Intrahepatic bile duct proliferation induced by 4,4'-diaminodiphenylmethane in rats

Role of the human N-acetyltransferase 2 genetic polymorphism in metabolism and genotoxicity of 4, 4'-methylenedianiline

4, 4'-Methylenedianiline (MDA) is used extensively as a curing agent in the production of elastomers and is classified as reasonably anticipated to be a human carcinogen based on sufficient evidence in animal experiments. Human N-acetyltransferase 1 (NAT1) and 2 (NAT2) catalyze the N-acetylation of aromatic amines and NAT2 is subjected to a common genetic polymorphism in human populations separating individuals into rapid, intermediate, and slow acetylator phenotypes. Although MDA is known to undergo N-acetylation to mono- and di-acetyl metabolites, very little is known regarding whether this metabolism is subject to the NAT2 genetic polymorphism. We investigated the N-acetylation of MDA by recombinant human NAT1, NAT2, genetic variants of NAT2, and cryoplateable human hepatocytes obtained from rapid, intermediate and slow acetylators. MDA N-acetylation was catalyzed by both recombinant human NAT1 and NAT2 exhibiting a fivefold higher affinity for human NAT2. N-acetylation of MDA was acetylator genotype dependent as evidenced via its N-acetylation by recombinant human NAT2 genetic variants or by cryoplateable human hepatocytes. MDA N-acetylation to the mono-acetyl or di-acetyl-MDA was highest in rapid, lower in intermediate, and lowest in slow acetylator human hepatocytes. MDA-induced DNA damage in the human hepatocytes was dose-dependent and also acetylator genotype dependent with highest levels of DNA damage in rapid, lower in intermediate, and lowest in slow acetylator human hepatocytes under the same MDA exposure level. In summary, the N-acetylation of MDA by recombinant human NAT2 and cryopreserved human hepatocytes support an important role for the NAT2 genetic polymorphism in modifying MDA metabolism and genotoxicity and potentially carcinogenic risk.

Systems toxicology of chemically induced liver and kidney injuries: histopathology-associated gene co-expression modules

Organ injuries caused by environmental chemical exposures or use of pharmaceutical drugs pose a serious health risk that may be difficult to assess because of a lack of non-invasive diagnostic tests. Mapping chemical injuries to organ-specific histopathology outcomes via biomarkers will provide a foundation for designing precise and robust diagnostic tests. We identified co-expressed genes (modules) specific to injury endpoints using the Open Toxicogenomics Project-Genomics Assisted Toxicity Evaluation System (TG-GATEs) - a toxicogenomics database containing organ-specific gene expression data matched to dose- and time-dependent chemical exposures and adverse histopathology assessments in Sprague-Dawley rats. We proposed a protocol for selecting gene modules associated with chemical-induced injuries that classify 11 liver and eight kidney histopathology endpoints based on dose-dependent activation of the identified modules. We showed that the activation of the modules for a particular chemical exposure condition, i.e., chemical-time-dose combination, correlated with the severity of histopathological damage in a dose-dependent manner. Furthermore, the modules could distinguish different types of injuries caused by chemical exposures as well as determine whether the injury module activation was specific to the tissue of origin (liver and kidney). The generated modules provide a link between toxic chemical exposures, different molecular initiating events among underlying molecular pathways and resultant organ damage. Published 2016. This article is a U.S. Government work and is in the public domain in the USA. Journal of Applied Toxicology published by John Wiley & Sons, Ltd.

Liver stem cells: a potential source of hepatocytes for the treatment of human liver disease

Severe liver injury often leads to the proliferation of oval cells, which differentiate along hepatocytic and biliary lineages. Because oval cells proliferate only when hepatocyte replication is impaired, they are considered to be the progeny of facultative liver stem cells (FLSCs). Identification and isolation of FLSCs has been hampered by the lack of markers that delineate these bipotential progenitors. We hypothesized that transition ductal cells are FLSCs because they are located in a unique anatomical niche sharing tight junctions with a neighboring hepatocyte and another terminal ductular cell. Alternatively, it has been proposed recently that bone marrow-derived stem cells are FLSCs since these cells differentiate along the hepatic lineage following colonization of the liver. The intent of this review is to provide insight into the nature and origin of liver stem cells and to explore the possibility that stem cell technology may lead to the development of clinical modalities for the treatment of human liver disease.

