* Biological Monitoring and Modeling, Pacific Northwest National Laboratory, 902 Battelle Boulevard, P7-56, Richland, Washington 99352; Toxicology & Environmental Research and Consulting, The Dow Chemical Company, Building 1803, Midland, Michigan 48674;
ENVIRON International, 602 E Georgia Ave., Ruston, Louisiana 71270;
U.S. EPA, Office of Research and Development, National Health Effects Research Laboratory, Experimental Toxicology Division, Pharmacokinetics Branch, U.S.E.P.A Mail Room, B14301, Research Triangle Park, North Carolina 27711
1 To whom correspondence should be addressed at Biological Monitoring and Modeling, Pacific Northwest National Laboratory, 902 Battelle Boulevard, P7-56, Richland, WA 99352. Fax: 5093769446. E-mail: justin.teeguarden{at}pnl.gov.
Received October 28, 2004; accepted February 22, 2005
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key Words: bisphenol A; PBPK model; endocrine; glucuronide; human; metabolism; pharmacokinetics; physiologically based pharmacokinetics; plasma protein binding; risk assessment.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
After ingestion of low doses in either rats or humans, the phase II metabolism of BPA to form the glucuronic acid conjugate BPA-monoglucuronide (BPAG) appears to be nearly complete as a result of "first pass" metabolism by uridine diphosphate-glucuronosyl transferase (UGT) in the intestine and liver (Inoue et al., 2003; Pottenger et al., 2000
; Upmeier et al., 2000
; Völkel et al., 2002
).
BPAG in rodents is eliminated principally in the bile (Inoue et al., 2001) and subsequently in the feces as BPA (Pottenger et al., 2000
), whereas in humans BPAG is eliminated principally in the urine (Völkel et al., 2002
). In rats, BPA eliminated via the bile may undergo enterohepatic recirculation, prolonging systemic exposure. Bisphenol A and BPAG pharmacokinetics have been characterized in several strains and sexes of rats, and in male and female humans, but they have been subject only to characterization by classical compartmental and non-compartmental analyses (Pottenger et al., 2000
; Sun et al., 2002
; Upmeier et al., 2000
; Völkel et al., 2002
; Yoo et al., 2000
).
The predominant concern in the safety assessment of this compound arose from studies that reported weak estrogenic effects of this compound in vivo in rodents (European-Commission, 2003; Pottenger et al., 2000
). The estrogenicity of BPA is attributed to its weak in vitro agonist activity, on the order of 1/10,000 of that of estrogen, and differences in the binding to ER
and ERß (Matthews et al., 2001
). The role of the estrogen receptor in the weak in vivo effects implies, but does not establish, thresholds for tissue effects and specificity of effects seen in tissues, based on the composition of receptors within various tissues, at least in adults (Andersen and Barton, 1999
). BPAG, the principal BPA metabolite, is not estrogenic (Matthews et al., 2001
; Snyder et al., 2000
; Twomey, 1998
). Uterine weight gain, an estrogenic effect attributable to hypertrophy, hyperplasia, and/or the imbibing of water, or some combination of these, has been observed in female rats administered high doses (800 mg/kg) of BPA (Twomey, 1998
; Yamasaki et al., 2000
).
Bisphenol A binds to plasma proteins in rodents, monkeys, and humans. The bound form represents 9095% and the free form
510% of the total (Csanady et al., 2002
; Kurebayashi et al., 2002
; Mayersohn, 2003
). Humans and monkeys appear to have modestly lower free fractions (5%) (Csanady et al., 2002
) than rats (
510%) (Mayersohn, 2003
). The impact of this protein binding on uptake into potential target tissues, especially those with low metabolic capacity where metabolism is not expected to affect uptake (e.g., uterus, brain), should be considered an important part of the pharmacokinetic (PK) characterizations of BPA, because binding may restrict access to the tissue and the target receptor population within the tissue (Mendel, 1992
).
The pharmacokinetics of BPA are complicated by interspecies differences in the biliary elimination of BPAG, the involvement of the gastrointestinal (GI) tract in glucuronidation, and enterohepatic recirculation. Although these processes may be expected to affect BPA pharmacokinetics, the extent to which the pharmacokinetics of BPAG influences those of BPA has not been evaluated. In addition to describing the bulk kinetics of BPA and BPAG, doseresponse models of weak estrogen receptor (ER) agonists should incorporate a description of the relationship between blood concentrations (free or total) and tissue ER binding (Plowchalk and Teeguarden, 2002). Such analyses are not easily approached with standard compartmental and non-compartmental models, but they may be more effectively addressed with physiologically based pharmacokinetic (PBPK) models. Physiologically based pharmacokinetic models describe the relationship between the "external" dose administered and the internal or target tissue concentrations and thus improve predictions of the doseresponse relationship. Several PBPK models have been used successfully in this capacity (Andersen et al., 1987
; Clewell et al., 1997
) because of their flexibility in incorporating information about the physicochemical characteristics of a chemical, as well as the physiological and biochemical processes (e.g., absorption, distribution, metabolism, elimination, protein and receptor binding) that control target tissue concentrations (Andersen et al., 1987
; Plowchalk and Teeguarden, 2002
). Such models can also leverage the available information about other key events (such as protein and receptor binding) that are related to either pharmacologic effects or adverse effects.
The principal objective of this work was to develop an oral-route pharmacokinetic model for BPA in rats and humans that (1) incorporated restrictions on the concentrations of unbound (free) BPA in the plasma that result from plasma protein binding and (2) predicts the degree of ER binding that may occur in the uterus. The model incorporates information on BPA pharmacokinetics after intravenous (i.v.) dosing and explores the reasons for route-specific differences in BPA kinetics. We also sought to assess the ability of a simple BPAG sub-model to describe BPAG pharmacokinetics and evaluate the extent to which BPAG pharmacokinetics may influence BPA pharmacokinetics.
