Assessment of mechanisms driving non-linear dose-response relationships in genotoxicity testing.

In genetic toxicology, risk assessment has traditionally adopted linear dose-responses for any compound that causes genotoxic effects. Increasing evidence of non-linear dose-responses, however, suggests potential cellular tolerance to low levels of many genotoxicants with diverse modes of action. Such putative non-linear dose-responses need to be substantiated by strong mechanistic data that identifies the mechanisms responsible for the tolerance to low doses. This can be achieved by experimental demonstration of cytoprotective mechanisms and by providing experimental support for the existence of tolerance mechanisms against low dose effects. By highlighting key experiments into low dose mechanisms, this review aims to clarify which mechanistic data are required to support the use of non-linear dose-response models in risk assessment. Such key experiments are presented and discussed for alkylating agents, oxidants, particulate matter, nucleoside analogues, topoisomerase inhibitors and aneugens and exemplify the use of gene knockout models or transgenic models as well as chemical modulators of key effectors of relevant pathways and their impact on dose-response relationships. In vitro studies are particularly valuable to elucidate mechanisms of low-dose protection or lack thereof, while in vivo experiments are most appropriate for deriving a safe dose. In order to evaluate the existence of non-linear dose-response relationships for genotoxicants, we suggest that careful attention should be given to the mode of genotoxic action, relevant biomarkers of exposure, as well as to the existence and impact of potential cytoprotective mechanisms like detoxifying metabolism and DNA repair.

Automatic analysis of the micronucleus test in primary human lymphocytes using image analysis.

The in vitro micronucleus test (MNT) is a well-established test for early screening of new chemical entities in industrial toxicology. For assessing the clastogenic or aneugenic potential of a test compound, micronucleus induction in cells has been shown repeatedly to be a sensitive and a specific parameter. Various automated systems to replace the tedious and time-consuming visual slide analysis procedure as well as flow cytometric approaches have been discussed. The ROBIAS (Robotic Image Analysis System) for both automatic cytotoxicity assessment and micronucleus detection in human lymphocytes was developed at Novartis where the assay has been used to validate positive results obtained in the MNT in TK6 cells, which serves as the primary screening system for genotoxicity profiling in early drug development. In addition, the in vitro MNT has become an accepted alternative to support clinical studies and will be used for regulatory purposes as well. The comparison of visual with automatic analysis results showed a high degree of concordance for 25 independent experiments conducted for the profiling of 12 compounds. For concentration series of cyclophosphamide and carbendazim, a very good correlation between automatic and visual analysis by two examiners could be established, both for the relative division index used as cytotoxicity parameter, as well as for micronuclei scoring in mono- and binucleated cells. Generally, false-positive micronucleus decisions could be controlled by fast and simple relocation of the automatically detected patterns. The possibility to analyse 24 slides within 65h by automatic analysis over the weekend and the high reproducibility of the results make automatic image processing a powerful tool for the micronucleus analysis in primary human lymphocytes. The automated slide analysis for the MNT in human lymphocytes complements the portfolio of image analysis applications on ROBIAS which is supporting various assays at Novartis.

Use of the alkaline comet assay for industrial genotoxicity screening: comparative investigation with the micronucleus test.

We evaluated the suitability of the alkaline comet assay as a screening test in industrial routine testing of new chemicals. Thirty-six pharmaceutical compounds with unknown genotoxic potential were tested comparatively in the comet assay and micronucleus test (MNT) using V79 Chinese hamster cells. The comparison of results is generally based on at least two independent experiments, each with two replicate cultures at a minimum of three concentrations. We found a high degree of concordance between results of the comet assay and MNT. All compounds with negative MNT results were also negative in the comet assay. All positive compounds in the comet assay were also positive in the MNT. However, 16 of 38 positive MNT results were negative in the comet assay. Some of the contrary findings may be due to aneugenic effects, which are detected in the MNT but not in the comet assay. However, the majority of the contrary results may be a consequence of cytotoxicity, which can induce elevated micronucleus frequencies but may not lead to positive effects in the comet assay. Additional data of 39 compounds tested in the Ames test and the comet assay were compared. Four of these compounds that were Ames positive were also positive in the comet assay. However, the comet assay also detected 16 compounds that were negative in the Ames test. We believe that the comet assay in vitro is a useful, fast screening system in mammalian cells that can be used in a test battery during drug development.

