Based on new scientific developments and experience of the regulation of chemical compounds, a working group of the Gesellschaft fuer Umweltmutationsforschung (GUM), a German-speaking section of the European Environmental Mutagen Society, proposes a simple and straightforward approach to genotoxicity testing. This strategy is divided into basic testing (stage I) and follow-up testing (stage II). Stage I consists of a bacterial gene mutation test plus an in vitro micronucleus test, therewith covering all mutagenicity endpoints. Stage II testing is in general required only if relevant positive results occur in stage I testing and will usually be in vivo. However, an isolated positive bacterial gene mutation test in stage I can be followed up with a gene mutation assay in mammalian cells. If this assay turns out negative and there are no compound-specific reasons for concern, in vivo follow-up testing may not be required. In those cases where in vivo testing is indicated, a single study combining the analysis of micronuclei in bone marrow with the comet assay in appropriately selected tissues is suggested. Negative results for both end points in relevant tissues will generally provide sufficient evidence to conclude that the test compound is nongenotoxic in vivo. Compounds which were recognized as in vivo somatic cell mutagens/genotoxicants in this hazard identification step will need further testing. In the absence of additional data, such compounds will have to be assumed to be potential genotoxic carcinogens and potential germ cell mutagens.

Describing the toxicological profile of a substance is the first step required for risk assessment. Although a wide range of in vitro methods are widely used to characterise toxicological properties including toxicokinetics, regulatory acceptance is mainly confined to in vitro tests which investigate genotoxic end-points. In vitro tests have been proposed for the endpoints acute toxicity, repeated dose toxicity and toxicity to reproduction which encompass the minimum requirements in the OECD SIDS programme. However, until now, limitations of the proposed tests preclude their application in a regulatory framework. Presently, in vitro tests play a major role in obtaining information on mechanism of toxicity with the perspective to be able to identify pathways of toxic responses by applying toxicogenomics techniques. Physiologically based toxicokinetic modelling is using data from in vitro studies to build up the model for a specific compound. Information from both areas is incorporated into the risk assessment to derive compound-specific safety factors, which account for species differences and for the variability among the human population, including possible sensitive subpopulations. Future developments to further enhance the use of in vitro methods in regulatory toxicology include the development of (Q)SAR approaches supplemented by mechanisms of toxicity, which can be addressed by developing methods of molecular toxicology.

Zusammenfassung Der Nachweis von Acrylamid in Lebensmitteln gibt zu erheblicher Besorgnis Anlass. Acrylamid ist ein toxischer Stoff, der im Tierversuch u. a.neurotoxisch und Krebs erzeugend wirkt; es gibt keinen Anlass anzunehmen, dass diese Wirkungen nicht auch im Menschen auftreten können.Die tägliche Belastung durch Acrylamid in Lebensmitteln wird für den durchschnittlichen Verzehrer in Deutschland auf ungefähr 0,6 μg/kg Körpergewicht geschätzt. Für Personen mit starkem Verzehr hoch belasteter Produkte kann die Acrylamidaufnahme mehrfach höher liegen. Es ist nicht anzunehmen, dass diese Belastungen beim Menschen zu neurotoxischen Wirkungen führen. Für die Krebs erzeugende Wirkung von Acrylamid muss aufgrund der ebenfalls vorhandenen mutagenen Wirkungsweise davon ausgegangen werden, dass kein Schwellenwert vorliegt.Daher kann ein Krebsrisiko durch den Verzehr acrylamidbelasteter Nahrung nicht ausgeschlossen werden.Die Acrylamidbelastung durch Lebensmittel sollte so weit gesenkt werden wie vernünftigerweise möglich.

