Transgenic (cyp19a1b-GFP) zebrafish embryos as a tool for assessing combined effects of oestrogenic chemicals

Endocrine disrupting chemicals and especially oestrogen receptor (ER) agonists have been extensively studied over the years due to their potential effects on sexual development and reproduction in vertebrates, notably fish. As ER agonists can exist as complex mixtures in the aquatic environment, evaluating the impact of combined exposure on oestrogenic effects has become increasingly important. Use of predictive models such as concentration addition (CA) and independent action (IA) has allowed assessment of combined estrogenic effects of complex multi-compound mixtures of ER agonists in various fish in vitro and in vivo experimental models. The present work makes use of a transgenic zebrafish strain, tg(cyp19a1b-GFP), which expresses the green fluorescent protein (GFP) under the control of the cyp19a1b (brain aromatase or aromatase B) gene to determine the oestrogenic potency of ER agonists alone or in mixtures. In these studies, tg(cyp19a1b-GFP) zebrafish embryos were exposed for four days (from one to five days post fertilization) to five different oestrogenic chemicals; 17α-ethinylestradiol (EE2), 17β-estradiol (E2), estrone (E1), bisphenol A (BPA) and 4-tert-octylphenol (OP), and three mixtures of up to four of these compounds. The mixture of BPA, OP and E2 was also tested with primary cultures of rainbow trout hepatocytes by analysing the ER-mediated induction of the oestrogenic biomarker vitellogenin in order to compare the performance of the two methods for assessing oestrogenic effects of complex mixtures. The three tested mixtures were predominantly acting in an additive manner on the expression of GFP. Additivity was indicated by the overlap of the 95% confidence interval of the concentration response curves for the observed data with the CA and IA prediction models, and model deviation ratios within a factor of two for a majority of the mixture concentrations. However, minor deviations determined as more than additive effects for the mixture of EE2, E1 and E2 and less than additive effects for the mixture of BPA, OP, EE2 and E1 were observed at the higher mixture concentrations tested. The successful prediction of additivity by CA and IA in tg(cyp19a1b-GFP) zebrafish embryos and deviations at high mixture concentrations seemed to correspond well to results obtained in the rainbow trout hepatocyte assay. The present results clearly show the usefulness of combining predictive modelling and use of in vitro bioassays for rapid screening of oestrogenic effects of complex mixtures and environmental samples.