Mutation of a major keratin phosphorylation site predisposes to hepatotoxic injury in transgenic mice.

Simple epithelia express keratins 8 (K8) and 18 (K18) as their major intermediate filament (IF) proteins. One important physiologic function of K8/18 is to protect hepatocytes from drug-induced liver injury. Although the mechanism of this protection is unknown, marked K8/18 hyperphosphorylation occurs in association with a variety of cell stresses and during mitosis. This increase in keratin phosphorylation involves multiple sites including human K18 serine-(ser)52, which is a major K18 phosphorylation site. We studied the significance of keratin hyperphosphorylation and focused on K18 ser52 by generating transgenic mice that overexpress a human genomic K18 ser52--> ala mutant (S52A) and compared them with mice that overexpress, at similar levels, wild-type (WT) human K18. Abrogation of K18 ser52 phosphorylation did not affect filament organization after partial hepatectomy nor the ability of mouse livers to regenerate. However, exposure of S52A-expressing mice to the hepatotoxins, griseofulvin or microcystin, which are associated with K18 ser52 and other keratin phosphorylation changes, resulted in more dramatic hepatotoxicity as compared with WT K18-expressing mice. Our results demonstrate that K18 ser52 phosphorylation plays a physiologic role in protecting hepatocytes from stress-induced liver injury. Since hepatotoxins are associated with increased keratin phosphorylation at multiple sites, it is likely that unique sites aside from K18 ser52, and phosphorylation sites on other IF proteins, also participate in protection from cell stress.

Hepatic damage and oxidative stress induced by Griseofulvin in mice.

Erythropoietic Protoporphyria (EPP) is a disease associated with ferrochelatase deficiency, which produces accumulation of protoporphyrin IX (PROTO IX) in erythrocytes, liver and skin. In some cases, a severe hepatic failure and cholestasis was observed. Griseofulvin (Gris) develops an experimental EPP with hepatic manifestations in animals. The aim of this work was to further characterize this model studying its effect on different metabolisms in mice Gris feeding (0-2.5%, 7 and 14 days). PROTO IX accumulation in liver, blood and feces, induction of ALA-S activity, and a low rate of Holo/Apo tryptophan pyrrolase activity was produced, indicating a reduction of free heme pool. The progressive liver injury was reflected by the aspect and the enlargement of liver and the induction of hepatic damage. Liver redox balance was altered due to porphyrin high concentrations; as a consequence, the antioxidant defense system was disrupted. Heme oxygenase was also induced, however, at higher concentrations of antifungal, the free heme pool would be so depleted that this enzyme would not be necessary. In conclusion, our model of Protoporphyria produced liver alterations similar to those found in EPP patients.

Susceptibility to hepatotoxicity in transgenic mice that express a dominant-negative human keratin 18 mutant.

Keratins 8 and 18 (K8/18) are intermediate filament phosphoglycoproteins that are expressed preferentially in simple-type epithelia. We recently described transgenic mice that express point-mutant human K18 (Ku, N.-O., S. Michie, R.G. Oshima, and M.B. Omary. 1995. J. Cell Biol. 131:1303-1314) and develop chronic hepatitis and hepatocyte fragility in association with hepatocyte keratin filament disruption. Here we show that mutant K18 expressing transgenic mice are highly susceptible to hepatotoxicity after acute administration of acetaminophen (400 mg/Kg) or chronic ingestion of griseofulvin (1.25% wt/wt of diet). The predisposition to hepatotoxicity results directly from the keratin mutation since nontransgenic or transgenic mice that express normal human K18 are more resistant. Hepatotoxicity was manifested by a significant difference in lethality, liver histopathology, and biochemical serum testing. Keratin glycosylation decreased in all griseofulvin-fed mice, whereas keratin phosphorylation increased dramatically preferentially in mice expressing normal K18. The phosphorylation increase in normal K18 after griseofulvin feeding appears to involve sites that are different to those that increase after partial hepatectomy. Our results indicate that hepatocyte intermediate filament disruption renders mice highly susceptible to hepatotoxicity, and raises the possibility that K18 mutations may predispose to drug hepatotoxicity. The dramatic phosphorylation increase in nonmutant keratins could provide survival advantage to hepatocytes.