Effects of nalidixic acid on hamster knee cartilage morphology and synovial fluid composition

Quinolone-induced changes were studied in the knee joints of 4-wk-old female hamsters given intraperitoneal doses of either nalidixic acid (400 mg/kg body weight) or vehicle on days 0 and 1. After euthanasia on day 4, synovial fluid was collected for cytologic evaluation and for analysis of concentrations of hyaluronan, proteoglycans, total protein, and collagen as hydroxyproline. Slides of formalin-fixed decalcified tissues were stained with hematoxylin-eosin or safranin O for histologic scoring of lesion severity. Nine of 10 hamsters treated with nalidixic acid had fissures within articular cartilage of the femur and reduced safranin O staining of matrix indicative of loss of proteoglycans. Synovial membranes from affected joints, however, were not inflamed. Synovial fluid cell counts and cytomorphology were unaffected by treatment. In synovial fluid from 5 of 10 treated hamsters, proteoglycans were elevated by more than 2 SDs above the control group, and individual animal levels correlated with the histologic severity score (r2 = 0.36; p = 0.02). The hyaluronan content of the synovial fluid from treated hamsters was mildly but significantly elevated (p = 0.005), and the histologic severity score again correlated with individual animal levels (r2 = 0.42; p = 0.01). Hydroxyproline was unaffected by treatment. Although synovial fluid changes and histologic changes were correlated on a group basis, interanimal variability was significant and the magnitude of biochemical changes were far smaller than those that occur during inflammation. Changes in synovial fluid composition are not sufficiently robust to predict cartilage changes in individual animals.

Comparison of liver progenitor cells in human atypical ductular reactions with those seen in experimental models of liver injury

The ultrastructural characteristics of liver progenitor cell types of human atypical ductular reactions seen in chronic cholestasis, in regenerating human liver after submassive necrosis, in alcoholic liver disease, and in focal nodular hyperplasia are compared with liver progenitor cell types seen during experimental cholangiocarcinogenesis in hamsters; during hepatocarcinogenesis in rats; and in response to periportal liver injury induced by allyl alcohol in rats. Three types of progenitor cells have been identified in human atypical ductular reactions: type I: primitive, has an oval shape, marginal chromatin, few cellular organelles, rare tonofilaments, and forms desmosomal junctions with adjacent liver cells; type II: bile duct-like, is located within ducts, has few organelles, and forms lateral membrane interdigitations with other duct-like cells; and type III: hepatocyte-like, is located in hepatic cords, forms a bile canaliculus, has tight junctions with other hepatocyte-like cells, prominent mitochondria and rough endoplasmic reticulum, and some have lysosomes and a poorly developed Golgi apparatus. Each type is seen during cholangiocarcinogenesis in hamsters, but the most prominent cell type is type II, duct-like. A more primitive cell type ("type 0 cell"), as well as type I cells, are seen in the intraportal zone of the liver within 1 to 2 days after carcinogen exposure or periportal injury in the rat, but both type II and type III are seen later as the progenitor cells expand into the liver lobule. After allyl alcohol injury, type 0 cells precede the appearance of type I and type III cells, but most of the cells that span the periportal necrotic zone are type III hepatocyte-like cells showing different degrees of hepatocytic differentiation. Some type II cells are also seen, but these are essentially limited to ducts. It is concluded that there is a primitive stem cell type in the liver (type 0) that may differentiate directly into type I and then into type II, duct-like or or type III hepatocyte-like cells. The terms oval cell, transitional hepatocyte, biliary hepatocyte, hepatocyte-like cell, atypical ductular cell, neocholangiole, etc., are used to describe these cells. Although these terms are useful as general descriptive terms for liver precursor cells at the light microscopic level, the cells included in these descriptive categories may be very different from one another biologically and ultrastructurally.

Modulation of taurine levels in the rat liver alters methylene dianiline hepatotoxicity

Methylene dianiline (DAPM) causes hepatic damage and bile duct necrosis in rats. This has been detected histologically and biochemically. The toxicity was dose related over the range 0-100 mg/kg but the dose response relationship showed a maximum at about 75-100 mg/kg. This was true for both histopathology and biochemical parameters of liver dysfunction. When animals were depleted of taurine using beta-alanine pretreatment, the toxicity of DAPM was increased. Conversely treatment of rats with taurine, significantly attenuated the rise in alanine transaminase (ALT). However depletion of taurine with guanidinoethanesulphonate (GES) attenuated rises in both transaminases. It is concluded that taurine may play a role in the toxicity of DAPM but that GES, although depleting taurine as does beta-alanine, causes additional effects such as increasing glutathione (GSH), perhaps leading to protection.

