Inhibitory effect of geniposide on aflatoxin B1-induced DNA repair synthesis in primary cultured rat hepatocytes.

We have previously demonstrated that geniposide (GP) inhibits the aflatoxin B1 (AFB1) induced-hepatotoxicity and hepatic DNA binding in rats. To address the mechanism of action, the effects of GP on AFB1-induced DNA repair synthesis and AFB1 biotransformation in cultured rat hepatocytes were investigated. By evaluation of unscheduled DNA synthesis (UDS), GP reduced AFB1-induced DNA repair synthesis in a dose-dependent manner in hepatocyte cultures. GP elevates the metabolism of AFB1 and decreases the formation of AFM1. The enzyme activities of glutathione S-transferase (GST) and GSH-peroxidase (GSH-Px) in AFB1-treated hepatocyte cultures are enhanced in the presence of GP. GP reduces AFB1-induced DNA repair synthesis through an increased AFB1 detoxication metabolism. It provides one possible mechanism for the chemopreventive activity of GP.     

Enhancing effect of ethanol on aflatoxin b1-induced DNA-damage in glutathione-depleted rat hepatocytes

The effect of reduced glutathione (GSH) and ethanol on aflatoxin B1 (AFB1)-induced DNA single strand breaks was studied in primary cultured hepatocytes. Buthionine sulfoximine (BSO) which decreased intracellular GSH to 13% of those of the control levels increased DNA fragmentation of AFB1-treated hepatocytes by over 17% of those without BSO. Thus, a decrease in hepatocyte GSH levels increased AFB1-induced DNA damage. Although ethanol in itself did not induce DNA damage, a combination of BSO and ethanol increased the percentage by over 23% of that with BSO only. Ethanol did not affect the amount of GSH, total cytochrome P-450 (P450), glutathione S-transferase (GST) and epoxide hydrolase (EHase) in cultured hepatocytes. However, GSH-depleted rat hepatocytes exposed to ethanol significantly increased the level of P450IIIA, which activates AFB1. The enhancing effects of ethanol in the presence of BSO are probably due to the induction of this isozyme in rat hepatocytes. The GSH-depleted hepatocytes are more susceptible to chemical carcinogens in the presence of ethanol.     

Natural chlorophyll inhibits aflatoxin B1-induced multi-organ carcinogenesis in the rat

Chemoprevention by chlorophyll (Chl) was investigated in a rat multi-organ carcinogenesis model. Twenty-one male F344 rats in three gavage groups (N = 7 rats each) received five daily doses of 250 microg/kg [(3)H]-aflatoxin B(1) ([(3)H]-AFB(1)) alone, or with 250 mg/kg chlorophyllin (CHL), or an equimolar amount (300 mg/kg) of Chl. CHL and Chl reduced hepatic DNA adduction by 42% (P = 0.031) and 55% (P = 0.008), respectively, AFB(1)-albumin adducts by 65% (P < 0.001) and 71% (P < 0.001), respectively, and the major AFB-N(7)-guanine urinary adduct by 90% (P = 0.0047) and 92% (P = 0.0029), respectively. To explore mechanisms, fluorescence quenching experiments established formation of a non-covalent complex in vitro between AFB(1) and Chl (K(d) = 1.22 +/- 0.05 microM, stoichiometry = 1Chl:1AFB(1)) as well as CHL (K(d) = 3.05 +/- 0.04 microM; stoichiometry = 1CHL:1AFB(1)). The feces of CHL and Chl co-gavaged rats contained 137% (P = 0.0003) and 412% (P = 0.0048) more AFB(1) equivalents, respectively, than control feces, indicating CHL and Chl inhibited AFB(1) uptake. However, CHL or Chl treatment in vivo did not induce hepatic quinone reductase (NAD(P)H:quinone oxidoreductase) or glutathione S-transferase (GST) above control levels. These results are consistent with a mechanism involving complex-mediated reduction of carcinogen uptake, and do not support a role for phase II enzyme induction in vivo under these conditions. In a second study, 30 rats in three experimental groups were dosed as in study 1, but for 10 days. At 18 weeks, CHL and Chl had reduced the volume percent of liver occupied by GST placental form-positive foci by 74% (P < 0.001) and 77% (P < 0.001), respectively compared with control livers. CHL and Chl reduced the mean number of aberrant crypt foci per colon by 63% (P = 0.0026) and 75% (P = 0.0004), respectively. These results show Chl and CHL provide potent chemoprotection against early biochemical and late pathophysiological biomarkers of AFB(1) carcinogenesis in the rat liver and colon.     

