Evaluation of the Carcinogenic Potential of Clofibrate in the p53 + / − Mouse

Abstract

This study was conducted as part of International Life Sciences Institute (ILSI) program to evaluate the carcinogenic potential of clofibrate, a nongenotoxic, peroxisome proliferator-activated receptor (PPAR) α agonist, following oral administration to p53^{+ / −} heterozygous mice for a minimum of 26 weeks. p-Cresidine, a urinary bladder carcinogen, was given orally at 400 mg/kg/day as a positive control. Initial clofibrate doses were 50, 250, and 400 mg/kg/day for males and 50, 200, and 500 mg/kg/day for females. Due to unexpected mortality during the first week of dosing, clofibrate doses were lowered to 25, 75, and 100 mg/kg/day for males and 25, 75, and 125 mg/kg/day for females. Clinical signs and mortality were greater in p53^{+ / −} than wild-type (WT) mice. With the exception of liver weights, no marked differences in any other parameters either between the sexes or between WT and p53^{+ / −} mice were noted. Moderate increases in liver weights noted in WT males given 100 mg/kg/day clofibrate were not associated with any microscopic changes. No neoplastic response was observed in p53^{+ / −}mice after 6 months of exposure to clofibrate at doses up to 100 mg/kg/day for males and 125 mg/kg/day for females. Transitional-cell hyperplasia and carcinoma of the urinary bladder were noted in both sexes given p-cresidine, demonstrating that the p53^{+ / −} mouse responded to a known mouse carcinogen as expected. Clofibrate produced non-neoplastic findings in the adrenals, pancreas, and prostate, whereas p-cresidine affected the kidney, liver, pancreas, and spleen.

Clofibrate is a nongenotoxic, peroxisome proliferator-activated receptor (PPAR) α agonist (NDA 1993). PPARα agonists comprise a diverse class of compounds that have therapeutic indications, including dyslipidemia (Yki-Jarvinen 2004). These chemicals bind to PPARs, which form a heterodimer with the retinoid X receptor (RXR) and interact with DNA to modulate gene regulation at a transcriptional level (Yki-Jarvinen 2004). Three PPAR subtypes α, β, and γ have been identified with a broad tissue distribution including the liver, heart, kidney, adipose tissue, and intestine (Fajas et al. 1997; Lee et al. 1995; Storer et al. 2001). Chronic administration of PPARα ligands typically leads to hepatocarcinogenesis in rats and mice (Kluwe et al. 1982; Reddy, Rao, and Mood 1976). For example, the PPARα agonists fenofibrate and clofibrate induce liver tumors in mice at doses of ≥60 mg/kg/day and ≥200 mg/kg/day, respectively (NDA 1993). However, humans are resistant to peroxisomal proliferation and the development of hepatocarcinomas after exposure to PPARα ligands (Klaunig et al. 2003). Although PPARα activation is required for hepatocarcinogenesis in rodents (Cheung et al. 2004; Ward et al. 1998), the mechanism of action is not completely understood; it likely involves a combination of cell proliferation, oxidative stress, and peroxisomal proliferation (Cheung et al. 2004; Hoivik et al. 2004). In addition, recent data suggest that structural differences between the human and rodent PPARα or tissue/species-specific coactivators/corepressors may be responsible for the species-specific difference in hepatocarcinogenesis (Cheung et al. 2004). These models included the Tg 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 the neonatal mouse model (Robinson and MacDonald 2001).

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. The p53 tumor suppressor gene is critical for cell cycle control and DNA repair and suppresses 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 is the model of choice for alternative carcinogenicity study protocols due to its exhibit accelerated tumorgenesis and generally is accepted for use in evaluating the carcinogenic potential of genotoxic agents (French, Storer and Donehower 2001; Robinson and MacDonald 2001). As of 2002, the p53^{+ / −} knockout mouse was the predominant model of choice for carcinogenicity study protocols received by the Food and Drug Administration (FDA) that contained an alternative animal model (MacDonald et al. 2004). Although the National Toxicology Program (NTP) and 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), there is still some concern regarding the limited amount of published data available in transgenic models related to their replacement of the 2-year rodent bioassays (Blain 2003). This study was conducted as part of the ILSI program and evaluated the carcinogenic potential of clofibrate following oral administration to p53^{+ / −} heterozygous mice for a minimum of 26 weeks. Subsidiary objectives of this study included comparisons of plasma levels, of micronucleus production, and of DNA strand breakage as detectable in a COMET assay between heterozygous p53^{+ / −}and wild-type (WT) mice for both clofibrate and p-cresidine. p-Cresidine was included as a positive control in all ILSI p53^{+ / −}studies because it is a known rodent urinary bladder carcinogen (National Toxicology Program 1979).

