mg/kg/day, one mouse had a mass (11 to 15 mm) in the spleen (hemangiosarcoma).

Macroscopic findings were noted in the stomach and thymus of NMU-treated mice (data not shown) and were associated with microscopic evidence of neoplasms. Nodules, pale (or white) areas and thickening of the nonglandular region of the stomach were noted and correlated with squamous papilloma. Thymic enlargement/masses (ranging in size from 6 to 15 mm) were associated with lymphoma. In addition, changes in the kidney (pallor, enlargement, and mottling), liver (enlargement), and spleen (enlargement) were associated with infiltration by lymphoma cells in one NMU-treated female mouse. One mouse also had mottling of the spleen (infiltrated by lymphoma cells) and a mass (6 to 10 mm) on the left lobe of the lungs (broncho-alveolar carcinoma). Non-neoplastic treatment-related macroscopic findings included opacity of the eyes in two animals treated with NMU.

Microscopic Pathology

Neoplastic Findings in Clofibrate-Treated Mice

Recent data from our laboratory indicate that clofibrate produced papillomas in Tg.AC mice after dermal application (Torrey et al. 2005b) but is noncarcinogenic in p53^+/– mice after 6 months of exposure (Torrey et al. 2005a) or in neonatal mice up to 1 year after treatment on litter days 9 and 16 (Nesfield et al. 2005) or in Tg.AC after 6 months of oral exposure (Torrey et al. 2005c). These data suggest that clofibrate was tumorgenic in most animal models that detect nongenotoxic carcinogens. However, recently published data suggests that structural differences between the human and rodent PPARα receptor and/or tissue/species-specific coactivators/corepressors may be responsible for the differential response noted amongst mouse strains/models (Cheung et al. 2004).

Treatment-related neoplastic findings in clofibrate-treated animals are shown in Table 6. Treatment of male transgenic mice (but not nontransgenic mice) with ≥100 mg/kg/day of clofibrate resulted in the appearance of hepatocellular adenomas and carcinomas possibly related to an increase in peroxisome proliferation (Amacher et al. 1997; Milton, Elcombe, and Gibson 1990; Simpson 1997). In transgenic females, neoplasms were also identified in the harderian glands (adenoma), lungs (bronchoalveolar adenoma and carcinoma), skin (squamous papilloma), tail (squamous-cell carcinoma), spleen (hemangiosarcoma), and uterus (hemangiosarcoma). Neoplasia occurred in nontransgenic females in the thymus (histiocytic sarcoma) and uterus (deciduoma). Of these findings, only the liver tumors in males and the harderian gland tumors in females are likely related to treatment with clofibrate. In general, the incidences of neoplasms found in the other tissues are within incidiences in historical controls (Tennant et al. 2001). These findings may be related to an increased incidence of peroxisome proliferation and/or other PPARα-related effects such as inhibition of apoptosis, oxidative stress, alteration of cell signaling molecules or gap communications (Klaunig et al. 2003). In contrast, no definitive neoplastic response was observed in p53^{+ / –} mice after 6 months of exposure to the nongenotoxic carcinogen clofibrate at doses up to 100 mg/kg/day for males and 125 mg/kg/day for females (Torrey et al. 2005a). These findings are consistent with the hypothesis that the p53^{+ / –} animal model is best suited for examining genotoxic carcinogens. In contrast, the rasH2 animal model is suited for examining both genotoxic and nongenotoxic compounds. In addition, a greater toxicity was observed in the p53^{+ / –} strain, with limited doses of clofibrate (Torrey et al. 2005a).

Non-Neoplastic Findings in Clofibrate-Treated Mice

Clofibrate treatment of both nontransgenic (200/250 mg/kg/day) and transgenic mice (≥100/150 mg/kg/day) resulted in non-neoplastic changes (Table 7) in the liver (male and females) and kidneys (male only).

