Search

Development of TAK-875 was discontinued when a small number of serious drug-induced liver injury (DILI) cases were observed in Phase 3 clinical trials. Subsequent studies have identified hepatocellular oxidative stress, mitochondrial dysfunction, altered bile acid homeostasis, and immune response as mechanisms of TAK-875 DILI and the contribution of genetic risk factors in oxidative response and mitochondrial pathways to the toxicity susceptibility observed in patients. We tested the hypothesis that a novel preclinical approach based on gene pathway analysis in the livers of Collaborative Cross mice could be used to identify human-relevant mechanisms of toxicity and genetic risk factors at the level of the hepatocyte as reported in a human genome-wide association study. Eight (8) male mice (4 matched pairs) from each of 45 Collaborative Cross lines were treated with a single oral (gavage) dose of either vehicle or 600 mg/kg TAK-875. As expected, liver injury was not detected histologically and few changes in plasma biomarkers of hepatotoxicity were observed. However, gene expression profiling in the liver identified hundreds of transcripts responsive to TAK-875 treatment across all strains reflecting alterations in immune response and bile acid homeostasis and the interaction of treatment and strain reflecting oxidative stress and mitochondrial dysfunction. Fold-change expression values were then used to develop pathway-based phenotypes for genetic mapping which identified candidate risk factor genes for TAK-875 toxicity susceptibility at the level of the hepatocyte. Taken together, these findings support our hypothesis that a gene pathway-based approach using Collaborative Cross mice could inform sensitive strains, human-relevant mechanisms of toxicity, and genetic risk factors for TAK-875 DILI. This novel preclinical approach may be helpful in understanding, predicting, and ultimately preventing clinical DILI for other drugs.

Administration of tamoxifen (TAM) has been shown to induce hepatocellular carcinogenesis and TAM-DNA adduct formation in rat liver. Here we present TAM-DNA adduct localization and semi-quantitation in hepatic tissue of rats by immunohistochemical staining followed by image analysis. We have also used a quantitative immunoassay to provide a validation for the immunohistochemical values. Rats were fed diets containing 0, 5, 50, 150 or 500 p.p.m. TAM for 45 weeks. Serial sections of paraffin-embedded liver were stained for TAM-DNA adducts using a polyclonal TAM-DNA antiserum. Subsequently, visualization of TAM-DNA adducts was performed by peroxidase-conjugated secondary antibody-mediated signal amplification using biotinyl tyramide followed by streptavidin-alkaline phosphatase and fast red. Semi-quantitation of nuclear color intensity was achieved with an Automated Cellular Imaging System (ACIS), with a detection limit of 1 TAM-DNA adduct per 10(7) nt for these experiments. In parenchymal cells of liver sections from TAM-exposed animals a dose-dependent increase in nuclear staining was observed by ACIS and the TAM-DNA adduct levels determined by ACIS were validated in liver DNA by quantitative chemiluminescence immunoassay (CIA). Comparison of semi-quantitative values determined by ACIS with quantitative values determined by CIA showed a strong correlation (r = 0.924) between the two methods. At 45 weeks of TAM exposure the liver cytoplasm contained placental glutathione S-transferase (GST-p)-positive foci, as indicated by new fuchsin staining. Staining of serial sections revealed a relative lack of TAM-DNA adducts within these enzyme-altered foci. In addition, some GST-p foci contained islands of cells that did not stain for GST-p but were positive for TAM-DNA adduct formation. This study validates the use of ACIS for TAM-DNA adduct formation and demonstrates that steady-state TAM-DNA adduct levels observed in livers of rats chronically fed TAM for several months increase in relation to dose. In addition, unlike the normal surrounding liver, preneoplastic GST-p-positive foci have virtually no TAM-DNA adducts.

Neoplastic development is a multistep process that involves the stochastic accumulation of heritable genetic alterations in proto-oncogenes, DNA repair genes, and tumor suppressor genes. Loss of heterozygosity (LOH) analysis has been used successfully to identify the genetic determinants of neoplastic development, including tumor suppressor genes, in several species and organs but not in the rat liver. We report the results of a sensitive genome-wide LOH analysis of rat hepatocellular carcinomas (HCCs). Heterozygous rats (Wistar-Furth x Fisher 344) were subjected to an Initiation-Promotion-Progression (IPP) protocol of hepatocarcinogenesis. Two weeks after initiation (by partial hepatectomy, 10 mg/kg diethylnitrosamine), the rats were placed on a diet containing 0.05% phenobarbital (PB). After 24 wk of PB promotion, the rats received either 100 or 1 50 mg/kg ethylnitrosourea. Hepatocellular tumors were resected after a total of 76wk of PB promotion. LOH analysis was completed on 26 HCCs by using 60 microsatellite markers covering all 20 rat autosomes and chromosome X. While 85% of the HCCs had one or more allelic imbalances, the average HCC had 3.3 allelic imbalances (range 0-9). A conditional hypothesis-testing method called the Hot-Cold model was used to determine the location of statistically significant elevations in the frequency of allelic imbalances. Elevated allelic imbalances were observed on chromosomes 1q, 6, 8, 11, 15, 17, and 20p. Together, these allelic imbalances suggest that the retinoblastoma and insulin-like growth factor genes as well as the resistance to chemical carcinogenesis (rcc) locus may be involved in HCC development in the rat but that LOH of the p53 gene is not. The elevated rate of allelic imbalances on chromosomes 8,11, and 17 may indicate the location of undiscovered tumor suppressor genes important to neoplastic development in rat liver. Microdissection-based LOH analysis of HCC revealed that contamination of non-neoplastic and nonhepatocellular tissue was not masking LOH in the whole-tumor analysis. There were no statistically significant differences in the frequency of allelic imbalances between HCC of any differentiation state (histological grade). To the degree that it does not reflect differences in etiological factors, the absence of allelic imbalances in chromosomal regions containing the p53 and mamose-6-phosphate/insulin-like growth factor II receptor tumor suppressor genes and the generally low frequency of allelic imbalances in these tumors, suggests that LOH and allelic imbalances play a less significant role in the molecular pathogenesis of HCC in rats than humans.