Concepts of threshold in mutagenesis and carcinogenesis.

Although the existence of a threshold in the dose effect relationship is well documented for many, if not most, types of toxicological effects the existence of a threshold for the mutagenic effects of ionising radiation and of certain chemicals has been questioned since the middle of the century and only recently the question of thresholds for radiation and chemical carcinogenesis has been addressed. The essential facts for the interpretation of threshold dose-response curves are common to all type of effects and are: (i) the number and the identity of the target; (ii) the type and sensitivity of the endpoint used to quantify the effect. We therefore will first try to model the type of interactions which may be expected between a mutagen and its target and define from this whether a threshold dose-effect can be expected; in a second step the concept will be extended to heritable mutations and carcinogenesis.

Analysis of chromosome loss and chromosome segregation in cytokinesis-blocked human lymphocytes: non-disjunction is the prevalent mistake in chromosome segregation produced by low dose exposure to ionizing radiation.

The aim of the present work was to examine in human lymphocytes, firstly, whether in vitro gamma-rays as compared with X-rays also induce chromatid malsegregation and at higher frequencies than chromosome loss and, secondly, whether the cytokinesis-blocked micronucleus assay combined with fluorescence in situ hybridization might be useful for the biomonitoring of individuals exposed to ionizing radiation. After irradiation, the relative frequencies of centromere-positive micronuclei decreased from 39.2% at 0.1 Gy to 21. 63% at higher doses. There was no statistically significant increase in MNCen+ frequencies at doses below 1 Gy (0.1, 0.25 and 0.5 Gy), but a statistically significant increase at 1 (P < 0.05) and 2 Gy (P < 0.001) was observed for all the donors. No significant differences in baseline and gamma-ray-induced non-disjunction frequencies for chromosomes 1 (P = 0.9) and 17 (P = 0.8) between individuals were detected. For radiation-induced non-disjunction, lower doses (0.1, 0. 25 and 0.5 Gy) of gamma-rays did not induce a statistically significant increase in non-disjunction frequencies whereas 1 Gy and above clearly induced a statistically significant increase in the total non-disjunction frequencies for all the donors (P < 0.05 at 1 Gy and P < 0.0001 at 2 Gy). The aneugenic effect of radiation is less clearly dose dependent at the lower doses, suggesting an apparent threshold below which no change could be demonstrated. At high radiation doses the major mechanism for gamma-ray-induced aneuploidy is related to chromosome loss through non-disjunction, as has been demonstrated using X-rays, and not through the formation of micronuclei.

p53-independent apoptosis and p53-dependent block of DNA rereplication following mitotic spindle inhibition in human cells.

We have studied the response of human transformed cells to mitotic spindle inhibition. Two paired cell lines, K562 and its parvovirus-resistant KS derivative clone, respectively nonexpressing and expressing p53, were continuously exposed to nocodazole. Apoptotic cells were observed in both lines, indicating that mitotic spindle impairment induced p53-independent apoptosis. After a transient mitotic delay, both cell lines exited mitosis, as revealed by flow-cytometric determination of MPM2 antigen and cyclin B1 expression, coupled to cytogenetic analysis of sister centromere separation. Both cell lines exited mitosis without chromatid segregation. K562 p53-deficient cells further resumed DNA synthesis, giving rise to cells with a DNA content above 4C, and reentered a polyploid cycle. In contrast, KS cells underwent a subsequent G1 arrest in the tetraploid state. Thus, G1 arrest in tetraploid cells requires p53 function in the rereplication checkpoint which prevents the G1/S transition following aberrant mitosis; in contrast, p53 expression is dispensable for triggering the apoptotic response in the absence of mitotic spindle.

Spindle poisons can induce polyploidy by mitotic slippage and micronucleate mononucleates in the cytokinesis-block assay.

