Mechanisms of Promotion and Progression of Preneoplastic Lesions in Hepatocarcinogenesis by DDT in F344 Rats

Transcription

1 TOXICOLOGIC PATHOLOGY, vol 31, no 1, pp 87 98, 2003 Copyright C 2003 by the Society of Toxicologic Pathology DOI: / Mechanisms of Promotion and Progression of Preneoplastic Lesions in Hepatocarcinogenesis by DDT in F344 Rats TAKANORI HARADA, 1 SATORU YAMAGUCHI, 1 RYOICHI OHTSUKA, 1 MAKIO TAKEDA, 1 HIDEKI FUJISAWA, 1 TOSHINORI YOSHIDA, 1 AKIKO ENOMOTO, 1 YUKO CHIBA, 1 JUNKO FUKUMORI, 1 SAYURI KOJIMA, 1 NARUTO TOMIYAMA, 1 MACHIKO SAKA, 1 MASAKAZU OZAKI, 2 AND KEIZO MAITA 1 1 Institute of Environmental Toxicology, Mitsukaido-shi, Ibaraki , Japan, and 2 Sumika Technoservice Corporation, Osaka , Japan ABSTRACT Time-related changes in potential factors involved in hepatocarcinogenesis by DDT were investigated in a 4-week and a 2-year feeding studies of p,p -DDT with F344 rats. In the 4-week study with males at doses of 50, 160, and 500 ppm, cell proliferation and gap junctional intercellular communication (GJIC) were examined after 1, 2, 3, 7, 14, and 28 days. Cell proliferation was enhanced within 3 days at any dose level, but returned to normal after 7 days, whereas GJIC was inhibited throughout the study. In the 2-year study with both sexes at doses of 5, 50, and 500 ppm, cell proliferation, GJIC, enzyme induction, and oxidative stress were investigated after 26, 52, 78, and 104 weeks. Males and females showed an inhibition of GJIC and increases in P450 isozymes (CYP2B1 and CYP3A2) in a dose-dependent manner at all time points, but no significant change in cell proliferation. Lipid peroxide for males at 50 and 500 ppm and 8-hydroxydeoxyguanosine for both sexes at 500 ppm were elevated throughout the study. Histologically, eosinophilic foci and hepatocellular adenomas increased in males at 50 ppm and both sexes at 500 ppm. Hepatocellular carcinomas also developed in males at 500 ppm. These results indicate that DDT may induce eosinophilic foci as a result of oxidative DNA damage and leads them to neoplasms in combination with its mitogenic activity and inhibitory effect on GJIC. Oxidative stress could be a key factor in hepatocarcinogenesis by DDT. Keywords. DDT; hepatocarcinogenesis; cell proliferation; intercellular communication; enzyme induction; oxidative stress; eosinophilic foci; F344 rat. INTRODUCTION DDT is a nonsystemic contact and stomach insecticide with a broad spectrum of insecticidal activity that has been used primarily in the prevention of malaria, yellow fever, and sleeping sickness, and also in agriculture for the control of insects since 1940 (14). DDT is still used in developing countries for the control of malaria and other insect-transmitted diseases, although the use of this compound has been banned or restricted to certain areas in many countries since the 1970s because of its chemical characteristics such as accumulation and bioconcentration in lipid systems of all animal species that may result in occurrence of potential adverse effects on humans and wild animals. DDT and its metabolites have been studied extensively for their toxicity and carcinogenicity in experimental animals and shown to have a carcinogenic effect on the liver (14). DDT is classified as a nongenotoxicmitogenic hepatocarcinogen based on the mode of action (4) and has been found to induce microsomal enzymes (6, 23, 24, 31) and GGT-positive foci (6) and to inhibit gap junctional intercellular communication (GJIC) (17, 19, 20, 39) in the rodent liver. In consideration of these findings, we have undertaken further mechanistic studies to examine time-related changes in cell proliferation, GJIC, enzyme induction, oxidative stress, and histological alterations including altered hepatocellular foci (AHF) during the carcinogenic process in the liver of F344 rats treated with DDT. This paper describes Address correspondence to: Takanori Harada, Institute of Environmental Toxicology, 4321 Uchimoriya-machi, Mitsukaido-shi, Ibaraki , Japan. Fax: +81 (297) ; harada@iet.or.jp the time-related observations on these parameters, and potential factors that might be involved in the hepatocarcinogenesis by DDT are discussed with a special attention to promotion and progression of preneoplastic lesions. MATERIALS AND METHODS Test Chemical and Route of Exposure Dichlorodiphenyltrichloroethane [1,1,1-trichloro-2,2-bis- (p-chlorophenyl)ethane] (p,p -DDT, purity > 98%) was obtained from Tokyo Kasei Kogyo Co, Ltd (Tokyo, Japan) and used as the test chemical in the present study. The test compound was administered orally by incorporating it into a basal diet (a standard laboratory chow, Certified MF Mash, Oriental Yeast Co, Ltd, Tokyo, Japan). The test diet was supplied ad libitum with water. Preparation of test diet was conducted at a 4-week interval and concentration of the test chemical in the diet was monitored periodically. Animals, Housing, and Clinical Observations Specific-pathogen-free (SPF) Fischer (F344/DuCrj) rats of both sexes were purchased from Charles River Japan Inc (Kanagawa, Japan) at 4 weeks of age. They were housed 5 per cage in stainless steel wire-mesh cages and kept in an animal room controlled at: temperature, 24 ± 2 C; humidity, 55 ± 15%; ventilation, more than 10 times/hour; and illumination, 12-hour light/dark cycle. After a 1-week acclimatization to the testing environment, these animals were subjected to treatment at 5 weeks of age. All animals were handled in accordance with the Guidelines for Animal Experimentation by the Japanese Association for Laboratory Animal Science (18). They were observed daily for clinical signs and /03$3.00+$0.00