Griseofulvin resistance mutation of Chinese hamster ovary cells that affects the apparent molecular weight of a congruent to 200,000-dalton protein.

A single-step griseofulvin-resistant mutant (GrsR-4) of CHO cells which exhibit very specific cross-resistance towards certain microtubule inhibitors showed the absence of a protein of molecular weight congruent to 200,000 (designated P5) and the concomitant presence of a new protein spot, M5, of lower molecular weight (Mr congruent to 180,000) which is not present in other cell lines. Peptide mapping studies showed that proteins P5 and M5 are related to each other and that M5 may be missing a peptide fragment present in P5. In GrsR-4 X GrsS cell hybrids, both P5 and M5 were present in equal amounts, which provided evidence against post-translation mechanisms in the origin of M5 and indicated that the GrsR-4 mutant most likely contains a nonsense mutation in the structural gene for protein P5, which causes its premature termination and leads to the formation of M5. Our studies also showed that in different Chinese hamster cell lines the two alleles of the protein P5 are nonidentical and make protein products which differ from each other in isoelectric points. It is suggested that protein P5 and its isoelectric variant P6 may constitute microtubule-associated proteins.

Comparative study of human neuronal and glial cell sensitivity for in vitro neurogenotoxicity testing.

Cell cultures from neuronal and glial origin have proven to be powerful tools for elucidating cellular and molecular mechanisms of nervous system development and physiology, and as neurotoxicity models to evaluate in vitro the possible effects of chemicals. But cellular heterogeneity of nervous system is considerable and these cells have been shown to respond diversely to neurotoxic insults, leading to disparate results from different studies. To shed more light on suitability of cellular models of nervous origin for neurotoxicity screening, the objective of this study was to compare the sensitivity to genetic damage induction of two nervous cell lines. To this aim, neurons (SH-SY5Y) and glial (A172) cells were treated with differently-acting genotoxic agents (bleomycin, actinomycin-D, methyl methanesulfonate, mitomycin C, and griseofulvin). After discarding cytotoxicity, genotoxicity was evaluated by a battery of assays encompassing detection of different genetic lesions. Results obtained showed that glial cells are generally more resistant to genotoxic damage induced by clastogenic agents, but more sensitive to aneugenic effects. These results highlight the need of proper design of in vitro neurotoxicology studies, especially for neurogenotoxicity screening, emphasizing the importance of employing more than one nervous cell type for testing the potential toxicity of a particular exposure.

SFTG international collaborative study on in vitro micronucleus test V. Using L5178Y cells.

In this report, results are presented from an international study of the in vitro micronucleus assay using mouse lymphoma L5178Y cells. This study was coordinated by an organizing committee supported by the SFTG (the French branch of the European Environmental Mutagen Society). Test chemicals included mannitol, bleomycin, 5-fluorouracil, colchicine and griseofulvin. Mitomycin C was used as a positive control. Each chemical was evaluated in at least two laboratories following a variety of different protocols (short and long exposures, varying recovery times, with and without cytochalasin B) in order to help determine a standard protocol for routine testing in mouse lymphoma L5178Y cells. Mannitol was the only exception, being tested in only one laboratory. Mannitol was negative, while bleomycin induced a concentration-dependent increase in micronucleated cells. Equivocal results were obtained for 5-fluorouracil, colchicine and griseofulvin. High levels of cytotoxicity interfered with the assessment of aneuploidy for colchicine and griseofulvin, preventing the ability to obtain clear results in all the treatment schedules. Experiments with 5-fluorouracil, colchicine and griseofulvin showed that both short and long treatment times are required as each compound was detected using one or more treatment protocol. No clear differences were seen in the sensitivity or accuracy of the responses in the presence of absence of cytochalasin B. It was also found that a recovery period may help to detect compounds which induce a genotoxicity associated to a reduction in cell number or cell proliferation. Overall, the results of the present study show that mouse lymphoma L5178Y cells are suitable for the in vitro micronucleus assay.