The human in vitro cytokinesis-blocked micronucleus (MN) assay has been extensively used for detection of clastogenic and aneugenic agents. In this test binucleate cells are generally considered to be the main target cell population for assessing genotoxic effect and almost no attention is paid to the biological information contained in mono-nucleate cells. In this study we analysed the frequencies of micronucleate mononucleates in a control population and after in vitro exposure to clastogens or aneugens. A clear increase in MN in mononucleates was found only after exposure to aneugenic compounds. By means of fluorescence in situ hybridization using a chromosome 1-specific probe we further analysed the proportion of mononucleate cells with and without MN which were tetrasomic (tetraploid) and would have been induced during aneugen treatment by mitotic slippage. The data indicate that treatment with nocodazole induces tetrasomy for chromosome 1 (tetraploidy) and an increase in MN frequency in mononucleate diploid and tetraploid lymphocytes. The results thus confirm that some mononucleates pass mitosis without chromatid segregation to daughter nuclei. These data suggest that MN in mononucleates may be useful to distinguish clastogens from aneugens and increase the sensitivity of the test.

The in vitro micronucleus test: a multi-endpoint assay to detect simultaneously mitotic delay, apoptosis, chromosome breakage, chromosome loss and non-disjunction.

Genotoxicity testing aims to detect a large range of genetic damage endpoints and evaluate such results in context of cell survival. The cytokinesis block micronucleus test offers the advantage to provide simultaneously information on both cell cycle progression and chromosome/genome mutations. Indeed, 1. frequencies of cytokinesis-blocked binucleated cells (and polynucleated) are good estimators of the mitotic rate; 2. frequencies of apoptotic figures in mononucleated and binucleated cells provide a measure for cell death before or after cell division; 3. combination of fluorescence in situ hybridization (FISH) for centromere/telomeres and micronucleus scoring allows discrimination between clastogenic and aneugenic events; 4. detection of FISH signals for chromosome specific sequences in both macronuclei and micronuclei, discriminates between aneuploidy due to chromosome non-disjunction or to chromosome loss. The cytokinesis block in vitro micronucleus test is thus a cytogenetic multi-test providing mechanistic information with a simple, rapid, objective, microscopical analysis.

Indication for thresholds of chromosome non-disjunction versus chromosome lagging induced by spindle inhibitors in vitro in human lymphocytes.

Risk assessment from exposure to spindle inhibitors should take into account the possibility of threshold concentration-response curves for aneuploidy induction. We analysed concentration-dependent induction of chromosome non-disjunction by well known spindle poisons (colchicine, carbendazim, mebendazole and nocodazole) and a reference clastogen, methyl methanesulphonate (MMS) in vitro in human lymphocytes; and integrated these findings with earlier results of chromosome loss in micronuclei. Chromosome non-disjunction was estimated on cytokinesis-blocked lymphocytes after simultaneous fluorescent in situ hybridization labelling with two chromosome-specific centromeric probes (chromosomes 1 and 17). The frequencies of spontaneous non-disjunction showed important inter-individual variations and were surprisingly high (7.04-15.39%). Lower concentrations of aneugens did not induce a statistically significant increase of non-disjunction frequencies over the respective control levels, whereas higher concentrations clearly induced a concentration-dependent increase in the non-disjunction frequencies with the four aneugens tested. On the contrary, even at high concentrations, MMS induced a slight increase in the frequency of non-disjunction but without being statistically significant when compared with the control frequencies. We estimated the inflection points, the first statistically significant concentrations, the last non-statistically significant concentrations and the number of events from concentration-response curves of chromosome non-disjunction and chromosome loss. A threshold-type of concentration-response for non-disjunction is highly probable for colchicine and nocodazole. For carbendazim and mebendazole the inflection point fell above the first statistically significant concentrations. But since it is obvious from dose-response curves where the inflection point/threshold lies, it appears that the model might be picking up some irregularities (possibly due to experimental variability in the dose-response curve at concentrations greater than the threshold). For accurate estimation of the threshold, analysis of more concentrations or more cells might be needed. Our data strongly indicate that in cultured human lymphocytes chromosome non-disjunction is a major mechanism of aneuploidy induction by spindle inhibitors and since non-disjunction occurs at lower concentration than chromosome loss, the aneuploidy threshold should be estimated on the basis of non-disjunction rather than on micronuclei frequencies (chromosome loss).

In vitro genotoxic effects of hard metal particles assessed by alkaline single cell gel and elution assays.