MATERIALS AND METHODS

Animals

Male and female (n = 15/sex/group) C57BL/6TacfBR-[KO]p53 N5 heterozygous p53^{+ / −} knockout and WT mice, approximately 8 to 9 weeks old, were obtained from Taconic (Germantown, NY, USA). Mice were housed in suspended, stainless-steel, 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.

Treatments and Data Collection

Clofibrate (2-(4-Chlorophenoxy)-2-methylpropanoic acid, ethyl ester; batch number 514) was obtained from Zeneca Pharmaceuticals (Cheshire, UK), prepared monthly as a suspension in 0.5% methylcellulose in water at concentrations ranging from 2.5 to 50 mg/ml, and stored refrigerated (2–8°C). p-Cresidine was obtained from Pfaltz and Bauer (Waterbury, CT, USA), prepared monthly in corn oil at 40 mg/ml, and stored refrigerated (2–8°C). Dose concentration (clofibrate and p-cresidine) and homogeneity (clofibrate only) testing over the course of the study demonstrated that the actual clofibrate dosing suspension concentrations were within 10% to 25% of the target concentrations (95% of samples were within 10%) and were homogeneous (data not shown). p-Cresidine dosing suspension concentrations were within 10% of the intended concentrations (data not shown).

Design of the study is shown in Table 1. Mice were administered initial doses of clofibrate at 50, 250, or 400 mg/kg/day for males and at 50, 200, or 500 mg/kg/day for females in 0.5% methylcellulose in water at 10 ml/kg by oral gavage (specified as day 1). Positive-control groups of male and female mice received p-cresidine in corn oil at 400 mg/kg/day. Vehicle-control animals received only 0.5% methylcellulose in water treated by reverse osmosis. The initial doses of clofibrate were selected based upon a 1-month range-finding study in WT mice at doses up to 700 (males) and 1018 (females) mg/kg/day. Dosing of high-dose clofibrate animals was suspended on day 3 due to death at doses from 200 to 500 mg/kg/day; dosing of all remaining groups was suspended on day 6. Following a washout period of 4 to 13 (high-dose clofibrate animals) or 7 to 13 days (all other treatment groups) dosing of all groups was resumed on day 14. Analysis of plasma collected before dosing on day 14 showed no detectable clofibrate level in any mouse (data not shown). Clofibrate doses were resumed at 25, 75, and 100 mg/kg/day for males and at 25, 75, and 125 mg/kg/day for females; administration of the original dose of p-cresidine was resumed at the same time. In-life data collected for 26 weeks after resumption of dosing included clinical observations, body weight, food consumption, and mass palpation. Toxicokinetic samples from satellite groups (1–3/sex/group/timepoint) were collected in tubes containing EDTA from the vena cava after administration of CO₂ (terminal sample). Collection was made on days 1 and 195 at 0 (predose), 1, 3, 6, and 24 h post dosing (clofibrate and p-cresidine) and at 0 (predose) and 24 h post dosing (vehicle control). Samples were placed on wet ice immediately after collection and 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 clofibric acid (clofibrate and vehicle-control groups) or p-cresidine (positive-control groups).

