In numerous cases, the German health-related indication value (HRIV) concept has proved its practicability for the assessment of drinking water relevant trace substances (Umweltbundesamt 2003). The HRIV is based on the toxicological profile of a substance. An open point of the HRIV concept has been the assignment of standardized test procedures to be used for the assessment. The level of the HRIV is at its lowest as soon as the genotoxicity of the substance is detected. As a single test on its own, it is not sufficient enough to assess the human toxicological relevance of a genotoxic effect or exclude it in the case of a negative result; a reasonable test battery was required, technically oriented towards the already harmonized international, hierarchical evaluation for toxicological assessment of chemicals. Therefore, an important aim of this project was to define a strategy for the genotoxicological assessment of anthropogenic trace substances. The basic test battery for genotoxicity of micropollutants in drinking water needs to fulfill several requirements. Although quick test results are needed for the determination of HRIV, a high degree of transferability to human genotoxicity should be ensured. Therefore, an in vitro genotoxicity test battery consisting of the Ames fluctuation test with two tester strains (ISO 11350), the umu test and the micronucleus test, or from the Ames test with five tester strains (OECD 471) and the micronucleus test is proposed. On the basis of selected test substances, it could be shown that the test battery leads to positive, indifferent, and negative results. Given indifferent results, the health authority and the water supplier must assume that it is a genotoxic substance. Genetically modified tester strains are being sensitive to different chemical classes by expression of selected mammalian key enzymes for example nitroreductase, acetyltransferase, and glutathione-S-transferase. These strains may provide valuable additional information and may give a first indication of the mechanism of action. To check this hypothesis, various additional strains expressing specific human-relevant enzymes were investigated. It could be shown that the additional use of genetically modified tester strains can enhance the detectable substance spectrum with the bacterial genotoxicological standard procedures or increase the sensitivity. The additional use provides orienting information at this level as a lot of data can be obtained quite quickly and with little effort. These indications of the mechanism of action should be however verified with a test system that uses mammalian cells, better human cells, to check their actual relevance. The selection of appropriate additional tester strains has to be defined from case to case depending on the molecular structure and also still requires some major expertise.