SFTG international collaborative study on in vitro micronucleus test IV. Using CHL cells.

In this report, are presented the results of an international collaborative study on the in vitro micronucleus assay, using CHL cells. Fourteen laboratories participated in this study which was coordinated by an organizing committee supported by the SFTG (the French branch of the European Environmental Mutagen Society). Nine coded substances, having different modes of action and at different levels were assessed in the in vitro micronucleus test, using a common protocol. Mitomycin C was used as a positive control. In order to help to define a standard protocol on CHL cells, short and long treatment periods followed by various recovery times, with or without cytochalasin B, were compared. After an evaluation of the acceptability of the assays, the tested chemicals were classified as negative, positive or equivocal. Mannitol and clofibrate were judged as negative in all treatment schedules. Bleomycin was positive in all the treatment schedules, with an increase in the number of micronucleated cells in both mononucleate and binucleate cells when using cytochalasin B. This was also shown for the aneugens colchicine, diethylstilboestrol and griseofulvin, as expected. Urethane was judged as equivocal only after long treatment with cytochalasin B, and negative in all other treatment schedules. In any case, no genotoxic compound would have been missed with schedules including a short and a long treatment time, whether the treatment was followed by a recovery period or not and whether cytochalasin B was used or not. Thus, these results show that CHL cells were suitable for accurately detecting clastogenic and aneugenic compounds of various types in the in vitro micronucleus test.

SFTG international collaborative study on in vitro micronucleus test I. General conditions and overall conclusions of the study.

This study, coordinated by the SFTG (French branch of European Environmental Mutagen Society), included 38 participants from Europe, Japan and America. Clastogens (bleomycin, urethane), including base and nucleoside analogs (5-fluorouracil and cytosine arabinoside), aneugens and/or polyploidy inducers (colchicine, diethylstilboestrol, griseofulvin and thiabendazole), as well as non-genotoxic compounds (mannitol and clofibrate), were tested. Four cell types were used, i.e. human lymphocytes in the presence of cytochalasin B and CHO, CHL and L5178Y cell lines, in the presence or absence of cytochalasin B, with various treatment-recovery schedules. Mitomycin C was used as a positive control for all cell types. Mannitol and clofibrate were consistently negative in all cell types and with all treatment-recovery conditions. Urethane, known to induce questionable clastogenicity, was not found as positive. Bleomycin and mitomycin C were found positive in all treatment-recovery conditions. The base and nucleoside analogs were less easy to detect, especially 5-fluorouracil due to the interference with cytotoxicity, while cytosine arabinoside was detected in all cell types depending on the treatment-recovery schedule. Aneugens (colchicine, diethylstilboestrol and griseofulvin) were all detected in all cell types. In this study, the optimal detection was ensured when a short treatment followed by a long recovery was associated with a long continuous treatment without recovery. There was no impact of the presence or absence of cytochalasin B on the detection of micronucleated cells on cell lines. Scoring micronucleated cells in both mononucleated and binucleated cells when using cytochalasin B was confirmed to be useful for the detection and the identification of aneugens. In conclusion, these results, together with previously published validation studies, provide a useful contribution to the optimisation of a study protocol for the detection of both clastogens and aneugens in the in vitro micronucleus test.

Effect of Griseofulvin on 5-aminolevulinate synthase and on ferrochelatase in mouse liver neoplastic nodules.