A standard ILSI study design was utilized (Robinson and MacDonald 2001), which included histological evaluation of tissues without required clinical pathology evaluations; however, individual participants deviated from the standards according to local preferences without overall impact on ILSI standards. Mice were anesthetized with CO₂ then euthanized by exsanguination via of the caudal vena cava. At termination, observations included macroscopic and microscopic observations, organ weights, genotyping, and micronuclei and COMET analyses. The following organs were weighed: adrenal glands, brain, heart, liver, lungs, kidneys, testes, pituitary gland, prostate gland, spleen, thymus, thyroid gland, 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 for all mice dying after day 8 in the vehicle control groups, the p-cresidine groups, and the high-dose clofibrate groups (400/100 [males] and 500/125 [females] mg/kg/day): adrenal glands, aorta, brain, cecum, colon, duodenum, epididymides, esophagus, eyes and optic nerves, femur/joint/bone marrow, harderian glands, head, heart, ileum, jejunum, gallbladder, kidneys, larynx, liver, lungs with bronchi, lymph nodes (mesenteric and mandibular), nasal cavity, ovaries, pancreas, peripheral nerve, pituitary gland, prostate gland, rectum, salivary glands, seminal vesicles, skeletal muscle, skin with mammary glands, spinal cord (cervical, thoracic, and lumbar), spleen, subcutaneous tissue from rump, sternum/bone marrow, stomach, tail, testes, thymus (thymic area), thyroid with parathyroid gland, tongue, trachea, tumors/masses, urinary bladder, uterus with cervix, and vagina. The animal identification site was collected but not examined. In addition, gross lesions were similarly collected, processed, and examined by light microscopy. When appropriate, microscopic evaluation was also done for the low and mid clofibrate dose groups.

Genotyping

Approximately 1 cm of the tail was collected from all mice except those in satellite toxicokinetic groups, frozen in liquid nitrogen, transported frozen on dry ice, and stored at or below approximately –70°C. Genotyping analysis for heterozygous and WT strains was conducted by Charles River Therion, Troy, NY, USA. Analysis of the tail tissue confirmed that all mice had the intended genotype (data not shown).

Micronucleus Analysis

Approximately 0.2 ml of blood was taken from each of the first 7 animals/micronucleus group/sex at terminal sacrifice. The blood samples were placed into test tubes containing sodium heparin solution and diluted 1:4 with fetal bovine serum. Slides with blood smears were air-dried, fixed with methanol, and stained with acridine orange solution (12 μg/ml in pH 6.8 phosphate buffer) for 1 min just prior to analysis. The frequency of polychromatic erythrocytes (PCEs) and the PCE-to-normochromatic erythrocyte (NCE) cell ratio were established to assess bone marrow cytotoxicity. The PCE/NCE ratio is commonly used to assess bone marrow toxicity (Heddle et al. 1991). The ratio of PCEs to NCEs provides an estimation of any perturbations in hematopoesis as an effect of the treatment (Gollapudi and McFadden 1995). An increase in PCEs indicates a stimulation of proliferative activity. An increase in NCEs signals the destruction of immature erythrocytes, a cytotoxic effect. The micronucleus frequency (expressed as percent micronucleated cells) was determined by analyzing the number of micronucleated PCEs from at least 2000 PCEs per animal. The frequency of the PCE-to-NCE ratio was determined by scoring the number of PCEs and NCEs observed in the optic fields while scoring at least the first 200 erythrocytes on the slide. The frequency of micronucleated NCEs was determined by recording the number of micronucleated NCEs in the total erythrocytes scored to establish the PCE-to-NCE ratio. The micronucleus scoring and analysis was conducted at Covance Laboratories (Vienna, VA, USA).