Dioxin and certain structurally related compounds increase the incidence of liver neoplasms in rodents upon chronic bioassay. Short-term studies indicate the lack of direct DNA-damaging effects including covalent binding to DNA; however, secondary mechanisms may be important in the observed carcinogenicity as these chemicals affect a number of pathways necessary for maintenance of normal growth control and differentiation status. Studies with TCDD in the mouse skin support a lack of initiating activity but an ability to promote the growth of previously initiated lesions indicative of a promoting agent. Mouse skin tumour promotion studies indicate that Ah receptor activation may be involved in promotion by TCDD and selected structurally related compounds. While the mechanism of carcinogenicity induced by TCDD is unknown, the processes involved have a no-effect level, which in the rat liver is at an exposure level below 10 ng TCDD/kg/day. At least for the rodent liver, the relative effective dose for cytochrome P450 induction is not a good indicator of promotion potency. Studies on liver tumour promotion in the female rat liver support a non-genotoxic mechanism for the induction of neoplasms by TCDD. The ability of TCDD to enhance proliferation and inhibit apoptotic processes in focal hepatic lesions further supports an indirect mechanism of carcinogenicity.

An albumin-simian virus 40 (SV40) large T-antigen (T-Ag) transgenic model and a chemically induced model of multistage hepatocarcinogenesis were created in our laboratory to study the molecular mechanisms involved in the genesis and progression of neoplasia in the rat liver. In the study presented here, these two models of rat hepatocarcinogenesis were used to perform a comparative mutational analysis of three tumor suppressor genes involved in hepatic neoplastic growth. By using polymerase chain reaction-single strand conformation polymorphism analysis and sequencing, exons 5-8 of the p53 tumor suppressor gene and a region between nt 4325 and 4479 of the rat mannose 6-phosphate/insulin-like growth factor 2 receptor (M6p/Igf2r) coding sequence were screened. The latter is homologous to the human M6P/IGF2r coding sequence which is mutated in human hepatocellular carcinoma. A complete single strand conformation polymorphism analysis of the entire coding region of the rat adenomatous polyposis coli (Apc) gene was also performed for the first time in rat tumorigenic samples. Twenty-six chemically induced rat hepatocellular carcinomas, 21 neoplasms from the livers of SV40 T-Ag animals, and five immortalized hepatic cell lines from the transgenic rats were evaluated. None of the hepatic tumors exhibited mutations in the regions analyzed. The albumin-SV40 T-Ag transgenic cell line L-60, derived from normal hepatic tissue, had two mutations in contiguous codons of exon 5 of the p53 gene: a GGT --> GTT missense transversion in codon 183 and a silent mutation in codon 184. The transversion, which may affect the DNA binding domain of the p53 protein, probably originated during cell culture and may have been positively selected because it gave a growth advantage to the mutated cells. The studied region of the M6p/Igf2r gene was not found to be mutated in these two models of rat hepatocarcinogenesis. Although M6p/Igf2r, Apc, and p53 have been shown to be mutated in a variety of human hepatic proliferative diseases, our results indicate that aberrations in these genes may not be necessary for liver carcinogenesis in the rat.

Previous work from this laboratory has reported on the effects of two sequential 5 day periods of fasting and subsequent refeeding on tumor promotion in multistage hepatocarcinogenesis in the rat (Carcinogenesis, 18, 159-166, 1997). In the present extension of the earlier study, the sequential fasting-refeeding regimen was begun at later time points (28 and 54 days post-initiation) than the first study. This was done to determine whether larger-sized altered hepatic foci (AHF) exhibited a depletion similar to that of the relatively small AHF in the published experiment and to study concomitant molecular changes during the fasting periods. Groups of animals were fasted in the presence and absence of 0.05% phenobarbital (PB) in the drinking water. During the fasting periods, both body and liver weights decreased dramatically, less in the fast begun at 54 days. This change was accompanied by a significant decrease in the bromodeoxyuridine (BrdU) labeling indices of hepatocytes within AHF. Apoptotic bodies increased dramatically in the non-focal (surrounding the AHF) hepatocytes during the fasting periods. These parameters were slightly lower in hepatocytes of rats administered PB during the fasting periods, most notably during the 54-66 day period. With the nick end-labeling method, the proportion of hepatocytes undergoing apoptosis was significantly higher in cells within AHF at the end of each of the fasting periods in all but one group. Concomitantly, the number of AHF and percentage of liver volume occupied by AHF decreased dramatically during the fasting periods. Refeeding caused a marked increase in BrdU labeling in hepatocytes within and surrounding AHF during the first week or two, most notably in animals not receiving PB during the fasting period. Both the number and volume percentage of liver AHF returned to control values within approximately 2 weeks of the refeeding regimen. Assays of nuclear DNA fragmentation with samples of whole liver indicated that a 'laddering' effect was most noticeable in livers of animals subjected to the fasting-refeeding regimen when phenobarbital was not present during the fasting period. Studies of the levels of mRNA of several genes in the total liver revealed that the expression of c-myc increased 3- to 9-fold during the fasting periods but rapidly returned to normal levels after refeeding. Levels of albumin and insulin-like growth factor I mRNAs decreased significantly during the fasting period, but rapidly reappeared on refeeding. These results indicate that the extensive loss of AHF during the short-term fasting periods occurs even when the number and volume of AHF are 10- to 50-fold greater at the beginning of the fast than the values published previously. Both the decrease in insulin growth factor I and the elevation of c-myc expression during the fasting period may indicate the role of these genes in the transcriptional regulation of hepatocyte apoptosis in both normal and preneoplastic hepatocytes in the rat.

A well characterized model of multistage carcinogenesis is that of hepatocarcinogenesis in the rat. The histopathology as well as the cell and molecular biology of the stages of initiation, promotion, and progression have been elucidated to varying degrees in this system. Putatively single initiated hepatocytes are identified by their expression of the ubiquitous marker of hepatocarcinogenesis, glutathione-S-transferase pi (GSTP). 0.5-1.0 x 10(6) GSTP-positive "initiated" hepatocytes developed within 14 days after initiation with a subcarcinogenic dose of diethylnitrosamine (DEN). Approximately 1% of these cells develop clonally into altered hepatic foci (AHF) in animals administered promoting agents, such as phenobarbital, chronically for 4-8 mo. Hepatocytes within AHF during the stage of promotion exhibit normal diploid karyotypes but various phenotypes depending on the chemical nature of the promoting agent. Continued administration of the promoting agent results in the infrequent development of hepatocellular carcinomas; however, administration of a complete carcinogen or a progressor agent during the stage of promotion results in substantial numbers of hepatic neoplasms. In order to quantitate the development of the stage of progression more accurately, markers selective for this stage have been sought. Transforming growth factor-alpha (TGF-alpha) appears to be such a marker of progression. About 500 TGF-alpha-positive lesions develop spontaneously following initiation and continued promotion, usually within GSTP-positive AHF, but administration of a single dose of a progressor agent such as ethylnitrosourea may increase this number 3-fold or more. Some agents such as gamma radiation and hydroxyurea, when administered as single or a few closely spaced multiple doses, result in no increased number in TGF-alpha-positive lesions but a markedly enhanced increase in their growth rate. By monitoring gene expression using quantitative stereology, the stages of hepatocarcinogenesis can be analyzed and quantified in sufficient detail so that the animal data can be utilized in biomathematical modeling to develop more accurate models for estimation of human cancer risks.