Accidental intoxication with methylene dianiline p,p'-diaminodiphenylmethane: acute liver damage after presumed ecstasy consumption

BACKGROUND

MDA is the abbreviation for methylene dianiline (p,p' diaminodiphenylmethane; 4,4'-methylenedianiline; CAS 101-77-9); and for methylendioxyamphetamine (MDMA, N, alpha-dimethyl-1,3-benzodioxole-5-ethanamine; CAS 42542-10-9). While the former is used for the production of polyurethane foams, the latter is a psychometric drug, which is becoming increasingly popular in the techno scene.

METHODS

We report six participants of a technoparty (1 female, 5 males, ages 17-25) who were admitted to the hospital with severe colicky abdominal pain and subsequently developed symptoms of hepatotoxicity. They had ingested an alcoholic beverage that had been spiked with a powdery substance they dubbed MDA.

RESULTS

All patients showed similar clinical symptoms, with an identical time course. Acute jaundice developed within 2 days after ingestion. Enzymes indicating cholestasis increased steadily over 7 days and reached peak values of 800 U/L (AP) and 380 U/L (GGT), whereas transaminases remained moderately elevated. Between days 5 and 7, all patients became febrile for one day, their body temperatures rising up to 40 degrees C. There was no evidence for hemolysis or an infectious hepatitis. Toxicological analysis revealed the presence of p,p'-diaminodiphenylmethane (4,4'-methylenedianiline) at a concentration of 130 mg/L in one of two urine extracts examined.

CONCLUSIONS

The analytical data indicate that the participants of the technoparty assumed the aniline-derivative, the cause of Epping Jaundice, was methylendioxyamphetamine because the same abbreviation, MDA, is used for both compounds. An overview of the acute liver toxicity of aniline derivatives is given and the possibility of amphetamine-induced liver damage is discussed.

Toxicity of quinolone antimicrobial agents

An approach to minimization of toxicity of a new compound is to elucidate the mechanisms of toxicity of analogous compounds and to clarify their structure-toxicity relationships. A problem with this approach, however, is that such elucidation remains difficult. For quinolones, some improvements in this mechanistic approach have been achieved in the central nervous system (CNS), particularly with regard to their interaction with non-steroidal anti-inflammatory drugs (NSAIDs), and in genotoxicity and phototoxicity studies, particularly in comparison with other toxicities, such as to the cardiovascular, gastrointestinal, bone, reproductive, and developmental systems. This review concentrates on a description of the known effects of quinolones on various organ systems in experimental animals and humans. Given the logarithmic increase in the synthesis of new quinolones, it is questionable whether these drugs share similar safety and efficacy. Nevertheless, this mechanistic approach to the investigation and minimization of toxicity has produced satisfactory results to date and deserves to be continued.

Distribution of glucose-6-phosphatase activity in normal, hyperplastic, and preneoplastic rat liver

The significance of glucose-6-phosphatase (G6P) expression by bile duct-like cells proliferating during hepatocarcinogenesis in the histogenesis of hepatocellular carcinoma is not clear. To this end, we measured the histochemical and biochemical activity of G6P in normal rat liver, and in rat livers in which bile duct-like proliferation was induced by either hyperplastic (bile duct ligation for 14 days or feeding alpha-naphthylisothiocyanate for 28 days) or neoplastic (feeding a choline-devoid diet containing 0.1% ethionine for 60 days) regimens. In normal, hyperplastic, and preneoplastic livers, G6P histochemical activity was confined to the hepatocytes; proliferated bile duct-like cells, like normal bile ducts, did not display visible G6P staining. When the enzyme activity was determined biochemically, however, hydrolysis of glucose-6-phosphate was observed in both parenchymal and nonparenchymal liver cells isolated from all experimental animals. In elutriated nonparenchymal fractions, G6P activity was directly proportional to the number of cells positive for gamma-glutamyl transpeptidase and cytokeratin no. 19 (markers of bile duct cells) and inversely proportional to the number of cells positive for vimentin (marker of mesenchymal cells). These results indicate that, while by light microscopy hepatic G6P histochemical activity is detectable only in the hepatocytes, the biochemical activity is also expressed in proliferating bile duct-like cells. However, the nonparenchymal activity is observed during both neoplastic and hyperplastic liver growth, thus indicating that the presence of this enzyme in bile duct-like cells proliferating during hepatocarcinogenesis should not necessarily be construed as supporting their stem cell nature nor their neoplastic commitment.