COMET Analysis

At terminal necropsy, samples of whole blood, liver, and duodenum from male mice were processed and sent frozen to Integrated Laboratory Systems (Research Triangle Park, NC) for COMET analysis. For each frozen minced liver or duodenum tissue sample, 10 μl of the cell suspension was used for an analysis of DNA damage using the alkaline (pH > 13) COMET assay (Tice et al. 2000). For the whole blood, samples were thawed, the supernatant removed, and the entire remaining pellet used for analysis. Each sample was coded prior to slide preparation. Eight slides were prepared per tissue from five male mice each from the vehicle control, positive control, and clofibrate at 100/125 mg/kg/day p53^{+ / −} groups, and the vehicle control and clofibrate 100/125 mg/kg/day WT groups. Two slides per sample were removed after 1 to 2 h of lysis, neutralized in 0.4 M Trizma base (pH 7.5), and fixed in 100% ethanol to analyze, in the absence of electrophoresis, the frequency of cells with extremely low molecular weight (ELMW) DNA. Following lysis, two additional slides per animal per tissue were exposed to alkali (pH > 13) for 20 min, followed by electrophoresis in the same buffer for an additional 20 min at approximately 0.9 V/cm and 300 mAmp. For each tissue type, replicate slides from each mouse were electrophoresed together within a single run. After electrophoresis, slides were immersed in an excess amount of 0.4 M Trizma base (pH 7.5) to neutralize any remaining alkali, immersed in 100% cold ethanol for fixation, and air dried. Prior to “blinded” scoring, the DNA was stained with SYBRGreen. For each sample, the frequency of cells with ELMW DNA was determined by scoring 100 cells at 250× magnification for levels of diffusion ranging from I (condensed DNA, no diffusion) to IV (no condensation, diffuse DNA). For each sample, the extent of DNA migration was determined by scoring 100 cells (50 cells per each of two slides) at 250× magnification using the Kinetic Imaging Komet 4.0 image analysis system (Liverpool, UK). The treatment group mean frequency of cells with ELMW DNA and the treatment group mean extent of DNA migration, based on % migrated DNA, tail length, and tail moment (fraction of migrated DNA × tail length), were compared.

Statistical Methods

Toxicological data are expressed as mean ± standard deviation of the mean. The micronucleus assay used an analysis of variance (Winer 1971) on untransformed proportions of cells with micronuclei per animal and on untransformed PCE:NCE ratios. For the COMET assay, the treatment group mean frequency of cells with ELMW DNA and the treatment group mean extent of DNA migration, based on % migrated DNA, tail length, and tail moment (fraction of migrated DNA × tail length), were compared between dose groups within p53^{+ / −} or WT mice using a nonparametric one-tailed Mann-Whitney U test, and between p53^{+ / −} and WT mice using a nonparametric two-tailed Kruskall-Wallis test. An alpha level of .05 was used to indicate statistical significance.

RESULTS AND DISCUSSION

Range-Finding Assay

To select doses for the 6-month study, a 1-month range-finding study was performed with clofibrate in WT mice at 68, 258, 491, and 699 (males) and 68, 292, 650, and 1018 (females) mg/kg/day. AUC increased with dose, decreased by day 30 compared to day 1, and did not differ by sex. AUC_0–24 values (μg · h/ml) on day 1 were 721, 5142, 7251, and 10962 for doses of 68, 258, 491, and 699 mg/kg/day in males, respectively; values for females were 1043, 5157, 17677, and 8882 for doses of 68, 292, 650, and 1018 mg/kg/day, respectively. Liver weights were increased at doses > 258 (males) and 292 (female) mg/kg/day. Hepatocellular hypertrophy was noted in males at doses > 491 mg/kg/day and at all doses in females; it was accompanied by an increase in hepatocellular eosinophilia in males only. Six of eight mice that died or were sacrificed before the end of the study had myocardial degeneration/necrosis. Body weights, food consumption, and macroscopic findings were unaffected by treatment with clofibrate (data not shown).

Toxicokinetics

Systemic exposure (AUC and C _max) of clofibric acid increased in approximate proportion to dose (Table 2). Although slight differences were noted in systemic exposures, no marked differences (greater than two-fold) were observed in any of the parameters either between the sexes or between WT and p53^{+ / −}mice. In addition, exposure (AUC and C _max) was lower on day 195 than day 1, even when the lowered doses are taken into account, whereas the T _max was largely unaffected over the course of the study.

Plasma concentrations of p-cresidine were found at the 1-hour sampling and generally diminished in a time-related manner thereafter (Table 3). Mean C _max were lower on day 195 than on day 1 with no marked differences between sexes. The decrease in exposure may be related to metabolic enzyme induction as indicated by hepatocyte hypertrophy, and is consistent with an adaptive response that can be seen as part of a spectrum of effects in response to administration of hepatotoxins (Delker et al. 2000).