Long-term treatment of breast cancer patients with tamoxifen has prompted concern over potential toxicity of this drug with chronic administration. Since tamoxifen has estrogenic action in the rat liver and estrogenic agents can increase hepatoma incidence in rats, tamoxifen and two non-isomerizable, fixed-ring analogs (FRT1 and FRT2) were evaluated as promoting agents in a two-stage model of hepatocarcinogenesis in female Fischer F344 rats. The rats were subjected to 70% partial hepatectomy and half of the animals were administered the initiating agent, diethylnitrosamine (DEN; 10 mg/kg body wt), while the other half were not initiated. Groups of initiated and uninitiated animals were allowed to recover for 2 weeks and were then administered tamoxifen or one of the fixed-ring analogs admixed into AIN-76A diet at 25, 100 or 250 mg/kg diet. After 6 months of anti-estrogen administration the rats were sacrificed and uterine weights, blood levels of anti-estrogen, and liver histopathology were assessed. Uterine weights were decreased 2- to 3-fold by each of the agents, consistent with an anti-estrogenic action in the rat. The serum levels in rats administered 250 mg anti-estrogen/kg diet for 6 months were 320+/-20 ng/ml for tamoxifen, 320+/-10 for FRT1 and 350+/-20 for FRT2. The liver levels after a 6 month administration of 250 mg anti-estrogen/kg diet were 13 870+/-860 ng/g for tamoxifen, 13 300 +/-860 for FRT1 and 26 900+/-1900 for FRT2. A dose-dependent increase in serum and liver level of each compound was noted when measured at the 6 month time period. The number and percentage of the liver occupied by altered hepatic foci (AHF) were determined by quantitative stereology. A dose-dependent increase above initiated controls was observed in the initiated, tamoxifen-treated rats. Both fixed-ring analogs also increased the number and size of AHF compared with initiated controls, but were less potent than tamoxifen, suggesting that tamoxifen has an intrinsic promoting action in the liver that is independent of its ability to isomerize to more potent estrogenic compounds. In addition, the fixed-ring analogs have a weaker promoting activity in the rat liver than does tamoxifen. This may be due to pharmacokinetic differences at the lower two doses, but it is independent of achieved serum level at the highest dose and hence may reflect differences in intrinsic activity of these compounds. Thus tamoxifen and the two fixed-ring analogs promote the development of rat hepatocarcinogenesis.

The experimental three-stage hepatocarcinogenesis protocol of initiation, promotion, and progression, coupled with the analytical technique of stereology, permits quantitative analysis of the carcinogenic process, including the derivation of biologically based risk assessment models. The aberrant expression of the placental isozyme of glutathione S -transferase (PGST) is an efficient marker for initiated, preneoplastic, and neoplastic hepatocytes. Putatively initiated cells and their clonal progeny can be identified, enumerated, and their growth characteristics determined on the basis of their aberrant expression of this protein. A lack of suitable markers has made the identification and quantitation of hepatocytes in the early stage of progression more difficult. One characteristic of cells in the stage of progression is the evolution of relatively autonomous growth. The alteration of growth factor signalling pathways may provide one mechanism for this observation. The expression of transforming growth factor α (TGF α ) is seen in many malignancies. The initiation-promotion-progression protocol has been used to induce progression in the rat liver. The focal expression of TGF α was found to correlate with areas of progression in rats subjected to this protocol. The ability to identify and quantitate cells in the stage of progression should facilitate application of the Moolgavkar-VenzonKnudson model for assessing human risk from carcinogens active at each of these three stages. Validation of this model will require determination of the number and growth characteristics of hepatocytes in the stage of progression.

Carcinogenesis is a multistep, multistage process that begins with irreversible, but heritable damage to a single cell. The partial hepatectomy/diethylnitrosamine (DEN) model of rat hepatocarcinogenesis has been well characterized and many aspects of the stage of initiation are known. Recently, it has been suggested that hepatocytes expressing the placental isozyme of glutathione S-transferase (PGST) may be one population of initiated cells. Male Fischer rats were subjected to a 70% partial hepatectomy and at the peak of cell proliferation 24 h later were administered either the solvent trioctanoin, or 10 mg DEN/kg. The rats were administered 100 mg bromodeoxyuridine (BrdU)/kg 1 h prior to death at various times after DEN administration. Since initiation of the carcinogenesis process requires the division of cells containing DNA damage to induce mutations, we examined the concentration of alkylated adducts and the labeling index at various times after DEN administration. In addition, the time course of hepatic PGST expression was determined concurrent with the adduct concentration and labeling index. During the first day after DEN or solvent administration to a rat subjected to a 70% partial hepatectomy, a diurnal variation in labeling index was observed. A recovery to postsurgical labeling index levels was demonstrated for both the solvent- and DEN-treated groups by 7 days. The concentration of three promutagenic lesions was maximal at 6 h after DEN administration. The detectable level of the O6EG adduct was negligible by 24 h after DEN administration, while the two O-alkylpyrimidines, O2ET and O4ET, were retained for much longer periods. Single hepatocytes expressing PGST were observed by 2 days after DEN administration, while small foci of PGST-expressing hepatocytes could be reliably detected by 2 weeks. Two phases of PGST expression in single hepatocytes were observed. The first phase was maximal at day 3 and complete by day 6, while the second reached a plateau by day 8 and was maintained for the 28 days of the study. The presence of the three O-alkylation adducts during a time of enhanced cellular proliferation suggests that all three promutagenic adducts may contribute to the initiation that results in the partial hepatectomy/DEN model of rat hepatocarcinogenesis.