2 88 HARADA ET AL TOXICOLOGIC PATHOLOGY mortality. Body weights and food consumption were recorded periodically. Experimental Design Two experiments were conducted for the present study: a preliminary 4-week feeding study in males and a 2-year feeding study in both sexes. In the 4-week feeding study, 4 dose levels of 0, 50, 160, and 500 ppm were selected and each dose group consisted of 30 males. After 1, 2, 3, 7, 14, and 28 days of treatment, 5 animals from each group were sacrificed to examine cell proliferation and gap junctional intercellular communication (GJIC) in the liver. In the 2-year feeding study, 4 dose levels of 0, 5, 50, and 500 ppm were chosen and each dose group consisted of main (40 males and 40 females) and satellite (20 males and 20 females) groups. In the satellite group, 6 males and 6 females for each dose level after 26 and 52 weeks of treatment and all surviving animals after 78 weeks of treatment were sacrificed to examine cell proliferation, GJIC, microsomal enzyme induction, oxidative stress, and histopathology in the liver. In addition, concentrations of DDT and its metabolites, DDE [1,1-dichloro-2,2-bis(pchlorophenyl)ethylene] (p,p -DDE) and DDD [1,1-dichloro- 2,2-bis(p-chlorophenyl)ethane] (p,p -DDD), in the liver were measured on 3 animals/sex/group at each sampling time. In the main group, all surviving animals were sacrificed after 104 weeks of treatment and subjected to measurements of the same parameters as the satellite group. The liver tissues from animals found dead or sacrificed in extremis during the study were also examined histologically. Analysis of DDT and its Metabolites in the Liver Liver tissues (right lobes) from 3 males and 3 females for each dose level after 26, 52, 78, and 104 weeks of treatment in the 2-year feeding study were subjected to analyses of DDT and its metabolites, DDE and DDD. Concentrations of p,p - DDT, p,p -DDE, and p,p -DDD in the liver tissues were determined by high-performance liquid chromatographic (HPLC) method (36). Measurement of Cell Proliferation in the Liver All animals from each group sacrificed on schedule after 1, 2, 3, 7, 14, and 28 days of treatment in the 4-week feeding study and 6 males and 6 females from each group after 26, 52, 78, and 104 weeks of treatment in the 2-year feeding study were subjected to measurement of cell proliferation activity in the liver. The liver (left lobe) and duodenum tissues fixed in methanol were embedded in paraffin, and the paraffin sections were subjected to immunohistochemistry for proliferating cell nuclear antigen (PCNA). The duodenum tissue was used as positive control for PCNA staining. According to the guidelines for measurement of cell proliferation activity in the liver (9), approximately 1,000 hepatocytes for each animal were examined using an image analyzer, Luzex III (Nireko Corporation, Tokyo, Japan) and number of PCNApositive (S-phase) cells (7) was counted. PCNA labeling index (LI) was determined as follows: PCNA LI (%) = number of PCNA-positive cells/number of cells examined (about 1,000 hepatocytes) 100. Measurement of Gap Junctional Intercellular Communication (GJIC) in the Liver The same animals used for cell proliferation analysis were subjected to measurement of GJIC. Frozen sections of the liver tissue (left lobe) embedded in a Tissue Mount (Chiba Medical Co, Ltd, Saitama, Japan) for each animal were subjected to immunohistochemistry for hepatic gap junction protein connexin 32 (Cx32) using rat monoclonal Cx32 antibody (Zymed Laboratories, Inc, San Francisco, CA, USA). Number of Cx32 spots per hepatocyte was counted for evaluation of GJIC. The count of Cx32 spots was performed on 100 hepatocytes in the centrilobular area for each animal using an image analyzer, Luzex III. Measurement of Hepatic Drug-Metabolizing Enzymes Analyses of microsomal enzyme activity and cytochrome P450 isozyme contents were performed on 6 males and 6 females after 26, 52, 78, and 104 weeks of treatment in the 2-year feeding study. Liver tissues (median lobe) from these animals were perfused with 1.15% KCL solution to remove blood as much as possible and then homogenized with icecold 1.15% KCL using a homogenizer, Polytron (KINEMAT- ICA GmbH, Luzern, Switzerland). The liver homogenate from each animal was subjected to stepwise centrifugation to obtain microsomal fraction. The obtained microsomal pellet was suspended in 0.1M Na/K-phosphate buffer (ph 7.4) and then subjected to determination of pentoxyresorufin O- dealkylase (PROD) activity (22) and P450 isozyme contents (CYP1A1, CYP2B1, CYP3A2, and CYP4A1) by Western blot analysis using goat polyclonal antibodies (Dako Japan, Kyoto, Japan) against these CYPs. The concentration of each CYP was quantified with an imaging densitometer, Model GS-700 (Nippon Bio-Rad Laboratories, Tokyo, Japan). Measurement of Oxidative Stress Markers in the Liver Liver tissues (median lobe) from the same animals subjected to measurement of drug-metabolizing enzyme P450 were used to determine oxidative stress markers of lipid peroxide (LPO) and 8-hydroxydeoxyguanosine (8-OHdG). For measurement of LPO, the liver tissues were homogenized and then subjected to analysis of peroxide (malondialdehyde) by the TBA method (27) using a spectrophotometer (UV-2200, Shimadzu Corp, Kyoto, Japan). As for 8-OHdG, DNA was extracted from the liver tissues using DNA Extractor WB kit (Wako Pure Chemical Industries, Ltd, Osaka, Japan). The extracted DNA samples were dissolved in distilled water at a concentration of 100 µg/µl, heat-denatured at 95 C for TABLE 1. Average feed consumption and DDT intake during the study with F344 rats. Feed (g/rat/day) DDT (mg/kg/day) Study Male Female Male Female 4-Week Year

3 Vol. 31, No. 1, 2003 HEPATOCARCINOGENESIS BY DDT IN F344 RATS 89 TABLE 2. Mean concentrations of DDT and its metabolites in the liver of F344 rats from 2-year feeding study. DDT at weeks: DDE at weeks: DDD at weeks: Sex M <0.1 < < F < M F M F minutes, and cooled on ice. The samples (100 µl each) were mixed with 1.5 µl of 2M sodium acetate buffer (ph 4.5) and 45 µg of nuclease P1, and then digested at 37 C for 1 hour. Then, 18 µl of 1M Tris-HCl buffer (ph 7.5) and 3 units of phosphate alkaline type III added to each DNA solution were incubated at 37 C for 1 hour. Levels of 8-OHdG were determined on DNA aliquots (50 µl) of the samples by ELISA method (28) using an 8-OHdG Check Kit (Japan Institute for the Control of Aging, Shizuoka, Japan). Histopathological Examination with Morphometry The liver tissues fixed in 10% neutral-buffered formalin from all animals in the 2-year feeding study were embedded in paraffin and paraffin sections were stained with hematoxylin and eosin (H&E) for histopathological examination. Hepatocellular proliferative lesions were classified according to the Guides (G1-5) for Toxicologic Pathology (10). Altered hepatocellular foci (AHF) and hepatocellular tumors identified in H&E-stained sections were classified as eosinophilic, basophilic and clear cell foci, and hepatocellular adenomas and carcinomas, respectively. In addition, selected serial sections were stained with toluidine blue, cresyl violet, and periodic acid-schiff (PAS) reaction to further characterize the AHF and hepatocellular tumors. As for morphometry, 2-dimensional quantitative analysis of AHF (15, 16) was performed on H&E-stained liver sections from all animals at interim sacrifices and 10 animals/sex/group at terminal sacrifice in the 2-year feeding study. The animals at terminal sacrifice were selected to be free of mononuclear cell leukemia that might influence the occurrence of AHF (13). The number