Treatment of mice with griseofulvin for 8 months induced hepatocellular nodules in the liver which persist after discontinuation of griseofulvin feeding. We investigated the porphyrogenic effect of griseofulvin on these nodules and surrounding nonneoplastic liver after renewed short-term exposure of tumor-bearing mice to this agent. Griseofulvin treatment for 4 days led to marked elevation of the activity of 5-aminolevulinate synthase in peritumoral (3.8-fold) and control (6-fold) liver. The increase in enzyme activity was much less pronounced in the nodules (1.5-fold). Ferrochelatase activity was markedly decreased under the same experimental conditions in both peritumoral and control livers (to 18 and 13.5%, respectively, of the pretreatment values), but the effect was considerably smaller in nodules (to 40% of the pretreatment value). These results may explain the lack of porphyrin accumulation in tumor tissue.

Effect of decreased ferrochelatase activity on iron and porphyrin content in mitochondria of mice with porphyria induced by griseofulvin.

The content of iron and protoporphyrin in liver mitochondria from mice with porphyria induced by griseofulvin was measured. The amount of porphyrin was 0.0076 +/- 0.0043, 4.11 +/- 0.58 and 22.2 +/- 6.8 nmol/mg protein (n = 5) in mitochondria from control animals and animals treated with griseofulvin for 3 days and 4-5 weeks, respectively. The energy coupling of the mitochondria was greatly diminished after 4-5 weeks of treatment, and the ferrochelatase activity was inhibited 80-90%, compared to that of control animals. Mitochondrial preparations isolated by differential centrifugation were contaminated with iron-containing lysosomes which could be removed by Percoll density-gradient centrifugation. In purified mitochondrial preparations no change in the amount of non-heme iron was found after griseofulvin feeding, representing 3.36 +/- 0.15, 3.97 +/- 0.40 and 3.59 +/- 0.23 nmol/mg protein for control animals, 3 days- and 4-5 weeks-treated animals, respectively (n = 4). A mitochondrial iron pool previously identified in rat liver mitochondria and shown to be available for heme synthesis in vitro (Tangerås, A. (1985) Biochim. Biophys. Acta 843, 199-207) was also present in mitochondria from mice. The magnitude of this iron pool, as well as its availability for heme synthesis, was not changed after treatment of the animals with griseofulvin. The fact that porphyrin, but not iron, accumulated in the mitochondria when ferrochelatase was inhibited is discussed with regard to our understanding of the process of heme synthesis and its regulation.

Detection of numerical chromosomal aberrations by flow cytometry: a novel process for identifying aneugenic agents.

Aneuploidy plays a significant role in adverse human health conditions including birth defects, pregnancy wastage, and cancer. Currently, there is no screening method sufficiently validated that can be used routinely to identify aneugenic agents in vitro because most conventional test systems rely on the labor-intensive microscopic assessment of the aneuploid cell population. Our laboratory has recently developed a flow cytometry-based procedure for assessing numerical chromosomal aberrations in mitotic populations of lymphocytes on the basis of DNA content. Studies were conducted in 24 h treated human lymphocyte cultures to determine the sensitivity of this flow cytometry-based procedure to detect aneugenic agents. A comparison between the microscopic and the flow cytometry-based procedures for scoring polyploidy shows a strong agreement exists between the two methods. Treatments with two known aneugenic agents, griseofulvin, and paclitaxel (taxol), resulted in a dose-related increase in the mitotic index, aneuploidy, and polyploidy. In contrast, results from the treatments with two known clastogenic agents, mitomycin-C, and etoposide, show a dose-related decrease in the mitotic index with a slight increase in the frequency of hypodiploidy at concentrations that produce severe chromosomal breakage. There were no increases in hyperdiploidy and polyploidy observed. In conclusion, the reproducibility of the results obtained in this study indicates that this flow cytometry-based procedure for assessing numerical chromosomal effects in mitotic populations on the basis of DNA content is promising for the routine detection and characterization of aneugenic agents.

A metabolomic perspective of griseofulvin-induced liver injury in mice.