Tamoxifen, an antiestrogen used in the treatment of breast cancer, was assessed for carcinogenic potential in the two-stage model of experimental hepatocarcinogenesis. Groups of female Fisher F344 rats were initiated with a non-necrogenic, subcarcinogenic dose of diethylnitrosamine (DEN; 10 mg/kg, po) and fed tamoxifen at a concentration of 250 mg per kg of AIN-76A diet for 6 or 15 months. The livers of these animals exhibited an increase in size and number of altered hepatic foci compared with those animals which were initiated with DEN but not exposed to tamoxifen. This finding indicates that tamoxifen may have a carcinogenic potential in the rat liver. After 6 months of treatment, neoplastic nodules were observed in 3/8 rats in the DEN-initiated, tamoxifen-treated group. In the initiated group provided with tamoxifen for 15 months, neoplastic nodules were observed in 7/8 rats and hepatocellular carcinomas in 3/8 rats. The serum level of tamoxifen in these rats was 200-300 ng/ml. The ratio of tamoxifen, 4-hydroxy tamoxifen, and N-desmethyl tamoxifen was 1:0.1:0.5-1 in the serum. When adjusted for age-related weight increases, the serum and liver levels of tamoxifen and its N-desmethyl metabolite did not change over the 15 months. In the rat liver, the level of tamoxifen and its N-desmethyl metabolite was 10-29 micrograms/g liver after 6 or 15 months of chronic dietary administration. The ratio of tamoxifen:4-hydroxy tamoxifen:N-desmethyl tamoxifen was 1:0.1.3-3.3 in the liver. Therefore, the liver had 20- to 30-fold more tamoxifen and 4-hydroxy tamoxifen and at least 100-fold more N-desmethyl tamoxifen than the serum (assuming 1 gram of tissue is equivalent to 1 ml of serum). These results indicate that tamoxifen is a promoting agent for the rat liver at serum levels found in patients given the usual therapeutic course of tamoxifen. The high concentrations of tamoxifen attained in the rat liver indicate that actions other than its known estrogenicity for liver could contribute to its promoting action. In addition, these results indicate that the pharmacodynamic differences in tamoxifen metabolism in rats and humans and at low versus high doses should be determined. Thus, the therapeutic indications for tamoxifen should be balanced by the potential risk it may present as a promoting agent in mammalian liver.

The potential relationship between mutagenesis and carcinogenesis has been examined in the Nagase analbuminemic rat treated with a single dose of benzo[a]pyrene, an incomplete liver carcinogen. The apparent mutation rate at the albumin locus was calculated by determining the number of hepatocytes expressing a cross-reactive product of albumin in analbuminemic rats treated with benzo[a]pyrene. The rate of initiation, the first stage in carcinogenesis, was determined by assessing the number of hepatocytes expressing the placental isozyme of glutathione S-transferase (PGST) after administration of benzo[a]pyrene. Since the expression of PGST may represent hepatocellular changes independent of initiation, promotion with phenobarbital was employed to clonally expand those putatively initiated hepatocytes expressing PGST. With immunohistochemical measures to assess changes in albumin expression, a threefold increase in the number of hepatocytes expressing albumin was detected after administration of benzo[a]pyrene in Nagase analbuminemic rats. A more than five-fold increase in altered hepatic foci (AHF) exhibiting increased PGST expression was observed in animals given benzo[a]pyrene treatment followed by phenobarbital, compared with those given benzo[a]pyrene alone. The number of albumin-expressing single hepatocytes detected was of the same order of magnitude as the number of individual hepatocytes and AHF expressing PGST, suggesting that similar events may be involved in their formation. Since 3 x 10(6) single hepatocytes expressing albumin were found in the analbuminemic rat liver after a single administration of benzo[a]pyrene, while less than 2 x 10(4) AHF expressing PGST were observed, formation of individual hepatocytes expressing albumin was a far more frequent event than clonal expansion of initiated hepatocytes in the Nagase analbuminemic rat. However, the number of loci of PGST expression including AHF and single hepatocytes is comparable to that of single hepatocytes expressing albumin.

Hepatocarcinogenesis was examined in an initiation--promotion protocol with a single initiating dose of 7,12-dimethylbenz[a]anthracene (DMBA) followed by promotion with phenobarbital (PB) in the Nagase analbuminemic rat (NA), the Sprague-Dawley rat (SD) and their F1 crosses. All rats received a 70% partial hepatectomy, followed at 24 h by 30 mg DMBA/kg body wt or the solvent. After a 2 week recovery following surgery, half of the solvent control and initiated groups received either basal diet or promotion with 0.05% PB mixed into the basal NIH-07 diet. After 12 weeks of promotion, the rats were killed and the livers perfused and fixed in situ with paraformaldehyde. Liver slices were paraffin embedded and stained for the placental isozyme of glutathione S-transferase (PGST). The number of altered hepatic foci (AHF) expressing PGST per liver was determined by quantitative stereology and used as an endpoint for comparison of initiation in the rat strains. The NA rat had a lower response to this initiation-promotion protocol than did the SD rat. The F1 progeny of the male NA and female SD rats were more similar to the NA parent in their responsiveness to initiation, whereas the F1 progeny of the female NA and male SD were similar to the SD parent in this respect. Putative mutagenesis and carcinogenesis were examined in the F1 progeny of the female NA and male SD rat. In these rats, serial liver sections were stained either for albumin to detect putative mutations at that locus, or PGST to identify putatively initiated hepatocytes. In the NA/SD F1, the number of single hepatocytes with a putative mutation at the albumin locus was the same (3.7 x 10(5)/liver) as those expressing a common marker of preneoplasia (PGST). The number of AHF expressing PGST was approximately 5% that of the single cells exhibiting an altered expression of albumin or PGST, indicating a possible quantitative correlation between initiation and mutation in vivo when individual hepatocytes with altered gene expression were counted. These studies also suggest that only a subpopulation of the putatively initiated hepatocytes expands clonally in the presence of the promoting agent, PB. The progeny of the female NA rat crossed with the SD male rat appears to provide a useful model in which to compare mutation and carcinogenesis simultaneously in vivo.

Differences in the distribution of phenotypic alterations in initiated and promoted cell populations reflect both the dose and the action of the specific initiating agent, as well as the mechanism of action of the promoting agent. The use of multiple phenotypic markers to characterize AHF should allow further experimentation on the characteristics of promoting agents in hepatocarcinogenesis, especially their ability to promote separate populations of initiated cells, and on those characteristics of initiated cells that enable certain ones to survive and multiply in the specific environment provided by a promoting agent. Studies of promoting agents with differing mechanisms of action should further allow questions of reversibility, substitution and additivity of the actions of promoting agents to be addressed. The possibility that individual promoting agents do not enhance the growth of the entire population of initiated cells indicates that study of combinations of promoting agents is an important future direction of research. Therefore, information on the characteristics of promoting agents, singly and in combination, is necessary to assess more accurately the contribution of promoting agents to the carcinogenic risk for humans distinct from the effects of initiating agents and complete carcinogens.

