Toward identification of fish specific estrogens through effect-directed analysis based on zebrafish in vitro and in vivo reporter gene assays

Estrogens are present in all vertebrates, from telesots to mammals, which can bind to and activate estrogen receptors (ERs) through a conserved mechanism. However, existence of cross-species differences as regards ER subtypes and the binding affinity of (xeno)estrogens on ERs, as well as the cellular context may lead to differential ER activation by environmental contaminants when assessing hazard on aquatic organisms. For this purpose, we previously developed, characterized and demonstrated the functionality of stable zebrafish (zf)-ER subtypes (zfERa, zfERb1, zfERb2) in the zebrafish liver (ZFL) cell line and transgenic cyp19a1b-GFP zf embryosto quantify estrogenic activity in complex mixtures. In the present study, we used zebrafish (zf)-based in vitro and in vivo bioassays to assess estrogenic activity in surface waters by investigating 20 French river sites using polar organic chemical integrative sampling (POCIS). Simultaneously, the human MELN bioassay (MCF-7 cells-based, expressing hERa) was used to study cross-species differences. POCIS-based bio-monitoring provide significant information on zf-based in vitro bioassays as zfERs and hERa assays significantly differed at some sites, identifying zf-specific estrogenic activities in surface waters. Higher numbers of POCIS extracts were found to be more active on zfERbs than on zfERa, which is in line with the better sensitivity of zfERbs to steroid estrogens. The responsiveness of zfERb2 to environmental samples is important as this ER subtype is present in fish species but not in humans. In complementary to in vitro assays, the estrogenic activities at the organism level were observed using zf embryo assay at sites that were highly active on in vitro bioassays. These in vivo results are relevant for hazard assessment as they reveal that detected active compounds target radial glial cells expressing brain aromatase which are known to be involved in neurogenesis. In addition, combining these zf-based tools within an EDA approach, we were able to isolate fractions that were active only on zfERb2 subtype, hence reinforcing the hypothesis of species-specificity for, yet unknown, ER ligands. Further investigation using mass spectrometry techniques is foreseen to identify fish-specific active compounds which will allow improving bioassay-based environmental risk assessment towards aquatic species