Biomarkers of genotoxicity and other end-points in an integrated approach to environmental risk assessment.

Risk is defined as the probability of a given toxicological hazard resulting in actual biological harm. This involves some form of mathematical relationship between exposure and toxic effects. Simplified models based on laboratory testing in surrogate species neglect potentially important factors in real life situations. Our own approach to the study of atmospheric and edaphic pollution, focused on realism, includes the use of sentinel species (animals as prospectors and integrators of information, along both the spatial and the temporal axes) and selected biomarkers. We aim to: (i) consider pollution as a complex mixture; (ii) take into account homeostasis of the environment and of living organisms; (iii) be realistic (all data obtained in the field; calculations based on actual effects; exposure measured as internal dose). The proposed test battery divides toxicological information into four blocks: systemic effects (serum biochemistry and histopathology in wild wood mice), reproduction (epididymis cell count in mice, malformations in amphibian larvae), genotoxicity (Comet test in mice and earthworms) and population effects (abundance and diversity in arthropods). Each block is represented by the sum of the results of the tests performed within the block (presented as a severity score from 0 to 3). A final value is obtained to represent the integrated toxicological harm (ITH) occurring at a given location. To assess exposure, taking into account bioavailability, we propose (i) for soil contamination studies, measuring EROD activity in liver; (ii) for atmospheric pollution, the gaseous fraction is taken from immission gases analysis, while the solid fraction is assessed through levels of metals in sentinel organisms, the values of both fractions then being combined. Finally, a regression line is established for exposure versus ITH in four to five locations with decreasing exposure levels, ranging from the immediate neighbourhood of the pollution focus to controls, following the main dissemination line. In this model we may interpolate new exposure data to find the corresponding predicted ITH. Such a prediction may be directly interpreted as a form of risk assessment or, alternatively, these pairs (toxicological harm/exposure) could then be related to a conventional scale of ecotoxicological risk.