Bioassay battery interlaboratory investigation of emerging contaminants in spiked water extracts - Towards the implementation of bioanalytical monitoring tools in water quality assessment and monitoring.

Bioassays are particularly useful tools to link the chemical and ecological assessments in water quality monitoring. Different methods cover a broad range of toxicity mechanisms in diverse organisms, and account for risks posed by non-target compounds and mixtures. Many tests are already applied in chemical and waste assessments, and stakeholders from the science-police interface have recommended their integration in regulatory water quality monitoring. Still, there is a need to address bioassay suitability to evaluate water samples containing emerging pollutants, which are a current priority in water quality monitoring. The presented interlaboratory study (ILS) verified whether a battery of miniaturized bioassays, conducted in 11 different laboratories following their own protocols, would produce comparable results when applied to evaluate blinded samples consisting of a pristine water extract spiked with four emerging pollutants as single chemicals or mixtures, i.e. triclosan, acridine, 17α-ethinylestradiol (EE2) and 3-nitrobenzanthrone (3-NBA). Assays evaluated effects on aquatic organisms from three different trophic levels (algae, daphnids, zebrafish embryos) and mechanism-specific effects using in vitro estrogenicity (ER-Luc, YES) and mutagenicity (Ames fluctuation) assays. The test battery presented complementary sensitivity and specificity to evaluate the different blinded water extract spikes. Aquatic organisms differed in terms of sensitivity to triclosan (algae > daphnids > fish) and acridine (fish > daphnids > algae) spikes, confirming the complementary role of the three taxa for water quality assessment. Estrogenicity and mutagenicity assays identified with high precision the respective mechanism-specific effects of spikes even when non-specific toxicity occurred in mixture. For estrogenicity, although differences were observed between assays and models, EE2 spike relative induction EC50 values were comparable to the literature, and E2/EE2 equivalency factors reliably reflected the sample content. In the Ames, strong revertant induction occurred following 3-NBA spike incubation with the TA98 strain, which was of lower magnitude after metabolic transformation and when compared to TA100. Differences in experimental protocols, model organisms, and data analysis can be sources of variation, indicating that respective harmonized standard procedures should be followed when implementing bioassays in water monitoring. Together with other ongoing activities for the validation of a basic bioassay battery, the present study is an important step towards the implementation of bioanalytical monitoring tools in water quality assessment and monitoring.