Reactive oxygen species participate in mdr1b mRNA and P-glycoprotein overexpression in primary rat hepatocyte cultures

P-glycoproteins encoded by multidrug resistance type 1 (mdr1) genes mediate ATP-dependent efflux of numerous lipophilic xenobiotics, including several anticancer drugs, from cells. Overexpression of mdr1-type transporters in tumour cells contributes to a multidrug resistance phenotype. Several factors shown to induce mdr1 overexpression (UV irradiation, epidermal growth factor, tumour necrosis factor alpha, doxorubicin) have been associated with the generation of reactive oxygen species (ROS). In the present study, primary rat hepatocyte cultures that exhibit time-dependent overexpression of the mdr1b gene were used as a model system to investigate whether ROS might participate in the regulation of intrinsic mdr1b overexpression. Addition of H2O2 to the culture medium resulted in a significant increase in mdrlb mRNA and P-glycoprotein after 3 days of culture, with maximal (approximately 2-fold) induction being observed with 0.5-1 mM H2O2. Furthermore, H2O2 led to activation of poly(ADP-ribose) polymerase, a nuclear enzyme activated by DNA strand breaks, indicating that ROS reached the nuclear compartment. Thus, extracellularly applied H2O2 elicited intracellular effects. Treatment of rat hepatocytes with the catalase inhibitor 3-amino-1,2,4-triazole (2-4 mM for 72 h or 10 mM for 1 h following the hepatocyte attachment period) also led to an up-regulation of mdrlb mRNA and P-glycoprotein expression. Conversely, antioxidants (1 mM ascorbate, 10 mM mannitol, 2% dimethyl sulphoxide, 10 mM N-acetylcysteine) markedly suppressed intrinsic mdr1b mRNA and P-glycoprotein overexpression. Intracellular steady-state levels of the mdrl substrate rhodamine 123, determined as parameter of mdr1-type transport activity, indicated that mdr1-dependent efflux was increased in hepatocytes pretreated with H2O2 or aminotriazole and decreased in antioxidant-treated cells. The induction of mdr1b mRNA and of functionally active mdr1-type P-glycoproteins by elevation in intracellular ROS levels and the repression of intrinsic mdrlb mRNA and P-glycoprotein overexpression by antioxidant compounds support the conclusion that the expression of the mdr1b P-glycoprotein is regulated in a redox-sensitive manner.     

Iron-induced oxidative DNA damage and its repair in primary rat hepatocyte culture

Iron-overload diseases frequently develop hepatocellular carcinoma. The genotoxic mechanism whereby iron is involved in hepatocarcinogenesis might involve an oxidative process via the intermediate production of reactive oxygen species. This was presently investigated by examining kinetics of formation and repair of DNA base lesions in primary rat hepatocyte cultures supplemented with the iron chelate, ferric nitrilotriacetate Fe-NTA (10 and 100 microM). Seven DNA base oxidation products have been identified in DNA extracts by gas chromatography- mass spectrometry, which showed a predominance of oxidized-purines (8- oxo-guanine, xanthine, fapy-adenine, 2-oxo-adenine) above oxidized pyrimidines (5-OHMe-uracil, 5-OH-uracil, 5-OH-cytosine) in control cultures. All these DNA oxidation products revealed a significant dose- dependent increase at 4 to 48 h after Fe-NTA supplementation, among which fapy-adenine showed the highest increase and 5-OH-cytosine was the least prominent. Involvement of iron in this oxidative process was established by a correlation between extent in DNA oxidation and intracellular level of toxic low molecular weight iron. DNA excision- repair activity was estimated by release of DNA oxidation products in culture medium. All the seven DNA oxidation products were detected in the medium of control cultures and showed basal repair activity. This DNA repair activity was increased in a time- and dose-dependent fashion with Fe-NTA. Oxidized-pyrimidines, among which was 5-OHMe-Uracil, were preferentially repaired, which explains the low levels detected in oxidized DNA. Since oxidized bases substantially differed from one another in terms of excision rates from cellular DNA, specific excision- repair enzymes might be involved. Our findings, however, demonstrate that even though DNA repair pathways were activated in iron-loaded hepatocyte cultures, these processes were not stimulated enough to prevent an accumulation of highly mutagenic DNA oxidative products in genomic DNA. The resulting genotoxic effect of Fe-NTA might be relevant in understanding the hepatocarcinogenic evolution of iron-overload diseases.      