Hepatic centrilobular granular eosinophilia was characterized by fine eosinophilic granules in the cytoplasm of centrilobular hepatocytes. Hepatic centrilobular granular eosinophilia is a recognized change with clofibrate treatment and is related to the induction of genes coding for cytochrome P450 and β-oxidation enzymes with the proliferation of peroxisomes and smooth endoplasmic reticulum (Amacher et al. 1997; Milton et al. 1990; Simpson 1997). This granularity is due primarily to the increased number and size of peroxisomes. Reduced glycogenic vacuolation was detected in all transgenic and nontransgenic female mice in the high-dose groups and in the majority of the intermediate-dose group (graded as very slight) and may be associated with decreased glycogen deposition or increased glycogenolysis. An increased incidence of hepatocellular mixed foci was noted only for transgenic mice treated with 100 or 200/250 mg/kg/day. The hepatocytes comprising these foci had vacuolated and slightly basophilic cytoplasm. This change is recognized as a potential preneoplastic lesion under certain circumstances (Jones, Popp, and Mohr 1997) and could represent initiated cells following carcinogen administration, but is also recognized as a spontaneous change in ageing rodents (Greaves 1990).

Renal cortical vacuolation in males was characterized by the presence of variable-sized, clear, discrete vacuoles in tubular epithelial cells.

Neoplastic Findings in NMU-Treated Mice

Neoplasms were identified in the Harderian gland (adenoma), lungs (bronchoalveolar adenoma and adenocarcinoma), stomach (squamous papilloma), thymus (lymphoma), and vagina (squamous papilloma) of males and/or females after a single intraperitoneal dose of NMU.

Multiple squamous papillomas of the nonglandular portion of the stomach were associated with localized squamous hyperplasia.

Thymic lymphoma was characterized by the ablation of the normal thymic architecture by a homogeneous population of neoplastic lymphocytes. A proportion of animals with thymic lymphoma had multiple additional tissues affected by an infiltrate of lymphoma cells. Neoplasms included thymic lymphoma and stomach squamous-cell carcinoma. Non-neoplastic proliferative lesions included stomach acanthosis, hyperkeratosis, and papillary hyperplasia, and adenomatous hyperplasia of the duodenum or ileum (Hoivik et al. 2005).

NMU induces various neoplasms based on the production of carcinogenic O ⁶-methylguanine (O ⁶-mG) lesions in DNA (Qin et al. 1999). The rasH2 transgenic mouse is also known to have an increased incidence of certain types of spontaneous neoplasms. There is a synergistic mechanistic effect of NMU treatment and the overexpression of ras protooncogenes in transgenic mice in the etiology of some tumor types (Mangues et al. 1994). NMU treatment of rasH2 transgenic mice has been associated with the induction of lung adenomas and carcinomas, stomach papillomas, skin papillomas, and thymic lymphomas at an increased incidence compared with untreated transgenic mice (Mitsumori et al. 1998; Yamamoto et al. 1997). The presence of many stomach papillomas and thymic lymphomas in the present study is in agreement with these reports. Other neoplasms seen are possibly related to the transgenic status and/or NMU treatment.

Non-Neoplastic Findings in NMU-Treated Mice

Treatment-related findings were seen in the eyes (retinal atrophy), mammary glands (duct hyperplasia), stomach (squamous hyperplasia and squamous erosion), harderian gland (acinar or focal hypertrophy/hyperplasia), lungs (bronchoalveolar hyperplasia), and thymus (lymphoid hyperplasia) of males and/or females after a single intraperitoneal dose of NMU. Similar NMU administration in p53^{+ / –} mice had comparable effects (Hoivik et al. 2005).

