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

David R. Plowchalk*,1 and Justin Teeguarden{dagger}

* DuPont Haskell Laboratory for Health and Biomedical Sciences, Newark, Delaware 19714; and {dagger} The K.S. Crump Group, Inc., ICF Consulting, Research Triangle Park, North Carolina 27709

Received February 15, 2002; accepted May 20, 2002


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
A physiologically based pharmacokinetic (PBPK) model for estradiol (E2) in rats and humans (male and female) was developed to provide a quantitative tool for evaluating the importance of physiological parameters on E2 blood and tissue concentration time-course and for predicting blood and tissue concentrations in rats and humans. A hepatic extraction model was developed to evaluate the significance of plasma protein binding on the hepatic extraction of E2 and the approach was integrated into the E2 model. Sufficient data was available to parameterize and validate oral and iv routes. The E2 model simulations of E2 blood and tissue concentrations compared well to experimental values. Estrogen receptor content strongly impacts distribution and elimination kinetics of E2 as well as tissue concentrations. The prolonged terminal elimination phase seen after iv bolus administration reflects the slow release of receptor bound E2 from tissues. E2 uptake behavior in the ovariectomized, but not intact rat uterus, was best described as diffusion-limited. Simulations with the hepatic extraction model predicted extensive binding of E2 to albumin (rat) and SHBG (sex-hormone binding globulin humans), although hepatic extraction does not appear to be restricted to the unbound fraction, implying that the total plasma E2 concentration is important when considering hepatic uptake. Important determinants of E2 disposition are tissue ER content and binding affinity, nonreceptor binding proteins, vascular permeability, partition coefficients, hepatic blood flow, and extrahepatic metabolism. As an integral part of a research program, the quantitative framework developed for E2 can be extended to other endocrine-active compounds (EACs) and used to evaluate the biological activity of EACs.

Key Words: PBPK model; endocrine system; estradiol; risk assessment; endocrine-active compound (EAC).


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Estradiol (E2) is a vital steroid hormone that regulates numerous endocrine functions through binding to two estrogen receptor (ER) isoforms, ER{alpha} and ERß. The concentration of these receptor subtypes and the corresponding cofactors vary with tissue type, leading to tissue specific regulation of estrogen response, and toxicity. The distinction between a physiological and a toxic response to estrogen exposure is determined by the magnitude, timing, and duration of exposure.

Estradiol is carried in the plasma in two forms, bound to plasma binding proteins, and to a lesser extent, as "free" estradiol. In the human, binding to the sex hormone binding globulin (SHBG) and albumin restrict the free fraction of estradiol in plasma. Albumin and {alpha}-fetoprotein (AFP) are the corresponding rat proteins. The free form is presumed to be the pharmacologically active fraction, which is capable of passing through the lipid bilayer and equilibrating with tissues. However, under specific conditions, protein bound E2 appears to be available for tissue uptake (Mendel, 1992Go; Pardridge and Mietus, 1979aGo). In vivo, free plasma concentrations of estradiol are regulated by feedback control (hypothalamic-pituitary-gonadal axis) and binding to albumin, SHBG or AFP (Andersen et al., 1997Go).

Interest has recently grown in the use of E2 as a prototypical estrogenic endocrine-active compound (EAC). EACs are a structurally diverse group of chemicals that have impacts on the endocrine system, some of which act specifically by binding to the estrogen receptor (Andersen and Barton, 1999Go). Comparatively, the biology of these ligands is poorly characterized, and given their number, efforts to evaluate their comparative binding or transactivating potency have centered on comparisons made to the endogenous ligand E2 (Danzo, 1997Go; Jobling et al., 1995Go; Klotz et al., 1996Go; Milligan et al., 1998Go; Shelby et al., 1996Go; Soto et al., 1995Go; Stahl et al., 1998Go). Rankings must be based on more than external dose since pharmacologically active plasma and tissue concentrations and potency of these compounds are determined by the interaction of a variety of processes that are compound specific: binding affinity to the ER and plasma binding proteins, ligand receptor competition, metabolic and total clearance rates, and tissue:blood partition coefficients among others. Pharmacodynamic processes (regulation of serum estrogenicity, enzyme induction, and regulation of tissue receptor levels) will also influence the nature and magnitude of toxicity following exposure. In addition, the relevant dose metrics (plasma, tissue, total, or free concentrations) must be determined.