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Pottenger and co-workers reported the time course of BPA in whole blood and the total radioactivity in plasma in adult male and female Fischer 344 (F344) rats after oral gavage dosing of 10 or 100 mg/kg of radiolabeled 14C-BPA. Male rats did not have detectable levels of BPA in blood after the lower dose, and peak concentrations in the females were only 24 times the limit of detection (0.04 µM) (Pottenger et al., 2000
). The data on the concentrations of BPA in the blood after the 100 mg/kg dose were the most complete and were therefore used for development of the model. Gender differences in the pharmacokinetics of BPA after oral administration were observed in the study by Pottenger and co-workers. The percentage of dose excreted in the urine and feces as a function of time was not published, but was obtained from the original study report prepared for good laboratory practice (GLP) purposes. Bisphenol A concentrations were measured in samples of whole blood, and radioactivity derived from the radiolabeled test material was measured in the blood plasma, necessitating the following additional calculations to arrive at blood BPAG concentrations: (1) the concentration of BPA in the blood was divided by 0.83 (rat blood:plasma partition coefficient (Mayersohn, 2003
) to calculate the BPA concentration in the plasma and converted to nmoles per gram of plasma (assuming a density equal to water); (2) total plasma radioactivity reported in µg-BPA-equivalents/g plasma was converted to nmol BPA-equivalents/g plasma; (3) the calculated nmol/g BPA in plasma (step 1) was subtracted from the nmol BPA-equivalents/g plasma (step 2) to estimate the nmol/g plasma of metabolites; (4) metabolites were assumed to be entirely BPAG because the metabolite characterization of the plasma at selected times demonstrated that nearly all of the radioactivity in the plasma after oral dosing was BPAG, with only minor contributions from other metabolites. The glucuronide represented >90% of total metabolites for all doses, sexes, and times, except the 100 mg/kg males 15 min after exposure, where the glucuronide was
70% of the total; (5) BPAG was assumed not to distribute to red blood cells, so the blood concentrations of BPAG were estimated as the plasma concentration x (1.0-hematocrit). The rat hematocrit was assumed to be 0.45 (Waynford and Flecknell 1992
).
Upmeier reported the plasma pharmacokinetics of BPA after oral (10 or 100 mg/kg) and i.v. (10 mg/kg) administration of BPA in female DA/Han rats (Upmeier et al., 2000). The authors refer to the unconjugated form of BPA measured in plasma as "free" BPA. This naming convention should not be confused with the one used here, and elsewhere, in which "free BPA" refers to the fraction of unconjugated BPA not bound to plasma proteins or tissue receptors, and the glucuronidated form is referred to directly (i.e., BPAG). The oral data from the Upmeier et al. study were not useful for confirming/validating parameterization of the model made using the data from Pottenger et al. (2000)
because of extremely high variability in the measured plasma BPA concentrations (standard deviations approximately equal to mean values). The authors note the variability and present individual animal plasma BPA concentrations 10 min after oral dosing that range from
5 to
90 ng/ml, and that rapidly fall or rapidly rise afterward, presenting no consistent pattern. The "composite design" used for this study, where samples were taken from each rat at several times, but not at all time points, may have contributed to the variability.
The pharmacokinetics of BPA and BPAG in humans after a low oral dose of BPA was recently described by Völkel and co-workers (2002). Three groups consisting of either three adult females, three adult males, or four adult males were administered 5 mg of d16-BPA in a hard gelatin capsule, and the concentrations of BPA and BPAG were determined by gas chromatography/mass spectometry (GC/MS) and liquid chromatography/mass spectrometry (LC-MS)/MS in plasma and urine up to 32 h after administration. The doses correspond to 5490 µg/kg body weight. To avoid confounding measured plasma BPA with observable background contamination of BPA, d16-BPA was used to characterize the pharmacokinetics of BPA. BPAG was the only compound that could be detected in the blood of the study volunteers. Bisphenol A was below the limit of detection in both urine (6 nM) and plasma (9 nM). There were no observable differences in the blood concentrations of BPAG in males and females, suggesting that there were no significant differences between males and females in BPA metabolism, although the sample size is small. Völkel reports a clearance rate of 0.13 l/min, which approximates renal clearance rates of creatinine in humans.
Plasma protein binding.
Blood was collected from adult (7683 d) male and female F344 rats (Harlan Sprague Dawley) via the hepatic vein and mixed with heparin (150 U/10 ml blood) (Mayersohn, 2003) under ether anesthesia. Plasma was obtained by centrifugation of blood pooled from 28 animals at 3000 x g for 10 min. The pH was adjusted to 7.4 with the addition of H3PO4. Bisphenol A was added to plasma at final concentrations of 1.33, 11.26, 26.16, 50.99, 75.82, 100.65, 150.30, and 199.96 µg/ml, each concentration containing a constant amount of radioactive BPA. After 15 min incubation at 37°C, two aliquots were removed and radioactivity was quantified (Beckman, liquid scintillation system) before (plasma) and after (filtrate) ultrafiltration centrifugation (Amicon MW cutoff 30,000, 2000 x g, 15 min). Nonspecific binding in the test system was characterized and found not to interfere with the accuracy of the binding assay. The amount of free BPA was calculated as 100x (filtrate counts/plasma counts) (see Table 1 in the Supplementary Material online).
Standard Scatchard analysis was used to estimate the BPA equilibrium binding association constant (Ka, and Kd [as 1/Ka]) and maximum binding (Bmax) for binding to rat plasma proteins. The ratio bound BPA/free BPA was plotted as a function of bound BPA (mM) (Fig. 1). Equation (1) describes the relationship of the plotted data:
![]() | (1) |
|
Studies of uterine responses to estrogens or xenoestrogens typically involve one or more doses per day for 3 days (Ashby and Tinwell, 1998; Barton, 1999
; O'Connor et al., 1996
; Odum et al., 1997
). The uterine wet weight increase after three daily doses of estradiol has been shown to be much greater than after a single dose, a finding that has led to adoption of this dosing regimen in uterotrophic assay protocols. However, the relationship between receptor occupancy and the uterine weight increase in a 3-day dosing regimen is not clear. The time course events leading to uterine weight gain can be generalized as occurring in two phases: early phase biosynthetic events occurring 06 h after exposure and later events culminating in increases in uterine weight gain 624 h after exposure. The available data for increases in wet and dry weight after exposure to BPA all use multi-day dosing regimens, so two dose metrics for receptor occupancy were estimated for comparison with the response data. These dose metrics are the percent receptor occupancy at 6 h after dosing (POCCU1) and the 024 h area under the curve (AUC) for the occupied ER in the uterus (AUCUB1). The model does not include dimerization of the serum estradiol ER complex or binding of the agonist-occupied ER dimer complex with the nuclear DNA, so neither of these dose metrics attempts to relate the degree of nuclear receptor occupancy to the uterotrophic response following a single dose of BPA.
Estrogen receptor occupancy dose metricsPOCCU1 and AUCUB1after oral administration of BPA were compared to the increases in uterine wet weights reported by Twomey (1998). Two comparisons were made, one that included the effect of plasma protein binding restrictions on free BPA and one that did not. In the Twomey study, BPA was administered for 3 consecutive days at dose rates of 10, 100, 200, and 800 mg/kg/day, after which uterine weight was assayed.
PBPK model structure.