Risk assessment of xenobiotics is a qualitative and quantitative assessment of toxic properties conventionally based on data resulting from tests in animals exposed to the substance. The assessment of dose-effect relationship includes evaluation of exposure at the site of action. More recently, emphasis is put on understanding the relationship between exposure at the site of action and the resulting effect, i.e. toxicodynamic. In this respect, results from genotoxicity studies may be a measure for exposure and at the same time of an effect. Results of toxicodynamic endpoints such as binding to receptors or release of hormones have been used when replacing default values for interspecies extrapolation. It may also be envisaged to use toxicodynamic endpoints in order to get an estimate of intraspecies variability. It was demonstrated that this approach may be helpful only if the relationship between the toxicodynamic endpoint and the definite endpoint is known by using the example of bisphenol A. Whereas there are clear effects of bisphenol A in in vitro and ex vivo studies, the classical two generation study has not been able to detect an effect on reproduction and/or fertility. Looking in the future development of toxicodynamic endpoints, gene profiling and the analysis of proteins ('proteomics') may be helpful tools employed in screening and being related to the mode of action are explored for their suitability in terms of toxicodynamic endpoints.

In recent years, there has been widespread interest in the relationship between carcinogenic exposure and mutation spectra in cancer-related genes. To evaluate potential benefits and/or limitations in the use of mutation spectra in genetic toxicology, a GUM working group has been established to discuss this subject. Based on methodological possibilities and limitations, the impact of mutation spectra in the interpretation of animal experiments and in the identification of etiological agents in human cancer has been considered. With respect to experimental animals, the analyses of mutation spectra within long-term rodent carcinogenicity studies may provide some additional information on the mode of action of the respective carcinogen, however, the interpretation of results should be done carefully and only in context with other toxicological data available. Regarding human exposure, the analysis of mutation spectra in p53 or ras genes supplies information on the genotoxic properties of the respective agent. Nevertheless, on the individual level, the presence or absence of defined mutations in cancer-related genes in human tumors does not permit a definite conclusion about the causative agent.

A collaborative study with 10 participating laboratories was conducted to evaluate a test protocol for the performance of the in vitro micronucleus (MN) test using the V79 cell line with one treatment and one sampling time only. A total of 26 coded substances were tested in this study for MN-inducing properties. Three substances were tested by all 10 laboratories and 23 substances were tested by three or four laboratories in parallel. Six aneugenic, 7 clastogenic and 6 non-genotoxic chemicals were uniformly recognised as such by all laboratories. Three chemicals were tested uniformly negative by three laboratories although also clastogenic properties have been reported for these substances. Another set of three clastogenic substances showed inconsistent results and one non-clastogenic substance was found to be positive by one out of three laboratories. Within the study, the applicability of the determination of a proliferation index (PI) as an internal cytotoxicity parameter in comparison with the determination of the mitotic index (MI) was also evaluated. Both parameters were found to be useful for the interpretation of the MN test result with regard to the control of cell cycle kinetics and the mode of action for MN induction. The MN test in vitro was found to be easy to perform and its results were mainly in accordance with results from chromosomal aberration tests in vitro.

In the regulation of chemical substances, it is generally agreed that there are no thresholds for genotoxic effects of chemicals, i.e. , that there are no doses without genotoxic effects. When classifying and labelling chemicals, dangerous properties of chemicals are to be identified. In this context, in general, the mode of action (threshold or not) is not considered for genotoxic substances. In the process of quantitative risk assessment, however, determination of the type of dose-effect relationships is decisive for the outcome and the type of risk management. The presence of a threshold must be justified specifically in each individual case. Inter alia, the following aspects may be discussed in this respect: aneugenic activity, indirect modes of action, extremely steep dose-effect relationships in combination with strong toxicity, specific toxicokinetic conditions which may lead to 'metabolic protection' prior to an attack of DNA. In the practice of the regulation of chemical substances with respect to their genotoxic effects, the discussion of thresholds has played a minor role. For notified new substances, there are, in general, no data available that would allow a reasonable discussion. Concerning substances out of the European programme on existing substances, so far 29 have been assessed in our institute with respect to genetic toxicity. Eight out of these have shown considerable evidence for genotoxicity. For two of them, a possible threshold is discussed: one substance is an aneugen, the other one is metabolised to an endogenic compound with genotoxic potential. In the practice of risk assessment of genotoxic substances, the discussion of the mode of action for genotoxicity is frequently associated with the evaluation of potential carcinogenic effects. Here, tissue-specific genotoxic effects in target organs for carcinogenicity are to be discussed. Moreover, the contribution of genotoxicity to the multifactorial process of tumour development should be assessed.