The age-dependent effects of diet and of phenobarbital (PB), a known promoter of hepatocarcinogenesis, on indigenous DNA adducts (I-compounds) were studied by the 32P-post-labeling technique. Late-gestation female Sprague-Dawley rats were fed either AIN-76A semisynthetic or Teklad cereal-based diet. At 21 days after birth, the male pups were weaned and continued on the diets, with half of each dietary group receiving 0.05% PB mixed into the diet for 2, 4 and 8 months. Age-dependent increases in I-compound levels were observed. In addition, both the levels of individual I-compounds and the overall number of I-compounds were greater in rats fed the Teklad cereal-based diet than in those fed the AIN-76A diet. Independent of the parent diet, PB administration reduced the levels of the majority of I-compounds in a time-dependent manner. This effect of PB was detected earlier in the Teklad-fed than in the AIN-76A-fed group. In contrast to the I-compounds, a second group of spots, termed reverse I-compounds, declined between 2 and 4 months and, especially in AIN-76A-fed animals, tended to increase when PB was administered. It is hypothesized that alterations of DNA modification patterns may play a role in diet-modified hepatocarcinogenesis promoted by PB.

Knowledge of environmental factors that are causative of human cancer has been gained primarily from epidemiologic studies. Since neoplastic disease occurs spontaneously in lower mammals and can be induced in at least one species by virtually all of the agents shown to be carcinogenic by epidemiologic studies (1), animals may be appropriate counterparts for the assessment of human disease. Since the earliest studies demonstrating the chemical induction of cancer in mice and rats, the use of such rodents for determining the carcinogenicity of chemical agents has represented the “gold standard” in classifying an agent as a carcinogen. However, since the numbers of new chemicals entering our environment every year are in the thousands (2), the “gold standard” cannot be applied to every chemical for reasons of expense and time.

Carcinogenesis can be induced experimentally by exposure to exogenous agents or it can occur spontaneously without intentional or active intervention. Carcinogenesis can be actively induced by chemicals, radiation, infectious biological agents, transgenesis, or selective breeding. In the human and occasionally when testing potential carcinogens in animals, cancer may result from passive exposure to carcinogens encountered in the ambient environment or from changes in the internal milieu of the animal. Many carcinogens alter the structure of DNA resulting in carcinogenesis, but a significant number of carcinogens do not appear to act through this mechanism. When the action of specific carcinogenic agents is considered in relation to the stages of cancer development, initiation, promotion, and progression, the mechanism of the induction of carcinogenesis by DNA-reactive agents that alter genomic structure can be reconciled with those agents that do not act in this manner. As some cells are fortuitously initiated by uncontrolled variables such as irradiation and through changes in normal processes, the stimulation of growth and altered genetic expression by nongenotoxic agents may result indirectly in cancer development. The final stage of carcinogenesis, progression, can occur spontaneously, enhanced by formation and propagation of genetic errors due to increased cellular proliferation associated with the promotion stage. In addition, chemical and viral agents that lack the capacity for initiation and promotion may actively convert cells in the stage of promotion to the stage of progression. Therefore, the diverse mechanisms of action of carcinogenic agents in relation to their effects on specific stages in the natural history of cancer development allow for greater congruence of many of the theories of carcinogenesis. The influence of the roles of nongenotoxic carcinogenic agents and the potential role of progressor agents on the carcinogenesis process allow a more accurate identification of the potential risk that specific carcinogenic agents pose for increasing human cancer.

Polychlorinated biphenyls (PCBs) are a group of industrial chemicals that are widely distributed in the environment. Because these compounds occur as mixtures, studies of their possible interactive effects are essential for an understanding of the mechanism of the toxicity of these mixtures. For the determination of a possible interaction of the effects in vivo of 2,5,2',5'-tetrachlorobiphenyl (TCB) and 3,4,3',4'-TCB, rats were exposed to a single dose of diethylnitrosamine (DEN) and subsequently to 0.1 p.p.m. 3,4,3',4'-TCB and/or 10 p.p.m. 2,5,2',5'-TCB in the feed for 1 year. The two major targets of PCB toxicity, the liver and the peripheral blood, were examined after these treatments. TCB treatment after DEN exposure caused a predominance of increased placental glutathione S-transferase (PGST) and deficiencies of ATPase as preneoplastic markers in focal hepatic lesions. When 0.05% phenobarbital (PB) was administered after DEN exposure, the distribution of markers in altered hepatic foci (AHF) was essentially equal for increased PGST and gamma-glutamyltranspeptidase (GGT) and for ATPase deficiency. Many of these AHF also exhibited increased P450 b/e expression. Our results demonstrated that the two PCB congeners interacted in vivo to produce an increase in AHF that were PGST positive and ATPase negative. PGST-positive and ATPase-negative AHF correlated best with focal areas of P450 b/e expression. The combination of the two PCBs caused a greater than additive decrease in the total number of lymphocytes and antibody-producing B-cells. Also the thymocyte-dependent T-helper cells isolated from the animals receiving the combination of TCBs demonstrated a morphologically abnormal subpopulation. The results indicate that the interaction of 2,5,2',5'-TCB and 3,4,3',4'-TCB in vivo induced much greater toxicity and mutagenicity in peripheral lymphocytes and hepatocytes than treatment with either congener alone.

The rates of cell proliferation and cell loss in conjunction with the differentiation status of a tissue are among the many factors contributing to carcinogenesis. Nongenotoxic (non-DNA reactive) chemicals may affect this balance by increasing proliferation through direct mitogenesis or through a regenerative response following loss of cells through cytotoxic (oncotic) or apoptotic necrosis. In a recent NTP study in Fischer rats and B6C3F(1) mice, the mycotoxin fumonisin B(1) caused renal carcinomas in male rats and liver cancer in female mice. In an earlier study in male BD-IX rats, fumonisin B(1) caused hepatic toxicity and hepatocellular carcinomas. An early effect of fumonisin B(1) exposure in these target organs is apoptosis. However, there is also some evidence of oncotic necrosis following fumonisin B(1) administration, especially in the liver. Induction of apoptosis may be a consequence of ceramide synthase inhibition and disruption of sphingolipid metabolism by fumonisin B(1). Fumonisin B(1) is not genotoxic in bacterial mutagenesis screens or in the rat liver unscheduled DNA-synthesis assay. Fumonisin B(1) may be the first example of an apparently nongenotoxic (non-DNA reactive) agent producing tumors through a mode of action involving apoptotic necrosis, atrophy, and consequent regeneration.