The model was initially developed for adult male and female rats and was later extended to humans. A nested model structure is used, with sub-models for BPA and BPAG comprising the overall model (Fig. 2). The BPA model consists of compartments representing blood, uterus, liver, the lumen of the GI tract, and a body compartment representing the remaining perfused tissues. In this model, BPA distributes to and from a non-metabolizing body compartment. The other tissues are formulated the same as the body compartment, but they have additional terms describing processes that affect tissue concentrations of BPA (metabolism, protein or ER binding, uptake). Tissue compartments are formulated as perfusion limited (well mixed, rapid equilibrium), as described elsewhere (Andersen, 1981). Tissues volumes, flows, and rates of uptake, metabolism, and elimination are scaled allometrically to adjust for differing body weights of the rats used in the PK studies. The body compartment is described by the following equations, where "1" refers to BPA, "2" to BPAG, "A" to amount (i.e., dA/dt), and "free" to BPA not bound to proteins or receptors:
![]() | (2) |
![]() | (3) |
![]() | (4) |
![]() | (5) |
![]() | (6) |
![]() | (7) |
|
The blood binding and uterine binding submodels are written in molar rather than milligram units (the remainder of the model) for consistency with the way binding constants are typically measured and reported. Unit conversions are placed directly in the equations shown below, although they are handled separately in the code. Binding to the uterine estrogen receptor is described as simple equilibrium binding of a single ligand to a single site on the receptor (equation 8):
![]() | (8) |
![]() | (9) |
![]() | (10) |
![]() | (11) |
The approach to partitioning (PU/FF) used for the uterine compartment is different from that used for other tissues in this model. The difference arises both from the method used to measure the partition coefficients (Csanady et al., 2002) and the need to model uterine tissue receptor binding as a function of free rather than total BPA. Csanady and co-workers measured blood:tissue partition coefficients for several tissues using blood directly rather than the standard blood:saline, saline:tissue protocol. The resulting partition coefficients account for the net affect of processes controlling apparent partitioning (e.g., binding to plasma proteins). These partition coefficients can be applied directly in the model for the tissues Csanady and co-workers measured if they are used in a fashion consistent with the way they were measured; that is, if they are applied to total blood BPA rather than free blood BPA.
The situation is different for the uterus for two reasons. First, in the absence of better data, we chose to use the partition coefficient reported by Csanady et al. for muscle tissue. As measured, this partition coefficient would reflect partitioning to a muscle-like tissue and the effect of binding to blood proteins, but not receptor binding in a tissue like the uterus. To properly apply the muscle partition coefficient to the uterus, we must separate the receptor binding and partitioning processes in the model. Second, ER binding of BPA in the tissue is expected to be a function of the free material in the blood, not the total. To accommodate these needs, we calculate free concentration of BPA in the blood and allow the tissue and receptor to equilibrate to the free concentration rather than total blood BPA. Because these equations are written based on free BPA concentrations, the muscle partition coefficient reported by Csanady et al. was adjusted by dividing by the free fraction of BPA. The adjustment is made in the model because the free fraction of BPA is dynamic, changing if plasma binding is saturated or additional compounds that compete for binding are added.
Free and total BPA are calculated in the plasma compartment according to equations 12 (total amount of BPA in blood), 13 (total concentration in blood), 14 (concentration in plasma), 15 (amount in plasma), 16 (free concentration in plasma), and 17 (free concentration of the plasma binding protein in µM) and 18 (free concentration in blood):
![]() | (12) |
![]() | (13) |
![]() | (14) |
![]() | (15) |
![]() | (16) |
![]() | (17) |
![]() | (18) |
Glucuronidation is the major pathway for metabolic elimination of BPA and the chemical species formed, BPAG, is inactive for receptor binding and estrogenic activity (Matthews et al., 2001; Snyder et al., 2000
). Therefore, the presence of BPAG in the body was described with a volume of distribution, and tissues were not explicitly described. The BPAG produced in the liver is passed to the GI tract lumen via the bile or directly to the volume of distribution in the BPAG sub-model. The fraction of BPAG passed to these two compartments is controlled by the term BILE, which is adjusted for species differences in the elimination of BPAG. In humans, BPAG is eliminated almost entirely by renal excretion, although some biliary elimination followed by complete reuptake cannot be ruled out. We represent human elimination of BPAG as occurring exclusively from the blood via the kidneys, whereas in the rat, BPAG elimination is principally via the bile.
The rate of change of BPAG in the body is determined by its rate of formation and elimination in urine (equation 19). The concentration of BPAG in blood was estimated with the assumption that the concentration in the volume of distribution was equivalent to the plasma concentration and no BPAG entered the red blood cells (equation 20). Therefore, the blood concentration was the product of the concentration in the volume of distribution and the fraction of the blood comprised of plasma (i.e., 1-hematocrit [HCT]):
![]() | (19) |
![]() | (20) |
|
![]() | (21) |
Model Parameterization
Physiological parameters.
The physiological parameters used in the rat and human models are summarized in Table 1. Body weights of rats were study specific (Pottenger et al., 2000). Tissue volumes and blood flows were obtained from Brown et al. (1997)
, except for the uterine compartment. Uterine tissue volumes were reported elsewhere for the rat (Bruce, 1976
) and human (Langlois, 1970
).
Tissue distribution parameters.
Csanady and co-workers measured the human tissue:blood partition coefficients of bisphenol A and the rat and human tissue:blood partition coefficients for the phytoestrogen daidzein using the vial equilibration technique (Csanady et al., 2002). The reported BPA partition coefficients in humans were 1.35 for muscle and kidney, 1.46 for liver, and 3.31 for fat. A value of 1.35 was assumed for the bulk tissue compartments for the human, and the liver value was used directly. Based on the finding that average tissue:blood partition coefficients for daidzein were not statistically significantly different between rats and humans, the human BPA partition coefficients were assumed to be a good proxy for rodent values, and were thus used in the rat model (Table 1). The blood:plasma partition coefficient has been measured in adult and 21-day-old male and female, as well as pregnant F344 rats and adult male and female Sprague-Dawley rats. The blood:plasma partition coefficient had a range of 0.720.94, with a mean value of 0.83 and a standard error of the mean (S.E.M.) of 0.015 (Mayersohn, 2003
). This value was used directly in the model.
A volume of distribution for BPAG has not been reported for rats. Völkel estimated a volume of distribution at steady state (VDss) for this glucuronide of 0.43 l/kg (Völkel et al., 2002) using a noncompartmental approach corresponding to a three-compartment model. The VDss is an overestimation of the VD0 (volume of distribution at Tzero) for models of greater than one compartment (O'Flaherty, 1981
). This was verified here through simulation of the Völkel data, and a much smaller VD of 0.17 l/kg was assumed for humans (Table 1). In the absence of other data, this value was adopted for rats as well (Table 1). The value compares well to the measured VD for morphine-6-glucuronide, which was reported to be between 0.28 and 0.43 l/kg in adult humans (Barrett et al., 1996
).
Metabolism, excretion, and enterohepatic recirculation.