Griseofulvin (GSF) causes hepatic porphyria in mice, which mimics the liver injury associated with erythropoietic protoporphyria (EPP) in humans. The current study investigated the biochemical basis of GSF-induced liver injury in mice using a metabolimic approach. GSF treatment in mice resulted in significant accumulations of protoporphyrin IX (PPIX), N-methyl PPIX, bile acids, and glutathione (GSH) in the liver. Metabolomic analysis also revealed bioactivation pathways of GSF that contributed to the formation of GSF-PPIX, GSF-GSH and GSF-proline adducts. GSF-PPIX is the precursor of N-methyl PPIX. A six-fold increase of N-methyl PPIX was observed in the liver of mice after GSF treatment. N-methyl PPIX strongly inhibits ferrochelatase, the enzyme that converts PPIX to heme, and leads to PPIX accumulation. Excessive PPIX in the liver results in bile duct blockage and disturbs bile acid homeostasis. The accumulation of GSH in the liver was likely due to Nrf2-mediated upregulation of GSH synthesis. In summary, this study provides the biochemical basis of GSF-induced liver injury that can be used to understand the pathophysiology of EPP-associated liver injury in humans.

Experimental protoporphyria: effect of bile acids on liver damage induced by griseofulvin.

The effect of bile acids administration to an experimental mice model of Protoporphyria produced by griseofulvin (Gris) was investigated. The aim was to assess whether porphyrin excretion could be accelerated by bile acids treatment in an attempt to diminish liver damage induced by Gris. Liver damage markers, heme metabolism, and oxidative stress parameters were analyzed in mice treated with Gris and deoxycholic (DXA), dehydrocholic (DHA), chenodeoxycholic, or ursodeoxycholic (URSO). The administration of Gris alone increased the activities of glutathione reductase (GRed), superoxide dismutase (SOD), alkaline phosphatase (AP), gamma glutamyl transpeptidase (GGT), and glutathione-S-transferase (GST), as well as total porphyrins, glutathione (GSH), and cytochrome P450 (CYP) levels in liver. Among the bile acids studied, DXA and DHA increased PROTO IX excretion, DXA also abolished the action of Gris, reducing lipid peroxidation and hepatic GSH and CYP levels, and the activities of GGT, AP, SOD, and GST returned to control values. However, porphyrin accumulation was not prevented by URSO; instead this bile acid reduced ALA-S and the antioxidant defense enzymes system activities. In conclusion, we postulate that DXA acid would be more effective to prevent liver damage induced by Gris.

Rearrangement of the hepatocyte cytoskeleton after toxic damage: involution, dispersal and peripheral accumulation of Mallory body material after drug withdrawal.

Large cytoplasmic aggregates (Mallory bodies) containing randomly arranged 10 to 20 nm thick filaments are characteristic structures of hepatocytes in certain liver diseases, especially alcoholic hepatitis of man and griseofulvin intoxication of mouse. Biochemical and immunological studies have shown that such Mallory bodies contain structural proteins related to the cytokeratins present in epithelial cells of various kinds, normal hepatocytes included. We have studied the mode of reorganization of the hepatocyte cytoskeleton during involution of Mallory bodies experimentally induced in mice by prolonged griseofulvin feeding after withdrawal of the drug, using immunofluorescence microscopy and electron microscopy. Dispersal of Mallory bodies into small granules accumulating in the hepatocyte periphery is a characteristic feature of Mallory body involution. Dispersed heaps of Mallory body material are seen in hepatocyte cortices at lateral (hepatocyte-to-hepatocyte) as well as at sinusoidal cell surfaces. Mallory body material dispersed to the cell periphery exhibits a filamentous ultrastructure an often shows conspicuous associations with desmosomes. Three different types of desmosome complexes can be discriminated in livers of mice allowed to recover from the intoxication: (i) desmosomes with normally looking bundles of tonofilaments; (ii) desmosomes associated with Mallory body filaments but few, if any, tono-filaments; and (iii) desmosomes lacking tonofilaments as well as Mallory body filaments. The associations between Mallory body filaments and desmosomes during Mallory body involution further strengthen the relationship between Mallory body filaments and tonofilaments. The observations suggest that Mallory body material, at least in part, may be utilized for rearrangement of the hepatocyte cytoskeleton, especially tonofilament-desmosome complexes.

