ToxSci Advance Access originally published online on July 28, 2004
Toxicological Sciences 2004 82(1):154-163; doi:10.1093/toxsci/kfh238
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Toxicological Sciences vol. 82 no. 1 © Society of Toxicology 2004; all rights reserved.
Xenoestrogenicity in In Vitro Assays Is Not Caused by Displacement of Endogenous Estradiol from Serum Proteins
* Institute for Risk Assessment Sciences (IRAS), Utrecht University, Yalelaan 2, 3584 CM Utrecht, the Netherlands; Biodetection Systems B.V., 1031 CM Amsterdam, the Netherlands; National Institute of Public Health and the Environment (RIVM), 3720 BA Bilthoven, the Netherlands
Received April 30, 2004; accepted July 26, 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX A. MODEL DERIVATIONREFERENCESAPPENDIX B. COMPETITION MODEL...REFERENCES |
|---|
The possibility that compounds tested for estrogenicity can compete for binding places on serum proteins and cause an increase of available and very potent endogenous estrogens is of great interest for both the in vitro assay results and the prediction of risk for humans. In in vitro assays, small amounts of estradiol remaining after the charcoal stripping of serum applied in the culture medium could be displaced by the tested compounds, leading to an estrogenic response that might be falsely attributed to the test compound. We have studied the stripping efficiency of charcoal and measured whether reported xenoestrogens can displace estradiol from serum in an in vitro assay using negligible depletion–solid phase microextraction (nd–SPME). Possible competition was also studied with a mathematical exposure model, from which the predictions were compared to the measurements. We found that the common charcoal stripping procedure removed 99% of initially present estradiol. Additionally, our results with charcoal adsorption indicate that charcoal is not useful for serum protein binding assays, as it adsorbs more than the free fraction of ligand. Although the competition model predicted a displacement of estradiol from the serum proteins at the higher applied doses of xenoestrogen, the measurements showed no displacement. Therefore, we conclude that estrogenic responses in the in vitro assay applied here are not caused by displacement of remaining estradiol in the stripped serum. The possibility remains, however, that our displacement hypothesis does apply for estrogen sulfates, as these are present in much higher concentrations than estradiol in stripped serum.
Key Words: displacement; serum proteins; estrogen; charcoal assay.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX A. MODEL DERIVATIONREFERENCESAPPENDIX B. COMPETITION MODEL...REFERENCES |
|---|
Since the announcement that exposure to industrial chemicals can impair animal reproduction (Colborn et al., 1993
; Jobling and Sumpter, 1993), a large research effort has been conducted to identify compounds that are indeed estrogenic both by in vivo and in vitro assays. One of the issues that has become clear from this research effort is that the availability of compounds is crucial for their activity. Binding to the serum proteins albumin and sex hormone binding globulin (SHBG) appears to have a large influence on the activity of (xeno-)estrogens (Crain et al., 1998; Nagel et al., 1997).
Furthermore, it has been reported that mixtures of compounds can also influence the availability of the individual compounds. Both Danzo (1997) and Déchaud et al. (1999) found that xenoestrogens can displace the endogenous steroids dihydrotestosterone (DHT), testosterone, and estradiol, from SHBG. This displacement might disrupt the normal hormone balance, forming a pathway of endocrine disruption alternative to the receptor-binding pathway. However, SHBG is only a minor binding protein for estrogens in serum. Whether the displacement of endogenous steroids by xenobiotics can occur in whole (human) serum, or on other serum proteins, is still a question. Villeneuve et al. (2002) found no displacement of estradiol or testosterone from carp serum proteins by polybrominated diphenyl ethers (PBDEs). In contrast, Ishihara et al. (2003) found interference of endocrine disrupting compounds such as diethystilbestrol and pentachlorophenol with the binding of a thyroid hormone with the plasma protein transthyretin.
The possibility that xenobiotics can compete for binding sites on serum proteins and cause an increase of the free concentration of very potent endogenous estrogens is of great interest for both the in vitro assay results and the prediction of risk to humans. Humans contain a large reservoir of bound endogenous estrogens and naturally have many of the binding places on the serum proteins already occupied by endogenous compounds. Exposure to a mixture of xenobiotics could therefore theoretically lead to a ready displacement of endogenous estrogens.
Competition may also occur in in vitro tests. Most in vitro estrogenicity assays apply a small percentage of serum in the culture medium of the cells. This serum is usually stripped with charcoal, to remove the majority of the endogenous estrogens. Our hypothesis, which formed the basis for this study, is that a small amount of protein-bound estradiol, remaining after the stripping procedure, can be displaced by a test compound in the in vitro assay. This could cause an estrogenic response that is then not directly attributable to the test compound. The high dosages necessary for most reported xenoestrogens to produce an effect support this possibility of displacement: a high dose may saturate the serum proteins, leading to competition.
We have investigated this hypothesis. First, we have determined how much estradiol remains in serum after the common stripping procedure with charcoal. Further, a mathematical exposure model was formulated to predict the occurrence of competition and displacement, and this model was validated with a series of competition experiments in 98.5% serum, where free concentrations were measured using nd-SPME, a method developed by Vaes et al. (1996). Lastly, experimental and calculation results were compared for the occurrence of competition in an in vitro estrogenicity assay with 5% serum. The combination of modeling and experimental work is extremely useful for generation and exploration of hypotheses and for interpretation of the outcome of experiments.