To better appreciate the physiologic responses produced by E2 in target tissues and their dose-response relationships, it is necessary to understand E2 pharmacokinetics and the factors controlling its disposition within the quantitative framework of a physiologically based pharmacokinetic model. Reliable PBPK models will be useful for conducting dose-response analysis, establishment of appropriate dose metrics for ER agonists, and understanding impacts of exposure route, dose, and timing of target tissue levels. While several pharmacokinetic analyses have been conducted for E2 in rat and human plasma using compartmental (Eisenfeld, 1967Go) and model-independent (Kuhnz et al., 1993Go) approaches, a PBPK model for estradiol has not been developed. Construction of a PBPK model for E2 is the initial and necessary step in the development of a biologically motivated quantitative tool for evaluating the dose-response characteristics of EACs.

There are three primary objectives for the work presented here: (1) Construct a PBPK model to describe the uptake, distribution, and clearance of E2 in the rat and human following oral and iv administration; (2) Evaluate the impact of plasma protein binding on tissue uptake in the liver and uterus and free plasma concentrations of E2; and (3) Identify physiological and biochemical parameters that have the most significant impact on E2 pharmacokinetics, plasma, hepatic, and uterine E2 concentrations, and use this information to establish a research agenda to collect the necessary data to develop a more complete PBPK model for E2. In addition, we sought to describe the framework for, and importance of using, a quantitative model that integrates the processes that influence EAC pharmacokinetics and tissue response (receptor binding affinity, receptor distribution, restriction of free EAC concentrations via plasma binding proteins, tissue kinetics, and clearance) for evaluating the biological activity of EACs.

The PBPK model developed here provides a flexible tool for integrating existing biological knowledge of the processes governing E2 pharmacokinetics, evaluating their significance and impact, as well as providing predictions of in vivo binding potency and tissue dosimetry. The final model will be central to ongoing efforts to characterize the toxicity and dose-response relationships for the prototypical EAC, E2, and by extension, other EACs.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Model structure.
The model comprised the various tissues, at an appropriate level of detail, important for absorption, distribution and storage, as well as metabolic clearance of estradiol (Fig. 1Go). The liver and plasma were treated as distinct compartments, while the remaining tissues (except target tissues) were lumped into typical richly and poorly perfused tissue compartments. In addition, 3 target organs were included, the uterus, ovary, and pituitary. These were modeled separately to allow prediction of target tissue concentrations. Each tissue compartment (except ovary and richly perfused compartments) is described as a diffusion-limited compartment containing a tissue plasma subcompartment, which represents the plasma in the tissue vasculature, and a tissue subcompartment separated by a diffusional barrier (Kedderis et al., 1993Go). Movement of E2 between the plasma and tissue is a function of diffusional clearance (PAT, where T is a specific tissue) and is limited to free E2 (i.e., E2 not bound to albumin, SHBG or ER). The remaining compartments were treated as rapid equilibrium, well-mixed compartments as described elsewhere (Andersen, 1981Go). A separate systemic plasma compartment includes protein binding (albumin and SHBG), extrahepatic metabolism, and input terms for intravascular administration of E2. The liver compartment includes an intrinsic clearance term, estrogen receptor (ER) binding sites and input terms for po administration of E2. A second binding term in the liver was used to address the estrogen-specific binding protein in male rats (Rogerson and Eagon, 1986Go) and served as a site for additional nonreceptor E2 binding in female rats.



View larger version (22K):
[in this window]
[in a new window]
 
FIG. 1. Diagram of PB-PK Model 2 used to simulate E2 pharmacokinetics in rats and humans.

 
Unbound E2 in the plasma is a function of its distribution between plasma binding proteins (albumin and SHBG) and was solved by approximation. Competition for protein binding sites with other endogenous steroids is assumed not to occur. For high clearance tissues such as the liver, if the plasma transit time through the organ (tissue plasma volume/tissue plasma flow rate) is long relative to the Koff for E2:Binding Protein, the effective free plasma concentrations can be greater than that under equilibrium conditions (Mendel, 1992Go). Rapid clearance from the tissue plasma compartment results in movement of additional E2 off the binding protein and into the tissue, resulting in effective free concentrations that are greater than predicted based on equilibrium binding equations. Each tissue plasma compartment includes an independent description of plasma protein binding so that an apparent Kd can be defined for each organ (Pardridge and Landaw, 1984Go). This allows the tissue plasma of each organ to have a different free E2 fraction. Additionally, simulations can be configured so that all E2 is free and available to the tissues (as in well-mixed compartments) by setting BMAXA and BMAXG (Bmax for albumin and SHBG) equal to zero and PAi (diffusional clearance) to a value exceeding plasma flow. Intravascular and po dosing routes were included in the model. These could be used together or separately to reflect experimental conditions. The apparent Kd approach is an approximation that replaces direct treatment of receptor ligand kinetics, tissue transit time and clearance, which would require data that is not available. Alternative approaches, the use of a constant to reduce the available concentrations, for example, would accomplish the same thing. ER binding in tissue compartments was described in terms of equilibrium binding:

where the bound E2 concentration (Ci, bound) in tissue i is a function of the tissue receptor content (Bmax,i) and dissociation constant (Kd). Percent receptor occupancy (POcci) is simply calculated as:

Parameterization of PB-PK model.
Models were written in ACSL simulation language (MGA Software, Concord, MA) using ACSL Graphic ModellerTM. Parameter estimations and sensitivity analyses were conducted with ACSL OptimizeTM and ACSL MathTM. Simulations of rat and human data were conducted with the same model, and differed only in species- and sex-specific physiologic parameters.

The various model parameters were either obtained directly from the available literature or inferred from experimental data. Where feasible, inferred parameters were validated using independent data sets. Validation of these parameter values is presented in the Results section.

Physiologic parameters.
The rat and human organ blood flows and volumes used for the PBPK model are listed in Table 1Go (ILSI, 1994Go). Uterine blood flow and organ size are reported for both intact and ovariectomized rats (Bruce, 1976Go; Kerr et al., 1992Go; Zhang et al., 1995Go). Although treatment of ovariectomized rats with E2 increases uterine blood flow greater than 100% in 2 h (Zhang et al., 1995Go), this parameters was assumed to be time-invariant since the iv bolus data sets have few data points past 2 h. Likewise, uterine volume was held constant since there is only approximately a 25% increase in uterine weight within 3 h of treatment with E2 (Kerr et al., 1992Go; Zhang et al., 1995Go). Uterine weight for nonpregnant women is a function of age and parity. The value used here is for women approximately 30 years of age who bore 1–3 children during their lifetime (Langlois, 1970Go). Uterine weight for nulligravidis women > 60 years old was used to approximate uterine volume in postmenopausal or ovariectomized women (Langlois, 1970Go). Cardiac output is reported in terms of plasma as opposed to whole blood since this is more consistent with the available data sets and simplified the treatment of plasma protein binding.


View this table:
[in this window]
[in a new window]
 
TABLE 1 Physiologic Parameters Used in the Estradiol PBPK Model for Rats and Humans
 
Partition coefficients.
In vitro measurements of tissue:plasma partition coefficients (PCs) for E2 are available for a limited number of tissues. E2 tissue:polycarbonate partition coefficients (Murphy et al., 1995Go) were used to calculate tissue:blood partition coefficients for fat, liver, and pituitary, and were assumed equivalent to tissue:plasma PCs (Table 2Go). The PC of 0.6 for diaphragm (Germain et al., 1978Go) was rounded to 1.0 and used for uterus and poorly perfused tissue compartments since these organs are comprised mostly of muscle tissue. Studies by De Hertogh et al.(1971a) found similar plasma and muscle E2 concentrations following ip administration of E2 suggesting a plasma:muscle PC of 1.0 is reasonable. Jehan et al.(1982) found concentrations of total radioactivity (which could include metabolites) in liver after iv administrations of radiolabeled E2 were ~3 times greater than plasma, which is consistent with an upper bound liver:plasma partition coefficient of 3.6.


View this table:
[in this window]
[in a new window]
 
TABLE 2 Model Parameters for Tissue Binding, Plasma Protein Binding, Metabolism and Absorption
 
The PC utilized in this model are estimated from ratios of total tissue E2:total blood concentrations reported in the literature and are "apparent" partition coefficients that do not distinguish between partitioning and specific binding to tissue (ER) receptors or plasma binding proteins (albumin, SHBG). As such, they are useful approximations that would benefit from revision through correction for the influence of specific binding. In several modeled tissues (uterus, liver, slowly perfused) tissue:blood partitioning was modeled as a function of free (accounts for binding) rather than total blood E2 concentrations, from which the partition coefficients were derived. Partition coefficients were assumed not to differ between species, therefore the same PCs were used for rats and humans unless otherwise noted.

