Résultats exploitables
An analytical method consisting of solid phase extraction, cleanup, derivatization and gas chromatography coupled to ion trap detection was developed for the determination of (synthetic) steroid hormones at very low concentrations, i.e. in the low ng/L range. The method was applied within the ACE project for the determination of actual exposure concentrations for in vivo studies, specifically for estradiol (E2) and ethynylestradiol (EE2). Detection limits were 0.2-1ng/L for E2 and 0.5-2ng/L for EE2; average recoveries for the corresponding deuterated internal standards (E2-d4 and EE2-d4) were 95% (n=220) and 99% (n=141), respectively. The method was extended to other target analytes such as the estrogenic steroids estrone, estriol, the synthetic estrogen mestranol and the xeno-estrogen bisphenol A. Furthermore, several androgenic as well as progestagenic hormones can be analysed with the developed method.
Extension to other matrices such as surface water, sewage treatment plant influents and effluents, sediments and sludge required some adaptation of the extraction and cleanup steps, but has proven to be feasible. Extraction of solids is done by Accelerated Solvent Extraction, while different cleanup steps include gel permeation chromatography and reversed phase HPLC fractionation.
The in vitro Estrogen Responsive Chemical Activated Luciferase Gene eXpression (ER-CALUX) reporter gene assay was used in this study to investigate combination effects. The ER-CALUX assay is a recombinant reporter gene assay which measures transactivation of the estrogen receptor following exposure to (xeno-) estrogens in T47D human breast cancer cells stably transfected with an estrogen-regulated luciferase reporter gene construct. In this study, 13 (xeno-) estrogenic compounds, including estradiol, estrone, ethinylestradiol, estriol, mestranol, diethylstilbestrol, bisphenol A, nonylphenol, octylphenol, resorcinol monobenzoate, benzophenone 3, 2,4-dihydroxybenzophenone, and, 4,4'-dihydroxybenzophenone, were tested. Compounds were first tested individually to examine dose-response relationships and determine NOEC, EC1 and EC50 levels. When tested together in ratios based on these levels, mixture effects could be well predicted by the concept of concentration addition, independent of the actual mixture ratio and the effect level under observation. The ratio between observed and predicted EC50 values never exceeded a factor of 1.2. In fact, significant mixture effects could be found even if every compound was present at concentrations that would provoke no significant effect if applied singly, e.g. individual EC1s or NOECs or fractions of these values. In this study, the ER-CALUX assay proved to be very sensitive and gave highly reproducible results, even for the multi-component mixtures.
A fast assay for binding of compounds to the estrogen receptor alpha in 96-well plate format was developed, based on the competition between fluorescent coumestrol and estrogenic compounds. Displacement of coumestrol was measured as a decrease in fluorescence intensity using a microplate reader. Competitive binding curves for the well-known estrogenic compounds 17a-estradiol (E2), ethinylestradiol, 4-nonylphenol, 4-octylphenol, genistein, bisphenol A, tamoxifen and diethylstilbestrol and 9 other estrogenic compounds were constructed by using 7-10 different concentrations of the compounds and a fixed concentration of ER-a-LBD (14nmol) and coumestrol (100nmol). IC50 values and relative potencies (compared to E2) of the estrogenic compounds were determined. The assay was validated by comparing the relative potencies to those from standard radioligand binding assays in the literature. Low CV values were obtained from both negative and positive controls analyzed both in an experiment day (3% and 12% resp.) and in different experiment days (5% and 7% resp.). Z' values of 0.73 and 0.66 obtained from within day and between day experiments, respectively, indicate the high quality of the present assay. The present fluorescent binding assay has proven to be fast and easy, and allows accurately quantifying the binding affinity of estrogenic ligands. The method is also suitable as a high-throughput screening assay for ER ligands in pharmaceutical as well as environmental research.
A modified Yeast Estrogen Screen (YES) that serves the special purpose of studying interactions of estrogenic agents with non-estrogenic toxicants has been developed, implemented and validated. Simultaneous measurability of both estrogenic activity and fungal toxicity in a single experimental set-up was achieved by using inhibition of yeast growth as toxicity parameter. Growth inhibition of yeast cultures is expressed as final biomass in relation to untreated controls and determined by measurements of optical density. The assay was optimised in terms of testing capacity, reproducibility of concentration response curves and sensitivity to both estrogenic and growth inhibiting effects. Details of the protocol, such as initial cell density, incubation time, measuring wavelengths etc. were adjusted accordingly. A serious of estrogenic (estradiol, estrone, estriol) and non-estrogenic compounds (Cycloheximide, Mercury, Dinitro-aniline, DMSO, LAS) were used as reference substances during the process of optimisation and validation. As a result, an optimised test protocol for the modified YES assay has been implemented for routine testing and is now ready for use in the mixture effect analyses.