Previous studies have shown that the presence of several hepatic I-compounds, i.e., age-dependent covalent DNA modifications, is related to the presence of a natural ingredient, i.e., oats, in the diet. To demonstrate the biological significance of these novel DNA modifications, the effect of oat lipids on tumor initiation and promotion was examined in a rat liver tumor model. Female Sprague-Dawley rats were treated with a single dose of diethylnitrosamine, a hepatic carcinogen, 24 hours after a 70% partial hepatectomy, then subjected to dietary phenobarbital promotion. Diets containing 10% oat lipids or corn oil were given during the initiation or the promotion stage of the tumorigenesis. At the end of the feeding, hepatic I-compounds were measured by 32P postlabeling, and the number and volume of enzyme-altered hepatic foci, which served as preneoplastic markers, were measured in serial sections of liver by the method of quantitative stereology. Rats receiving oat lipids-supplemented diets had five- to sixfold higher levels of I-compounds in their liver DNA than those receiving control diets. Meanwhile, rats receiving diets containing oat lipids during promotion had significantly smaller numbers and reduced volume of altered hepatic foci compared with those fed the control diet containing corn oil. These observations support the hypothesis that some I-compounds, e.g., the oats-specific I-compounds, are novel DNA modifications related to nutrient metabolism. The diet containing oat lipids may have chemopreventive activities, as demonstrated in this model system.

A modified initiation-selection procedure for neonatal male and female rat hepatocarcinogenesis were examined utilizing the methods of quantitative stereology. In this study, diethylnitrosamine (10 mg DEN/kg) was given a few days after birth. At weaning, the rats were fed 0.02% 2-acetylaminofluorene (AAF) for 2 weeks with a mitotic stimulus [70% partial hepatectomy (PH)] after 1 week on the diet. Quantitative stereological analyses in conjunction with the use of several enzyme markers were used to determine the number and volume of altered hepatic foci (AHF) detected at 1 week, 3 months and 7 months after the selection procedure. This format resulted in an equivalent number of AHF in male and female rats. The AHF were three times larger in males than in females 1 week after discontinuation of AAF administration. Three months after the selection procedure, the number of AHF had decreased by at least a third and their volume percentage was the same in male and female rats. After 7 months, the number and volume fraction of detectable AHF in females were comparable to those which had been observed at 1 week after selection. In the male, the number but not the volume fraction were similar at 7 months compared with 1 week after selection. Both initiation with DEN and selection with AAF/PH contribute independently to the total population of AHF in male and female rats. At least half of the AHF detected 7 months after the selection protocol were due to DEN administration alone. Rats receiving only the AAF/PH selection exhibited one third of the number of AHF observed with the complete protocol. Administration of a non-necrogenic dose of DEN to neonatal rats when coupled with the AAF/PH selection procedure resulted in a significant promotion of the growth of initiated hepatocytes at 1 week, 3 months or 7 months after the selection procedure. These studies demonstrated that (i) the number of AHF detected after a non-necrogenic dose of DEN during the first week of life with subsequent AAF/ PH selection after weaning decreases within the first 3 months after the selection procedure, but can re-develop with a promotion stimulus; (ii) the AAF/PH selection procedure itself may initiate hepatocytes in the absence of DEN administration; (iii) the AAF/PH selection procedure is equally effective with respect to the number of AHF observed after phenobarbital promotion in weaning male and female rats initiated near birth.

The stages in the carcinogenesis process include initiation, promotion and progression. Although many characteristics of tumor promotion and promoting agents have been reported, relatively few studies on the effects of combinations of promoting agents have been detailed. For study of the combined effects of phenobarbital (PB) and mestranol (MS) in multistage rat hepatocarcinogenesis, an initiation-promotion protocol was developed. Female rats were injected i.p. at 5 days of age with either diethylnitrosamine (DEN) (10 mg/kg) or the solvent tricaprilyn. At weaning, approximately 10 rats from both the DEN-initiated and the solvent control groups were provided basal diet alone, PB (10, 100 or 500 mg/kg diet), MS (0.02 or 0.2 mg/kg diet) or various combinations of both PB and MS in the basal diet. At 8 months of age, the rats were killed and the livers removed, sectioned and fixed in ice-cold acetone. Sections of 5 microgram thickness were stained for placental glutathione S-transferase (PGST) expression, and the volume fraction of liver occupied by altered hepatic foci was determined by stereology. In addition, incorporation of bromodeoxyuridine into nuclei of PGST-expressing (focal) and non-PGST-expressing (non-focal) hepatocytes was determined. Administration of the highest dose of PB resulted in a significant decrease in non-focal hepatocyte labeling index, with a 4-fold differential between the focal and non-focal hepatocyte labeling index. Administration of 0.2 mg/kg diet MS resulted in effective promotion. The non-focal labeling index was increased and the focal labeling index was further enhanced (3-fold) relative to the non-focal index by this dose of MS. Combination of the lower MS dose with PB resulted in at least an additive promoting effect; however, a lower volume fraction was noted for the combination of low MS dose plus the highest PB dose. Combination of the higher MS dose with PB resulted in an elevation of volume fraction only for the middle PB dose. These findings indicate that the potency of promotion by mixtures is modulated by the dose of each component as well as by pharmacodynamic and pharmacokinetic properties of each component of the mixture.

Since tamoxifen is efficacious for the prevention of second primary breast neoplasms in humans and has a low reported incidence of acute side effects, several structurally related compounds have been developed for the treatment of breast cancer including toremifene and idoxifene. We have compared the karyotypic alterations that occur after a single per os administration of 35 mg/kg of tamoxifen, toremifene or idoxifene to female Sprague-Dawley rats. One day following treatment, the rats were sacrificed and the hepatocytes isolated and cultured. After 47 h in culture, colcemid was added for 3 h prior to harvest of the hepatocytes for karyotypic evaluation. At least 100 metaphase spreads were examined for each of five rats per treatment. Toremifene resulted in aneuploidy in 50 +/- 7% of the cells examined and idoxifene induced a 57 +/- 4% aneuploidy compared with the 85 +/- 7% level induced by tamoxifen. Since the level of aneuploidy in solvent-treated rats was 3 +/- 3 %, the induction of aneuploidy in at least 50% of the cells from rats treated with tamoxifen, toremifene or idoxifene was highly significant. Analysis of electron micrographs of cultures treated with these antiestrogens demonstrated a range of phenotypes including multipolar spindles in toremifene-treated rats and condensed chromosomes in the presence of an intact nuclear envelope in occasional idoxifene-treated rat hepatocytes. The exclusion of chromosomes from the spindle apparatus and the lagging of some chromosomes on the metaphase plate correlate with the high rate of induction of aneuploidy in the rat liver as determined by karyotypic analysis of hepatocytes from rats treated with these triphenylethylenes.