Besides its application in the stripping of serum, charcoal is also commonly used to separate bound and free ligands in protein binding studies as performed in this work. These studies mainly include receptor-binding studies [e.g., Campbell and Clark (1984); EORTC (1973); Skovgaard Poulsen (1982)], but charcoal is also applied as a separation method in binding studies with serum proteins [e.g., Danzo (1997); Tollefsen (2002); Villeneuve et al. (2002)], including studies on competition. Pitfalls and sources of variability of this method have already been reported (Campbell and Clark, 1984; Peck and Clark, 1977; Pettersson et al., 1985; Skovgaard Poulsen, 1981; Thorpe, 1987), and the absorption time of the charcoal appears to be critical (Laidley and Thomas, 1994; Peck and Clark, 1977). As the affinities for serum proteins are generally much lower than those for receptors, dissociation rates will be faster for these proteins. Thus, the charcoal will start to adsorb dissociated ligands sooner in the case of serum proteins than in the case of receptors. As we had doubts concerning the efficiency and removal rate in the charcoal assay, we verified whether only the free fraction is removed by charcoal in serum protein solutions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX A. MODEL DERIVATIONREFERENCESAPPENDIX B. COMPETITION MODEL...REFERENCES |
|---|
General Procedures
Apparatus and reagents. [2,4,6,7–3H]17ß-estradiol (37 MBq/ml) was purchased from New England Nuclear (Boston, MA). Specific activity and radiochemical purity are specified for each experiment. 4-n-octylphenol (99%) was purchased from Aldrich (Steinheim, Germany), PCB #138 (2,2',3,4,4',5-hexachlorobiphenyl; 99%) from Riedel-deHaën (Seelze, Switzerland), and 2,3,4-trichlorophenol (TCP; 99%) from Aldrich. Ethylacetate (99.8%), ethanol (99.8%), and the derivatization agent MTBSTFA (N-methyl-N-(tert-butyldimethylsilyl)-trifluoroacetamide; >97%) were purchased from Lab-Scan (Dublin, Ireland), Riedel-deHaën, and Fluka (Buchs, Germany), respectively. 7-µm polyacrylate (PA) and 7-µm polydimethylsiloxane (PDMS) fiber for nd-SPME was purchased from Supelco (Bellefonte, PA), as well as the 1.8-ml vials and 250-µl inserts. Fetal calf serum (FCS) was obtained from Bodinco (Alkmaar, the Netherlands).
Cell culture. Human embryonal kidney 293 (HEK293) cells were obtained from the American Type Culture Collection (ATCC; Rockville, MD) and cultured as described by Lemmen et al. (2002).
DCC-stripping of serum. A dextran-coated charcoal (DCC) slurry with 0.5% charcoal (Norit SA-3; Boom, Meppel, the Netherlands) and 0.05% dextran (Sigma) in 50 mM Tris buffer (pH 7.4) was prepared and left to stir overnight at 4°C. Fetal calf serum was heated to 56°C for 30 min in a water bath. A volume of slurry similar to the volume of serum to be stripped was centrifuged for 10 min at 1500 x g (Megafuge 1.0 of Heraeus Sepatech, Osterode/Harz, Germany). The supernatant was discarded and the FCS was added to the remaining pellet. This slurry was gently shaken for 45 min in a water bath of 45°C, and subsequently centrifuged for 20 min at 1500 x g. The supernatant FCS was then transferred to fresh DCC pellets and the stripping procedure was repeated. The resulting supernatant FCS was filtered over a 0.2 µm filter.
Analytical procedure. [3H]estradiol in water samples and PA SPME fibers was extracted with scintillation liquid and analyzed as described in Heringa et al. (2002). Total [3H]estradiol concentrations were derived from the water samples by calibration with a series of known amounts of [3H]estradiol dissolved in scintillation liquid. Free concentrations were derived from the concentrations in the fiber (Cf, determined using the same calibration series) using equation 1.
 | (1) |
were used: k1 = 64 min–1 and k2 = 0.014 min–1. In the experiments performed in in vitro culture plates, however, these constants could not be used because the sample volume was different, and this affects the value of these rate constants. Therefore, free aqueous concentrations were derived from fiber measurements by a separate calibration series of fiber concentrations sampled in known free aqueous concentrations.
Octylphenol was extracted from PDMS SPME fibers and water samples with ethylacetate. The ethylacetate extracts were derivatized with MTBSTFA and spiked with two internal standards (PCB #138 and TCP) and then analyzed on a Varian GC-MS as in Mol et al. (2000). Total concentrations were derived from the water samples by calibration with a series of known concentrations of octylphenol. Free aqueous concentrations were derived from fiber measurements by a separate calibration series of fiber concentrations sampled in known free aqueous concentrations.
Mathematical model for competition. A mathematical model was formulated to predict potential competition in serum, to identify the critical factors, and to support our experimental set-up. The essence of this model was formed by the mass balances given in equation 2 (see Appendix A for their derivation). In this equation, A, S, E, and X are albumin, SHBG, estradiol, and competitor concentrations, respectively, with the subscripts free and total indicating free or total concentrations. Two letters together, such as EA, denote the concentration of the complex of the specified compound with the specified protein, in this case estradiol bound to albumin. Ka is the affinity constant, with the capital letters in the subscript indicating for which compound and protein this constant is applicable. The model included both albumin and sex-hormone binding globulin (SHBG) as binding proteins, as practically all binding of estradiol (and probably also xenoestrogens) in serum is accountable to these two proteins (Dunn et al., 1981).
 | (2) |
|
Experiments
Stripping efficiency of DCC. DCC-stripped serum was spiked with [3H]estradiol (2.627 TBq/mmol; >95% radiochemical purity) to a concentration of 7.0 x 10–9 M and left to equilibrate overnight at 4°C. Three control vials and six other vials were filled with 1.6 ml of this spiked serum. From the controls, 100-µl samples were counted and the remaining solution was subjected to nd-SPME for 2 h as described in Heringa et al. (2002)
to determine the free fraction of estradiol. The other six samples were subjected to the stripping procedure as described above, excluding the filtration step. Three of the six vials were also excluded from the heating step, to determine its effect. The resulting supernatants were transferred to clean vials and centrifuged again, after which the supernatants were again transferred to clean vials. 100-µl samples of these supernatants were counted. Total concentrations of [3H]estradiol of the stripped sera samples were divided by those of the controls to obtain the stripping efficiency.
