Endocrine activity in the aquatic environment - Establishment of a bioassay battery to detect endocrine disruptive substances in various sample matrices

摘要

Earth’s accessible freshwater resources are under constant pressure. Most of the 35 million cubic kilometers freshwater are bound in glaciers (over 68 %), permanent snow und not accessible groundwater resources. Accessible freshwater is crucial for mankind as well as ecosystems functions in general. At the same time, over 50 % of all renewable and accessible freshwater flows are already used by mankind. In Europe only 21 % of the available freshwater are used and most of it is not consumed but restored. Therefore, Europe’s problems related to freshwater supply do not depend on absolute quantity but on distribution, alterations of the hydromorpholgy and an increasing number of (anthropogenic) pollutants. Because of their potential to disrupt the reproduction over altering the endocrine system one important and heavy discussed group of micropollutants are the endocrine disruptive compounds (EDCs). In the last years more and more (potential) EDCs have been detected in the environment as well as organic tissues of humans and wildlife. For the prediction of a substance’s endocrine activity a first screening with QSAR tools (e.g. OECD QSAR Toolbox, VirtualToxLab) is useful. These predictions actually are good in case of receptor binding properties if the chemical structure is known. To screen samples for known EDCs, chemical analysis is a fast and sensitive approach. For example, to date it is possible to detect estrogens in the low ng∙L-1 range (e.g. 0.3 ng∙L-1 17ß-Estradiol) in groundwater samples with a LC system coupled to a triple-stage quadrupole mass spectrometer with electrospray ion source (LC-ESI-MS2). Nevertheless, to detect unknown endocrine disruptive potentials of a substance or to screen an environmental sample with unknown composition, bioassays are the methods of choice. In the present study, the two recombinant receptor-reporter assays lyticase assisted yeast estrogen screen (LYES) and “estrogen receptor mediated chemical activated luciferase gene expression” (ER CALUX®) assay as well as the H295R steroidogenesis assay, that allows the detection of effects on the whole steroid synthesis pathway, were applied. Additionally, a variation of the in vivo reproduction test with the New Zealand mudsnail Potamopyrgus antipodarum (Pa-Repro) has been used. This bioassay battery therefore covers receptor-mediated activities as well as activities on the process of steroid genesis and impacts on whole organisms as well as populations. To get insights into the development as well as international validation of knew test systems for the detection of EDCs the validation of the newly developed H295R was scientifically attended. It could be shown that the H295R successfully detected all suspected or known EDCs with good correlations between the different laboratories. Nevertheless, beside the number of advantages of this test system some disadvantages were also figured. Those were the high variability of steroid concentration between different passages as well as the high costs for the detection of produced steroids. To overcome the disadvantages several approaches to improve the test system were pursued. At first cells were adapted to chemical defined media which lead to a more stable production of steroids. With the development of a new ELISA method for the detection of steroids in the exposure medium both costs as well as working time could be saved. Hereafter, the selected bioassay battery was applied to various sample matrices as well as single substances. Both LYES and ER CALUX® were used to assess the estrogenic activity of heterocyclic aromatic hydrocarbons (hetero-PAHs). While none of the hetero-PAHs showed estrogenicity in the LYES, nine (indole, 1 benzothiophene, benzofuran, 2-methylbenzofuran, 2,3-dimethylbenzofuran, quinoline, 6 methylquinoline, carbazole, dibenzothiophene, dibenzofuran, acridine, and xanthene) out of twelve showed significant estrogenic activity in the ER CALUX®. With the help of liquid chromatography with diode array detection (LC-DAD) and high resolution tandem mass spectrometry (LC-HRMS2), it could be demonstrated that the used T47Dluc cell line was capable for metabolizing of the parent substances. To assess the endocrine disrupting potential of the biological larvicide VectoBac®, which is extracted out of Bacillus thuringienses var. israelensis (Bti), the whole bioassay battery was used. With both receptor-mediated assays LYES and ER CALUX® the previously detected estrogenic activity of three commercial available tablet formulas could be confirmed. From the subsequently investigated active substance VectoBac® TP as well as the formula WDG only TP showed dose-dependent estrogenic activity in both assays. Neither for TP nor WDG clear influences on the steroidogenesis could be detected in the H295R. The investigation of surface and groundwater samples which were taken after the application of both TP and WDG also showed no estrogenic activity in the receptor-mediated bioassays. To elucidate the potential effect of the suspected xenoestrogen VectoBac® TP on a complex aquatic community a microcosm study was performed. Within this microcosm, the impact on phytoplankton and zooplankton as well as the grazer P. antipodarum could be studied. While no effects on the zooplankton were detected, a clear shift from green to blue algae could be detected at concentrations which also showed significant estrogenic activity in the LYES. No clear effects on the reproduction of P. antipodarum could be observed, but it is assumed that this was a result of the limited food supply due to the community shift within the phytoplankton. Nevertheless, the concentrations used in the microcosm study were ten to 50 times higher than it would be expected in the field and therefore, a shift in the ecosystem structure of routinely treated areas seems unlikely. In the last approach the three selected in vitro bioassays were used to assess the performance of advanced sewage treatment processes with regard to the removal of pharmaceuticals from hospital sewage before discharge into the public sewerage. The high estrogenic activity of the raw sewage sample detected in the LYES could be significantly reduced by a full scale membrane bioreactor (MBR). After subsequent ozonation the estrogenicity was reduced to the level of the solvent controls in both receptor-mediated assays. By means of liquid chromatography-tandem mass spectrometry (LC-MS2) it could be verified that after ozonation concentrations of estrone (E1) were significantly reduced. In contrast, there was a significant increase of 17ß-estradiol production as well as aromatase activity in H295R cells after ozonation. It is hypothesized that this is partly due to formation of active by-products during ozonation. Overall, the selected in vitro bioassays were suitable for the detection of endocrine activity in various sample matrices as well as single substances. The receptor-mediated bioassay ER CALUX® was able to detect estrogenic activity in all matrices used. These included groundwater, surface water and sewage as well as single substances (hetero-PAHs). Additionally, in case of the hetero-PAHs, the estrogenic activity of single substances which require metabolic activation could be detected. With the LYES investigation of medium to highly contaminated water samples was possible, low contaminated water samples could not be investigated due to the high limit of detection (LOD) as well as the relatively high estrogenicity of the solvent controls. The H295R was applicable for medium contaminated samples as well as single substances. Performance of the bioassays markedly varied throughout the present study. The ER CALUX® provided the most stable performance (very good), followed by the H295R (good), the LYES (good to moderate) and P. antipodarum (moderate). Problems occurred with the consistent production of testosterone (H295R), the background of the solvent controls (LYES) as well as food supply and response to the positive control (P. antipodarum). Concluding, it can be stated that the selected bioassay battery is suitable to detect (potential) endocrine disrupting substances in different water sample matrices such as groundwater, surface water and wastewater. Additionally, even single substances that require metabolic activation to exert their estrogenic activity can be investigated.