and area of each type of AHF and the total area of each liver section examined were measured using a color image processor, IPAP system (Sumika Technoservice Corp, Osaka, Japan) (37). In addition, the area fraction of liver occupied by AHF (mm 2 /cm 2 ) was calculated and expressed as percentage (%). Statistical Analysis The following statistical methods (8) were used to determine the significance of differences in the results between treated and control groups: Dunnett s multiple comparison test for PCNA LI, microsomal enzyme activity (PROD), protein contents of P450 isozymes, values of stress markers (LPO, 8-OHdG), and number and area fraction of AHF; Student s t-test or Aspin-Welch test for size (area) of AHF; Mann-Whitney s U-test for number of GJIC protein Cx32 spots; and Fisher s exact test for the incidence of histopathological lesions. RESULTS Clinical Sign, Mortality, and Body Weight In the 4-week feeding study, there were neither abnormalities in clinical signs nor deaths in any dose group. Mean body weights of all treated groups were comparable to controls during the study. In the 2-year feeding study, males and females in the highdose (500 ppm) group showed a whole body tremor in the late stage of treatment (weeks 70 to 104) that was more evident in females. However, there were no significant differences in mortality between treated and control groups. The mortality rates (%) in the 0, 5, 50, and 500 ppm groups at termination were 5/40 (12.5), 10/40 (25.0), 4/40 (10.0), 7/40 (17.5) for males and 7/40 (17.5), 13/40 (32.5), 8/40 (20.0), 7/40 (17.5) for females, respectively. Mean body weights of males and females in the high-dose group were reduced by 12% and 25% during the study when compared to controls, but those of other dose groups were comparable to controls. Feed Consumption and DDT Intake Average feed consumption and DDT intake in the 4-week and 2-year feeding studies are shown in Table 1. DDT intake (mg/kg/day) was calculated from the data on feed TABLE 3. PCNA LI (%) in the liver of male F344 rats from 4-week feeding study. PCNA LI (mean ± SD) on days: ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0.48 ( ): Number of rats examined.

4 90 HARADA ET AL TOXICOLOGIC PATHOLOGY TABLE 4. PCNA LI (%) in the liver of F344 rats from 2-year feeding study. PCNA LI (mean ± SD) at weeks: Sex M 0.70 ± ± ± ± 1.03 (6) (6) (6) (5) F 0.25 ± ± ± ± 0.32 (6) (6) (6) (5) 5 M 0.67 ± ± ± ± 0.19 (6) (6) (6) (5) F 0.13 ± ± ± ± 0.07 (6) (6) (6) (6) 50 M 0.77 ± ± ± ± 0.21 (6) (6) (6) (4) F 0.87 ± ± ± ± 0.25 (6) (6) (6) (6) 500 M 0.32 ± ± ± ± 0.48 (6) (6) (6) (6) F 0.45 ± ± ± ± 0.06 (6) (6) (6) (4) ( ): Number of rats examined. consumption, nominal dose level, and body weight. In the 2-year study, feed consumption by males at 500 ppm tended to be higher than that by controls during the study. DDT intake tended to be higher in females than males in the mg/kg/day basis. Concentrations of DDT, DDE, and DDD in the Liver In the 500 ppm group, the concentration of DDD was consistently higher in males than females throughout the study and those of DDT and DDE also showed a similar trend after 52 weeks of treatment and thereafter (Table 2). In other dose groups, however, no apparent trend was noted. Cell Proliferation in the Liver In the 4-week feeding study, significant increases in PCNA LI were observed in the 500 ppm group after 1, 2, and 3 days of treatment, 50 ppm group after 2 and 3 days, and 5 ppm group after 2 days (Table 3). However, there were no significant changes in PCNA LI in any dose group after 7 days of treatment and thereafter when compared to controls. In the 2-year feeding study, there were no significant changes in PCNA LI in any dose group of either sex throughout the study when compared to controls (Table 4). Gap Junctional Intercellular Communication (GJIC) In the 4-week feeding study, significant and dosedependent decreases in GJIC protein Cx32 were observed in the treated groups from the beginning (day 1) until the end (day 28) of the study (Table 5). In the 2-year feeding study, consistent and/or significant decreases in GJIC protein Cx32 were observed in the 50 and 500 ppm groups of both sexes during the study (Table 6). This change was dose-dependent. In the 5 ppm group, a significant decrease in GJIC protein Cx32 was noted in males after 78 weeks of treatment, but not at other time points. Hepatic Microsomal Enzyme Activity and P450 Isozyme Contents Significant increases in pentoxyresorufin O-dealkylase (PROD) activity were observed in males at 5, 50, and 500 ppm and females at 50 and 500 ppm (Tables 7 and 8). This change, however, was most evident in the 50 ppm group of both sexes and the increase in PROD was not significant in males at 500 ppm after 52, 78, and 104 weeks of treatment. With respect to P450 isozyme contents, significant and dose-dependent increases in CYP2B1 and CYP3A2 were observed in the 50 and 500 ppm groups of both sexes (Tables 7 and 8). The relative increased levels (treated group/control group) of CYP2B1 and CYP3A2 when compared to the respective controls tended to be higher in females than males. Statistically significant increases or decreases in CYP1A2 and CYP4A1 were noted in the treated groups of both sexes, but those changes were not consistent during the study. Oxidative Stress in the Liver Significant increases in hepatic lipid peroxide (LPO) contents were consistently observed in males at 50 and 500 ppm during the study (Table 9). In females, however, such changes were not consistent and significant increases in LPO were noted at 50 ppm after 26 and 104 weeks of treatment and at 500 ppm after 26 weeks. As for hepatic 8-OHdG levels, significant increases were observed only in the 500 ppm group of both sexes, which were more evident in males than females (Table 10). Histopathology of the Liver with Morphometry DDT-related changes were observed in various nonneoplastic and neoplastic hepatocellular lesions including centrilobular hepatocellular hypertrophy, altered hepatocellular foci (AHF) and hepatocellular tumors. Centrilobular hepatocellular hypertrophy (Figure 1) was observed in males at 5, 50, and 500 ppm and females at 50 and 500 ppm. The incidence and severity of the lesion were dose-dependent and more evident in males than females (Table 11). The AHF observed in the present study were morphologically classified as eosinophilic (Figure 2), tigroid basophilic TABLE 5. Number of GJIC protein Cx32 spots in the liver of male F344 rats from 4-week feeding study. No. of Cx32 spots (mean ± SD) on days: ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0.79 ( ): Number of rats examined.