The values of parameters for metabolism, excretion, and enterohepatic recirculation in rats were estimated by fitting the oral gavage data of Pottenger et al. (2000). The Km for hepatic glucuronidation was assumed to be 10 µM, the middle of the range of values reported for rats and humans after measurement in microsomal preparations and isolated hepatocytes (Kuester and Sipes, 2003
). The reported sex differences in pharmacokinetics were accommodated by estimating parameter values separately for male and female rats (Table 1). The close interrelationship between absorption, metabolism, and enterohepatic recirculation prevented determination of unique values of the corresponding model parameters by either visual or formal optimization (ACSL Optimize, AEGIS technologies). However, different aspects of the blood BPA or BPAG time course varied in their sensitivity to these model parameters. For example, rates of enterohepatic recirculation do not affect the BPA Tmax or Cmax, but do influence later blood concentrations. The rates of metabolism and uptake both influence BPA Tmax and Cmax, but the rate of uptake has a stronger influence on BPA Tmax. These differences in sensitivity allow simultaneous estimates of these parameters to be made with the given data sets.
Simulating i.v. route pharmacokinetics in rats required the use of different rates of hepatic glucuronidation, enterohepatic recirculation, and fecal elimination. Table 2 summarizes parameters that are different between the oral route and i.v. route simulations. This discrepancy is addressed further in the Discussion.
|
Limited data are available for estimating the affinity constant (KD1) of BPA for the rat estrogen receptor. Bisphenol A appears to have a greater affinity for ERß than for ER (Kuiper et al., 1997
, 1998
; Matthews et al., 2001
). Estimates of its relative affinity for ER
compared to E2 range from 0.006% to 0.05% (Kuiper et al., 1998
; Matthews et al., 2001
; Nagel et al., 1998
). A KD of 400 nM (representing a relative affinity of 0.05% compared to the KD of 0.2 nM for E2) was used here (Teeguarden and Barton, 2004
). The KD used for E2:ER binding was 0.2 nM (Nagel et al., 1998
; Zava and Duwe, 1997
). This value is plausible based on a review of published values, which are highly variable (Plowchalk and Teeguarden, 2002
).
Modeling Approach
The rat PBPK model for BPA and BPAG was developed to simulate blood and uterine concentrations after exposure to BPA by routes relevant to human exposure (oral) and experimental characterization of BPA pharmacokinetics (i.v). Bisphenol A metabolism was attributed solely to the liver, although the ability of intestinal tissues to glucuronidate BPA has recently been demonstrated (Inoue et al. 2003). Because BPA administered by the i.v. route should be subject to predominantly hepatic rather than intestinal metabolism, i.v. BPA kinetic data (Upmeier et al., 2000
) were used to estimate the hepatic metabolism of BPA, as well as the BPA volume of distribution. The oral route kinetic data were more complete, containing blood BPA and BPAG concentration time course data and cumulative elimination of total radioactivity in the feces and urine of male and female rats; these data were used for fitting the remaining parameters. Rates of uptake, fecal excretion and enterohepatic recirculation as well as urinary clearance of the glucuronide were fitted to the available oral (Pottenger et al., 2000
) rat pharmacokinetic data. Simulation of these data required the estimation of oral-routespecific rates of glucuronidation, which were still attributed to the liver.
The model was extended to humans by means of physiological parameters obtained from the literature and fitting of the remaining parameters to oral-route pharmacokinetic data obtained from male and female volunteers (Völkel et al., 2002).
Plasma protein binding of BPA was incorporated into both the rat model and the human model, and simulations of free and protein bound BPA were compared to published data (Csanady et al., 2002; Mayersohn, 2003
). The rat model was used to simulate ER binding in the uterus after oral exposure to BPA, and various dose metrics were tested for correlations with uterine weight gain observed in separate oral-route experiments (Twomey, 1998
).
Sensitivity Analysis
The numerical values of measured and estimated model parameters are not known with absolute certainty. An evaluation of the impact of uncertainty in the parameters on model estimates of blood BPA and BPAG concentrations was performed by conducting a sensitivity analysis. The analysis was carried out by measuring the change in model output for a 1% change in a particular model parameter when all the other parameters were held fixed. A normalized sensitivity coefficient of 1 indicates that there is a one-to-one relationship between the fractional change in the parameter and model output. A positive value for the normalized sensitivity coefficient indicates that the output and the corresponding model parameter are directly related and a negative value indicates they are inversely related. Time series sensitivity coefficients are reported to convey the time dependence of the various model parameters on BPA and BPAG kinetics.
The rat sensitivity analysis was conducted under the conditions of the female oral 100 mg/kg oral dose study, and the human sensitivity analysis was conducted under conditions of the human oral study (5 mg/person, females). These doses were selected to allow evaluation of model behavior in the range of the experimental range. Normalized sensitivity coefficients were reported as a function of time and presented here for all parameters with sensitivity coefficients greater than 0.1.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
|
|
|
|
|
Human Pharmacokinetics
After oral administration of 5 mg BPA, the human model accurately simulated plasma BPAG concentrations for most of the time course (<20 h) in both males and females, In general, however, it under-predicted BPAG at the 2448 h post-exposure interval (Fig. 7). Adjusting the rate of urinary elimination affected simulated concentrations at all but the earliest time points and did not improve overall fits to these data. The model accurately predicted the cumulative urinary elimination of BPAG in both males and females (Fig. 8).
|
|
Sensitivity Analysis
With few exceptions, the rat and human models showed significant sensitivity (>0.1) to the same parameters. Sensitivity coefficients for the rat are presented here, and important differences with the human model are noted. The modeled BPA concentrations showed sensitivity to parameters that can be grouped into three categories: those that control uptake, distribution, and elimination. Sensitivities varied significantly with time (Fig. 9). Parameters controlling uptake and metabolismthe rate of uptake, metabolic constants, liver blood flow, cardiac output, and the body:blood partition coefficienthad a strong impact on BPA concentrations at early time points (1 h or less). Bisphenol A concentrations at later time points were sensitive to parameters controlling elimination, including Vmax and enterohepatic recirculation (Fig. 9). The human model was not sensitive to the rate of fecal elimination because biliary elimination of BPAG and enterohepatic recirculation are not as quantitatively important as they are in the rat.