Applicability of charcoal assay for albumin or serum binding. In this experiment, we tried to determine at what absorption time the charcoal starts adsorbing more than the free fraction of ligand in serum protein solutions. A solution of 1.4 x 10–9 M [3H]estradiol (2.627 TBq/mmol; >90% radiochemical purity) and 5.01 x 10–4 M BSA (Sigma; 98% pure, essentially fatty-acid free, essentially 
-globulin free) in 50 mM Tris buffer (pH 7.4) and one of 7.0 x 10–9 M [3H]estradiol in 100% DCC-FCS was prepared and left to equilibrate at 4°C for 2 h and overnight, respectively.
Twenty-four 1.8-ml vials, containing a pellet from 1.6 ml DCC-slurry, were filled with 1.6 ml of the BSA solution and shaken. After adsorption times ranging from 30 s to 27 h, duplicates of the vials were centrifuged at 1500 x g for 2 min, after which 100 µl of supernatant was quickly transferred to a counting vial and analyzed. Two empty vials were also filled with 1.6 ml BSA solution (controls), from which 100 µl was counted and the rest was subjected to 33 min of nd-SPME (Heringa et al., 2002). These experimental procedures were performed at room temperature (20° ± 2°C). Total concentrations of [3H]estradiol in the stripped samples were divided by those of the controls and set out against the DCC adsorption time. From the fiber measurements, the free fraction of [3H]estradiol in the controls was calculated.
A similar procedure was followed for the FCS, apart from the batch of [3H]estradiol (2.627 TBq/mmol; >95% radiochemical purity) and the DCC adsorption times, which ranged from 20 s to 7 h. Further, three controls were prepared instead of two, and the centrifugation time was only 1 min. The experiment with FCS was repeated at 4°C, to study the effect of temperature on adsorption speed.
Calibration and validation of competition model. Three competitive binding assays in 98.5% serum were performed with three competitors for estradiol, to calibrate and validate the competition model. For each assay, 11.7 µl of [3H]estradiol stock solution (2.627 TBq/mmol,; >97% radiochemical purity) was added to 23.0 ml of FCS. After mixing, 1.576 ml of this mixture was transferred to 1.8-ml vials and 16 µl of a dilution of the competitor in ethanol was added to obtain total competitor concentrations of 1.0 x 10–9, 1.0 x 10–8, 1.0 x 10–7, 1.0 x 10–6, 1.0 x 10–5, and 1.0 x 10–4 M in duplicate, and a [3H]estradiol concentration of 7.2 x 10–9 M (physiological human concentration according to Saeki et al. (1991)). After equilibration for at least 2 h, 100 µL was sampled from each of the 12 vials and counted, and the remaining mixtures were subjected to nd-SPME for 2 h as in Heringa et al. (2002). The competitors studied were bisphenol-A (Sigma), ß-endosulfan (98.7%, Riedel-de Haën), and non-labeled 17ß-estradiol (98%, Sigma). The free fraction of [3H]estradiol was plotted against the nominal concentration of the competitor. The model was manually fitted to the data of the experiment with non-labeled estradiol, by adjusting only Ka,SE, Atotal, and Stotal, because the values for these parameters were uncertain. The resulting calibrated model was then compared to the data of the experiments with bisphenol-A and endosulfan for validation.
Competition of octylphenol in in vitro assay with cells. This experiment was performed to determine whether the response of octylphenol in an in vitro estrogenicity assay could be caused by displacement of a low concentration of estradiol, for example the estradiol remaining in serum after stripping. The in vitro estrogenicity assay chosen as an example assay was the estrogenicity reporter gene assay with ERß-stably transfected 293 HEK cells as described by Lemmen et al. (2002). The cells were plated in four 24-well tissue culture plates (Corning Inc., Corning, NY) with 80,000 cells and 200 µl suspension per well. After 48 h, the medium was replaced by 2.5 ml/well of fresh phenol-red–free medium containing 5% DCC-FCS and the desired concentrations of [3H]estradiol (3.52 TBq/mmol; >94% radiochemical purity) and/or octylphenol. Dose–response curves of estradiol alone, octylphenol alone, and octylphenol with a low amount of estradiol were created. A concentration of 1 x 10–11 M estradiol was chosen for the combination assay, as this was the lowest concentration at which the free fraction could be detected.
Plate-lids containing SPME fibers were then placed on all plates and the plates were placed in the incubator. Exactly 24 h later, the lids with fibers were removed and the fibers were processed as described above to measure free concentrations of estradiol and octylphenol. Samples of the culture medium were taken and analyzed as described above to obtain total estradiol and octylphenol concentrations. From these two parameters, the free fractions were calculated.
The estrogenic response (luciferase activity) of the cells in plates IV–VII was then analyzed as described by Lemmen et al. (2002). Free fractions of [3H]estradiol were plotted against the nominal octylphenol or estradiol concentration and compared with the prediction of the competition model. Values for the parameters needed in the model were taken from Table 1, except for Stotal and Ka,SE, which were obtained from the model validation. For Atotal, the minimum and maximum values from Table 1 were both modeled, as this parameter is uncertain and at the same time very sensitive for the outcome.