In-Life Findings

The incidence of mortality for all mice (except the micronucleus groups) Days 1 to the end of the study is summarized in Table 4. On days 1 and 2, clofibrate doses at ≥200 mg/kg/day caused deaths preceded by clinical signs of decreased activity, dehydration, hypothermia, ataxia, hunched posture, and labored breathing. Due to severe clinical signs of toxicity, dosing of the high-dose clofibrate groups was suspended on day 3 and of the remaining groups on day 6. The p53^{+ / −} strain was more sensitive to the effects of clofibrate at doses ≥250 mg/kg/day, as shown by mortality and clinical signs. Recently published data suggest 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). Dosing resumed for all groups on day 14, using lower doses for the clofibrate groups. No cumulative toxicity was evident as dosing progressed. Additional signs related to treatment with clofibrate included ataxia, tremors, and tiptoe gait. The incidence of clinical signs decreased over time. Decreased activity and rough coat were noted in mice given 400 mg/kg/day p-cresidine throughout the dosing period; the incidence was greater in males than in females.

No treatment-related changes in body weight or food consumption were noted in p53^{+ / −} or WT mice dosed with clofibrate (data not shown). Food consumption was lower for mice given 400 mg/kg/day p-cresidine and generally correlated with the body weight decrease noted (Table 5). This difference was greater in males than in females. These findings may associated with the nasal toxicity of p-cresidine (Reznik et al. 1981) resulting in a decreased sense of smell resulting in reduced food consumption, and corresponding reduced absorption/systemic exposure of the chemical (Table 3).

Organ Weights

Clofibrate-related organ weight changes were limited to the liver in males at 100 mg/kg/day: 10% and 11% increases (p53^{+ / −}), and 14% and 13% increases (WT) above control value for absolute and relative weights, respectively. However, these modest changes were not associated with any unequivocal microscopic changes and may represent prototypical PPARα ligand–induced hypertrophic responses in hepatocytes that were below the limit of detection by light microscopy. Both sexes showed increased relative and absolute spleen weights in p53^{+ / −}mice treated with p-cresidine (data not shown).

Macroscopic Pathology and Lesions

p-Cresidine–related macroscopic findings were noted in mice of both sexes (Table 6) and were limited to urinary bladder mucosal surface changes that included thickening, raised areas, and masses, which correlated microscopically to transitional-cell hyperplasia or carcinoma. There were no findings related to treatment with clofibrate. This is consistent with the nongenotoxic mechanism of carcinogenicity of clofibrate and the sensitivity of this transgenic model to genotoxic carcinogens.

Treatment-related neoplastic lesions for p-cresidine are summarized in Table 7. No lesions were present in WT mice either treated with the vehicle-control material or clofibrate at 100/125 mg/kg/day (data not shown). Two mice in the p53^{+ / −} groups treated with the vehicle control had neoplastic lesions. One male mouse had an adenocarcinoma of the pancreas, and one female mouse had malignant lymphoma in the thymus, thymic lymph node, heart, lung, spleen, and liver (data not shown). These findings correlate with the historical incidence of background tumors in p53^{+ / −} knockout mice (Storer et al. 2001). p53^{+ / −} mice treated with the nongenotoxic carcinogen, clofibrate, did not exhibit a neoplastic response, which is consistent with this model being sensitive to genotoxic carcinogens. Neoplastic findings were limited to a basal-cell skin carcinoma in one low-dose female and to a single basophilic focus of hepatocellular alteration in one high-dose male (data not shown). Thus, no definitive neoplastic response was observed in p53^{+ / −} mice after 6 months of exposure to clofibrate at doses up to the maximum tolerated dose (100 mg/kg/day for males and 125 mg/kg/day for females). Previous 2-year bioassays using conventional mouse strains have found that clofibrate induces tumors at doses of ≥200 mg/kg/day (NDA 1993). The reason for these differences is likely related to the primary mechanism of carcinogenicity in the p53^{+ / −} model (mutation by genotoxins), as opposed to a mechanism related to proliferative effects on liver cells, which were not observed in the present study. In addition, the present study used lower doses, necessitated by p53^{+ / −} strain-specific toxicity. There may also be strain-specific differences in the mechanism of PPAR-induced carcinogenesis that may have led to a nonprototypical liver response. Recent data from our laboratory indicates that clofibrate is hepatocarcinogenic in rasH2 mice after 6 months of exposure (Nesfield et al. 2005a) and produced papillomas in Tg.AC mice after dermal application (Torrey et al. 2005a), but is noncarcinogenic in neonatal mice up to 1-year after treatment on litter days 9 and 16 (Nesfield et al. 2005b) or in Tg.AC after 6 months of oral exposure (Torrey et al. 2005b).