Existing evidence from mixture studies in other bioassays (especially from the PREDICT and BEAM EU-projects) as well as conceptual/mathematical arguments have been analysed in order to identify the major issues, that facilitate or hamper the successful completion of the mixture studies with the fathead minnow and the sea bass in ACE.
The following topics have been identified as being of major importance:
(a) If the single substance concentration-response relationships are of low accuracy, the assessment of the predictive power of both concepts (Concentration Addition and Independent Action) is severely limited. A low accuracy can either have experimental reasons (a low overall reproducibility of the concentration-effect data) or biometrical reasons (biased, inadequate modelling of the concentration-effect curve).
(b) Predictions according to the concept of Concentration Addition are restricted to the range of effects, which can be described for all mixture components. Hence, if clear differing maximum and/or minimum effects are observed, the applicability of the concept is limited.
(c) Independent Action assumes, that all concentration-effect curves can be reasonably scaled to a common minimum and maximum effect.
(d) The precise estimation of low-effect levels (whose determinations have a high demand in terms of experimental capacity and necessary test animals) is of paramount importance for the application of the concept of Independent Action to multi-component mixtures.
(e) Non-monotonic concentration-effect relationships (as they might be observed for chemicals that show endocrine effects and a high toxicity simultaneously) would clearly violate the conceptual basis of the prediction concepts.
(f) Long time intervals between the first single substance experiments and the final mixture experiments may introduce additional uncertainties.
(g) The number of possible mixture components is only restricted by the available experimental capacity. A high number of mixture components doesn't per se diminish the predictive power.
A question of considerable relevance for the regulation of chemicals and their mixtures is whether combination effects occur when each chemicals is present at low effect concentrations. This issue was addressed experimentally by using a variety of in vitro assays for estrogenicicty, including estrogen receptor binding, the yeast estrogen screen, the E-Screen and the ER-CALUX assay. "Low effect concentration" was operationalised in terms of concentrations equal to EC01 or no-observed effect levels. By combining sufficiently large numbers of estrogenic chemicals it was possible to demonstrate that mixture effects occurred even when each individual component was present at levels around EC1 or NOEC. In the E-Screen assay it was not possible to estimate EC01 for many chemicals because their concentration-response relationships exhibited a very shallow slope in the low dose range, levelling off at the 1% effect level. In these cases, EC2.5 were used.
A freshwater fish species (fathead minnow), and a marine species (sea bass) were used to assess the predictability of the vitellogenin response to a mixture of environmental estrogens. Fish were exposed to chemicals using a flow-through system. Firstly, concentration-response curves were determined for each individual chemical. Mixture effects were predicted on the basis of the potency of each chemical using the model of concentration addition (CA).
Actual mixture effects were then determined by combining the chemicals according to their relative potencies. Fish were then exposed to various dilutions of the mixture using a fixed ratio design (based on the EC50 of each chemical). The level of agreement between the observed and predicted mixture effects was analysed using biometrical modelling techniques. A second mixture experiment using the fathead minnow was designed to investigate the potential for mixture effects at low doses of the individual mixture components. The doses tested were equivalent to one-fifth of the EC50 of each chemical. This dose was selected on the premise that it would not be sufficient to induce a significant effect when present individually, but that combined exposure to all five chemicals at this concentration would induce a 50% response.
The results obtained from the freshwater fish mixture studies demonstrated an excellent agreement between predicted and observed effects for the 5-component mixture. There was no statistical deviation between the best fit of the observed and predicted mixture effect, with the prediction lying within the 95% confidence limits along the full length of the curve. The results of the sea bass studies also demonstrated an excellent agreement between predicted and observed effects with both a 5- and a 3-component mixture. These findings provide evidence that estrogenic chemicals act together according to the principles of CA in fish.
