Behavior as an indicator - Development of a monitoring system for pharmaceuticals in aquatic systems.

Pollution is a key environmental problem and a major contributor to global change. Ecotoxicology combines ecology and toxicology to study the effects of pollutants on organisms and ecosystems as a whole, in an attempt to mitigate deleterious effects of environmental pollution. To determine the environmental safety of a chemical compound and to promote the sustainability of ecosystems, ecotoxicological studies use a battery of standardised short-term exposure tests to assess harmful, stressful or lethal effects of pollution on study organisms. Such tests are performed on different types of organisms and rapidly provide valuable data for regulatory decision making but are often offset by a low ecological relevance. Moreover, more subtle behavioural effects are not assessed despite having direct and indirect ecological consequences. Therefore, and although safety margins are adopted, environmental risks of pollutants may not be correctly estimated. This is especially true for emerging contaminants such as pharmaceutical compounds. To advance the assessment of risks associated with such pollutants, standardised tests to quantify sub-lethal behavioural effects over ecologically relevant time periods are needed. However, such tests are not compatible with the relatively long lifespan of traditional vertebrate model organisms.

In this thesis, we explore the potential of the short-lived Turquoise killifish Nothobranchius furzeri as a vertebrate model for behavioural studies, specifically with regard to ecotoxicological testing of whole-life and multigenerational exposure to persistent contaminants.

Firstly, we studied what constitutes normal behaviour in N. furzeri to characterise its behavioural baseline in Chapter 1, 2 and 3. We found intrinsic behavioural variation in multiple traits at the population- and individual level, and N. furzeri males and females were shown to be dissimilar in their behavioural profile. In addition to intrinsic behavioural variation, we also showed that conditions of the rearing environment shape the development of behaviour in N. furzeri. These findings are directly relevant for animal welfare as well as for the development of optimised husbandry- and experimental protocols for behavioural studies in ecotoxicology with N. furzeri.

We then performed single-stressor exposure studies on N. furzeri with fluoxetine as a common neuroactive pharmaceutical pollutant, starting with a short-term exposure assay (Chapter 4) and advancing to a more ecologically relevant exposure regime (Chapter 5). Finally, we performed a multigenerational mixed-stressor exposure study (Chapter 6) on sensitive life-history and behavioural endpoints to incorporate higher levels of ecological realism. Not only did we show that exposure to a neuroactive contaminant induced both behavioural and life-history alterations in N. furzeri, but we also demonstrated that mixed-stressor exposure studies over multiple generations are needed and possible in a time- and cost-efficient way when using N. furzeri as a model.

            Overall, our results contribute to an increased understanding of N. furzeri behaviour and demonstrate the potential of N. furzeri as a model organism to advance ecotoxicological assessment.