Analysis of the six additional chemicals for in vitro assays of the European Economic Communities' EEC aneuploidy programme using Saccharomyces cerevisiae D61.M and the in vitro porcine brain tubulin assembly assay.

We tested six additional chemicals (acetaldehyde, benomyl, diethylstilboestrol, diethylstilboestrol dipropionate, griseofulvin, and mercaptoethanol) for in vitro systems of the coordinated programme to study aneuploidy induction sponsored by the Commission of the European Communities in two in vitro test systems. Using Saccharomyces cerevisiae D61.M (mitotic chromosomal malsegregation assay), benomyl showed a dose-dependent increase in the frequency of chromosomal malsegregation with a lowest effective dose tested (LEDT) of 30 micrograms/ml (0.1 mM). Diethylstilboestrol (DES) showed solvent-dependent effects. DES dissolved in ethanol induced an increase in chromosomal malsegregation as well as in the frequency of total resistant colonies (mutations and recombinations) with a LEDT around 13 micrograms/ml (0.048 mM). Using dimethylsulfoxide as the solvent, no increases were observed with DES up to 333 micrograms/ml (1.24 mM). Acetaldehyde induced an increase in chromosomal malsegregation with the cold treatment protocol (LEDT: 1.25 microliters/ml (21 mM) and 0.75 microliters/ml (13 mM), respectively) but no increase with the overnight protocol (highest dose tested (HDT): 1.75 microliters/ml; 30 mM). Concerning the frequency of total cycloheximide-resistant colonies (mutations and recombinations) increases were obtained with both protocols. The other three compounds were negative when tested up to toxic doses (survival below 10%), up to the maximum solubility in the solvent used or up to heavy precipitation in the incubation mix. The HDT were 333 micrograms/ml (0.88 mM) for diethylstilboestrol dipropionate, 1,600 micrograms/ml (4.5 mM) for griseofulvin and 0.5 microliters/ml (7 mM) for mercaptoethanol. Concerning effects on porcine brain tubulin assembly in vitro, diethylstilboestrol and griseofulvin inhibited the assembly process. The IC30% (30% inhibition concentration) values were 12.5 microM and 100 microM for DES and griseofulvin, respectively. Mercaptoethanol showed no effects up to 50 mM.

Variation of mouse oocyte sensitivity to griseofulvin-induced aneuploidy and meiotic delay during the first meiotic division.

The effects of varying the time of chemical treatment on the induction of aneuploidy and meiotic delay in metaphase II (MII) oocytes were studied by administering 1,500 mg/kg griseofulvin (GF) at 0, 2, 4, 6, or 8 hr after an injection of human chorionic gonadotrophin (HCG). The results show that the oocytes have a different sensitivity to GF-induced aneuploidy and meiotic delay during the course of meiotic maturation. Although not restricted to a particular period of meiotic maturation, the frequency of aneuploidy was highest (P < 0.05) when GF was given at 2, 4, or 6 hr after HCG. The maximum frequency of hyperploidy (42.4%) occurred at the 4-hr treatment time. Also, GF treatment resulted in the induction of meiotic delay as demonstrated by ovulated metaphase I (MI) and polyploid MII oocytes. The meiotic delay data depict a period of relative resistance between two periods of sensitivity in that the percentages of ovulated MI oocytes were 53.3, 21.3, 3.5, 6.7, and 25.7 when GF was given at 0, 2, 4, 6, and 8 hr after HCG, respectively. Also, at these treatment times the percentages of polyploid oocytes were 0.6, 1.7, 7.7, 20.1, and 15.4, respectively. Therefore, the oocytes seem to be more sensitive to GF-induced meiotic delay during the periods preceding and following meiotic spindle assembly. In conclusion, the results demonstrate that the time of chemical treatment influences the frequency of aneuploidy and the degree of meiotic delay. Also, the results emphasize that to thoroughly characterize the aneugenic potential of a specific chemical several treatment times may be needed.