5 Vol. 31, No. 1, 2003 HEPATOCARCINOGENESIS BY DDT IN F344 RATS 91 TABLE 6. Number of GJIC protein Cx32 spots in the liver of F344 rats from 2-year feeding study. No. of Cx32 spots (mean ± SD) at weeks: Sex M 7.07 ± ± ± ± 1.15 (6) (6) (6) (5) F 6.08 ± ± ± ± 0.23 (6) (6) (6) (5) 5 M 6.87 ± ± ± ± 0.37 (6) (6) (6) (5) F 5.67 ± ± ± ± 0.42 (6) (6) (6) (6) 50 M 5.53 ± ± ± ± 0.35 (6) (6) (6) (4) F 4.80 ± ± ± ± 0.33 (6) (6) (6) (6) 500 M 5.39 ± ± ± ± 0.28 (6) (6) (6) (6) F 4.20 ± ± ± ± 0.47 (6) (6) (6) (4) ( ): Number of rats examined. (Figure 3), and clear cell (Figure 4) foci based on general cytological features and tinctorial properties. No other specific types of AHF were noted. The eosinophilic AHF typically contained hepatocytes with pale pink or ground glass appearance cytoplasm (Figure 5) and sometimes had cytoplasmic clear spaces that may represent glycogen deposits. They tended to be located in the region close to or within the centrilobular hypertrophic area in the treated rats (Figures 2 and 5). The eosinophilic AHF were first noted as a small focus (Figure 2) in the 500 ppm group (males at week 26, females at week 52) and then gradually increased in size with duration of exposure. Large eosinophilic AHF often contained a small number of basophilic phenotypes in addition to eosinophilic cells (Figure 6). Eosinophilic AHF were also observed in control animals, but the incidence was low and its appearance was much later (first noted at week 78) as compared to the high-dose group. The incidence of eosinophilic AHF was significantly increased in the 50 and 500 ppm groups of both sexes and that in the 500 ppm group reached nearly 100% after 104 weeks of treatment (Table 11). In contrast, the incidences of clear cell and tigroid basophilic AHF were significantly decreased in males and females at 500 ppm, respectively (Table 11). Quantitative morphometrical evaluation of AHF revealed that the number, size (area), and area fraction of eosinophilic AHF increased in correlation with duration of exposure and dose levels (Tables 12 14). These changes were more evident in males than females and statistically significant increases in these parameters were noted in males at 50 ppm and both sexes at 500 ppm. In contrast to eosinophilic AHF, tigroid basophilic AHF decreased in number, size, and area fraction in females at 500 ppm. Clear cell AHF were also reduced in number in males at 500 ppm, but the size tended to be larger and the area fraction of liver occupied by clear cell AHF was comparable to controls. With respect to neoplasia, hepatocellular tumors were classified as adenomas and carcinomas. The incidence of hepatocellular adenomas was significantly increased in males at 50 and 500 ppm and females at 500 ppm (Table 15). The hepatocellular adenomas were first noted in the 500 ppm group of both sexes after 78 weeks of treatment (6/7 in males and 1/8 in females). They contained typically eosinophilic hepatocytes that were morphologically similar to those of eosinophilc AHF described previously and often had basophilic phenotypes in small numbers (Figure 7). Basophilic cell type of adenomas was also noted but only in a few animals. The overall incidences of hepatocellular adenomas in these treated groups during the study were 5/40 (12.5%) for males at 50 ppm, 22/40 (55%) for males at 500 ppm, and 16/40 (40%) for females at 500 ppm. In addition, hepatocellular carcinomas were observed in the 500 ppm group after 104 weeks of treatment and the incidence in males and females was 14/40 (35%) and 2/40 (5%), respectively. Morphologically, the hepatocellular carcinomas contained hepatocytes with eosinophilic or basophilic cytoplasm and typically had admixture of both phenotypes (Figure 8). The population of basophilic cells in the carcinomas was much higher than that in adenomas or large eosinophilic AHF. In other dose groups, no carcinoma was found in either sex. TABLE 7. Hepatic microsomal enzyme activity and P450 isozyme contents in male F344 rats from 2-year feeding study. (weeks) P450 isozyme contents (pmol/mg protein) PROD (pmol/min/mg) CYP1A2 CYP2B1 CYP3A2 CYP4A ± 2 (6) 4.24 ± 0.97 (6) 10.4 ± 1.0 (6) 72.3 ± 15.3 (6) 21.9 ± 2.1 (6) ± 2 (6) 3.22 ± 0.97 (6) 9.75 ± 1.02 (6) 83.4 ± 6.9 (6) 35.4 ± 2.8 (6) 78 5 ± 2 (6) 6.97 ± 1.42 (6) 11.4 ± 1.7 (6) 69.6 ± 16.3 (6) 40.9 ± 7.5 (6) ± 2 (6) 6.37 ± 1.90 (6) 12.4 ± 2.7 (6) 10.1 ± 5.9 (6) 17.5 ± 4.7 (6) ± 9 (6) 6.88 ± 1.39 (6) 66.7 ± 8.9 (6) 175 ± 23 (6) 19.4 ± 1.6 (6) ± 18 (6) 7.05 ± 1.15 (6) 61.9 ± 8.0 (6) 175 ± 18 (6) 36.3 ± 5.4 (6) ± 15 (6) 5.55 ± 1.50 (6) 74.9 ± 9.0 (6) 94.0 ± 36.0 (6) 39.1 ± 12.4 (6) ± 29 (6) 8.99 ± 4.39 (6) 76.2 ± 17.8 (6) 56.6 ± 31.6 (6) 18.4 ± 7.7 (6) ± 25 (6) 3.25 ± 0.70 (6) 366 ± 34 (6) 368 ± 34 (6) 35.8 ± 2.4 (6) ± 42 (6) 2.65 ± 1.19 (6) 398 ± 50 (6) 364 ± 56 (6) 44.4 ± 7.0 (6) ± 95 (6) 3.44 ± 3.06 (6) 172 ± 56 (6) 150 ± 82 (6) 42.0 ± 15.8 (6) ± 45 (6) 7.79 ± 4.04 (6) 179 ± 33 (6) 60.4 ± 16.4 (6) 20.4 ± 8.7 (6) ± 26 (6) 1.78 ± 0.36 (6) 607 ± 71 (6) 534 ± 50 (6) 36.5 ± 3.9 (6) ± 15 (6) 3.96 ± 0.87 (6) 476 ± 70 (6) 540 ± 92 (6) 33.4 ± 3.3 (6) ± 15 (6) 3.73 ± 2.64 (6) 222 ± 38 (6) 189 ± 74 (6) 30.4 ± 6.2 (6) ± 26 (6) 5.04 ± 1.62 (6) 271 ± 65 (6) 176 ± 42 (6) 18.8 ± 9.1 (6) ( ): Number of rats examined.