Female rats were subjected to a 70% partial hepatectomy and administered either diethylnitrosamine (10 mg/kg) or the solvent, trioctanoin. After a 2 day recovery from the surgery, the rats were placed on basal diet alone or containing phenobarbital (500 mg/kg diet), mestranol (0.2 mg/kg diet), tamoxifen (250 or 500 mg/kg diet) or toremifene (250, 500 or 750 mg/kg diet) for 6 or 18 months prior to killing. The liver and kidneys were prepared for pathological diagnoses. In addition, sections of liver from the 6 month killing were frozen and serially sectioned. The sections were stained for expression of the placental isozyme of glutathione S-transferase (GST), gamma glutamyl transpeptidase (GGT), canalicular ATPase (ATP) and glucose 6-phosphatase (G6P) and scored by quantitative stereology for number and volume fraction of liver occupied by altered hepatic foci (AHF) with alterations in these markers individually and combined (ANY). Each of the agents increased the volume fraction of liver occupied by AHF when the ANY category was used. Statistical increases in both the GGT-positive and G6P-deficient AHF populations were observed in the spontaneously as well as DEN-initiated groups treated with tamoxifen or toremifene. After 18 months of administration, the highest concentration of tamoxifen increased the incidence of malignant hepatic neoplasms in non-DEN-initiated rats. Toremifene, at the highest tested dose, increased the incidence of hepatocellular carcinomas in the DEN-initiated groups to a level one-third that observed with tamoxifen administration to DEN-initiated rats. Both tamoxifen and toremifene increased the incidence of hypernephromas in previously DEN-initiated rats. While both tamoxifen and toremifene are effective promoting agents for DEN-initiated lesions, tamoxifen is more potent than toremifene in the induction of rat hepatocarcinogenesis.

Mathematical descriptions of complex biological phenomena, such as cancer, require an experimental format that faithfully recapitulates the biological process. In addition, the biological process must dictate the parameters in the mathematical formula. Evidence from the epidemiology of several human cancers and from experimental carcinogenesis in several organ systems indicates that cancer is a multistage process. The initiation-promotion-progression format of experimental carcinogenesis mimics the development of cancer in humans and other animals. In rats, the altered hepatic focus model of hepatocarcinogenesis has been well characterized and, coupled with the method of quantitative stereology, permits accurate determination of the number and the volume fraction of such altered foci per liver. The placental isozyme of glutathione S-transferase (PGST) is reportedly the best single marker of preneoplasia in the rat liver. Recently, single hepatocytes expressing PGST have been proposed as putatively initiated cells. Quantitation of individual hepatic cells and altered hepatic foci expressing PGST in the livers of rats subjected to an initiation-promotion protocol permits determination of the congruence of the Moolgavkar-Venzon-Knudson (MVK) model with experimental data. The best fit of the MVK model for the preneoplastic stages of hepatocarcinogenesis assumes that all hepatocytes are susceptible and that single hepatocytes expressing PGST are the initiated cell population for the focal lesions that express PGST. Further refinement of the initiation-promotion-progression model to permit accurate quantitation of early malignant conversion should allow a more complete analysis of the congruence of the MVK model for human cancer risk determination. In addition, the MVK model may be extended to other model systems and to human cancers in which early preneoplasia can be quantitated. Furthermore, the use of a more biologically based risk-assessment protocol, such as the MVK model rather than the stochastic one-hit model presently used, would permit incorporation of the present knowledge on the pathogenesis of cancer. To apply experimental data to a mathematical model that reflects the biological processes underlying human cancer development will require integration of the cell kinetics and experimental data to a mathematical model that reflects the biological processes underlying human cancer development including the pharmacokinetic and pharmacodynamic properties of the treatment chemicals.

Carcinogenesis develops in stages that have been operationally defined as initiation, promotion and progression. Although morphological end points have been described for detection and quantitation of these stages, to date initiation has been assessed only in the context of clonal growth in response to certain promoting agents. Initiated cells are morphologically indistinguishable from surrounding cells and early changes at the cellular level during initiation have not been clarified. One commonly used end point for the detection of preneoplastic hepatic lesions i their aberrant expression of the placental isozyme of glutathione S-transferase (PGST). Because single hepatocytes expressing PGST have been detected in aged rats and in those administered hepatocarcinogens, it has been suggested that such cells constitute a population of putatively initiated hepatocytes. In order to further elucidate the characteristics of single PGST-positive hepatocytes, we analyzed the number of these cells 2 and 18 weeks after various doses (0-100 mg/kg) of diethylnitrosamine (DEN) and of dimethylbenz[a]anthracene (DMBA). When determined 14 days after carcinogen administration, the number of single hepatocytes expressing PGST was greater after DEN administration (ranging from 0.8 +/- 0.3 per cm2 transection of liver at 1 mg/kg to 33.0 +/- 4.7 at 100 mg/kg) than after DMBA administration (ranging from 0.25 +/- 0.14 at 10 mg/kg to 3.03 +/- 0.5 at 100 mg/kg); none were detected in control rats of the same age. Additional rats were maintained on a basal diet or a basal diet plus phenobarbital for a further 4 month period. Whereas individual PGST-positive hepatocytes were only sporadically detected in rats treated with DMBA and maintained on a basal diet for 18 weeks, those rats placed on phenobarbital for 16 weeks had an even higher number of such PGST-positive hepatocytes than at 2 weeks after DMBA administration. In contrast, the dose-response curve observed for DEN-treated rats 18 weeks after carcinogen administration was similar to that observed 2 weeks after carcinogen treatment for both phenobarbital- and non-phenobarbital-treated rats. In addition, the number of single PGST-positive hepatocytes detected at 2 weeks was directly parallel to the number of altered hepatic foci expressing PGST 18 weeks after DEN administration. The dose-dependent induction of PGST-positive single hepatocytes after treatment with two hepatocarcinogens, the dose-dependent growth of altered hepatic foci (AHF) expressing PGST with phenobarbital administration and the parallel dose-response curve of single hepatocytes expressing PGST and later of AHF expressing PGST argue strongly for a precursor role of single PGST-positive cells in the development of AHF expressing PGST.

Cell proliferation is pivotal to all stages of the carcinogenesis process and is one of the primary characteristics of the promotion stage of cancer development. Both a two-stage model of initiation and promotion for analysis of early preneoplasia and a three-stage initiation-promotion-progression model of hepatocarcinogenesis were used to address the effect of the liver tumor-promoting agent phenobarbital (PB) on hepatic cellular proliferation. Male rats were subjected to a 70% partial hepatectomy and 10 mg diethylnitrosamine (DEN)/kg or the solvent alone and were administered PB for 4-8 months. Analysis of bromodeoxyuridine (BrdU) incorporation (1 h pulse) in liver within (focal) and not within (non-focal) altered hepatic foci (AHF) demonstrated a labeling index in AHF of 2% in DEN-initiated rats; the non-focal labeling index of placental glutathione S-transferase expressing hepatocytes was 0.3-0.6%. The focal labeling index was constant over the 8 month period of promotion. Inasmuch as one characteristic of promotion is the reversibility of the induced effects on clonal expansion of initiated cells, groups of rats initially promoted with PB were maintained in the absence of continued promotion for 4 or 8 months prior to being killed. Assessment of the focal labeling index after cessation of PB treatment indicated a drop in the index from 2.3% to 0.7%. When a progressor agent, ethylnitrosourea, was given at the time PB was discontinued for 4 or 8 months, a significant change in focal labeling index was not observed relative to the index in AHF when the animals were killed immediately after 8 months of PB promotion. Thus, cell proliferation plays an integral role in both the promotion and progression stages of multistage rat hepatocarcinogenesis and is influenced by administration of promoting and progressor agents.

