Intérêt d’une approche combinant bioessais in vitro et in vivo chez le poisson zèbre pour l’identification de perturbateurs endocriniens dans l'environnement aquatique

The aim of this PhD thesis was to develop and assess the potential of novel zebrafish (zf)- based in vitro and in vivo reporter gene assays as bio-analytical tools to monitor estrogenic activity and their implementation in effect-directed analysis (EDA) approach to identify fishspecific
estrogenic compounds in complex mixtures. For this purpose, we first characterized the response of the assays towards a panel of (xeno)estrogens, revealing differences in the affinity of zebrafish estrogen receptor (zfER) subtypes to (xeno)-estrogens. Comparison with human cell-based (MELN-hERα) assay further highlighted inter-species differences showing different chemical ranking towards different classes of known ER ligands. Then, application of these tools to different environmental matrices demonstrated for the first time their functionality to detect and quantify estrogenic activity in complex mixtures, highlighting the zfERβ2 assay as the most sensitive among the different in vitro zfER assays. The above in vitro estrogenic activity was also confirmed in vivo at the most contaminated sites by using the EASZY assay, further adding eco-toxicological relevance. Interestingly, we also reported zebrafish-specific activities at several sites that were not active by the human MELN assay, suggesting the occurrence of fish-specific ER ligands. To address this hypothesis, we applied our zebrafish tools in specific higher tier-EDA studies that allowed isolating zebrafish-specific active fractions by multi-step sample fractionation procedures. Chemical analyses of these specific fractions so far identified several candidate compounds, of which few showed higher selectivity towards zfERβ2 than hERα, hence confirming inter-species differences. This work supports recommendations for the integration of these effect-based tools in future water monitoring strategies.