The in vitro micronucleus test (MNT) is a useful assay for the detection of mutagenic events on both the chromosomal and the genomic level. The main disadvantage for introducing the in vitro MNT into official test guidelines seems to be the disparity of existing protocols. To contribute to the aim of standardisation, three different methodological approaches of the in vitro MNT with V79 cells were compared: the standard assay using an asynchronically growing mixed cell population, the cytokinesis block (CB) assay and a modified MNT, the so-called mitotic shake-off (MSO) method. V79 cells were thus treated with two known aneugens (colcemide and griseofulvin) and two clastogens (mitomycin C and cyclophosphamide) over various time periods. The cultures of the CB assay were additionally exposed to cytochalasin B (Cyt-B), an inhibitor of cell, but not of nucleus division. After treatment, the cells were harvested and analysed for the appearance of micronuclei (MN). All three assays yielded positive results for all test substances. These results support the suitability of the MNT with V79 cells with regard to the ability to detect the genotoxic potential of both clastogens and aneugens independent of the test protocol applied. Thus, all three methods are appropriate for MN detection, but due to the fact that the application of Cyt-B has no advantages for a cell line like V79 in which nearly all cells undergo a normal cell cycle, its use is not recommended.

According to regulations in the European Union, new chemical substances must be notified before they can be introduced onto the market. One of the prerequisites for notification is that toxicological properties, including mutagenicity, are examined. In this paper, a report on routine in vitro mutagenicity testing is given for 776 new substances notified in Germany between 1982 and 1997. In general, the methodological quality of testing was in line with internationally accepted guidelines. Bacterial gene mutation tests (Bact) were conducted for nearly all of the substances, 13.4% were positive. Of the Bact-positive substances, 36 were also tested in the in vitro chromosomal aberration test (CAbvit) and the mammalian cell gene mutation test (MCGM). Twenty-six of these (72.2%) were negative in both mammalian cell tests indicating that the genotoxic potentials of the substances are not relevant for man. Of all new substances, 333 were tested in CAbvit, here the percentage of positive findings was 25.2%. More than 80% of the in vitro clastogens were negative in the Bact. With respect to a sensitive detection of genotoxic potentials of substances, the combination `Bact+CAbvit' is appropriate for basic testing. In our database CHL cells were more sensitive to clastogenic effects than other cell types. Only very few clastogens were identified as `high toxicity clastogens'. MCGM tests were performed for 118 substances, quite often as follow-up in case of positive Bact tests. In total, 12.7% of the substances were positive in the MCGM. However, there was a clear difference in the frequencies of positive findings in HPRT tests (5.5%) and mouse lymphoma assays (MLA; 37.0%). None of the MCGM-positive substances was a `unique positive', i.e., negative in Bact and CAbvit.

In order to license a pharmaceutical or chemical, a compound has to be tested for several genotoxicity endpoints, including the induction of chromosomal aberrations in vitro. A working group within the GUM has evaluated published data on the in vitro micronucleus test with the aim of judging its suitability as a replacement for the in vitro chromosomal aberration test. After strict rejection criteria were applied, a database including 96 publications and 34 compounds was obtained. For 30 of these compounds, data on both tests were available. For 24 of the 30, concordant results in both test systems were obtained (80% correlation). The discordant results in 6 compounds can be explained by a known or suspected aneugenic potential of these compounds. Considering that cell types and test protocols were extremely heterogeneous, this correlation is rather encouraging. Comparison of the different protocols, and experience established within the working group yielded several recommendations for the routine use of the in vitro micronucleus test. Although many cell lines are suitable, those most often used in genotoxicity testing (e.g. CHL, CHO, V79, human lymphocytes, L5178Y mouse lymphoma cells) are recommended. Cytochalasin B may be used in the case of human lymphocytes; however, the possibility of its interaction with aneugenic test compounds should be considered. For continuously dividing cell lines, cytochalasin B is not recommended by the working group. Although, there seems to be flexibility in the choice of treatment and sampling times, the average generation time of the chosen cell line of choice should be taken into account when determining sampling time, and treatment of cells for at least one cell cycle duration is recommended. The use of appropriate cytotoxicity tests is strongly recommended. Although studies on some parameters of the test protocol may be useful, the introduction of the in vitro micronucleus test into genotoxicity testing and guidelines should not be delayed. Even in its present state, the in vitro micronucleus is a reliable genotoxicity test. Compared with the chromosomal aberration test, it detects aneugens more reliably, it is faster and easier to perform, and it has more statistical power and the possibility of automation.