The dose-response characteristics of initiation of hepatocarcinogenesis by diethylnitrosamine (DEN) was investigated in the neonatal female rat by means of the quantitative stereologic estimation of altered hepatic foci (AHF) expressing multiple markers. At 5 days of age, female Sprague-Dawley rats were given a single i.p. dose of DEN (0.1-30 mg/kg body wt) or the vehicle (trioctanoin). The semisynthetic AIN-76A diet was provided to half of the rats in each treatment group, while the remainder received this diet containing 500 mg phenobarbital (PB)/kg for 8 months from weaning until the animals were killed. To ascertain more exactly the dose-response relationship for initiation by DEN, the number, volume percentage and phenotypes of the resulting AHF were determined by quantitative stereological analysis on serial sections of frozen tissue, each stained for one of four markers of preneoplasia. A linear relationship was observed between the dose of DEN (0-30 mg/kg) and the number and volume percentage of AHF detected, with each single marker or the total number of AHF detected when the placental isozyme of glutathione S transferase, gamma-glutamyl transpeptidase (GGT), canalicular adenosine triphosphatase, or glucose-6-phosphatase was used as the marker. For each dose, PB administration increased the number and volume of AHF scored compared with similarly initiated rats that did not receive a promoting stimulus. This was, in part, owing to enhanced GGT expression in AHF with PB administration. Promotion by PB resulted in a distribution of AHF phenotypes altered from that observed in rats not receiving PB. Initiation of AHF in neonatal female rats by DEN was linear with doses from 0 to 30 mg/kg for all four of the phenotypic markers employed. In addition, while PB administration stimulated the growth of all AHF phenotypes, the growth of a subset of AHF that expressed the widest variation in preneoplastic markers was specifically enhanced by PB administration.

Previous studies have suggested that 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) acts as a promoting agent in various organ systems including the rat liver. Since a major characteristic of the stage of tumor promotion is its operational reversibility, we have assessed whether TCDD-induced promotion is reversible in a two-stage model of hepatocarcinogenesis. In this model, female Fischer F344 rats were administered a single, intragastric dose of the initiating agent, diethylnitrosamine (DEN, 10 mg/kg), at the peak of proliferation induced by a partial hepatectomy. TCDD was then administered biweekly (0.14 micrograms/kg, s.c.) for 1, 3 or 5 months. One group of animals was killed at each of these time points, while a second group was maintained for each time point for an additional 6 months in the absence of further TCDD. Four serial frozen sections of liver were each stained with a different enzyme marker of altered hepatic foci (AHF). The AHF were identified and the number and volume fraction determined by quantitative stereology. Exposure to TCDD resulted in an increase in the number and size of AHF in the initiated relative to the uninitiated rats. Increasing the duration of promotion with TCDD led to an increase in the number of AHF per liver, the volume fraction of the liver occupied by AHF and the number of markers expressed aberrantly by a single AHF. Discontinuation of TCDD administration for 6 months before killing the animals resulted in a decrease in the total number of AHF observed, but those AHF that remained increased in size with an overall increase in volume fraction of AHF. Analysis of the size class distribution for AHF for each of the periods of TCDD promotion revealed an increase in the larger AHF but a decrease in the smaller, thereby resulting in an overall decrease in number of AHF with an increase in the volume fraction of AHF. Increasing the duration of the TCDD exposure prior to its withdrawal led to an increased AHF size, phenotypic complexity and number of AHF remaining after cessation of TCDD administration. Although the levels of TCDD in livers of rats 6 months after cessation of TCDD administration were still greater than background, they were markedly reduced compared to immediately after administration. Thus, cessation of exposure to TCDD after a brief duration led to a reversal of its promotional effects on the majority of AHF, while prolonged exposure led to maintained promotion of a minority of AHF.

The effect of changing the format of administration as well as the total dose of the promoting agent phenobarbital (PB) on the development of altered hepatic foci (AHF) was determined in an initiation-promotion protocol with female rats fed the purified AIN-76 diet. Effects on the total number of AHF and the volume percentage of liver occupied by AHF were determined for four histochemical markers, the placental form of glutathione S-transferase, gamma-glutamyl-transpeptidase canalicular ATPase, and glucose-6-phosphatase after 16 and 60 weeks of promotion with varying doses and formats of PB, as well as for a further 16-week period in which no PB was administered. At the 16-week point, animals fed 0.1% PB continuously exhibited the largest number and volume percentage of AHF, whereas rats fed 0.1% PB for 4 days followed by 10 days of no PB with continuous repetition of this pattern during the 16-week treatment period exhibited no increase in the number of AHF over control and only a slight increase in volume percentage. Rats fed a continuous repetition of 0.2% PB for 2 days followed by 12 days of no PB exhibited an intermediate increase in the number of AHF as well as the volume percentage fraction after 16 weeks of this regimen. After 60 weeks of feeding PB by these three different formats, the numbers of AHF observed in these groups were equivalent and had increased above those seen after 16 weeks of feeding. The volume percentage occupied by the AHF in these three groups was also similar, although animals receiving 0.2% PB intermittently showed a significantly lower volume percentage than animals receiving 0.1% PB continuously for 60 weeks. When animals were maintained for an additional 16 weeks without PB feeding, the numbers of AHF decreased dramatically, much more so in animals fed PB intermittently, whereas the volume percentage fraction of AHF in livers of animals receiving 0.1% PB continuously for 60 weeks almost doubled. In contrast, the volume percentage fraction of AHF in livers of animals receiving PB intermittently for 60 weeks followed by 16 weeks of no PB was slightly less. Examination of the individual size classes of AHF showed little change in their distribution at 16 and 60 weeks, but after 16 weeks of PB withdrawal (76 weeks total time), the distribution of AHF in animals that had received 0.1% PB continuously for 60 weeks exhibited a decidedly greater shift to larger AHF than animals receiving PB intermittently for the 60-week period.(ABSTRACT TRUNCATED AT 400 WORDS)