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Bisphenol A is glucuronidated in both the liver and intestinal tissues (Inoue et al., 2003). First pass metabolism of BPA therefore involves both intestinal and hepatic tissues, and oral route pharmacokinetic (PK) data cannot be used to determine the individual contribution of each of these tissues to metabolic clearance. However, i.v. route PK data can be used to determine the contribution of the liver to metabolic clearance of BPA. Several i.v. route PK data sets were available, each from a different laboratory (Sun et al., 2002
; Upmeier et al., 2000
; Yoo et al., 2000
), and in different strains of rats. These PK data are not consistent with one another, and some systematic review was required to determine which data were most appropriate for use in developing the PBPK model. The discrepancies in these data take the form of large differences in apparent volumes of distribution (VD). Yoo and co-workers administered BPA at four dose levels to male Sprague-Dawley rats, 0.2, 0.5, 1, and 2 mg/kg, and reported serum BPA concentrations over approximately 2 h (Yoo et al., 2000
). The non-compartmental analyses of these data give an average VDss of 5.2 l/kg, greater than 5 times body weight. The VD0 calculated as Dose/C0 is
2.5 l/kg. Sun et al. administered BPA to male Wistar rats in doses of 10 or 20 mg/kg and reported BPA concentrations in plasma (Sun et al., 2002
). They did not report PK characteristics for BPA, but, we determined that these data have the same characteristics as those reported by Yoo et al., a VD0 of
2.5 l/kg. In contrast, Upmeier et al. reported plasma concentrations
4 times higher than those reported by Sun and colleagues for the same 10 mg/kg dose, with a correspondingly lower VD0 (0.60 l/kg). The difference in VD, almost fourfold, is unexpected and difficult to attribute to physiological differences in the various strains of rats. Csanady et al. (2002)
found essentially no differences in the volumes of distribution estimated for rats and humans from measured tissue partition coefficients and tissue volumes: the rat value was 1.4 l/kg and the human value was 1.5 l/kg. This approach would approximate a VDss; the VD0 would be expected to be lower. The ratio of VD0 to VDss reported by Yoo et al. is
0.5. This suggests that the VD0 based on the Csanady et al. (2002)
approach would be near 0.7 (1.4 l/kg*0.5), very close to the 0.6 l/kg observed by Upmeier and co-workers (2000)
. Finally, we found it impossible to simulate the Yoo et al. and Sun et al. i.v. data sets with the PBPK model, even using the much higher partition coefficients and clearances reported by these authors (data not shown), whereas the Upmeier et al. data were simulated well using the characteristics of partitioning reported by Csanady et al. (2002)
, which are also successfully used here to model the oral-route data sets. It is not clear to what the differences in these data sets should be attributed, only that they exist and are most likely the result of experimental differences. None of the authors report recovery of BPA from whole blood, nor do they evaluate the stability of BPAG in their samples during processing. The consistency between the tissue distribution of BPA implied by the Upmeier et al. i.v. data, the Pottenger et al. oral-route data, and those measured by Csanady et al. (2002)
led us to rely on the Upmeier et al. data for parameterization of the PBPK model and estimation of the hepatic contribution to total metabolic clearance.
Rates of BPA glucuronidation estimated from fitting the i.v. data were 420 times lower than the corresponding rates fitted to the oral data. Because strain and sex differences in oral-route and i.v. route studies confound direct comparison of these rates, it is difficult to reach conclusions as to the exact magnitude of the differences. However, the intrinsic metabolic clearance rates reported by Yoo et al., 100 ml/min/kg, are the same as those used to simulate the Upmeier et al. data (95 ml/min/kg), which are in female rather than male rats of a different strain. This consistency supports our estimation of these rates and suggests that strain differences alone do not account for the differences in BPA glucuronidation rates observed in the oral and i.v. studies. Instead, the difference can probably be attributed to glucuronidation of BPA in the GI tract tissues. The rate constant for enterohepatic recirculation used to fit the i.v. data was much lower than for the oral route data. Although experimental variability cannot be ruled out, it seems unlikely that route differences would be as high as a factor of 7. Instead, it is likely that, like the larger oral route Vmax for glucuronidation, the difference is attributable to oversimplification of hepatointestinal handling of BPA and BPAG.
Although successfully describing BPA pharmacokinetics after oral administration only requires a representation of total (GI tract plus hepatic) metabolic clearance, the less exact simulation of oral-route BPAG pharmacokinetics by the model may be the result of oversimplification of the GI tract compartment in the model. The under-representation of peak BPAG concentrations in female rats, for instance, may be the consequence of attributing BPA glucuronidation solely to the liver, where it is eliminated in the bile before returning to the blood. Specifying metabolism in the GI tract would allow production of BPAG in these tissues and passage directly to the blood, increasing peak concentrations after oral exposure. We confirmed this hypothesis by simulation, but we found blood concentrations of BPAG were then significantly over-predicted. This observation is consistent with the findings of Inoue et al., who, using isolated perfused intestinal segments, showed that rat intestinal tissues not only glucuronidate BPA but also secrete a portion of the BPAG produced back into the intestinal lumen (Inoue et al., 2003), reducing the amount delivered to the blood. However, these more complex descriptions of GI tract handling of BPA and BPAG proved difficult to implement. Even the simplest of the alternative model structures we considered (see Model 2 in the Supplementary Material online), added three to four parameters to the model, expanding the parameter space considerably. Model 2 was a revision of the initial model that included first-order glucuronidation of BPA in the GI tract, movement of the glucuronide from the GI tract to both blood and the GI tract lumen, hepatic clearance of blood BPAG, and hydrolysis of BPAG in the GI tract lumen (see Figure 1 in the Supplementary Material online). Using the hepatic Vmax derived from the i.v. data (Upmeier et al., 2000
), and adding GI tract metabolism of BPAG, simulations of blood BPA and BPAG (female rat 100 mg/kg dose group [Pottenger et al., 2000
]) were improved relative to the initial model (Fig. 10) while maintaining fits to the urine and feces data (see Figure 2 in the Supplementary Material online). This model structure also allowed a significant reduction in the first order rate constant for oral uptake (from 10 to
1). Simulations of the male 100 mg/kg and female 10 mg/kg dose oral route pharmacokinetic data (Pottenger et al., 2000
) were less satisfying (data not shown); attempts to fit these data required variation of several model parameters including the rate of BPAG hydrolysis in the GI tract, hepatic and GI tract glucuronidation of BPA, and/or the fraction of BPAG produced in the GI tract passing to the blood and GI tract lumen. Unique values of the added parameters were not obtainable, and the parameter values appeared to vary in unexpected ways, such as the requirement for a higher hepatic Vmax for the female 10 mg/kg dose group than the 100 mg/kg dose group. Although Model 2 shows greater fidelity to the underlying physiology and biochemistry relevant to BPA pharmacokinetics and improved ability to fit one of the oral-route rat data sets, implementing this version of the model with complete success for both oral doses would require additional experimental data necessary to obtain better estimates for the additional parameters. It also appears that the initial model, by lumping some processes such as GI tract and hepatic glucuronidation, masks some variability in the underlying processes, allowing a more consistent parameter set to be obtained. The simpler model was retained to keep the level of model detail consistent with the experimental data and to ensure that the need for additional research on GI tract handling of BPA and BPAG be clear. Attractive targets for additional research include in vitro characterization of BPA glucuronidation in hepatocytes and enterocytes in both humans and rats, hepatic clearance of blood BPAG, and further characterization of enterocyte processing of BPAG.