Hard metals (WC-Co) are made of a mixture of cobalt metal (Co, 5-10%) and tungsten carbide particles (WC, >80%). Excessive inhalation of WC-Co is associated with the occurrence of different lung diseases including an excess of lung cancers. The elective toxicity of hard metal is based on a physico-chemical interaction between cobalt metal and tungsten carbide particles to produce activated oxygen species. The aim of the present study was to assess the genotoxic activity of hard metal particles as compared with Co and WC alone. In human peripheral lymphocytes incubated with Co or WC-Co, a dose- and time-dependent increased production of DNA single strand breaks (ssb) was evidenced by alkaline single cell gel electrophoresis (SCGE) and modified alkaline elution (AE) assays. Addition of 1 M formate, a hydroxyl radical scavenger, had a protective effect against the production of ssb by both WC-Co or Co alone. On the basis of an equivalent cobalt-content, WC-Co produced significantly more ssb than Co. WC alone did not produce DNA ssb detectable by the AE assay, but results obtained with the SCGE assay may suggest that it either allows some uncoiling of the chromatin loops or induces the formation of slowly migrating fragments. Overall, this in vitro study is the first demonstration of the clastogenic property of cobalt metal-containing dusts. The results are consistent with the implication of an increased production of hydroxyl radicals when Co is mixed with WC particles. The SCGE results also suggest that WC may modify the structure of the chromatin, leading to an increased DNA sensitivity to clastogenic effects. Both mechanisms are not mutually exclusive and may concurrently contribute to the greater clastogenic activity of WC-Co dust. This property of WC-Co particles may account for the excess of lung cancers observed in hard metal workers.

The detection and evaluation of aneugenic chemicals.

Although aneuploidy makes a significant contribution to both somatic and inherited disease the mechanisms by which environmental chemicals may induce numerical chromosome aberrations are only poorly defined. The European Union Project was aimed to further our understanding of those chemical interactions with the components of the mitotic and meiotic cell division cycle which may lead to aneuploidy and to characterise the parameters such as cellular metabolism which may influence the activity of aneugenic chemicals. C-mitosis can be induced by the highly lipophilic polychlorinated biphenyl and the completion of mitosis and cleavage can be modified by agents which deplete cellular levels of reduced glutathione. Modifications of the fidelity of chromosome segregation were produced by inhibiting the functioning of topoisomerase II during chromatid separation. In contrast, the modification of centromere integrity resulted in chromosome breakage as opposed to disturbance of segregation. Modifiers of tubulin assembly and centriolar functioning in somatic cells such as acrylamide, vinblastine and diazepam reproduced their activity in rodent bone marrow and male germ cells. The analysis of chromosome malsegregation in Aspergillus nidulans by a structurally related series of halogenated hydrocarbons was used to develop a QSAR model which had high predictive value for the results of fungal tests for previously untested related chemicals. Metabolic studies of potential aneugens in genetically engineered human lymphoblastoid cells demonstrated the detoxification of the aneugenic activity of chloral hydrate and the activation of 2,3-dichlorobutane, 1,1,2-trichloroethane and trichloroethylene by Phase I biotransforming enzymes. Cell transformation studies in Syrian hamster dermal cultures using a panel of 22 reference and or potential aneugens indicated that 15 of the 22 produced positive results following single exposures. Five of the aneugens which were negative following single exposures produced positive results where cultures were continuously exposed for up to 6 weeks to low concentrations following a single non-transforming exposure to the mutagen dimethyl sulphate. The transformation studies indicate that a significant proportion of chemical aneugens are potential complete carcinogens and/or co-carcinogens. To optimise the enumeration of chromosomes following exposure to potential chemical aneugens whole chromosome paints and centromere specific probes suitable for use in fluorescence in situ hybridisation (FISH) were developed for the rat, mouse and Chinese hamster and selected human probes evaluated for their suitability for routine use. Molecular chromosome probes were used to develop protocols for enumerating chromosomes in metaphase cells and centromeres and micronuclei in interphase cells. The analysis of segregation of specific centromeres in binucleate cells following cytochalasin B treatment was shown to be a potentially valuable system for characterising non-disjunction following chemical exposure. Whole chromosome paints and centromere specific probes were used to demonstrate the presence of dose-response thresholds following treatment with a reference panel of spindle inhibiting chemicals. These data indicate that the FISH technology is suitable for evaluating the relative hazards of low-dose exposures to aneugenic chemicals.

Clastogenic and aneugenic effects of three benzimidazole derivatives in the in vitro micronucleus test using human lymphocytes.