Cell responses (E) were plotted against the logarithm of the nominal concentration and equation 3 (a four-parameter logistic equation, allowing for a variable slope) was fitted through the data points to obtain the EC50 value of the dose–response curves.
 | (3) |
Precautions. National guidelines for handling and disposing of radiolabeled compounds and genetically modified cell cultures were followed.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX A. MODEL DERIVATIONREFERENCESAPPENDIX B. COMPETITION MODEL...REFERENCES |
|---|
Stripping Efficiency of DCC
Table 2 shows that the stripping procedure removes almost all (99%) estradiol present initially, whereas 95% of it was bound.
|
Applicability of Charcoal Assay for Albumin or Serum Binding
Figure 1 shows the fraction of estradiol left with increasing DCC treatment times. The drawn line represents the fraction that was initially bound to protein, which was about 95%. The data show that after only 30 s, DCC absorbs much more than only the free fraction of estradiol in serum or a BSA solution. The similarity between the results in serum and the BSA solution can be explained by the dominancy of albumin in the binding of estradiol in serum. Decreasing the temperature decreases the adsorption efficiency, but even at 4°C much more than only the free fraction is removed after 20 s.
|
Calibration and Validation of Competition Model
Because unlabeled estradiol was the only competitor producing a distinct increase in the free fraction of labeled estradiol, and because its binding affinities were best known, the model was first fit to the data of unlabeled ("cold") estradiol (Fig. 2) for calibration. The model gave a good fit at a BSA concentration of 1.6 ± 10–4 M, an SHBG concentration of 3.0 x 10–8 M, and an estradiol-SHBG binding affinity of 1.0 x 108 M–1, which are reasonable values according to literature (Table 1). The hump in the model output at 1 ± 10–6 M unlabeled estradiol is caused by the displacement of labeled estradiol from SHBG, which occurs at a lower competitor concentration than the displacement from albumin (at 1 ± 10–4 M unlabeled estradiol).
|
The competition with bisphenol-A and ß-endosulfan was modeled with the same parameter values for Atotal, Stotal, and Ka,SE (the total albumin and SHBG concentration and affinity constant of estradiol for SHBG) found for unlabeled estradiol, to validate the calibrated model. The data and the output of the model are shown in Figure 3A and 3B. Although the model predicts an increase in the free fraction of estradiol at a concentration of about 1.0 ± 10–5 M competitor, the experimental data do not follow that trend. Bisphenol-A shows a slight competition, but the endosulfan data do not follow the model's prediction at all in this respect. Perhaps if measurements could have been possible at higher competitor concentrations, competition would appear at higher concentrations than the model predicted. However, the low solubility of these compounds limits such study.
|
Competition of Octylphenol in In Vitro Assay with Cells
The free fraction of [3H]estradiol increased slightly with the addition of octylphenol (Fig. 4), suggesting a displacement from the proteins. However, because of the large standard deviations, this does not seem a significant increase. Furthermore, when these free fractions are compared with those at various nominal [3H]estradiol concentrations (without octylphenol present), the increase appears to be a measurement variability rather than a displacement.
|
). Error bars denote standard deviations (n = 3). The first point of the octylphenol set represents a control, containing only the [3H]estradiol.Comparison of the supplier's value for Atotal with the minimal literature value revealed that the supplier's value gave the best prediction of the free fractions of octylphenol. Therefore, this value (1.85 x 10–5 M BSA) was used to compare the measured free fractions of estradiol with the model's predictions (Fig. 5). Clearly, this prediction does not fit the measured data well: the model underestimates the free fraction and predicts that displacement of [3H]estradiol should occur within the tested competitor concentrations range, whereas the measurements do not show any displacement.
|
The dose–response curves of octylphenol do not show any effect of the presence of [3H]estradiol, either (Fig. 6). Cell response after addition of octylphenol is therefore not caused by a release of bound estradiol, but by octylphenol itself.
|
), or octylphenol in presence of 1 x 10–11 M [3H]estradiol (
). Error bars denote standard deviations (n = 3). The first point of both curves (indicated by *) represents a control, containing only ethanol; the second point of the combination set (indicated by **) represents another control, containing only the [3H]estradiol. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX A. MODEL DERIVATIONREFERENCESAPPENDIX B. COMPETITION MODEL...REFERENCES |
|---|
Stripping Efficiency of DCC
The heating step in the stripping procedure is meant for denaturation of the SHBG, probably aiming for dissociation of the SHBG-bound ligands and increase of the stripping efficiency. Indeed Siiteri et al. (1982)
found higher free fractions of estradiol after heat treatment of serum. In contrast, our data show that the heating step slightly decreases the stripping efficiency. The cause of this discrepancy is unclear.
These results indicate that roughly 1% of the initially present estradiol remains in the serum, which corresponds to the results of Vignon et al. (1980). As the physiologic concentration of unconjugated estradiol in serum of fetal calves is 50–130 pM (Peterson et al., 1975), stripped FCS will contain 0.5–1.3 pM of unconjugated estradiol. If all of this remaining estradiol were displaced from serum proteins by other compounds, this would not be enough to produce a distinct estrogenic response in luciferase reporter gene assays (Lemmen et al., 2002), the E-SCREEN (Soto et al., 1995), or the YES (Arnold et al., 1996).
However, serum also contains a 1000-fold larger concentration of sulfated estradiol: 170 nM in FCS (Fairclough et al., 1976). Sulfated estradiol can readily be taken up by cells and transformed to the unconjugated form, producing estrogenic effects (Vignon et al., 1980). The stripping efficiency of these conjugates from serum with DCC could be much less than that of unconjugated estradiol, as Vignon et al. (1980) found that 10–15% of sulfated estrone (another endogenous estrogen) remains after the general stripping procedure. Assuming that sulfated estradiol is stripped with the same efficiency, 17–25 nM of estradiol sulfates would remain in DCC-stripped FCS. In fact, Vignon et al. (1980) measured 22.5 nM of total estrogen in DCC-stripped FCS. If only 1% of these sulfates were displaced from proteins and hydrolyzed by cells, that 1% would already be enough to cause an estrogenic effect in all three assay-types mentioned above.