It should be noted that the doses for this study were chosen from a 1-month range-finding assay in WT mice. It was expected that increases in liver weights and hepatocellular hypertrophy would be seen in one or more dose groups. However, because the p53^{+ / −} strain was much more sensitive than the WT to mortality and clinical signs caused by clofibrate, the doses in the 6-month study had to be lowered. At the new doses, the expected effects on liver were not observed in the p53^{+ / −} mice, and the only liver effect in WT mice was a moderate increase in liver weight in males only. It should be noted that for WT mice, the day 1 exposure in the range-finding test at 491 mg/kg/day was similar to that produced on day 1 in either strain in the 6-month study at 400 mg/kg/day (subsequently lowered to 100 mg/kg/day); females had similar results. Thus, the drug exposure was consistent from study to study in the p53^{+ / −} and WT mice, but toxicity was much greater in the p53^{+ / −} mice.

Transitional-cell carcinoma of the urinary bladder was present in 5 of 15 male mice and 4 of 15 female p53^{+ / −} mice treated with p-cresidine at 400 mg/kg/day (Table 7). These findings are consistent with previous studies (National Toxicology Program 1979). Diffuse transitional-cell hyperplasia was also present in 14 of 15 male mice and 14 of 15 female mice in this group. Almost all the mice with carcinomas also exhibited transitional-cell hyperplasia. Other neoplasms in mice treated with p-cresidine included an alveolar/bronchiolar adenoma of the lung in one male mouse, an osteosarcoma of the humerus in one female mouse, and an osteosarcoma in the inguinal area of a second female mouse. Thus, the p-cresidine positive control performed as expected, assuring that the model under the test conditions was capable of detecting a carcinogenic response.

Treatment-related nonproliferative lesions are summarized in Tables 8 and 9. Non-neoplastic changes in p53^{+ / −} mice treated with clofibrate were limited to adrenal glands (incidence of spindle-cell proliferation slightly increased in males and decreased in females, both only in the high-dose groups), pancreas (incidence of mononuclear-cell infiltration increased in females at 125 mg/kg/day), and prostate (incidence of both mononuclear-cell infiltration and prostatitis slightly increased in males at 100 mg/kg/day).

The incidence of nonproliferative lesions in the kidney, liver, pancreas, and spleen increased after treatment with p-cresidine. The renal changes, including papillary necrosis, have been observed with p-cresidine treatment in other studies that used p53^{+ / −} mice (Delker, Yano, and Gollapudi 2000; Petruska et al. 2002). Toxic hepatopathy was recorded when any of the following changes occurred in the liver: hepatocyte hypertrophy (often centrilobular), hepatocyte necrosis, hepatocyte apoptosis, and/or hepatocyte vacuolation. This nonspecific group of lesions is often present after the administration of hepatotoxins (Delker, Yano, and Gollapudi 2000). In general, the non-neoplastic changes present after the administration of p-cresidine had a higher incidence in male mice (Table 9).

Micronucleus Analysis

Neither clofibrate nor p-cresidine produced micronuclei in peripheral blood of male or female transgenic or WT mice. Cytotoxicity (based on changes in the PCE:NCE ratio) was also not observed for any group (data not shown). These findings are not surprising, given the inactivity of p-cresidine in the mouse bone marrow micronucleus and in unscheduled DNA synthesis (UDS) assays (Ashby et al. 1991), and in in vitro UDS assays (Kasper, and Muller 1993). They confirm the lack of clastogenic activity found in p53^{+ / −} male mice treated for 7 weeks at p-cresidine doses up to 800 mg/kg/day (Delker et al. 2000). Clofibrate is generally considered to be nongenotoxic in a number of systems, including the lacZ plasmid–based transgenic mouse mutation assay (Boerrigter 2004).