Dietary carotenoids inhibit aflatoxin B1-induced liver preneoplastic foci and DNA damage in the rat: role of the modulation of aflatoxin B1 metabolism

To study the effects of carotenoids on the initiation of liver carcinogenesis by aflatoxin B1 (AFB1), male weanling rats were fed beta- carotene, beta-apo-8'-carotenal, canthaxanthin, astaxanthin or lycopene (300 mg/kg diet), or an excess of vitamin A (21000 RE/kg diet), or were injected i.p. with 3-methylcholanthrene (3-MC) (6 x 20 mg/kg body wt) before and during i.p. treatment with AFB1 (2 x 1 mg/kg body wt). The rats were later submitted to 2-acetylaminofluorene treatment and partial hepatectomy, and placental glutathione S-transferase-positive liver foci were detected and quantified. The in vivo effects of carotenoids or of 3-MC on AFB1-induced liver DNA damage were evaluated using different endpoints: liver DNA single-strand breaks (SSB) induced by AFB1, and in vivo binding of [3H]AFB1 to liver DNA and plasma albumin. Finally, the modulation of AFB1 metabolism by carotenoids or by 3-MC was investigated in vitro by incubating [14C]AFB1 with liver microsomes from rats that had been fed with carotenoids or treated by 3- MC, and the metabolites formed by HPLC were analyzed. In contrast to lycopene or to an excess of vitamin A, both of which had no effect, beta-carotene, beta-apo-8'carotenal, astaxanthin and canthaxanthin, as well as 3-MC, were very efficient in reducing the number and the size of liver preneoplastic foci. In a similar way as 3-MC, the P4501A- inducer carotenoids, beta-apo-8'-carotenal astaxanthin and canthaxanthin, decreased in vivo AFB1-induced DNA SSB and the binding of AFB1 to liver DNA and plasma albumin, and increased in vitro AFB1 metabolism to aflatoxin M1, a less genotoxic metabolite. It is concluded that these carotenoids exert their protective effect through the deviation of AFB1 metabolism towards detoxication pathways. In contrast, beta-carotene did not protect hepatic DNA from AFB1-induced alterations, and caused only minor changes of AFB1 metabolism: seemingly, its protective effect against the initiation of liver preneoplastic foci by AFB1 is mediated by other mechanisms.      

Sequential histological and histochemical study of the rat liver during aflatoxin B1-induced carcinogenesis.

Male Wistar rats were given 50 mug of aflatoxin B1 twice a week for 4 weeks, and thereafter 75 mug twice a week for 10 weeks. Their livers were investigated histologically and histochemically for glycogen, RNA, fat, alkaline and acid phosphatases, adenosine triphosphatase, 5'-nucleotidase, glucose-6-phosphatase, glucose-6-phosphate dehydrogenase, succinic dehydrogenase, and alkaline and acid nucleases. No significant lesions occurred before 15 weeks. During this period, the liver was histochemically unchanged except for a periportal decrease of alkaline phosphatase and adenosine triphosphatase. Scattered hepatocytes with a strong glucose-6-phosphatase activity appeared. These changes represent toxic effects of aflatoxin B1 and are irrelevant to carcinogenesis. From 15 weeks onward, three types of liver cell hyperplastic foci and nodules developed. Histologically, and with respect to glycogen, fat, and RNA content, only two of these types were considered as potential precursors of hepatocarcinomas. However, all types exhibited a decrease or absence of the enzymes studied. Both histological and histochemical changes stressed the complex heterogeneity existing between and within hepatic foci and nodules. From 11 months on, hepatocarcinomas developed. The tumors disclosed similar histochemical changes. This similarity further supports the "precarcinomatous" nature of hyperplastic foci and nodules. It appears that focal changes in surface as well as in cytoplasmic and nuclear enzymes are intimately and very early linked to the carcinogenic process. Whether they are fundamental or only represent an epiphenomenon remains unclear.     

Testing for induction of DNA synthesis in human hepatocyte primary cultures by rat liver tumor promoters.