Applicability of Charcoal Assay for Albumin or Serum Binding
Laidley and Thomas (1994) found a comparable effect of time on absorption of [3H]testosterone from diluted trout plasma: remaining testosterone decreased by a factor of two when the charcoal adsorption time was increased from 1 min to 2 min. However, they did not measure the total or free amount of testosterone in the control (i.e., before charcoal adsorption), as we did, so it is unclear how much of the bound fraction was adsorbed in their experiment.
Our data clearly show that the DCC treatment for removing the free fraction in binding studies with serum proteins requires caution. Further evaluation is necessary for definititive conclusions, including studies with proteins that have different dissociation rates. For example, the low affinity for albumin (8.9 x 104 M–1), in comparison with the estrogen receptor (1–10 x 109 M–1 (Kuiper et al., 1997, 1998)), leads to a faster release of albumin-bound estradiol than receptor-bound estradiol. It would also be interesting to verify whether the results of serum binding assays using charcoal deviate from the same studies using a different separation technique. In the meantime, we prefer nd-SPME for binding and competition studies.
Calibration and Validation of Competition Model
The albumin binding appears to be the most significant factor in the occurrence of competition, which is logical, because estradiol is easily displaced from this protein as a consequence of its relatively low affinity for it; in addition, albumin has a large portion of the estradiol bound to it. This means that the possible displacement of endogenous hormones from SHBG by xenobiotics (Danzo, 1997; Déchaud et al., 1999; Meulenberg, 2002) has no significant consequence for the overall displacement in serum: the small amount of estradiol that can be released by SHBG will instantly be bound by the vast number of available albumin molecules.
The model appears to predict the free fractions quite well, but the turning point (where competition starts) was difficult to validate, because of the concentration limits imposed by the solubility of the competitors. The discrepancy between the model output and the experimental data could be caused by a possible substantial deviation of the estimates for the affinity constants of bisphenol-A and endosulfan for albumin (see Table 1) from the true values. Furthermore, the discrepancy could be related to the presence of other binding proteins in the serum, which are not included in the model. Serum is known to contain other binding proteins, such as 
1-acid glycoprotein and corticosteroid-binding globulin (CBG) (Englebienne, 1984). Moreover, FCS contains -fetoprotein (Abe et al., 1976), which is similar to albumin (Dobrila, 1996). Lastly, the competitors may be able to bind to more than one binding site on albumin, whereas only one binding site is assumed. Further validation and also further expansion of the model, including other binding proteins and ligands, would be valuable.
The measurements show that estrogenic effects due to competition in 100% serum, as in the in vivo situation, are not likely to occur below a total xenoestrogen concentration of 1 x 10–4 M in the serum. This is similar to the finding of Milligan et al. (1998) that displacement from human and fish plasma will occur only at very high concentrations of estrogenic agents. It is doubtful whether such a concentration will be reached, as literature values for human blood levels of xenoestrogens are, for example, 9 nM bisphenol-A (Ikezuki et al., 2002), 0.41–46 nM p,p'-DDE (Krstevska-Konstantinova et al., 2001), and 2.8 nM hexachlorobenzene (Charlier et al., 2003). However, considering that all the endogenous compounds also occupy binding sites in vivo, the sum of all the exogenous compounds might rise to a high enough level to cause competition. Competition in vivo is therefore still of interest in the field of endocrine disruption.
Competition of Octylphenol in In Vitro Assay with Cells
Again, the model predictions on the concentration of octylphenol at which competition occurs and the actual measurements differ. An explanation additional to those discussed above, is that it is imaginable that octylphenol can bind to the fatty acid binding sites of albumin, as the structure of octylphenol is quite similar to that of a fatty acid. If this indeed happens, then saturation of the estradiol binding sites will occur at a higher concentration of octylphenol than the model predicts.
We do not know the reason for the underestimation of the free fractions of estradiol in Figure 5. As the model predicts the free fractions of octylphenol quite well (results not shown), the lack of evaporation, cell membrane partitioning, or binding to other proteins by octylphenol in the model does not seem to matter much. Also, in the validation experiment, the estradiol-data (Fig. 2) could be predicted quite nicely, giving the impression that the model also describes well the relevant processes for estradiol. In Figure 3, however, the measured free fraction of [3H]estradiol is different in the two plots, which theoretically should be the same. Additionally, repetition of the experiment leading to Figure 5 resulted in measured free fractions of estradiol that varied per culture plate, sometimes leading to a good prediction of the free fraction of estradiol by the model. Therefore, the cause of the underestimation in Figure 5 must be sought in the measurements.
As there appears to be more sulfated than unconjugated estradiol in stripped serum, the question arises of whether the sulfates might still be subjected to competition at relevant doses of xenoestrogen. Inclusion of estradiol sulfate in the competition model, with an albumin affinity constant of 5 x 105 M–1 (Rosenthal et al., 1975) and a very low SHBG affinity constant [1 x 102 M–1, as SHBG does not appear to be involved in the conjugate binding (Rosenthal et al., 1975)], shows that displacement of the sulfates will not occur at lower competitor concentration than the displacement of unconjugated estradiol (Fig. 7A). However, Figure 7B shows that when concentrations of octylphenol are added that give an increasing cell response, then the free concentration of the sulfates also increases to levels at which estradiol gives an increasing response, This could indicate that the response seen at 1 x 10–6 to 1 x 10–5 M of octylphenol might be due (in part) to released estradiol sulfate. However, as this is the outcome of the competition model that has been shown to predict the start of displacement too early for all our experiments, caution is necessary with these calculation outcomes.