The second phase of mixture studies involving the fathead minnow vitellogenin assay involved the evaluation of the response to the five-component mixture, where each individual chemical was present at a low effect concentration. This demonstrated that although each chemical on its own did not induce a significant response, the 5-component mixture was capable of significantly raising VTG concentrations (to between 50-60% of the maximum response) in the test organism. There was good agreement once more between the observed effect of the mixture and the prediction of CA, with the prediction falling within the confidence limits of the observed effects. This confirms that the combined effects of these chemicals do not deviate from additivity in the low-effect concentration range, and reinforces the theory that allowable limits calculated for environmental concentrations of such chemicals may be inadequate if they are based on low-effect concentrations of individual chemicals. It is clear that, as a result of data such as these, risk assessment procedures, which are currently based on single substance exposure studies, will require extensive review.
A very fast LC-MS method was developed for a rapid (max. 15 minutes analysis time), high-throughput determination of non steroidal, natural (genistein) and synthetic (octylphenol (OP), nonylphenol (NP), nonylphenol mono- and diethoxylates (NP1EO and NP2EO), nonylphenol monoethoxylate carboxylate (NP1EC)) endocrine disrupter compounds (EDCs) in water samples. MS detection is performed by using an Electrospray Interface (ESI) in positive or negative ionization mode. Obtained Method Detection limits (MDLs) are between 0.03µg/L for carboxylated nonylphenol monoethoxylate and 20µg/L for nonylphenol monoethoxylate. The method permits to markedly speed-up the analysis of EDCs mixtures because no extraction/clean up step is necessary, in comparison with traditional methods for the extraction of EDCs from water samples, which are time-consuming (approx. 6-8h/sample) even if fully automated. Possible applications of this method lie mainly in those situations where a large number of samples must be analysed in a short time, and lowest concentration level are in the 0.03-20µg/L level. Toxicity tests or biodegradation experiments could greatly benefit from this technique. In addition, the monitoring of wastewater could be performed by applying this protocol. The proposed method can be easily extended with minor modifications also to other EDCs and organic polar pollutants
The interaction between toxic and estrogenic substances has been studied in a modified Yeast Estrogen Screen, the so-called 2-phase-YES (2P-YES) and the E-Screen assay. We investigated to which extent the apparent estrogenicity of a chemical or a chemical mixture can be masked or enhanced by the simultaneous presences of a non-estrogenic toxicant and whether this is dependent on the molecular mechanism of action of the toxicant. Furthermore, the impact of toxic confounders on the predictability of the joint action of estrogens by the concept of Concentration Addition was analysed. We selected heavy metals (Mercury, Lead), two organic solvents (DMSO, 2,4-Dinitroanline), a surfactant (LAS) and a specifically acting antibiotic (Cycloheximide) as toxicants. Estrogens were selected from a variety of substances, mainly comprising of natural and synthetic estrogens such as 17ß-Estradiol, Estrone or Ethinylestradiol.
Concentrations that provoked severe toxic effects if applied singly always also clearly diminished the apparent estrogenicity of both, single substances and mixtures. That is, the maximum observable estrogenic effect is lowered if a toxicant is present simultaneously to the estrogen(s) and effect concentrations such as the EC50 values are shifted to the right (i.e., the cells seem to react less sensitive to the estrogens). This pattern was observed in both assays (E-Screen and 2P-YES) and with all test chemicals. Additionally, the impacts of low concentrations of toxicants, that do not provoke any visible effects if applied singly, were also investigated n the 2P-YES. A clear dependence on the specifically applied test substance became apparent here. The apparent estrogenic activity was reduced in the presence of low non-toxic concentrations of 2,4-Dinitroaniline and Cycloheximide, but was enhanced by the presence of low concentrations of LAS or DMSO.
The apparent estrogenicity of the mixtures of estrogens was affected in direct proportion to the single chemicals. Consequently the predictability of the joint action by Concentration Addition was compromised, if the presence of confounder was not already accounted for on the level of single substances. Both under- and overestimations by Concentration Addition occurred, depending on the impact on the apparent estrogenicity of the mixture components. As soon as this is accounted for, i.e. the toxic confounders are present in equal amounts during the single substance and the mixture experiments; predictability by Concentration Addition is restored.