Three benzimidazole compounds thiabendazole (TBZ), carbendazim (MBC) and mebendazole (MEB) were analysed with the in vitro cytochalasin-B micronucleus test on human lymphocytes. TBZ was tested in isolated lymphocyte cultures and MBC and MEB were tested in both isolated lymphocyte and whole blood cultures. TBZ was tested up to 300 microM with and without S9-mix. Although signs of toxicity, without S9, were observed by a decrease in the division index at 300 microM, an increase in the frequency of micronucleated binucleates was not found with or without S9. MBC and MEB induced a statistically significant concentration-dependent increase in the micronucleus frequency. The effective concentration range for MEB (0.3-1.5 microM) was ten times lower than for MBC (5-25 microM). By means of fluorescence in situ hybridization with a 30 nucleotide oligomer of the alpha centromeric regions, common for all chromosomes, on the induced micronuclei MBC and MEB were found to induce a significant increase of centromere positive micronuclei in a dose-dependent manner. MBC and MEB are poorly soluble in water and therefore have a low bioavailability in vivo. However, increased micronucleus frequencies were found in this in vitro micronucleus study at doses comparable to in vivo plasma levels in mice and should, therefore, not be neglected in the risk evaluation of those compounds.

Indications for a threshold of chemically-induced aneuploidy in vitro in human lymphocytes.

The possible existence of a threshold for compounds inducing chromosomal loss was investigated for four known aneugens (colchicine, COL; carbendazim, MBC; mebendazole, MEB; nocodazole, NOC) and two clastogens (methyl methanesulfonate, MMS; mitomycin C, MMC) using the micronucleus (MN) test in human lymphocytes. The presence of a whole chromosome in the MN was studied by fluorescent in situ hybridization (FISH) using a synthetic pancentromeric oligonucleotide probe. FISH was applied on two different MN preparations: cytokinesis-blocked MN (MNCB) assay, and MN sorted by flow cytometry. At subtoxic concentrations analyzed by MNCB and FISH, COL, MEB, MBC, and NOC induced a concentration-dependent increase in centromere-positive MN (MNCen+). MMC seemed to induce an increase in both types of MN (MNCen- and MNCen+), while MMS induced only MNCen-. On the sorted micronuclei (in a wide range of low to subtoxic concentrations), the concentration-effect profile for MNCen+, with the four aneugens tested, showed a statistically nonsignificant increase over a range of concentrations, followed by a second range of high concentrations with a statistically significant increase. To analyze the existence of a threshold, a piecewise linear regression was applied to the data. The first concentration that showed a statistically significant increase in MNCen+ was chosen as a breakpoint (0.037 microM for COL, 2.62 microM for MBC, 0.27 microM for MEB, and 0.066 microM for NOC). The statistical correlation between observed and predicted values showed a high correlation (r = 0.99), indicating a clear threshold for aneuploidy induction. However, for MMS the concentration-effect profile for MNCen+ showed a continuous concentration-dependent decrease with no threshold. With the two cytotoxicity assays used (Bio-Rad and MTT), no significant reduction was detected either in the protein content or in mitochondrial succinate dehydrogenase activity with all chemicals tested for MN induction. Therefore, our data suggest that the observed thresholds were not due to indirect toxic effects but to real aneugenic effects.

Metabolic differences between whole blood and isolated lymphocyte cultures for micronucleus (MN) induction by cyclophosphamide and benzo[a]pyrene.

In order to study the metabolic differences between whole blood and isolated lymphocyte cultures, two indirectly acting mutagens cyclophosphamide (CP) and benzo[a]pyrene (B[a]P) were assessed for their potential to induce micronuclei (MN) in the presence and absence of S9 microsomal fractions. In isolated lymphocyte cultures supplemented with S9, CP and B[a]P induced a statistically significant increase in MN which was not observed in whole blood cultures. However, the direct-acting agent methyl methanesulphonate (which was used as a positive control) showed an increase in MN frequency in a dose-dependent manner in both culture methods. The effect of erythrocytes was then investigated by treating isolated lymphocyte cultures simultaneously with CP and S9 mix in the presence of purified erythrocyte concentrate (PEC). A clear reduction in the MN frequency was observed compared to the frequencies of MN induced in isolated lymphocyte cultures treated with CP and S9 mix in the absence of PEC. Thus, isolated lymphocyte cultures may represent a more sensitive test system for the evaluation of potential indirect-acting mutagens. However, whole blood cultures may reflect the 'real life' situation more accurately as a consequence of the presence of erythrocytes.