|
In conclusion, charcoal stripping removes so much endogenous estradiol that it is unlikely that competition results in detectable effects. Although the mathematical model predicted otherwise, we have not detected any displacement of estradiol from serum proteins by addition of bisphenol-A, endosulfan, or octylphenol in soluble concentrations. Considering the high affinity of octylphenol for albumin, in comparison with other compounds, it is unlikely that other xenoestrogens will present different results. Therefore, our initial hypothesis that estrogenicity of xenobiotics found in in vitro assays might be due to displacement of endogenous estradiol, must be rejected. The displacement of estradiol sulfate, and also of estrone sulfate (which is present at higher levels (Reed et al., 1986
)), however, remains of great interest for further research. Although the experimental data did not agree well with the outcome of the model, we still believe that the combination of experimental work with modeling is very useful in toxicology. |
APPENDIX A. MODEL DERIVATION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX A. MODEL DERIVATIONREFERENCESAPPENDIX B. COMPETITION MODEL...REFERENCES |
|---|
Assuming a single binding site per protein, the binding reaction of a ligand (L) to a protein (P) can be formulated as
 | (A.1) |
; Rowland and Tozer, 1995).
 | (A.2) |
 | (A.3) |
 |
).
 | (A.4) |
), which corresponds with the Langmuir equation:
 | (A.5) |
 | (A.6) |
Lfree is calculated first by the modeling software, using the lower formula of equation A.6 (for example by the GUESS...ROOTS function in Berkeley Madonna). With the obtained Lfree, the upper formula of equation I.6 can be used to calculate Ka or any other desired parameter.
In general, if there are n different ligands Li, i = 1, ..., n, which bind with binding association constant Ka,ij to m different proteins Pj, j = 1,..., m, the mass balances become:
 | (A.7) |
 | (A.8) |
 |
 |
 |
 |
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX A. MODEL DERIVATIONREFERENCESAPPENDIX B. COMPETITION MODEL...REFERENCES |
|---|
Rang, H. P., Ritter, J. M., and Dale, M. M. (1998). Pharmacology, Churchill Livingstone, New York.
Rowland, M., and Tozer, T. N. (1995). Clinical Pharmacokinetics, Willimas & Wilkins, Media (Philadelphia).
|
APPENDIX B. COMPETITION MODEL SCRIPT |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX A. MODEL DERIVATIONREFERENCESAPPENDIX B. COMPETITION MODEL...REFERENCES |
|---|
;model for competition between estradiol and a competitor (xenoestrogen) on ;serum proteins
METHOD RK4
RENAME TIME = logXtot ; is log of total competitor concentration (M)
RENAME STARTTIME = logXtot0
RENAME STOPTIME = logXtotf
RENAME DTMIN = dmin
RENAME DTMAX = dmax
RENAME DTOUT = dout
logXtot0 = –10.0
logXtotf = 0.0
dmin = 1.e–6
dmax = 1.
dout = 0.01
Ka,AE = 8.9e+4 ; is affinity constant of estradiol for albumin (M-1)
Ka,AX = 8.9e+4 ; is affinity constant of competitor for albumin (M-1), ;default is estradiol
Ka,SE = 1.0e+8 ; is affinity constant of estradiol for SHBG (M-1)
Ka,SX = 1.0e+8 ; is affinity constant of competitor for SHBG (M-1), ;default is estradiol
Atot = 1.8e–5 ; is total albumin concentration (M)
Stot = 3.0e–8 ; is total SHBG concentration (M)
Etot = 7e–9 ; is total estradiol concentration (M)
ffE = Ef/Etot ; is free fraction of estradiol
ffX = Xf/Xtot ; is free fraction of competitor
ffA = Af/Atot ; is free fraction of albumin
ffS = Sf/Stot ; is free fraction of SHBG
fbE = (Etot–Ef)/Etot ; is bound fraction of estradiol
fbX = (Xtot–Xf)/Xtot ; is bound fraction of competitor
Xtot = 10.00**logXtot
Af = (1.0–(Ka,AE*Ef+Ka,AX*Xf)/(1.0+Ka,AE*Ef+Ka,AX*Xf))*Atot ; is free conc. of albumin (M)
Sf = (1.0–(Ka,SE*Ef+Ka,SX*Xf)/(1.0+Ka,SE*Ef+Ka,SX*Xf))*Stot ;is free conc. of SHBG (M)
GUESS Ef = Etot/2. ; is free concentration of estradiol (M)
GUESS Xf = Xtot/2. ; is free concentration of competitor (M)
ROOTS Ef = (1.0+Ka,AE*Atot/(1.0+Ka,AE*Ef+Ka.AX*Xf)+Ka,SE*Stot/(1.0+Ka,SE*Ef+Ka,SX*Xf))*Ef–Etot
ROOTS Xf = (1.0+Ka,AX*Atot(1.0+Ka,AE*Ef+Ka,AX*Xf)+Ka,SX*Stot/(1.0+Ka,SE*Ef+Ka,SX*Xf))*Xf–Xtot
LIMIT Ef> = 0.
LIMIT Ef< = Etot
LIMIT Xf> = 0.
LIMIT Xf< = Xtot
|
ACKNOWLEDGMENTS |
|---|
The authors are grateful to Paul van der Saag, for permission to perform the in vitro assays in his laboratory, and to Richard Schreurs for the practical help with the load of in vitro work. We also appreciate the work of Frans Busser in developing the GC-MS analysis method of octylphenol. Co-author Dr. van der Burg has an affiliation with a company of interest, but his collaboration has not affected the outcome of this work, which was funded by Utrecht University.
|
NOTES |
|---|
1 To whom correspondence should be addressed at IRAS, Yalelaan 2, 3584 CM Utrecht, the Netherlands. Fax: +31-30–2535077. E-mail: j.hermens{at}iras.uu.nl.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONAPPENDIX A. MODEL DERIVATIONREFERENCESAPPENDIX B. COMPETITION MODEL...REFERENCES |
|---|
Abe, T., Komatsu, M., Takeishi, M., and Tsunekane, T. (1976). The alpha-fetoprotein level in the sera of bovine fetuses and calves. Jpn. J. Vet. Sci. 38, 339–345.