These results clearly show that toxic confounders need to be considered during the assessments of the estrogenic potential of any substance, chemical mixture of complex environmental sample. Otherwise there is a severe risk of over- or underestimating its "true" estrogenicity. This could for example lead to the seemingly contradictory situation that the estrogenicity of an environmental sample is higher after a remediation than before - due to a decrease in the concentration of a toxic confounder. It has to be pointed out that at the moment no model is at hand that could quantitatively predict the impact that a given concentration of a toxic confounder has. Hence, it is of paramount importance to carefully consider the presence of toxic confounders while testing any sample for estrogenicity. For this purpose suitable toxicity indicators are needed.
The predictability of mixture effects of multi-component mixtures of estrogenic chemicals was assessed for a number of in vitro assays measuring estrogenicity. The assays included an estrogen receptor binding assay (HRS), the yeast estrogen screen (YES), the E-Screen and the ER-CALUX assay. Five chemicals, i.e. ethynylestradiol, estradiol, bisphenol A, nonylphenol and octylphenol were tested in all assays. Concentration-response relationships were recorded, and this information was used to predict quantitatively the effects of combinations of all five agents. The mixture effect predictions were calculated using the concept of concentration addition. In all assays, there was good agreement between prediction and observation, although slight deviations from expectation were noted in the E-Screen. These results show that the joint effect of all chemicals is (concentration) additive in most cases. These studies were extended by examining mixtures composed of up to 19 components in the four in vitro assays. In most cases it was possible to predict with accuracy the resulting joint effects, by using the concentration addition concept.
Concordance between assays was analysed for single substances, as well as for mixture prediction. For single substance concordance, broadly speaking, in vivo responses are predicted by in vitro assays, for the five chemicals E2, EE2, BPA, OP and NP. Where anomalies were encountered, they were either not significant in a regulatory context (e.g. small fluctuations of relative potencies of xenoestrogens between assays), or were predictable (e.g. the reduced potency of E2 in vivo).
The analyses of mixture effects in the suite of assays employed in ACE demonstrated the capacity for estrogenic chemicals to act together in an additive manner both in vivo and in vitro, according to the principles of CA. This corroborates the working hypothesis of ACE, which states that additive effects are mirrored at increasing levels of biological complexity.
Evidence that estrogenic chemicals behave according to the principles of CA at different levels of biological complexity demonstrates the value of utilising in vitro approaches when trying to establish combination effects of mixtures of estrogenic chemicals. There is considerable potential for extrapolation of this in vitro data to assess whether there may be a risk of significant effects in vivo. This may help to avoid some of the costly and time-consuming whole animal studies used in risk assessment procedures.
The wide variety of chemical structures and their chemical/physical properties render the determination of endocrine disrupters (EDC) mixtures a challenge for analytical chemistry. Many analytical methods by both GC and HPLC have been recently proposed in literature for the specific determination of EDCs in environmental waters (river waters, sea waters, final WWTP effluents), but they are limited to the determination of individual compounds (estradiol, ethynylestradiol, bisphenol-A) or simple classes (alkylphenols, steroidal EDCs) of EDCs. The major drawback of GC techniques is that often-preliminary derivatization steps are needed in order to increase volatility of analytes, thus increasing analysis time and avoiding the concurrent detection of structurally different compounds. HPLC on the other hand, joining the absence of derivatization steps and the large variety of compounds that can be analysed, especially the polar ones, is becoming the most suitable technique for the determination of EDCs.
A new analytical method was developed for the simultaneous determination of EDCs in fresh and salt-water samples. The covered EDCs include the steroidal compounds estradiol, ethynylestradiol, estriol, estrone and mestranol, as well as the nonsteroidal octylphenol, nonylphenol, nonylphenol monoethoxylate carboxylate, bisphenol-A, benzophenone, genistein and diethylstilbestrol. The method is based on solid phase extraction (SPE) followed by LC-ESI-MS analysis.
Recovery efficiencies range between 70 and 99%, with the exception of estriol and benzophenone, which show recoveries of 60 and 50%, respectively. The method detection limits (MDLs) were between 0.1 ng/L for nonylphenol monoethoxylate carboxylate to 4ng/L for mestranol. The developed analytical method is enough sensitive, robust and rapid to be applied to the routine determination of natural and synthetic EDCs in natural waters, thus permitting a more comprehensive evaluation of the exposure to EDCs in the aquatic environment.
Recherche de données OpenAIRE...
Une erreur s’est produite lors de la recherche de données OpenAIRE
Aucun résultat disponible