Methylene dianiline: acute toxicity and effects on biliary function

4,4'-Methylene dianiline (4,4'-diaminodiphenylmethane, DAPM), which is used in the polymer industry, causes hepatobiliary damage in exposed humans. Our objectives were to characterize the acute toxicity of DAPM in liver, particularly on secretion of biliary constituents and on biliary epithelial cell gamma-glutamyl transpeptidase (GGT) activity. Biliary cannulas were positioned in Sprague-Dawley male rats under pentobarbital anesthesia. After 1 hr of control bile collection, each rat was given 250 mg DAPM/kg (50 mg/ml) po in 35% ethanol or 35% ethanol only; bile was collected for a further 4 hr. Groups of rats were also examined for liver injury and biliary function at 8 and 24 hr after DAPM. Four hours after DAPM administration, main bile duct cells were severely damaged with minimal damage to peripheral bile ductule cells. Focal periportal hepatocellular necrosis and extensive cytolysis of cortical thymocytes occurred by 24 hr. Serum indicators of liver injury were elevated by 4 hr and continued to rise through 24 hr. By 4 hr, biliary protein concentration was increased 4-fold while concentrations of biliary bile salt, bilirubin, and glutathione were decreased by approximately 80, 50, and 200%, respectively. DAPM also induced a striking effect on biliary glucose with an approximately 20-fold increase. Histochemical staining of main bile duct GGT was absent by 8 hr after DAPM. Bile flow was diminished by 40% at 4 hr; three of five rats had no bile flow by 8 hr and none had any bile flow by 24 hr. These results indicate that DAPM rapidly diminishes bile flow and alters the secretion of biliary constituents and is highly injurious to biliary epithelial cells.

Quinolone arthropathy--acute toxicity to immature articular cartilage

A class effect of quinolone antibacterial agents observed during animal toxicity testing is a specific arthropathy (QAP). Despite the growing list of laboratory animals susceptible to QAP and reports of arthralgia in patients treated with quinolones, the potential for QAP development in humans remains unknown. This review discusses current concepts in the biology of articular cartilage and how these concepts elucidate QAP pathogenesis. Biomechanical forces within synovial joints and toxicokinetic properties of quinolones contribute to QAP induction. Since a limited number of mechanistic pathways exist for acute articular damage, QAP may serve as a research tool to probe the pathobiology of injury to articular cartilage.

Ultrastructural changes in articular cartilages of immature beagle dogs dosed with difloxacin, a fluoroquinolone

The ultrastructural features of quinolone-induced arthropathy were studied in the humeral and femoral heads of nine skeletally immature Beagle dogs (3 months old) that were dosed orally with difloxacin at 300 mg/kg body weight and euthanatized 24, 36, or 48 hours later in groups of three. Three age-matched dogs were given a placebo and euthanatized after 48 hours. Mitochondria in chondrocytes had significantly greater cross-sectional areas (P less than 0.05) in electron micrographs from dogs euthanatized after 48 hours of treatment than did those in other groups. There was also a significantly greater percentage of chondrocytes with swollen mitochondria in treated dogs than in the controls (P less than 0.05). These changes preceded the necrosis observed in some chondrocytes in the dogs of the 48-hour group. Disruption of extracellular matrix was first observed in the pericellular matrix of necrotic chondrocytes, indicating that this change was secondary to the changes in chondrocytes. Fissures within cartilages apparently resulted from the loss of the normal association of proteoglycans with collagen fibrils.

Morphologic and biochemical changes in articular cartilages of immature beagle dogs dosed with difloxacin

Quinolones are efficacious antibacterial compounds, but they have been associated with arthralgia in human patients; experimentally, they have caused lesions in articular cartilages of immature animals. The earliest morphologic and biochemical changes induced in articular-epiphyseal cartilage complexes by difloxacin, a fluoroquinolone, were investigated in 27 3-month-old Beagle dogs that were dosed orally with the drug at 300 mg/kg body weight per day. Paraffin-embedded sections of humeral and femoral heads that were stained with either hematoxylin and eosin or toluidine blue and fast green were evaluated histologically, and lesions were scored according to established criteria. Although morphologic changes were not observed in cartilages of the control dogs or of the treated dogs in the 24-hr group, the severity of lesions, as represented by mean scores for lesions, increased during the 36-48 hr after dosing. The initial morphologic change, observed in cartilages from the treated dogs of the 36- and the 48-hr groups, was necrosis of chondrocytes that was rapidly followed by disruption of extracellular matrix and formation of fissures. Although glycosaminoglycan was aggregated along the margins of fissures, its concentration was not reduced in cartilages of any group of treated dogs. Collagen, however, was depleted from the cartilages of the dogs that were euthanized 36 or 48 hr after the first dose of difloxacin. Because degenerative changes were observed ultrastructurally in chondrocytes by 24 hr in a previous study, it was concluded that collagen was lost from affected cartilages as an early sequel to the degeneration of chondrocytes.