6 92 HARADA ET AL TOXICOLOGIC PATHOLOGY TABLE 8. Hepatic microsomal enzyme activity and P450 isozyme contents in female F344 rats from 2-year feeding study. (weeks) P450 isozyme contents (pmol/mg protein) PROD (pmol/min/mg) CYP1A2 CYP2B1 CYP3A2 CYP4A ± 1 (6) 16.8 ± 6.5 (6) 6.48 ± 0.94 (6) 10.7 ± 1.7 (6) 18.8 ± 2.5 (6) 52 8 ± 2 (6) 21.6 ± 6.2 (6) 3.81 ± 1.11 (6) 12.7 ± 2.1 (6) 14.4 ± 2.0 (6) 78 7 ± 1 (6) 10.9 ± 2.9 (6) 4.48 ± 0.93 (6) 12.1 ± 2.7 (6) 20.7 ± 5.1 (6) ± 3 (6) 17.9 ± 8.2 (6) 5.86 ± 0.77 (6) 7.98 ± 5.28 (6) 17.9 ± 5.4 (6) ± 16 (6) 22.6 ± 4.3 (6) 32.5 ± 3.6 (6) 16.4 ± 1.7 (6) 15.8 ± 2.2 (6) ± 26 (6) 29.6 ± 9.9 (6) 16.9 ± 4.6 (6) 25.8 ± 4.4 (6) 20.1 ± 1.3 (6) ± 13 (6) 12.0 ± 3.5 (6) 37.4 ± 3.7 (6) 29.2 ± 5.2 (6) 19.2 ± 4.7 (6) ± 18 (6) 17.9 ± 4.8 (6) 23.8 ± 6.3 (6) 28.6 ± 8.4 (6) 22.2 ± 6.5 (6) ± 74 (6) 23.6 ± 2.4 (6) 263 ± 39 (6) 53.5 ± 9.8 (6) 19.5 ± 2.9 (6) ± 44 (6) 31.3 ± 6.2 (6) 157 ± 17 (6) 82.3 ± 8.9 (6) 26.1 ± 3.0 (6) ± 23 (6) 11.5 ± 4.5 (6) 183 ± 39 (6) 80.8 ± 12.4 (6) 21.4 ± 0.5 (6) ± 39 (6) 14.5 ± 4.0 (6) 104 ± 25 (6) 63.1 ± 11.3 (6) 21.4 ± 1.7 (6) ± 44 (6) 20.6 ± 4.4 (6) 371 ± 46 (6) 171 ± 25 (6) 14.8 ± 4.3 (6) ± 30 (6) 31.0 ± 7.2 (6) 350 ± 29 (6) 155 ± 21 (6) 20.3 ± 2.9 (6) ± 10 (6) 18.8 ± 3.5 (6) 218 ± 29 (6) 171 ± 35 (6) 15.7 ± 3.5 (6) ± 25 (5) 20.6 ± 1.5 (5) 210 ± 34 (5) 189 ± 28 (5) 15.6 ± 4.8 (5) ( ): Number of rats examined. DISCUSSION The present study disclosed time-related changes in various factors that might be involved in hepatocarcinogenesis by DDT in F344 rats. Toxicokinetic data indicated that the concentrations of DDT and its metabolites, DDE and DDD, in the liver tended to be higher in males than females in the high dose group, although DDT intake (mg/kg/day) was higher in females. This difference might contribute in part to the sex difference in hepatic tumor development described later. PCNA labeling index (LI) in the liver from rats treated with DDT showed that cell proliferation was enhanced within 3 days of treatment but returned to normal thereafter. This cell proliferation pattern was consistent with that by nongenotoxic mitogenic hepatocarcinogens (4, 29, 30). It is generally known that the hepatic cell proliferation response to nongenotoxic mitogenic hepatocarcinogens typically occurs through an initial burst of enhanced DNA synthesis followed by enhanced mitosis (3, 29, 30). The enhanced cell proliferation, however, ceases after a few days even if treatment is continued. It is considered that an effective feedback mechanism (checkpoint function such as G1 or G2 arrest in cell cycle) may prevent excessive cell multiplication in the normal liver even if the growth stimulatory signals are steadily present due to continuous chemical treatment (30). It has been postulated that nongenotoxic chemicals with mitogenic activity TABLE 9. Hepatic lipid peroxide (LPO) contents in F344 rats from 2-year feeding study. Dose group LPO (nmol/g tissue) at weeks: No. of rats Sex examined M ± ± ± ± 29 F ± ± ± ± 53 5 M ± ± ± ± 122 F ± ± ± ± M ± ± ± ± 332 F ± ± ± ± M ± ± ± ± 68 F ± ± ± ± 101 may provide a selective growth advantage to spontaneously initiated precancerous cells over normal hepatocytes and lead them to neoplasms (4, 29). In the present study, however, eosinophilic AHF were not observed up to 78 weeks in the controls and at 2 lower dose levels of DDT, whereas an increased incidence of these AHF was seen as early as 26 weeks after starting the treatment with the highest dose of DDT. This result indicates that the occurrence of initiated cells could be accelerated by DDT in the high-dose group, which is considered to be due to hepatocellular DNA damage caused by oxidative stress described later. It is possible that the accelerated occurrence of initiated cells may result in the early appearance and increased incidence of eosinophilic AHF. In addition, the mitogenic activity of DDT may also contribute to the growth of initiated cells. Quantitative analysis of gap junctional intercellular communication (GJIC) in the liver from rats treated with DDT demonstrated a persistent decrease in GJIC protein Cx32 from the beginning to the end of treatment. GJIC in the liver has been shown to be inhibited by various nongenotoxic tumor-promoting agents including phenobarbital and DDT in vivo and in vitro (2, 6, 17, 19, 20, 38, 40). The inhibition of GJIC by tumor promoters may be produced in several ways (39). Because DDT is highly lipophilic and accumulates in cell membranes, it could interfere directly with the function of GJIC, whereas phenobarbital, which is not highly lipophilic, may inhibit GJIC in a different way. It TABLE 10. Hepatic 8-OHdG levels in F344 rats from 2-year feeding study. Dose group 8-OHdG (mg/mg DNA) at weeks: No. of rats Sex examined M ± ± ± ± 0.8 F ± ± ± ± M ± ± ± ± 0.7 F ± ± ± ± M ± ± ± ± 1.0 F ± ± ± ± M ± ± ± ± 9.9 F ± ± ± ± 0.9