The effect of griseofulvin on the heme pathway. Studies on tissue explant cultures.

The effect of griseofulvin on porphyrin biosynthesis was studied in vitro and in vivo using liver and skin explants from mice. Neither porphyrin nor precursors accumulation was detected after in vitro treatment of explants with griseofulvin. Culture medium porphyrins formed from added delta-amino-laevulinic acid (ALA) were increased after oral chronic intoxication. Similar results were obtained by cutaneous and oral intoxication. Consequently griseofulvin topical application is proposed as an alternative method to produce experimental erythropoietic protoporphyria. Probable intoxication mechanisms are proposed to explain the porphyrinogenic action of griseofulvin.

Bile secretion and liver microsomal mixed function oxidase system in mice with griseofulvin-induced hepatic protoporphyria.

Administration of 2.5% griseofulvin in the diet to male CD1 mice produced protoporphyria and cholestasis. Protoporphyria became evident as early as after 10 days of treatment, whereas cholestasis, expressed in terms of total bile flow reduction, developed only after 45 days of griseofulvin. Bile flow impairment was due both to the length of treatment and to the severity of liver protoporphyria. Griseofulvin administration was also associated with a significant modification of the relative amounts of hepatic microsomal cytochromes P-450 and b5, a loss in concentration/mg of protein of cytochrome P-450 and a concomitant increase of b5. Despite these changes, the activity of aniline hydroxylase expressed per mg of microsomal protein, assessed in vitro, was not modified.

Griseofulvin: generation time and ATP changes in the ciliate Tetrahymena pyriformis.

Tetrahymena pyriformis GL populations (20,000 cells/ml and 50,000 cells/ml) were studied at various times after exposure to griseofulvin. Growth rates were significantly different between controls and treated samples 90 min after treatment. ATP levels decreased in treated samples after 45 min. No differences were observed between controls and populations treated with acetone used to dissolve griseofulvin.

Indirect mechanisms of genotoxicity.

Indirect mechanisms of genotoxicity correspond to interactions of mutagens with non-DNA targets, and are expected to show threshold concentration-effect response curves. If these thresholds can be proven experimentally they may provide a third alternative for risk assessment, besides the No Effect Level/Safety Factor approach and the low dose linear extrapolation method. We contributed significantly to the in vitro assessment of thresholds in human lymphocytes exposed to the spindle inhibitors nocodazole and carbendazim showing dose dependency and existence of lower thresholds for induction of non-disjunction as compared to chromosome loss. Micronuclei correlated with p53-independent or p53-dependent apoptosis and elimination of aneuploid cells. Extrapolation from in vitro threshold values to the in vivo situation remains unsolved. Comparing the in vitro threshold values for griseofulvin in human and rat lymphocytes with in vivo NOAEL/LOAEL in bone marrow/gut/erythrocytes suggests that the in vitro human system is the most sensitive. The threshold for induction of non-disjunction in in vitro maturing, nocodazole-exposed mouse oocytes was in the same low range. Regulators (UK Committee on Mutagenicity, http://www.doh.gov.uk/com/com.htm) considered the importance of thresholds for indirect mechanisms of genotoxicity. Acceptance of a non-linear extrapolation for mutagens requires mechanistic studies identifying the mutagen/target interactions. Moreover appropriate risk evaluation will require additional studies on individual susceptibility for indirect mutagenic effects and on interactions of aneugens in complex mixtures.