The Working Group (WG) dealt with the harmonisation of routine methodologies of tests for unscheduled DNA synthesis (UDS) both in vitro and in vivo. In contrast to existing guidelines from OECD, EPA and EC on in vitro UDS tests (there is no Japanese UDS guidline), the Working Group recommends that in general in vitro UDS tests should be performed with primary hepatocytes. For routine applications any other cell types would need special justification. Hepatocytes from male rats are prefeable, unless there are contra-indications on the basis of e.g. toxicokinetic data.According to the OECD, EPA and EC guidelines, UDS may be analysed by means of autoradiography (AR) or liquid scintillation counting (LSC). The WG recommends use of AR. LSC is less suitable due to the problem of differentiation between UDS activity and replicative DNA synthesis, and the disadvantage that cells cannot be analysed individually. Since a specific cell type was recommended by the WG, methodological aspects could be described in more detail than in the present guidelines. For in vitro tests, it was agreed that the initial viability of freshly isolated hepatocytes should be at least 70%.With regard to the need for confirmatory experiments in the event of a clear-cut negative result, the majority view was that confirmation by a second (normally not identical) experiment is still needed; this is in line with the present OECD and EC guidelines.Evaluation of results from UDS tests should be based primarily on net nuclear grain (NNG) values, although it is recognised that nuclear and cytoplasmic grains result from different biological processes. Since grain counts are influenced by a number of methodological parameters, no global threshold NNG value can be recommended for discrimination of positive and negative UDS results. For in vitro assays, the criteria for positive findings go beyond those of the present guidelines and two alternative approaches are given which are based on (1) dose-dependent increases in NNG values and (2) reproducibility, dose-effect relationship and cytotoxicity.At present there is no official guideline on the performance of in vivo UDS tests. Some fundamental recommendations given for in vitro methodology also apply to the in vivo assay. For routine testing with the in vivo UDS test, again the general use of hepatocytes from male rats is recommended. However, concerning the requirement to use one or two sexes, consistency with other in vive genotoxicity assays (e.g. the micronucleus assay) would be preferable. As for the in vitro methodology, AR is preferred rather than LSC. For in vivo UDS tests, a minimum viability of 50% is considered to be sufficient. Sampling of cells 12-16 h after treatment and, if this is negative, 2-4 h is recommended. At least three animals per treatment group should be used.Evaluation of results should again be done on the basis of NNG values. The fundamental criterion for a positive result is given by an increase of NNG values for at least one experimental group. This NNG increase should be evaluated by consideration of (1) lab-specific historical controls or adequate statistics, or (2) interanimal variation, dose-effect relationship and cytotoxicity.

The in vivo rodent bone marrow micronucleus assay (Mnt) has assumed a pivotal role in screening strategies for the identification of substances potentially carcinogenic to humans. The analysis of the results of the current international 5-year effort to provide toxicological data for high production volume chemicals will play a crucial role in developing future strategies for identifying health hazards. As part of that program, consideration is being given to accepting either in vitro genotoxicity data or results of the Mnt. The present analyses indicate that for hazard identification purposes that, in fact, in vitro genotoxicity test results, such as those derived from the Salmonella mutagenicity assay, may be an acceptable alternative.