7 Figures 1 4 FIGURE 1. Centrilobular hepatocellular hypertrophy in a male F344 rat treated with DDT at 500 ppm for 26 weeks. H&E Small eosinophilic focus located in the region close to the centrilobular hypertrophic area in a male F344 rat treated with DDT at 500 ppm for 26 weeks. H&E Tigroid basophilic foci with slightly tortuous hepatic cords in a female control F344 rat at interim sacrifice after 78 weeks. Note cytoplasmic basophilia in clumps or linear arrangements. H&E Clear cell focus in a male control F344 rat at interim sacrifice after 78 weeks. The clear cytoplasm represents area of glycogen dissolved in the aqueous fixative. H&E

8 Figures 5 8 FIGURE 5. Eosinophilic focus containing hepatocytes with pale pink or ground glass appearance cytoplasm that is located within the centrilobular hypertrophic area in a male F344 rat treated with DDT at 500 ppm for 78 weeks. H&E Large eosinophilic focus in a male F344 rat treated with DDT at 500 ppm for 104 weeks. A small number of basophilic cells are also seen within the lesion. H&E Hepatocellular adenoma comprising eosinophilic hepatocytes with pale pink or ground glass appearance cytoplasm in a male F344 rat treated with DDT at 500 ppm for 104 weeks. A small number of basophilic cells are also seen within the lesion. H&E Hepatocellular carcinoma containing both eosinophilic and basophilic phenotypes of hepatocytes with trabecular pattern in a male F344 rat treated with DDT at 500 ppm for 104 weeks. H&E

9 Vol. 31, No. 1, 2003 HEPATOCARCINOGENESIS BY DDT IN F344 RATS 95 TABLE 11. Incidence of hepatocellular hypertrophy and AHF in F344 rats from 2-year feeding study. Hypertrophy Eosinophilic AHF Tigroid basophilic AHF Clear cell AHF (weeks) Male Female Male Female Male Female Male Female /6 0/6 0/6 0/6 2/6 2/6 0/6 0/6 52 0/6 0/6 0/6 0/6 3/6 4/6 5/6 0/6 78 0/8 0/8 1/8 0/8 8/8 8/8 5/8 3/ /35 0/33 23/35 6/33 34/35 30/33 30/35 8/33 Total a 0/40 0/40 24/40 8/40 37/40 37/40 30/40 9/ /6 0/6 0/6 0/6 1/6 1/6 1/6 0/6 52 6/6 0/6 0/6 0/6 1/6 6/6 4/6 0/6 78 4/8 1/8 2/8 1/8 8/8 8/8 6/8 2/ /30 1/27 18/30 8/27 29/30 26/27 24/30 0/27 Total 15/40 1/40 21/40 10/40 35/40 38/40 24/40 0/ /6 6/6 0/6 0/6 1/6 3/6 1/6 0/6 52 6/6 6/6 0/6 0/6 0/6 6/6 4/6 0/6 78 8/8 7/7 8/8 1/7 8/8 7/7 8/8 0/ /36 31/32 30/36 16/32 34/36 30/32 27/36 8/32 Total 35/40 34/40 34/40 19/40 38/40 38/40 27/40 8/ /6 6/6 4/6 0/6 0/6 2/6 0/6 0/6 52 6/6 6/6 6/6 2/6 0/6 0/6 0/6 0/6 78 7/7 8/8 7/7 8/8 3/7 6/8 0/7 0/ /33 31/33 33/33 31/33 29/33 23/33 13/33 11/33 Total 38/40 37/40 38/40 37/40 34/40 27/40 14/40 11/40 a Total number of animals examined (No. of scheduled sacrifices after 104 weeks + No. of unscheduled deaths). TABLE 12. Number of AHF in F344 rats from 2-year feeding study. No. of AHF (No./cm 2 ): mean ± SD No. of rats examined Eosinophilic Tigroid basophilic Clear cell (weeks) Male Female Male Female Male Female Male Female ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 1.3 TABLE 13. Mean area of AHF in F344 rats from 2-year feeding study. Mean area of AHF (mm 2 ): mean ± SD (No. of rats with AHF) No. of rats examined Eosinophilic Tigroid basophilic Clear cell (weeks) Male Female Male Female Male Female Male Female ± (2) 0.03 ± (2) ± 0.05 (3) 0.13 ± 0.09 (4) 0.02 ± 0.02 (5) (1) 0.09 ± 0.03 (8) 0.11 ± 0.03 (8) 0.05 ± 0.02 (5) 0.04 ± 0.01 (3) ± 0.16 (8) 0.46 ± 0.18 (3) 0.16 ± 0.04 (9) 0.17 ± 0.08 (10) 0.08 ± 0.04 (9) 0.09 ± 0.08 (5) (1) 0.02 (1) 0.03 (1) (1) 0.09 ± 0.03 (6) 0.03 ± 0.01 (4) ± 0.63 (2) 0.04 (1) 0.09 ± 0.05 (8) 0.08 ± 0.03 (8) 0.05 ± 0.03 (6) 0.05 ± 0.02 (2) ± 0.57 (7) 0.38 ± 0.39 (4) 0.11 ± 0.04 (9) 0.14 ± 0.03 (10) 0.08 ± 0.04 (10) (1) 0.03 ± (3) 0.03 (1) ± 0.02 (6) 0.05 ± 0.02 (4) ± 0.17 (8) 0.32 (1) 0.25 ± 0.49 (8) 0.10 ± 0.04 (7) 0.09 ± 0.06 (8) ± 0.54 (9) 0.35 ± 0.21 (2) 0.11 ± 0.06 (9) 0.15 ± 0.03 (10) 0.18 ± 0.24 (10) 0.05 ± 0.03 (2) ± 0.03 (4) 0.07 ± 0.06 (2) ± 0.03 (6) 0.06 ± 0.02 (2) ± 0.18 (7) 0.26 ± 0.13 (8) 0.07 ± 0.02 (3) 0.09 ± 0.06 (6) ± 0.43 (10) 1.02 ± 0.70 (10) 0.17 ± 0.10 (10) 0.08 ± 0.02 (8) 0.41 ± 0.53 (4) 0.03 ± 0.01 (4) Not available because of no AHF.