|
After submission of this manuscript, a rat PBPK model for BPA was published by Shin and co-workers (2004). The Shin et al. model represents a larger number of tissues directly (e.g., spleen, heart, lung, testis), although it does not include a uterine compartment, blood binding of BPA, receptor binding of BPA, or enterohepatic recirculation of BPAG. An unusual approach was taken to represent BPA GI tract tissue concentrations. This tissue was represented as being in equilibrium with blood with a partition coefficient of 46, although the high concentrations of BPA observed in the GI tract are the result of both partitioning and enterohepatic recirculation from a compartment external to the body. Given the much lower partition coefficients for other tissues reported by these authors (
15), it is likely that enterohepatic recirculation, rather than equilibrium with blood is responsible for the high measured BPA concentrations in the GI tract. Representing the GI tract as being in equilibrium with blood adds a compartment with an effective BPA volume of distribution of
524 ml for a 250-g rat and results in an inappropriately high volume of distribution for BPA. We confirmed our expectation that a model structured like the Shin et al. model, but without the high partitioning in the gut compartment, does not accurately simulate these or other i.v. route data (data not shown) because the volume of distribution is too low.
Shin et al. present simulations of blood and tissue BPA concentrations for a single repeat dose i.v. data set. These simulations agree well with the observed data. No effort appears to have been made to test the model against the larger body of i.v. data, which include steady-state blood concentrations after i.v. infusion, as measured and reported by these authors and several other published i.v. infusion or i.v. bolus studies. Attempts to simulate the Upmeier i.v. data (Upmeier et al., 2000) would have revealed both the need to include enterohepatic recirculation and a BPAG submodel, and the apparent difference between these data and their own. Similarly, efforts to simulate the rat oral-route data (Pottenger et al., 2000
) would have also revealed the need to include a submodel for BPAG and enterohepatic recirculation. Without demonstrating an ability to simulate rat (Pottenger et al., 2000
) or human (Völkel et al., 2000) oral-route pharmacokinetic data, and without using published human tissue partition coefficients for BPA (Csanady et al., 2002
), the model was scaled and used to predict human blood concentrations after oral exposure. The confidence in such predictions is very low, and great care should be exercised in interpreting the results of the human simulations.
Völkel reported blood concentrations of BPA in human volunteers exposed orally to 5 mg BPA as below the analytical detection limit of 9 nM even though this was a dose considerably higher than worst case estimates of daily human exposures from the European Union risk assessment of 0.6 mg/day (European Commission, 2003). This worst-case estimate of human exposure is highly conservative and is not consistent with other estimates of less than 1 µg/person/day based on biomonitoring data (Arakawa et al., 2004
). The human PBPK model was fitted to this upper bound (9 nM) on BPA blood concentration and so can provide only upper bound estimates of BPA concentrations after low oral exposure. The accurate predictions of blood BPAG concentrations and urinary elimination of BPAG by this model, however, raise the possibility of its use as a tool for exposure assessment. Both blood BPAG and urinary BPAG are biomarkers of exposure that have been used to estimate human exposure to BPA (Arakawa et al., 2004
; Ikezuki et al., 2002
; Ouchi and Watanabe, 2002
). Bisphenol A is rapidly eliminated as the glucuronide through the urine after oral ingestion, with
90% of the dose present in urine 6 h after ingestion. Reports of total BPA in spot samples of human urine are often accompanied by crude estimates of daily BPA exposure made by multiplying urinary BPA concentrations by an estimated urine output of 2l or by multiplying urinary BPA/g creatinine by an assumed daily creatinine elimination of 1.2 g/day (Arakawa et al., 2004
; Ouchi and Watanabe, 2002
). It is clear, given the rapid urinary elimination of orally ingested BPA, that the amount of BPA found in urine spot sampling studies would be related to short-term exposure rates and patterns, and the time between ingestion of the prior meal, previous voiding of the bladder, and the time of urine collection, rather than longer term, average BPA exposures. These simple (i.e., spot sampling) approaches do not account for the pharmacokinetics of BPA in humans, and they do not address the influence of ingestion and sampling patterns on these predictions. Limitations in these simple approaches to exposure estimation may be overcome through the use of a PBPK model that directly accounts for these important processes. Given urinary concentrations and some information about the exposure pattern and sampling time post-exposure, BPA exposures resulting in the observed urine concentration could be estimated with the model. An alternative experimental approach that would also overcome limitations of the spot sampling method would be to collect urine over a full 24 h period and report total BPA excreted.
The tissue-level response to exposure to weakly estrogenic compounds is determined by the magnitude, timing, and duration of exposure (Plowchalk and Teeguarden, 2002), and it is expected to correlate best with dose metrics closely related mechanistically to the response being studied. Uterine ER binding has been proposed as a relevant dose metric for the effects of weakly estrogenic compounds such as BPA (Plowchalk and Teeguarden, 2002
). Recently, Degen et al. (2002)
showed a poor correlation between blood concentrations of several weakly estrogenic compoundsBPA, genistein, and octylphenoland uterine wet weight in the standard uterotrophic assay, demonstrating the limitations of blood dose metrics for predicting response to this class of compounds (Degen et al., 2002
). In light of these findings, we have adopted the approach proposed by Plowchalk (Plowchalk and Teeguarden, 2002
), and correlated simulated estrogen receptor binding with uterine wet weight after oral exposure to BPA in rats. The results demonstrate clear correlations between receptor occupancy and uterine tissue response only when the "free" pharmacologically active BPA is allowed to partition into uterine tissue. This finding highlights the importance of incorporating the key processes controlling response to these compoundsabsorption, tissue distribution, metabolic and other clearances, receptor binding, and restriction of free concentrations by plasma protein bindinginto pharmacokinetic models expected to be used in safety assessments.
![]() |
SUPPLEMENTARY MATERIAL |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
ACKNOWLEDGMENTS |
---|
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Andersen, M. E., and Barton, H. A. (1999). Biological regulation of receptor-hormone complex concentrations in relation to dose-response assessments for endocrine-active compounds. Toxicol. Sci. 48, 3850.