Arnold, S. F., Robinson, M. K., Notides, A. C., Guillette, L. J., Jr., and McLachlan, J. A. (1996). A yeast estrogen screen for examining the relative exposure of cells to natural and xenoestrogens. Environ. Health Perspect. 104, 544–548.[Web of Science][Medline]
August, G. P., Tkachuk, M., and Grumbach, M. M. (1969). Plasma testosterone-binding affinity and testosterone in umbilical cord plasma, late pregnancy, prepubertal children, and adults. J. Clin. Endocrinol. Metab. 29, 891–899.
Campbell, P. S., and Clark, J. H. (1984). Effects of chlorpromazine on the inhibition and artifactual elevation of [3H]estradiol binding to estrogen receptors in rat uterine cytosol. J. Steroid Biochem. 21, 475–476.[CrossRef][Web of Science][Medline]
Charlier, C., Albert, A., Herman, P., Hamoir, E., Gaspard, U., Meurisse, M., and Plomteux, G. (2003). Breast cancer and serum organochlorine residues. Occup. Environ. Med. 60, 348–351.
Colborn, T., vom Saal, F. S., and Soto, A. M. (1993). Developmental effects of endocrine-disrupting chemicals in wildlife and humans. Environ. Health Perspect. 101, 378–84.[Web of Science][Medline]
Crain, D. A., Noriega, N., Vonier, P. M., Arnold, S. F., McLachlan, J. A., and Guillette, L. J., Jr. (1998). Cellular bioavailability of natural hormones and environmental contaminants as a function of serum and cytosolic binding factors. Toxicol. Ind. Health 14, 261–73.
Danzo, B. J. (1997). Environmental xenobiotics may disrupt normal endocrine function by interfering with the binding of physiological ligands to steroid receptors and binding proteins. Environ. Health Perspect. 105, 294–301.[Web of Science][Medline]
Déchaud, H., Ravard, C., Claustrat, F., de la Perriere, A. B., and Pugeat, M. (1999). Xenoestrogen interaction with human sex hormone–binding globulin (hSHBG). Steroids 64, 328–34.[CrossRef][Web of Science][Medline]
Dobrila, N.-L. (1996). Albumin Binding Proteins, Identification and Functional Characterization (thesis). Rijksuniversiteit Leiden, Leiden (the Netherlands).
Dunn, J. F., Nisula, B. C., and Rodbard, D. (1981). Transport of steroid hormones: binding of 21 endogenous steroids to both testosterone-binding globulin and corticosteroid-binding globulin in human plasma. J. Clin. Endocrinol. Metab. 53, 58–68.
Englebienne, P. (1984). The serum steroid transport proteins: Biochemistry and clinical significance. In Molecular aspects of medicine (H. Baum, J. Gergely and B. L. Fanburg, eds.), Vol. 7. Pergamon Press, New York.
EORTC, B. c. c. g. (1973). Standards for the assessment of estrogen receptors in human breast cancer. Eur. J. Cancer 9, 379–381.
Fairclough, R. J., Hunter, J. T., and Welch, R. A. (1976). Preparturient changes in plasma concentrations of estrone, estradiol-17beta and estradiol-17alpha in the fetal calf: Effects of dexamethasone infusion. Steroids 28, 881–888.[CrossRef][Web of Science][Medline]
Heringa, M. B., Pastor, D., Algra, J., Vaes, W. H. J., and Hermens, J. L. M. (2002). Negligible depletion solid-phase microextraction with radiolabeled analytes to study free concentrations and protein binding: an example with [3H]estradiol. Anal. Chem. 74, 5993–5997.[Medline]
Ikezuki, Y., Tsutsumi, O., Takai, Y., Kamei, Y., and Taketani, Y. (2002). Determination of bisphenol A concentrations in human biological fluids reveals significant early prenatal exposure. Hum. Reprod. 17, 2839–2841.
Ishihara, A., Sawatsubashi, S., and Yamauchi, K. (2003). Endocrine disrupting chemicals: interference of thyroid hormone binding to transthyretins and to thyroid hormone receptors. Mol. Cell. Endocrinol. 199, 105–117.[CrossRef][Web of Science][Medline]
Jobling, S., and Sumpter, J. P. (1993). Detergent components in sewage effluent are weakly oestrogenic to fish: An in vitro study using rainbow trout (Oncorhynchus mykiss) hepatocytes. Aquat. Toxicol. 27, 361–372.[CrossRef]
Krstevska-Konstantinova, M., Charlier, C., Craen, M., Du Caju, M., Heinrichs, C., de Beaufort, C., Plomteux, G., and Bourguignon, J. P. (2001). Sexual precocity after immigration from developing countries to Belgium: Evidence of previous exposure to organochlorine pesticides. Hum. Reprod. 16, 1020–1026.
Kuiper, G. G., Carlsson, B., Grandien, K., Enmark, E., Haggblad, J., Nilsson, S., and Gustafsson, J. A. (1997). Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors alpha and beta. Endocrinology 138, 863–870.
Kuiper, G. G., Lemmen, J. G., Carlsson, B., Corton, J. C., Safe, S. H., van der Saag, P. T., van der Burg, B., and Gustafsson, J. A. (1998). Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor beta. Endocrinology 139, 4252–4263.