10 96 HARADA ET AL TOXICOLOGIC PATHOLOGY TABLE 14. Area fraction of liver occupied by AHF in F344 rats from 2-year feeding study. Area fraction occupied by AHF (%): mean ± SD No. of rats examined Eosinophilic Tigroid basophilic Clear cell (weeks) Male Female Male Female Male Female Male Female ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0.03 is known that GJIC involves the passage of low molecular weight substances between adjacent cells via gap junctions and its function includes the possible regulation of cellular division through cell-to-cell exchange of replication signal molecules (19, 21, 40). Therefore, inhibition of GJIC may isolate initiated cells from the growth regulatory signals of neighboring cells and permit the clonal expansion of initiated cells. This suggests an important role for GJIC in the process of tumor formation. Analyses of hepatic microsomal enzymes revealed significant increases in pentoxyresorufin O-dealkylase (PROD) activity and P450 isozyme contents of CYP2B1 and CYP3A2 in the DDT-treated rats, which were consistent with previous works (6, 19, 20, 25). The increases in CYP2B1 and CYP3A2 were dose-dependent, whereas the elevation of PROD activity TABLE 15. Incidence of hepatocellular tumors in F344 rats from 2-year feeding study. Adenoma Carcinoma (weeks) Male Female Male Female /6 0/6 0/6 0/6 52 0/6 0/6 0/6 0/6 78 0/8 0/8 0/8 0/ /35 0/33 0/35 0/33 Total a 0/40 0/40 0/40 0/ /6 0/6 0/6 0/6 52 0/6 0/6 0/6 0/6 78 0/8 0/8 0/8 0/ /30 0/27 0/30 0/27 Total 0/40 0/40 0/40 0/ /6 0/6 0/6 0/6 52 0/6 0/6 0/6 0/6 78 0/8 0/7 0/8 0/ /36 0/32 0/36 0/32 Total 5/40 0/40 0/40 0/ /6 0/6 0/6 0/6 52 0/6 0/6 0/6 0/6 78 6/7 1/8 0/7 0/ /33 16/33 13/33 2/33 Total 22/40 16/40 14/40 2/40 a Total number of animals examined (No. of scheduled sacrifices after 104 weeks + No. of unscheduled deaths). was most evident in the mid-dose group and not significant in males of the high-dose group after 52, 78, and 104 weeks. This suggests that microsomal enzyme activity is not always consistent with its associated protein content and there seems to be an almost inverse correlation between the increase in PROD at different time points and the concurrent incidence of preneoplastic eosinophilic foci and hepatocellular tumors. As for P450 isozyme contents, the increases in CYP2B1 and CYP3A2 tended to be more evident in females than males. A similar result (preferential induction of CYP3A2 in females) also has been reported in Wistar rats treated with the technical grade DDT, a mixture of p,p -DDT (85%), o,p -DDT (15%) and o,o -DDT (trace amount) (31). Because CYP3A2 is androgen-dependent and not normally expressed in adult female rats (26), the induction of CYP3A2 by DDT in female rats suggests that DDT is able to modulate sexual metabolic dimorphism by affecting regulatory sites of hepatic metabolism (31). The preferential induction of CYP2B1 and CYP3A2 by DDT in female rats indicates an endocrine disrupting potential of DDT because these CYPs are involved in steroid metabolism. Measurements of hepatic oxidative stress markers disclosed significant increases in LPO in the mid- and high-dose groups and 8-OHdG in the high-dose group that were more evident in males than females. These results indicated that hepatocytes in the DDT-treated livers were exposed to oxidative stress and had cellular and DNA damages. It is postulated that the metabolic activation with enzyme induction of P450 by DDT may result in formation of reactive oxygen radicals (25, 32). The increased 8-OHdG level (an evidence of oxidative DNA damage) could play an important role in hepatocarcinogenesis by DDT. It has been shown that 8-OHdG leads to base mispairing (mutation) on DNA replication (5). Histologically, treatment with DDT induced centrilobular hepatocellular hypertrophy and increased eosinophilic AHF, hepatocellular adenomas and carcinomas. The severity of hepatocellular hypertrophy was dose-dependent and associated with microsomal enzyme induction. The eosinophilic AHF noted in the liver of DDT-treated rats tended to be located in the region close to or within the hypertrophic area.