Andersen, M. E., Clewell, H. J., III, Gargas, M. L., Smith, F. A., and Reitz, R. H. (1987). Physiologically based pharmacokinetics and the risk assessment process for methylene chloride. Toxicol. Appl. Pharmacol.y 87, 185205.[CrossRef]
Arakawa, C., Yoshinaga, J., Imai, H., Serizawa, S., and Shiraishi, H. (2004). Daily urinary excretion of bisphenol A. Env. Health Prev. Med. 9, 2226.[CrossRef]
Ashby, J., and Tinwell, H. (1998). Uterotrophic activity of bisphenol A in the immature rat. Environ. Health Perspect. 106, 719720.[ISI][Medline]
Barrett, D. A., Barker, D. P., Rutter, N., Pawula, M., and Shaw, P. N. (1996). Morphine, morphine-6-glucuronide and morphine-3-glucuronide pharmacokinetics in newborn infants receiving diamorphine infusions. Br. J. Clin. Pharmacol. 41, 531537.[CrossRef][ISI][Medline]
Barton, H. A. (1999). Development of a physiologically based pharmacokinetic model for bisphenol A. K. S. Crump Group, RTP.
Barton, H. A., and Andersen, M. E. (1997). Dose-response assessment strategies for endocrine-active compounds. Regul. Toxicol. Pharmacol. 25, 292305.[CrossRef][ISI][Medline]
Bendridi, N., Mappus, E., Grenot, C., Lejeune, H., Yves Cuilleron, C., and Pugeat, M. (2002). Intravenous injection of human sex steroid hormonebinding globulin in mouse decreases blood clearance rate and testicular accumulation of orally administered [2-125I]iodobisphenol A. Steroids 67, 637645.[CrossRef][ISI][Medline]
Brown, R. P., Delp, M. D., Lindstedt, S. L., Rhomberg, L. R., and Beliles, R. P. (1997). Physiological parameter values for physiologically based pharmacokinetic models. Toxicol. Ind.l Health 13, 407484.
Bruce, N. (1976). The distribution of blood flow to the reproductive organs of rats near term. J. Reprod. Fertil. 46, 359362.[ISI][Medline]
Bulayeva, N. N., Gametchu, B., and Watson, C. S. (2004). Quantitative measurement of estrogen-induced ERK 1 and 2 activation via multiple membrane-initiated signaling pathways. Steroids 69, 181192.[CrossRef][ISI][Medline]
Clark, J. H., and Peck, E. J. Jr. (1979). Female Sex Steroids. Receptors and function. In Monographs on Endocrinology (F. Gross, A. Labhart, T. Mann, and J. Zander, Eds.), p. 245. Springer-Verlag, New York.
Clewell, H. J., III, Allen, B. C., Covington, T. R., and Gearhart, J. M. (1997). Development of a Physiologically-Based Pharmacokinetic Model for Trichloroethylene and Its Metabolites for Use in Risk Assessment. EPA Office of Research and Development, Washington, DC.
Csanady, G. A., Oberste-Frielinghaus, H. R., Semder, B., Baur, C., Schneider, K. T., and Filser, J. G. (2002). Distribution and unspecific protein binding of the xenoestrogens bisphenol A and daidzein. Arch. Toxicol. 76, 299305.[CrossRef][ISI][Medline]
Dechaud, H., Ravard, C., Claustrat, F., de la Perriere, A., and Pugeat, M. (1999). Xenoestrogen interaction with human sex hormonebinding globulin (hSHBG). Steroids 64, 328334.[CrossRef][ISI][Medline]
Degen, G. H., Janning, P., Wittsiepe, J., Upmeier, A., and Bolt, H. M. (2002). Integration of mechanistic data in the toxicological evaluation of endocrine modulators. Toxicol. Lett. 127, 225237.[CrossRef][ISI][Medline]
European-Commission (2003). European Union Risk Assessment Report 4,4'-isopropylidenediphenol (bisphenol A), Environment and Quality of Life Series,Volume 37. (S. J. Munn, R. Allanou, K. Aschberger, F. Berhault, J. d. Bruijn, C. Musset, S. O'Conner, S. Pakalin, G. Pellegrini, S. Scheer and S. Vegro, Eds.), p. 290. Office for Official Publications of the European Communities, Luxembourg.
Gaido, K. W., Leonard, L. S., Lovell, S., Gould, J. C., Babal, D., Portier, C. J., and McDonnell, D. P. (1997). Evaluation of chemicals with endocrine modulating activity in a yeast-based steroid hormone receptor gene transcription assay. Toxicol. Appl. Pharmacol. 143, 205212.[CrossRef][ISI][Medline]
Goodson, A., Summerfield, W., and Cooper, I. (2002). Survey of bisphenol A and bisphenol F in canned foods. Food Addit. Contam. 19, 796802.[CrossRef][ISI][Medline]
Guyton, A. C. (1991). Female physiology before pregnancy and the female hormones. In Textbook of Medical Physiology (A. C. Guyton and J. Hall, Eds.), pp. 899914. W. B. Saunders Company, Philadelphia.
ICRP (International Commission on Radiological Protection). (1975). Report of the Task Group on Reference Man. IRCP Publication No. 23. Elsevier, New York.
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, 28392841.
Inoue, H., Yokota, H., Makino, T., Yuasa, A., and Kato, S. (2001). Bisphenol A glucuronide, a major metabolite in rat bile after liver perfusion. Drug Metab. Dispos. 29, 10841087.
Inoue, H., Yuki, G., Yokota, H., and Kato, S. (2003). Bisphenol A glucuronidation and absorption in rat intestine. Drug Metab. Dispos. 31, 140144.
Ishihara, A., Nishiyama, N., Sugiyama, S., and Yamauchi, K. (2003). The effect of endocrine disrupting chemicals on thyroid hormone binding to Japanese quail transthyretin and thyroid hormone receptor. Gen. Comp. Endocrinol. 134, 3643.[CrossRef][ISI][Medline]
Jensen, E. V. (1965). Metabolic fate of sex hormones in target tissues with regard to tissue specificity. In Proceedings of Second International Congress of Endocrinology, Vol. 83, pp. 420433. Excerpta Medica Foundation, Amsterdam.
Kerr, M. B., Marshall, K., and Senior, J. (1992). A comparison of the uterotrophic response in cyclic and ovariectomized normotensive and spontaneously hypertensive rats. J. Endocrinol. 135, 263269.[Abstract]
Kuester, R., and Sipes, I. G. (2003). In Vitro Metabolism of Bisphenol A (BPA): Age and Species Differences. The University of Arizona, Department of Pharmacology and Toxicology for the Bisphenol A. Task Group of the Society of Plastics Industry, Inc., Washington, DC.
Kuiper, G., Lemmen, J., Carlsson, B., Corton, J., Safe, S., van der Saag, P., van der Burg, B., and Gustafsson, J. (1998). Interaction of estrogenic chemicals and phytoestrogens with estrogen receptor beta. Endocrinology 139, 42524263.
Kuiper, G. G. J. M., Carlsson, B., Grandien, K., Enmark, E., Haggblad, J., Nilsson, S., and Gustafsson, J.- Å. (1997). Comparison of the ligand binding specificity and transcript tissue distribution of estrogen receptors and ß. Endocrinology 138, 863870.