Mouse strain differences in susceptibility to sporidesmin-induced biliary tract injury

Biliary tract injury was examined in four inbred strains of mice orally dosed with 500 micrograms of the fungal toxin sporidesmin. Semiquantitative histological analysis was used to assess the grade of necroinflammatory changes in the gall bladder, intra- and extrahepatic biliary tree and lobular parenchyma. Injury was greatest in the C57BL/6 and C3H strain mice and was least in SJL/J mice. In these strains injury was greatest at 4 days and had regressed by 10 days. In BALB/c mice the damage, although similar to that in SJL/J mice at 4 days, persisted at the same severity at day 10 and was accompanied by periductal fibrosis and occasionally by obliteration of ducts typical of sclerosing cholangitis. Analysis of the time-course of development of the lesions in C57BL/6 mice showed that the primary target for the toxin is the biliary epithelium. The severity of the lesions within the liver increased centripetally and the worst affected ducts were found at the confluence of the lobar ducts with the common bile duct. The variation in the degree of damage and rate of healing between strains may be due to differences in sporidesmin excretion in bile or interactions with biliary epithelial cells and/or efficacy of protective cellular repair mechanisms.

Origin, pattern, and mechanism of bile duct proliferation following biliary obstruction in the rat

Proliferation of bile duct-like structures is a hepatic cellular reaction observed in most forms of human liver disease and in a variety of experimental conditions associated with liver injury. Yet the origin, means of initiation, and significance of this hyperplasia are unknown. To clarify these issues we induced bile duct proliferation in rats by ligating the common bile duct and studied (a) hepatic incorporation of [3H]thymidine by histoautoradiography, (b) hepatic morphometry, (c) biliary tree volume using [3H]taurocholate as a marker of biliary transit time, (d) immunohistochemical expression of cytokeratin no. 19, (e) the effect of indomethacin, and (f) the role of increased biliary pressure, in the absence of physiological and biochemical evidence of cholestasis, on [3H]thymidine incorporation by the bile-duct cells. The results have demonstrated that (a) the proliferating bile duct-like cells are products of the extant biliary epithelium and retain its characteristics; (b) bile duct cells divide irrespective of the size of the duct in which they are located and form a system with a lumen continuous with the preexisting one; (c) bile duct proliferation results mainly in elongation, not in circumferential enlargement or sprouting of side branches; (d) portal macrophage infiltration does not play a role in the hyperplastic reaction, and (e) increased biliary pressure is the initiating factor in bile duct cell division. Our results provide evidence that under the present conditions, ductular metaplasia of hepatocytes does not occur and there is no functioning stem cell for biliary epithelial growth segregated in any particular duct size or within the portal connective tissue.

Determination of 4,4'-methylenedianiline in hydrolysed human urine using liquid chromatography with UV detection and peak identification by absorbance ratio

A liquid chromatographic method using multi-wavelength UV detection (258 and 285 nm) is presented for the determination of 4,4'-methylenedianiline (MDA) in hydrolysed human urine. The method is based on hydrolysis under strongly acidic conditions followed by derivatization with pentafluoropropionic anhydride. The perfluoro fatty acid amide derivative formed was analysed on a bonded octadecylsilyl column using isocratic elution with acetonitrile-water (67:33, v/v) as mobile phase. The overall recovery for urine samples containing 115 micrograms/l of MDA was 97 +/- 3%. The calibration graph was linear in the investigated range (12-122 micrograms/l) with a correlation coefficient of 0.998. The precision was 2.3% for urine samples containing 122 micrograms/l and the detection limit was 8 micrograms/l. The chromatograms were evaluated using a combination of retention time data and absorbance ratio by the simultaneous monitoring of the wavelengths 285 and 258 nm. The absorbance ratio (285/258 nm) was virtually constant (0.28 +/- 0.04) in the range 78-10,000 micrograms/l. The precision for the absorbance ratio was 6.1% for urine samples containing 124 micrograms/l and the lowest amount of MDA to give an absorbance ratio was 50 micrograms/l. The procedure for the hydrolysis of urine spiked with MDA and N,N'-diacetyl-MDA and urine from skin-exposed workers was studied under strongly acidic, weakly acidic and basic conditions. MDA was found in hydrolysed urines from skin-exposed epoxy resin workers in the concentration range 8-700 micrograms/l.