Parenchymal liver cells were isolated from human liver pieces of surgical waste as well as from rat livers. DNA synthetic activity was measured after different times in primary culture by [3H]thymidine incorporation and autoradiography. Labeling of control cultures of human hepatocytes at densities between 8,000 and 15,000 cells/cm2 was very low (0.4 to 1.3%). Human recombinant epidermal growth factor increased labeling 2- to 4-fold (P less than 0.01). Treatment with known inducers of liver growth in rats, namely, cyproterone acetate, alpha-hexachlorocyclohexane, nafenopin, phenobarbital, and rifampicin did not increase the number of labeled human liver cells. In some of the experiments, a 24-h exposure to the chemicals of rat or human hepatocytes was followed by a 24-h treatment with epidermal growth factor (EGF). In rat hepatocytes, incorporation rates were significantly increased. Cyproterone acetate and EGF acted in an additive manner, alpha-hexachlorocyclohexane and EGF were clearly overadditive, and phenobarbital had little effect. In human hepatocytes, little alteration in labeling indices was found; in some cases labeling was, rather, found to be lower than in cultures treated with EGF alone. These results show that human hepatocytes cultured in vitro are sensitive to stimulation of DNA synthesis by EGF; they differ from rat hepatocytes in their response to some drugs which show liver growth-promoting activity in rodents.     

Suppression of arachidonic acid cascade-mediated apoptosis in aflatoxin B1-induced rat hepatoma cells by glucocorticoids

It has been shown that hypophysectomy protects aflatoxin B1 (AFB1) hepatocarcinogenesis and the prevention of apoptosis is a critical process for tumorigenesis. In this paper, we analyzed the cell death of AFB1-induced rat hepatoma Kagura-2 (K2) cells elicited by an estrogen antagonist, tamoxifen (TAM), and transforming growth factor-beta1 (TGF- beta1) to elucidate the function of endocrine factors in AFB1 hepatocarcinogenesis. TAM and TGF-beta1 induced a typical apoptosis in K2 cells. The apoptotic cell death was efficiently suppressed by glucocorticoids (GCs), but not by other steroid compounds including 17beta-estradiol (E2). Cyclo-oxygenase (COX) inhibitors such as aspirin (ASP) and indomethacin (IND) also inhibited the apoptosis, while inhibitory effects of general lipoxygenase (LOX) inhibitors such as nordihydroguaiaretic acid (NDGA) and 5,8,11-eicosatrienoic acid (ETI) were not observed. TAM and TGF-beta1 enhanced the release of [3H]arachidonic acid (AA) from pre-labeled K2 cells, which was inhibited by dexamethasone (DEX). Furthermore, cytosolic phospholipase A2 (cPLA2) activity in K2 cells treated with TAM for 2 h was higher than that in the control. Prostaglandin J2 (PGJ2) and delta12-PGJ2, AA metabolites formed in the COX pathway, induced K2 cell death. These results suggest that AA metabolites are involved in apoptotic K2 cell death elicited by TAM and TGF-beta1, and GCs could act as a tumor promoter in AFB1 hepatocarcinogenesis through the prevention of apoptosis induced by AA metabolites formed in vivo.      

The binding of aflatoxin B1 to rat liver nuclear proteins and its effect on DNA-dependent RNA synthesis

This paper reports studies on the binding of aflatoxin B1 (AFB1) to rat liver nuclear proteins in vivo and in vitro, and its effect on RNA synthesis. Two hours after rats (200 g) were given a single i.p. injection of 300 micrograms AFB1 containing 50 microCi [3H]AFB1/100 g body wt, AFB1 was found bound to the free nuclear proteins (29.7 pmol/mg protein), histones (20.3 pmol/mg protein) and chromatin-bound non-histone proteins (13.8 pmol/mg protein). The binding of AFB1 to histones was further studied in vitro. We found that for a given type of histone, the binding level varied greatly depending on the conditions used. Under both in vivo and in vitro conditions, however, H3 was always the most efficient substrate, and H4/H2B always the least efficient substrates for AFB1 binding. These results suggest that the binding preference was mainly related to the intrinsic properties of the histone type, and was little affected by the geometric arrangement of the histones in chromatin. Using nuclear proteins added to the RNA synthesizing system in vitro, we found that only the histone fraction had a strong inhibitory effect. Further studies, however, indicated that this inhibition was not due to histones per se, but rather to poly-ADP-ribosylated histones present in the histone preparations. No detectable difference in effect was found between control and AFB1-bound nuclear proteins on RNA synthesis. Moreover, higher levels of AFB1 binding to histones did not potentiate the inhibitory effect. We therefore conclude, and in direct support to our previous correlation studies (see the preceding paper), that the binding of AFB1 to nuclear proteins has no inhibitory effect on RNA synthesis.     