Laidley, C. W., and Thomas, P. (1994). Partial characterization of a sex-steroid binding-protein in the spotted sea-trout (Cynoscion nebulosus). Biol. Reprod. 51, 982–992.[Abstract]
Lemmen, J. G., van den Brink, C. E., Legler, J., van der Saag, P. T., and van der Burg, B. (2002). Detection of oestrogenic activity of steroids present during mammalian gestation using oestrogen receptor alpha–and oestrogen receptor beta–specific in vitro assays. J. Endocrinol. 174, 435–446.[Abstract]
Meulenberg, E. P. (2002). A new test to identify endocrine disruptors using sex hormone-binding globulins from human serum. Eur. J. Lipid Sci. Technol. 104, 131–136.[CrossRef]
Milligan, S. R., Khan, O., and Nash, M. (1998). Competitive binding of xenobiotic oestrogens to rat alpha-fetoprotein and to sex steroid binding proteins in human and rainbow trout (Oncorhynchus mykiss) plasma. Gen. Comp. Endocrinol. 112, 89–95.[CrossRef][Web of Science][Medline]
Mol, H. G. J., Sunarto, S., and Steijger, O. M. (2000). Determination of endocrine disruptors in water after derivatization with N-methyl-N-(tert.-butyldimethyltrifluoroacetamide) using gas chromatography with mass spectrometric detection. J. Chromatogr. A 879, 97–112.[CrossRef][Web of Science][Medline]
Nagel, S. C., vom Saal, F. S., Thayer, K. A., Dhar, M. G., Boechler, M., and Welshons, W. V. (1997). Relative binding affinity-serum modified access (RBA-SMA) assay predicts the relative in vivo bioactivity of the xenoestrogens bisphenol A and octylphenol. Environ. Health Perspect. 105, 70–76.[Web of Science][Medline]
Peck, E. J., Jr., and Clark, J. H. (1977). Effect of ionic strength on charcoal adsorption assays of receptor-estradiol complexes. Endocrinology 101, 1034–43.
Peterson, A. J., Hunter, J. T., Welch, R. A. S., and Fairclough, R. J. (1975). Oestrogens in bovine fetal and maternal plasma near term. J. Reprod. Fertil. 43, 179–181.
Pettersson, K. S., Vanharanta, R. M., and Soderholm, J. R. (1985). Pitfalls in the dextran-coated charcoal assay of estrogen receptors in breast cancer tissue. J. Steroid Biochem. 22, 39–45.[CrossRef][Web of Science][Medline]
Reed, M. J., Cheng, R. W., Beranek, P. A., Few, J. D., Franks, S., Ghilchik, M. W., and James, V. H. (1986). The regulation of the biologically available fractions of oestradiol and testosterone in plasma. J. Steroid Biochem. 24, 317–20.[CrossRef][Web of Science][Medline]
Rosenthal, H. E., Ludwig, G. A., Pietrzak, E., and Sandberg, A. A. (1975). Binding of the sulfates of estradiol-17 beta to human serum albumin and plasma. J. Clin. Endocrinol. Metab. 41, 1144–54.
Saeki, T., Cristiano, A., Lynch, M. J., Brattain, M., Kim, N., Normanno, N., Kenney, N., Ciardiello, F., and Salomon, D. S. (1991). Regulation by estrogen through the 5'-flanking region of the transforming growth factor alpha gene. Mol. Endocrinol. 5, 1955–1963.
Siiteri, P. K., Murai, J. T., Hammond, G. L., Nisker, J. A., Raymoure, W. J., and Kuhn, R. W. (1982). The serum transport of steroid hormones. Recent Prog. Horm. Res. 38, 457–510.[Web of Science][Medline]
Skovgaard Poulsen, H. (1981). Oestrogen receptor assay-limitation of the method. Eur. J. Cancer 17, 495–501.
Skovgaard Poulsen, H. (1982). Estrogen receptors in human breast cancer: comparative features of the hydroxylapatite- and dextran-coated charcoal assay. Eur. J. Cancer Clin. Oncol. 18, 1075–1079.[CrossRef][Web of Science][Medline]
Södergard, R., Bäckström, T., Shanbhag, V., and Carstensen, H. (1982). Calculation of free and bound fractions of testosterone and estradiol-17ß to human plasma proteins at body temperature. J. Steroid Biochem. 16, 801–810.[CrossRef][Web of Science][Medline]
Soto, A. M., Sonnenschein, C., Chung, K. L., Fernandez, M. F., Olea, N., and Serrano, F. O. (1995). The E-SCREEN assay as a tool to identify estrogens: An update on estrogenic environmental pollutants. Environ. Health Perspect. 103(Suppl 7), 113–122.
Thorpe, S. M. (1987). Steroid receptors in breast cancer: Sources of inter-laboratory variation in dextran-charcoal assays. Breast Cancer Res. Treat. 9, 175–189.[CrossRef][Web of Science][Medline]
Tollefsen, K. E. (2002). Interaction of estrogen mimics, singly and in combination, with plasma sex steroid-binding proteins in rainbow trout (Oncorhynchus mykiss). Aquat. Toxicol. 56, 215–225.[CrossRef][Web of Science][Medline]
Vaes, W. H. J., Urrestarazu Ramos, E., Verhaar, H. J. M., Seinen, W., and Hermens, J. L. M. (1996). Measurement of the free concentration using solid-phase microextraction: binding to protein. Anal. Chem. 68, 4463–4467.[CrossRef]
Vesanen, M. (1993). Bovine uterine, cervical and ovarian cytosol estrogen and progesterone receptor concentrations in cystic ovarian disease. Acta Vet. Scand. 34, 35–43.[Web of Science][Medline]
Vignon, F., Terqui, M., Westley, B., Derocq, D., and Rochefort, H. (1980). Effects of plasma estrogen sulfates in mammary cancer cells. Endocrinology 106, 1079–1086.
Villeneuve, D. L., Kurunthachalam, K., Priest, B. J., and Giesy, J. P. (2002). In vitro assessment of potential mechanism-specific effects of polybrominated diphenyl ethers. Environ. Toxicol. Chem. 21, 2431–2433.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  C. Ankarberg-Lindgren and E. Norjavaara A purification step prior to commercial sensitive immunoassay is necessary to achieve clinical usefulness when quantifying serum 17 -estradiol in prepubertal children Eur. J. Endocrinol., January 1, 2008; 158(1): 117 - 124. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||