Tools to detect estrogenic activity in environmental waters

Abstract

The occurrence of estrogenic endocrine disruptors in water is of international concern because of potential adverse health effects on wildlife and humans. Chemical analysis and quantification of estrogenic compounds in water is problematic due to the great range of compounds with endocrine activity. Also, the ultra-low concentrations that can cause biological effects make it clear that additional methods are needed for this type of analysis. Bioanalytical methods have become increasingly popular and are seen as a possible screening tool for measuring estrogenic activity in water. Bioassays generally have significantly lower detection limits than chemical methods, provide an integration of potency and dose and, most importantly, require no prior knowledge of the specific chemical nature of a sample. Several in vitro bioassays have emerged over the past decade to test the estrogenicity of environmental samples. There are, however, concerns about their reproducibility, robustness, interlaboratory variability and their ability to integrate into a regulatory framework based on individual chemicals. This report describes an international effort to evaluate the performance of five in vitro bioassays to assess estrogenic activity in a variety of water matrices (http://www.edctoolbox.info). The project was jointly funded by members of the Global Water Research Coalition (GWRC), and evaluated a selected set of bioassays, including yeast estrogen screen (YES), ER-CALUX, MELN, T47D-KBluc and E-Screen assays. Spiked artificial (tap water spiked with known estrogenic chemicals such as hormones, alkylphenols, phthalates, pesticides and phytosterols) and real samples from sewage, river, groundwater and drinking water were tested. The results indicate that the ER-CALUX and E-Screen assays in this study successfully detected estrogenicity in environmental water samples even at very low levels of estrogenicity (from 0.1 to 320 ng/L EEq). The estrogenic activity measured in these bioassays could be correlated to the predicted estrogenic activity based on comprehensive chemical analysis (GC/MS, GC/ECD, and HPLC/MS/MS), suggesting that either of these two bioassays could be used as initial screening tools to detect estrogenicity in environmental water samples. The KBluc assay was very similar to the ER-CALUX, but these conclusions are based on a more limited dataset, and should be considered critically. The YES performed well with highly polluted environmental samples (such as sewage samples) but its relatively high detection and quantification limits meant that it was unable to measure low-level estrogenicity (eg ground and river water). With artificial samples, the performance of the YES assay was also significantly affected by octylphenol. The MELN assay tested in this study provided good qualitative data, clearly identifying low and high estrogenic activity in the samples. However, accurate quantification was more problematic, possibly due to matrix interference from complex matrices (such as sewage) in this assay. This study shows that some bioassay techniques are now sufficiently advanced that they can be used either as a cost-effective first-pass detection system or in combination with standard analytical methods to measure estrogenic pollutants in environmental waters. Standardization of bioassay data analysis was identified as a crucial step forward towards accurate bioassay-derived estrogenicity measurements.