11 Vol. 31, No. 1, 2003 HEPATOCARCINOGENESIS BY DDT IN F344 RATS 97 Such a spatial relationship was also reported in rats after continuous administration of low doses of N-nitrosomorpholine, which resulted in centrilobular perivenular hepatocellular hypertrophy, being closely related to the later development of preneoplastic hepatic foci including clear/acidophilic foci as well as hepatocellular neoplasms (33). The number and size of eosinophilic AHF were increased in correlation with duration of exposure and dose levels and its appearance was earlier in males than females. The incidence of eosinophilic AHF in the high-dose group after 104 weeks of treatment reached nearly 100% for both sexes. As the first appearance of eosinophilic AHF in the high-dose group (at week 26) was much earlier than that in the controls (at week 78), the eosinophilic preneoplastic lesions could be induced by DDT as a result of DNA damage of hepatocytes exposed to oxidative stress described previously. It is conceivable that DDT may accelerate the occurrence of initiated cells by oxidative stress, leading to the early appearance of eosinophilic AHF and promotes the growth of eosinophilic preneoplastic lesions through its mitogenic activity in combination with the inhibitory effect on GJIC as mentioned before. In other types of AHF, tigroid basophilic AHF decresed in number and size in females in the high dose group. A similar result was also reported in rats treated with phenobarbital (38). As to neoplasia, the overall incidences of hepatocellular adenomas and carcinomas in the high-dose group during the study were 55% and 35% for males and 40% and 5% for females, respectively. Because the hepatocytes within the majority of adenomas were morphologically similar to those of eosinophilic AHF and there was no increase in other types of AHF, it was suggested that the eosinophilic AHF could develop into neoplasms without passing an intermediate stage (1). The presence of basophilic cells in the large eosinophilic AHF and adenomas might be an indication of malignant transformation of hepatocytes within the lesions as the population of basophilic phenotypes was highest in the hepatocellular carcinomas (1). A similar type of nongenotoxic hepatocarcinogen, phenobarbital (PB), also has been shown to induce eosinophilic altered foci and hepatocellular tumors containing eosinophilic cells (GGT positive) (38). However, it should be recognized that the occurrence of eosinophilic AHF is not limited to PB and DDT because many other chemicals including genotoxic agents also induced such eosinophilic preneoplastic lesions (1, 12). With respect to sex difference in tumor development, it is generally believed that males have a higher incidence of hepatic tumors than females in rodents as well as in humans (35). Factors contributing to this sex difference have not been clearly demonstrated, but it may be due to the difference of hormonal pattern that is of primary importance for determining the metabolic activation of carcinogens and also due to the sex chromosome in its role as a carrier of genetic messages (35). In the present study, the incidence of hepatocellular tumors, especially carcinomas, was much higher in males than females. This sex difference might be partly due to the higher concentrations of DDT and its metabolites (DDE and DDD) in the liver of males that may result in higher production of oxidative stress through metabolic activation. In fact, the levels of LPO and 8-OHdG were much higher in males than females. In addition, the higher incidence of spontaneous eosinophilic foci in male F344 rats (11) may contribute to the sex difference. In other words, the male liver may have a preferable microenvironment for occrrence of eosinophilic AHF. The reduced body weight gain in the high-dose group that was more evident in females than males was also considered as an influential factor for tumor development, but it seems to have no correlation with the occurrence of hepatocellular tumors in F344 rats (34). In conclusion, the present study suggests that DDT may induce eosinophilic AHF as a result of oxidative DNA damage of hepatocytes and promote the progression of the preneoplastic lesions into hepatocellular tumors in combination with its mitogenic activity and inhibitory effect on GJIC. The production of oxidative stress (increases in LPO and 8-OHdG) that might be secondary to metabolic activation could be a key factor in hepatocarcinogenesis by DDT and may play an important role in tumor promotion and progression (malignant transformation). We are now conducting a further mechanistic study with microarray on the DDTinduced hepatocellular lesions including centrilobular hypertrophy, eosinophilic AHF, adenomas, and carcinomas. The results will be reported elsewhere. ACKNOWLEDGMENTS We thank Drs. Yasushi Taira, Yoshifumi Chazono, and Masahiko Harada in Sumika Technoservice Corporation, Osaka, Japan for their technical support to quantitative evaluation of AHF by IPAP system. We are also grateful to Drs. Shoji Teramoto, Koichi Ebino, Yasuhiro Kato, Eiko Nagayoshi, Yoshitsugu Odanaka, Yukiko Koma, Nobuaki Nakashima, Maki Kuwahara, and Yukiko Takeuchi, and Mss. Mutsumi Kumagai, Takako Kazami, and Yukari Tateyama in the Institute of Environmental Toxicology (IET) for their valuable suggestions, cooperation, and/or technical support in this work. REFERENCES 1. Bannasch P, Zerban H, Hacker HJ (1985). Foci of altered hepatocytes, rat. In: Monographs on Pathology of Laboratory Animals, Digestive System, Jones TC, Mohr U, Hunt RD (eds). Springer-Verlag, Berlin, pp Beer DG, Neveu MJ (1990). Proto-oncogene and gap-junction protein expression in rodent liver neoplasms. In: Mouse Liver Carcinogenesis: Mechanisms and Species Comparisons, Stevenson DE, McClain RM, Popp JA, Slaga TJ, Ward JM, Pitot HC (eds). Wiley-Liss, New York, pp Büsser M-T, Lutz WK (1987). Stimulation of DNA synthesis in rat and mouse liver by various tumor promoters. Carcinogenesis 8: Butterworth BE, Conolly RB, Morgan KT (1995). A strategy for establishing mode of action of chemical carcinogens as a guide for approaches to risk assessments. Cancer Lett 93: Clayson DB, Mehta R, Iverson F (1994). Oxidative DNA damage The effects of certain genotoxic and operationally non-genotoxic carcinogens. Mutation Res 317: Flodström S, Hemming H, Wärngård L, Ahlborg UG (1990). Promotion of altered hepatic foci development in rat liver, cytochrome P450 enzyme induction and inhibition of cell-cell communication by DDT and some structurally related organohalogen pesticides. Carcinogenesis 11: Foley J, Ton T, Maronpot R, Butterworth B, Goldsworthy TL (1993). Comparison of proliferating cell nuclear antigen to tritiated thymidine as a marker of proliferating hepatocytes in rats. Environ Health Perspect 101(Suppl 5): Gad SC, Weil CS (1994). Statistics for toxicologists. In: Principles and Methods of Toxicology, Hayes AW (ed). Raven Press, New York, pp