Histologic and histochemical changes in articular cartilages of immature beagle dogs dosed with difloxacin, a fluoroquinolone

The histologic and histochemical features of quinolone-induced arthropathy were studied using 14 skeletally immature Beagle dogs (3 to 4 months old) dosed orally with difloxacin at 300 mg/kg body weight once daily for 1, 2, 5, or 7 days. A placebo was given to eight other age-matched Beagle dogs that served as controls. A scoring technique that included lesion size and histologic features was used to determine the progression of lesions. Articular-epiphyseal cartilage complexes on the femoral and humeral heads and tibial tarsal bone were identified as predilection sites. Within predilection sites on femoral and humeral heads, lesions developed in specific areas. Lesions appeared within 2 days of the onset of treatment, and lesion scores increased with time. Grossly, the lesions were raised, fluid-filled vesicles on the articular surface. Histologic changes included vesicle formation with loss of proteoglycan, clumping of unmasked collagen, and degeneration and necrosis of chondrocytes. In lesions with higher scores, chondrocytes were often in clusters or they were undergoing metaplasia toward spindle-shaped cells. Although dissolution of matrix and necrosis of chondrocytes were typical of all lesions, smaller lesions had histologically normal chondrocytes adjacent to small vesicles. In sections stained with toluidine blue, proteoglycan was aggregated with collagen fibrils or was absent from the matrix adjacent to vesicles. Unique features, such as biomechanical forces, may predispose specific areas of articular cartilage to develop lesions.

Cellular events during hepatocarcinogenesis in rats and the question of premalignancy

The cellular, biochemical, and genetic changes that occur in the liver of rats exposed to chemical hepatocarcinogens are reviewed. Multiple new cell types appear in the liver of carcinogen-treated rats including foci, nodules, ducts, oval cells, and atypical hyperplastic areas. The application of phenotypic markers for these cell types suggests that hepatocellular carcinomas may arise from more than one cell type, including a putative liver stem cell that proliferates following carcinogen exposure. Study of DNA, RNA, and proteins produced by hepatocellular carcinomas and putative premalignant cells has so far failed to identify a gene or gene product clearly associated with the malignant or premalignant phenotype. Understanding the cellular lineage from normal cell through putative premalignant cell to cancer is critical to understanding the process of carcinogenesis. Application of new immunological (monoclonal antibody, transplantation) and molecular biological (gene cloning, oncogene identification) approaches to this problem holds promise that the process of hepatocarcinogenesis will be better known in the near future.

Methylene dianilene: a new toxic cause of visual failure with hepatitis

A 28-year-old man ingested methylene dianilene in potassium carbonate and gamma-butyrolactone. He developed toxic optic neuritis, with severe visual dysfunction (not previously reported in humans), prolonged toxic hepatitis, with disturbed liver-function tests 18 months after the incident, and other more transient effects. The course of his illness is described and the literature on methylene dianilene toxicity is reviewed.

Biliary cirrhosis and tumors induced by chronic administration of thiobenzamide to rats

Thiobenzamide (TB), a thiono-containing compound, was administered for 38 weeks to female Sprague-Dawley rats at a dose of 1 g/kg standard diet; the resulting liver pathology was followed up to 8 months after withdrawal of the compound from the diet. TB administration induced the appearance of biliary cirrhosis. In the first weeks of intoxication the progressive distortion of the liver architecture was mainly due to significant proliferation of the bile ductules. Later, the liver assumed a macronodular appearance. In addition to regenerative and degenerative changes of the hepatocytes, preneoplastic lesions were also detected, and some enzymic markers of the mixed-function monooxygenase system were decreased. Cholangiofibrotic areas were evident, and many biliary tubules within them showed mucous metaplasia. At the end of the intoxication period, as well as 4 months after drug suspension, large portions of the liver or entire lobes were substituted by connective tissue surrounding nests of bile ductules and atrophied hepatocellular nodules. Four months later, in the virtual absence of cirrhotic changes, each animal harboured one or more tumors (mainly cholangiomas).