COMET Analysis

Neither clofibrate nor p-cresidine produced DNA damage detectable in a COMET assay in whole blood, liver, or duodenum of male mice of either strain. For all three tissues, the group mean extent of DNA migration was not significantly different between the two animal strains or within each animal model and was not significantly increased in 100 mg/kg/day clofibrate–treated mice compared to the concurrent control group. The positive control, p-cresidine, did not induce a significant increase in DNA migration compared to the vehicle control in the p53^{+ / −} heterozygous mice. p-Cresidine has been reported previously to be inactive in an assay measuring DNA single strand breaks (Ashby et al. 1991). However, a COMET assay on CD-1 male mouse liver, lung, kidney, brain, and bone marrow after a single dose of p-cresidine at 595 mg/kg produced DNA damage, although only in bladder mucosa (Sasaki et al. 1998). In a COMET assay using hepatocytes in vitro, clofibrate caused DNA strand breaks at concentrations of 50 μM and higher, perhaps caused by the acyl glucuronide of clofibrate (Ghaoui et al. 2003).

Although the target organ for clofibrate carcinogenesis in a conventional bioassay is the liver (NDA 1993), treatment of p53^{+ / −} or WT male mice at 100 mg/kg/day for at least 26 weeks did not result in a significantly increased level of DNA damage in this tissue or in blood leukocytes or duodenum, as measured by the alkaline COMET assay. Light microscopy likewise revealed no damage to these tissues after clofibrate treatment in the present study.

CONCLUSION

This study was conducted in support of the ILSI alternative carcinogenicity models initiative to evaluate the carcinogenic potential of clofibrate, a nongenotoxic, PPARα agonist, following oral administration to p53^{+ / −} mice for a minimum of 26 weeks. No neoplastic response was observed with the nongenotoxic carcinogen, clofibrate, at doses up to 100 (males) or 125 (females) mg/kg/day, whereas p-cresidine at 400 mg/kg/day produced urinary bladder carcinogenesis. In contrast, clofibrate treatment (250 mg/kg/day) of female rasH2 mice was also associated with a slight increase in the incidence of various nonhepatic neoplasms compared with untreated transgenic mice and with similarly treated nontransgenic mice (Nesfield et al. 2005a). The reason for this difference is unknown, but may be related to unexpected toxicity in the p53^{+ / −} study, strain-specific differences in the mechanism of PPARα-induced carcinogenesis, or degree of peroxisomal proliferation as indicated by hepatic centrilobular granular eosinophilia in the rasH2 mice study. In addition, the findings are consistent with the hypothesis that the p53^{+ / −} animal model is best suited for examining genotoxic carcinogens and clofibrate is considered to be nongenotoxic and that the rasH2 animal model is suited for examining both genotoxic and nongenotoxic compounds. The lack of prototypical liver response to clofibrate may be related to the lower doses tested in these studies due to unexpected toxicity, as well as strain-specific differences in the mechanism of PPARα-induced carcinogenesis. Clofibrate produced non-neoplastic findings in the adrenals, pancreas, and prostate, whereas p-cresidine affected kidney, liver, pancreas, and spleen. The presence of transitional-cell hyperplasia and carcinoma in both sexes in the positive control group demonstrates that the p53^{+ / −} heterozygous mouse responds appropriately to p-cresidine, a known mouse carcinogen. Neither compound was genotoxic in vivo in the mouse peripheral blood micronucleus assay using p53^{+ / −}and wild-type mice of both sexes, or in a COMET assay to detect DNA strand breakage. Thus, treatment-related apoptosis or necrosis at this sample time did not affect interpretation of the DNA migration results. Toxicokinetic investigations with clofibrate showed no strain or sex differences; exposure for both compounds was decreased at the end of the study compared to day 1 for both clofibrate and p-cresidine.