Kurebayashi, H., Harada, R., Stewart, R. K., Numata, H., and Ohno, Y. (2002). Disposition of a low dose of bisphenol a in male and female cynomolgus monkeys. Toxicol. Sci. 68, 3242.
Langlois, P. L. (1970). The size of the normal uterus. J. Reprod. Med. 4, 220228.[ISI][Medline]
Matthews, J. B., Twomey, K., and Zacharewski, T. R. (2001). In vitro and in vivo interactions of bisphenol A and its metabolite, bisphenol A glucuronide, with estrogen receptors alpha and beta. Chem. Res. Toxicol. 14, 149157.[CrossRef][ISI][Medline]
Mayersohn, M. (2003). Bisphenol A: Characterization of Plasma Protein Binding in Rats. College of Pharmacy, University of Ariszona, Tuscon, Arizona.
Mendel, C. M. (1992). The free hormone hypothesis: distinction from the free hormone transport hypothesis. J. Androl. 13, 107116.
Milligan, S., Balasubramanian, A., and Kalita, J. (1998). Relative potency of xenobiotic estrogens in an acute in vivo mammalian assay. Environ. Health Perspect. 106, 2326.[ISI][Medline]
Nagel, S., vom Saal, F., and Welshons, W. (1998). The effective free fraction of estradiol and xenoestrogens in human serum measured by whole cell uptake assays: Physiology of delivery modifies estrogenic activity. Proc.e Soc. Exp. Biol. Med. 217, 300309.[ISI]
O'Connor, J. C., Cook, J. C., Craven, S. C., Van Pelt, C. S., and Obourn, J. D. (1996). An in vivo battery for identifying endocrine modulators that are estrogenic or dopamine regulators. Fundam. Appl. Toxicol. 33, 182195.[CrossRef][ISI][Medline]
O'Flaherty, E. J. (1981). Toxicants and drugs: Kinetics and dynamics. In Institute of Environmental Health, p. 398. University of Cincinnati Medical Center, Cincinnati.
Odum, J., Lefevre, P. A., Tittensor, S., Paton, D., Routledge, E. J., Beresford, N. A., Sumpter, J. P., and Ashby, J. (1997). The Rodent uterotrophic assay: Critical protocol features, studies with nonyl phenols, and comparison with a yeast estrogenicity assay. Regul. Toxicol. Pharmacol. 25, 176188.[CrossRef][ISI][Medline]
Ouchi, K., and Watanabe, S. (2002). Measurement of bisphenol A in human urine using liquid chromatography with multi-channel coulometric electrochemical detection. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 780, 365370.[ISI][Medline]
Owens, J. W., and Ashby, J. (2002). Critical review and evaluation of the uterotrophic bioassay for the identification of possible estrogen agonists and antagonists: In support of the validation of the OECD uterotrophic protocols for the laboratory rodent. Organisation for Economic Co-operation and Development. Crit. Rev. Toxicol. 32, 445520.[ISI][Medline]
Plowchalk, D. R., and Teeguarden, J. (2002). Development of a physiologically based pharmacokinetic model for estradiol in rats and humans: A biologically motivated quantitative framework for evaluating responses to estradiol and other endocrine-active compounds. Toxicol. Sci. 69, 6078.
Pottenger, L. H., Domoradzki, J. Y., Markham, D. A., Hansen, S. C., Cagen, S. Z., and Waechter, J. M., Jr. (2000). The relative bioavailability and metabolism of bisphenol A in rats is dependent upon the route of administration. Toxicol. Sci. 54, 318.
Shin, B., Kim, C., Jun, Y., Kim, D., Lee, B., Yoon, C., Park, E., Lee, K., Han, S. Y., Park, K., Kim, H., and Yoo, S. (2004). Physiologically based pharmacokinetics of bisphenol A. J. Toxicol. Environ. Health A 67, 19711985.[CrossRef][Medline]
Snyder, R. W., Maness, S. C., Gaido, K. W., Welsch, F., Sumner, S. C., and Fennell, T. R. (2000). Metabolism and disposition of bisphenol A in female rats. Toxicol. Appl. Pharmacol. 168, 225234.[CrossRef][ISI][Medline]
Stock, M., and Metcalfe, J. (1994). Maternal physiology during gestation. In The Physiology of Reproduction (E. Knobilll and J. Neill, Eds.), Vol. 2, pp. 947983. Raven Press, New York.
Sun, Y., Nakashima, M. N., Takahashi, M., Kuroda, N., and Nakashima, K. (2002). Determination of bisphenol A in rat brain by microdialysis and column switching high-performance liquid chromatography with fluorescence detection. Biomed. Chromatogr. 16, 319326.[CrossRef][ISI][Medline]
Teeguarden, J. G., and Barton, H. A. (2004). Computational modeling of serum-binding proteins and clearance in extrapolations across life stages and species for endocrine active compounds. Risk Anal. 24, 751770.[CrossRef][ISI][Medline]
Twomey, K. (1998). Bisphenol A: Uterotrophic Assay in Immature Rats (Oral Dosing), p. 49. Central Toxicology Laboratory, Alderley Park, Macclesfield, Cheshire, UK.
Upmeier, A., Degen, G. H., Diel, P., Michna, H., and Bolt, H. M. (2000). Toxicokinetics of bisphenol A in female DA/Han rats after a single i.v. and oral administration. Arch. Toxicol. 74, 431436.[CrossRef][ISI][Medline]
Völkel, W., Colnot, T., Csanady, G. A., Filser, J. G., and Dekant, W. (2002). Metabolism and kinetics of bisphenol A in humans at low doses following oral administration. Chem. Res. Toxicol. 15, 12811287.[CrossRef][ISI][Medline]
Waynford, H. B., and Flecknell, P. A. (1992). Experimental and Surgical Technique in the Rat. Academic Press, London.
Yamasaki, K., Sawaki, M., and Takatsuki, M. (2000). Immature rat uterotrophic assay of bisphenol A. Environ. Health Perspect. 108, 11471150.[ISI][Medline]
Yoo, S. D., Shin, B. S., Kwack, S. J., Lee, B. M., Park, K. L., Han, S. Y., and Kim, H. S. (2000). Pharmacokinetic disposition and tissue distribution of bisphenol A in rats after intravenous administration. J. Toxicol. Environ. Health A 61, 131139.[CrossRef][Medline]
Zava, D., and Duwe, G. (1997). Estrogenic and antiproliferative properties of genistein and other flavonoids in human breast cancer cells in vitro. Nutr. Cancer 27, 3140.[ISI][Medline]
Zhang, R.-S., Buth, P. H., Scremin, O. U., and Chaudhuri, G. (1995). H2 gas clearance technique for separating rat uterine blood flow into endometrial and myometrial components. Am. J. Physiol. 268, R569R575.[ISI][Medline]
|