Suppressive effect of geniposide on the hepatotoxicity and hepatic DNA binding of aflatoxin B1 in rats.

The effects of geniposide pretreatment on both hepatic aflatoxin B1 (AFB1)-DNA binding and AFB1 hepatotoxicity in rats has been examined. For these studies, male Sprague-Dawley rats were treated with AFB1 (2 mg/kg) by i.p. administration, and the different degrees of hepatic damage were revealed by the elevations of levels of serum marker enzymes such as aspartate aminotransferase (AST), alanine amino-transferase (ALT) and gamma-glutamyltranspeptidase (gamma-GT). After pretreatment of animals with geniposide (10 mg/kg) daily for 3 consecutive days, the enzyme elevations were significantly suppressed. This suggested that the geniposide possessed chemopreventive effects on the early acute hepatic damage induced by AFB1. Under these experimental conditions, consistent elevation of the activities of glutathione S-transferase (GST) and gamma-glutamylcysteine synthetase but not glutathione peroxidase (GSH-Px) and gamma-glutamyltranspeptidase were observed. Treatment of rats with geniposide significantly lowered hepatic GSH and GSSG levels, but the ratio of GSH to GSSG was not changed. Geniposide treatment also decreased AFB1-DNA adduct formation in AFB1-treated animals. From these results, we suggest that the protective effect of geniposide on AFB1 hepatotoxicity in rats might be due to the hepatic tissues' defense mechanisms that involve the enhanced GST activity for AFB1 detoxication and induction gamma-glutamylcysteine synthetase for GSH biosynthesis.     

Influence of dietary fat and selenium in initiation and promotion of aflatoxin B1-induced preneoplastic foci in rat liver

Aflatoxin B₁-induced hepatic -glutamyl transpeptidase-positivefoci were quantified in rats fed different levels of fat andselenium during either initiation or early promotion. Male Sprague-Dawleyrats were divided into 12 groups. One of six experimental dietswere fed to groups 1–6 prior to and during aflatoxin B₁exposure (initiation, weeks 1–4.5) and to groups 7–12during weeks 4.5–15 (promotion). The six experimentaldiets contained 2 or 20% corn oil, each with either <0.02,0.15 or 2.5 (or 1.9) p.p.m. selenium. When not fed the experimentaldiets, rats were fed a modified MN 76A diet. In groups 1–6,0.03% phenobarbital was added as a promoter to the AIN-76A diet.Individual and interactive effects of selenium and fat weredependent on the stage of carcinogenesis. High dietary fat fedwith either <0.02 or 0.15 p.p.m. selenium during initiationresulted in a significant increase in the number and size offoci when compared with low fat groups. In rats fed 20% fatand 2.5 p.p.m. selenium during initiation, preneoplastic developmentwas reduced below all low fat groups. In contrast, seleniumstatus but not dietary fat level influenced focal formationduring promotion. Rats fed <0.02 p.p.m. selenium had a significantlygreater percentage of liver section occupied by foci than ratsfed either 0.15 or 1.9 p.p.m. seleniwn. Feeding 1.9 p.p.m. seleniumduring promotion did not afford greater protection above the0.15 p.p.m. level. Hepatic glutathione peroxidase activity atweek 15 was significantly diminished in animals fed <0.02p.p.m. selenium during promotion. Feeding 1.9 p.p.m. seleniumwhen compared with 0.15 p.p.m. did not result in a consistentincrease in enzyme activity. Although differences were observedin growth due to dietary treatment, there were no significantcorrelations between preneoplastic foci and body weight, foodconsumption or food effidency. These findings indicate an interactionbetween dietary fat and selenium during initiation, but notduring early promotion. Furthermore, dietary selenium and fatmay function by different mechan isms at different stages ofcarcinogenesis.     