Bilateral retinal atrophy of the eyes was characterized by the loss of cells in the photoreceptor and outer nuclear layers. Bilateral retinal atrophy is a recognized change with NMU treatment and is related to the induction of apoptosis in retinal cells (Yuge et al. 1996). Mammary gland duct hyperplasia was characterized by increased numbers of epithelial layers above the duct basement membrane, with variable degrees of intraductal papillary change. Hyperplastic duct epithelial changes were particularly notable within the deep dermis layer but also affected ducts within the acinar mammary tissue. This finding is a recognized preneoplastic lesion that has been associated with the development of mammary carcinoma (Greaves 1990). Furthermore, NMU administration has been demonstrated to cause mammary carcinoma in transgenic mice (Mangues et al. 1994). Localized squamous hyperplasia of the nonglandular portion of the stomach consisted of increased number of epithelial layers and hyperkeratosis. The change was often associated with papillomas. Localized hypertrophy/hyperplasia of the Harderian gland, lymphoid hyperplasia of the thymus, and interstitial-cell hyperplasia of the testes are recognized spontaneous findings in aged mice, but they have also been associated with the proliferative spectrum of change that includes neoplasia (Greaves 1990; Maronpot 1999).

CONCLUSION

Clofibrate is a nongenotoxic PPARα agonist that causes hepatocellular neoplasms in rodents (NDA 1993). PPARα agonists comprise a diverse class of compounds that have therapeutic indications, including dyslipidemia (Yki-Jarvinen 2004). In general, chronic administration of many PPARα ligands induce hepatocarcinogenesis in rats and mice (Kluwe et al. 1982; Reddy et al. 1976). For example, the PPAR α agonists fenofibrate and clofibrate induce liver tumors in mice at doses of ≥60 mg/kg/day and ≥200 mg/kg/day, respectively (NDA 1993). Recent data from our laboratory indicate that clofibrate is produced papillomas in Tg.AC mice after dermal application (Torrey et al. 2005b) but is noncarcinogenic in p53^{+ / –} mice after 6 months of exposure (Torrey et al. 2005a) or in neonatal mice up to 1 year after treatment on litter days 9 and 16 (Nesfield et al. 2005) or in Tg.AC after 6 months of oral exposure (Torrey et al. 2005c). However, humans are resistant to peroxisomal proliferation and presumably the development of hepatocarcinomas after exposure to PPARαigands (Klaunig et al. 2003). In this study, male mice were administered doses of clofibrate at 50, 100, or 200 mg/kg/day, and female mice were administered 50, 150, or 250 mg/kg/day, for 27 weeks. In addition, nontransgenic male and female mice were treated with 200 or 250 mg/kg/day, respectively, of clofibrate for 27 weeks. C _max values for clofibric acid at the end of 6 months were 265 μg/ml for rasH2 males treated at 200 mg/kg/day and 362 μg/ml for rasH2 females treated at 250 mg/kg/day; there was no difference between transgenic and nontransgenic groups. Hepatocellular neoplasms, possibly related to an increase in peroxisome proliferation (Amacher et al. 1997; Milton et al. 1990; Simpson 1997), were observed in clofibrate-treated rasH2 male mice after 6 months of treatment, but not in nontransgenic males or females. Clofibrate treatment (250 mg/kg/day) of female rasH2 mice was associated with a slight increase in the incidence of various neoplasms (harderian gland, lungs, skin, spleen, tail, thymus, and uterus) compared with untreated transgenic mice and with similarly treated nontransgenic mice. These findings may be related to an increase incidence of peroxisome proliferation and/or other PPAR α-related effects such as inhibition of apoptosis, oxidative stress, and alteration of cell signaling molecules or gap communications (Klaunig et al. 2003). The positive control NMU produced an increased incidence of stomach papillomas and thymic lymphomas, mammary gland duct hyperplasia, stomach squamous hyperplasia and erosion, harderian gland hypertrophy/hyperplasia, bronchoalveolar hyperplasia, testicular interstitial-cell hyperplasia, retinal atrophy, and thymus changes. These findings are consistent with previous studies (Hoivik et al. 2005; Mitsumori et al. 1998; Yamamoto et al. 1997). The data suggest that the rasH2 mice are a good model for testing epigenetic carcinogens in a shorter timeframe than conventional mouse carcinogenicity bioassays.