Susceptibility to aflatoxin B1-induced carcinogenesis correlates with tissue-specific differences in DNA repair activity in mouse and in rat

To investigate the mechanisms responsible for species- and tissue-specific differences in susceptibility to aflatoxin B(1) (AFB(1))-induced carcinogenesis, DNA repair activities of nuclear extracts from whole mouse lung and liver and rat liver were compared, and the ability of in vivo treatment of mice with AFB(1) to alter repair of AFB(1)-DNA damage was determined. Plasmid DNA containing AFB(1)-N(7)-guanine or AFB(1)-formamidopyrimidine adducts were used as substrates for the in vitro determination of DNA repair synthesis activity, detected as incorporation of radiolabeled nucleotides. Liver extracts from CD-1 mice repaired AFB(1)-N(7)-guanine and AFB(1)-formamidopyrimidine adducts 5- and 30-fold more effectively than did mouse lung, and approximately 6- and 4-fold more effectively than did liver extracts from Sprague-Dawley rats. The susceptibility of mouse lung and rat liver to AFB(1)-induced carcinogenesis correlated with lower DNA repair activity of these tissues relative to mouse liver. Lung extracts prepared from mice treated with a single tumorigenic dose of 50 mg/kg AFB(1) i.p. and euthanized 2 hours post-dosing showed minimal incision and repair synthesis activities relative to extracts from vehicle-treated mice. Conversely, repair activity towards AFB(1)-N(7)-guanine damage was approximately 3.5-fold higher in liver of AFB(1)-treated mice relative to control. This is the first study to show that in vivo treatment with AFB(1) can lead to a tissue-specific induction in DNA repair. The results suggest that lower DNA repair activity, sensitivity of mouse lung to inhibition by AFB(1), and selective induction of repair in liver contribute to the susceptibility of mice to AFB(1)-induced lung tumorigenesis relative to hepatocarcinogenesis.     

Evidence for an indirect mechanism of aflatoxin B1 inhibition of rat liver nuclear RNA polymerase II activity in vivo

Previous studies suggested multiple sites of action of aflatox-inB₁ (AFB₁) in vivo to inhibit rat liver nuclear RNA synthesis- it impairs nucleolar DNA template function and inhibits RNAPolymerase II activity. We have previously shown that AFB₁ activatedin vitro inhibits nucleolar RNA synthesis. The question is whetherAFB₁ can inhibit RNA polymerase II under these in vitro conditions.Male Sprague-Dawley rats, 200 g, were injected i.p. with 0.6mg AFB₁ and liver nuclei were isolated 2 h later. When the totalnuclear free RNA polymerases were extracted and assayed in theabsence and presence of a-amanitin (3.2 µg/ml), we foundthat only -amanitin-sensitive (i.e., RNA polymerase II) activitywas inhibited (97%). DEAE-Sephadex chromatography confirmedthis result. When total nuclear free RNA polymerases were incubatedwith AFB₁ activated in vitro under conditions producing 70%inhibition of nucleolar RNA synthesis, no inhibition was observedfor either -amanitin-sensitive or -resistant activities. Similarresults were obtained with low and high (28 and 167 µg/ml)concentrations of AFB₁. This was further confirmed using highlypurified RNA polymerase II. We conclude that AFB₁ inhibitionof RNA polymerase II activity in vivo is not a result of directinteraction of AFB₁ to the enzyme.     

Effect of high and low dietary protein on the dosing and postdosing periods of aflatoxin B1-induced hepatic preneoplastic lesion development in the rat

Aflatoxin B1-induced liver lesion development is readily modified by dietary protein intake. Earlier work had shown that low-protein diets enhanced the acutely toxic lesion but depressed the carcinogenic lesion. This study examined the emergence of these lesions as a function of dietary protein intake, particularly with respect to whether the protein modification occurred during or after the aflatoxin B12 dosing period. High (20%) and low (5%) casein diets were fed to growing Fischer 344 rats during the dosing and postdosing periods of aflatoxin B2-induced hepatic preneoplastic lesion development. Focal areas of hepatocellular alteration were identified and quantitated by staining sections of liver for gamma-glutamyl transferase (GGT). Animals fed low casein diets during the dosing period displayed a characteristic spectrum of lesions including hepatomegaly, severe bile duct proliferation, cholangiofibrosis, and a tendency for developing large remodeling GGT-positive foci. These lesions were regarded as symptomatic of acute hepatoxicity. Animals fed high-protein diets during the dosing period had small, densely stained, GGT-positive foci, with only mild bile duct proliferation and no cholangiofibrosis, hepatomegaly, or large, remodeling GGT-positive foci. During the postdosing period, protein modulation markedly influenced the total number of foci. Animals fed high casein diets during this period exhibited an approximate 6-fold increase in the number of foci, regardless of the level of protein fed during the earlier dosing period. The marked increase in foci number (as well as area of liver occupied) in high casein diet animals during the postdosing period is regarded as an increased tendency for tumor development.