Tissue and plasma binding.
Model descriptions of ERs and ER binding were simplified to be consistent with the level of detail of the experimental data and the availability of parameter values. The following assumptions were made: (1) ER concentrations in tissues are time and treatment invariant; (2) ERs exist as a single type and are not separated into {alpha} and ß subtypes (Verheugen et al., 1984Go); (3) Type II estrogen binding sites (Clark and Mani, 1994Go) do not affect E2 disposition; and (4) both cytosolic and nuclear ER receptor are equally available in tissues.

Total estrogen receptor content (Bmax) and binding affinities (Kd) were obtained for most organs in rats, however, information for humans is somewhat limited (Table 3Go). When necessary, values for human tissues were extrapolated from rat values based on organ weight. ER content in some tissues is dependent on sexual maturity and hormonal status, therefore several values are reported for comparison purposes. As evident from Table 3Go, there was a significant amount of variability in published tissue receptor content, which is probably due to differences in strain, age, and hormonal status as well as methods used for tissue preparation and analysis. ER content in adult male and female rat liver ranged between 0.02–0.04 nmol/liver. Uterine and anterior pituitary ER content ranged between 0.002–0.012 nmol/uterus and 0.07–0.2 fmol/pituitary in adult rats, respectively. An ER dissociation constant of 0.25 nM was used in the model and is consistent with the range of values (0.1–0.5 nM) reported in the literature (Table 3Go). Values for Kd (~1–2 nM) reported by Notides (1970) and Clark and Peck (1979) are somewhat higher than the more commonly accepted range (0.1–0.5 nM) and were therefore considered outliers.


View this table:
[in this window]
[in a new window]
 
TABLE 3 Estrogen Receptor Content (Bmax) and Binding Affinities (Kd) for Various Tissues in Rats and Humans
 
E2 binds to both albumin and SHBG in human plasma resulting in only 2–5% free hormone in circulation (Dunn, 1983Go; Pardridge, 1986Go). In adult rats, albumin is the only significant plasma protein that binds E2, therefore, {alpha}-fetoprotein, the rat high affinity E2 binding protein found in immature rats (Masseyeff et al., 1975Go) was excluded from the rat model. A range of binding affinities and plasma concentrations have been reported for albumin and SHBG in rats and humans. Human plasma albumin concentrations are reported to be between 4 x 105 (pregnant) and 7.6 x 105 nM with an E2 binding Kd of 1.67 x 104 nM (Dunn, 1983Go, 1984Go; Manni et al., 1985Go; Masseyeff et al., 1975Go; Pardridge and Landaw, 1985Go). Human plasma SHBG concentrations are 19.6–38 nM in males, 37–70 nM in females, and 400 nM in pregnant females with an E2 binding Kd of 1.4–4.5 nM (Dunn 1983Go, 1984Go; Dunn et al., 1980Go; Nisula and Dunn, 1979Go; Pardridge and Landaw, 1985Go; Renoir et al., 1980Go; Rosner, 1991Go; Rosner and Smith, 1975Go). The Bmax and Kd for albumin used in the systemic plasma compartment for all rat simulations were 4.1 x 105 and 2.3 x 104 nM, respectively (Pardridge and Mietus, 1979aGo; Waynford and Flecknell, 1992Go). This parameterization led to predictions of plasma E2 free fraction, which were in close agreement to those reported in the literature. The Kd for albumin in the tissue plasma compartments of the uterus, pituitary, and poorly perfused tissue (Kd(app)) was varied to modify the apparent free E2 concentration presented to the tissue, as required by fitting experimental data. The albumin concentration and Kd for the systemic plasma compartment in humans (both men and women) were set to 5 x 105 and 1.67 x 104 nM, respectively. The binding capacity of SHBG in the plasma compartment was set to 20 and 40 nM for men and women, respectively, and a Kd of 1.5 nM was used for both sexes (Table 2Go). The Bmax values obtained for muscle, bone, and fat were used to calculate an initial estimate for ER content in the slowly perfused tissue compartment; however, this value was usually fitted. The final ER binding constants used in the E2 model are listed in Table 2Go.

Metabolic clearance of estradiol.
The metabolism of E2 in rats and humans is similar in that hepatic microsomal enzymes transform E2 into estrone as well as other hydroxylated metabolites (Martucci and Fishman, 1993Go). Excretion of these products occurs through the urine or bile as sulfate or glucuronide conjugates and very little systemic conversion (~5%) back to E2 occurs (Lobo and Cassidenti, 1992Go). Clearance of estradiol in the model was attributed solely to metabolism and assumed to take place in the liver and central plasma compartment. In addition, it was assumed that there is no endogenous production of E2, either through de novo synthesis or regeneration from conjugates, estrone metabolites, or enterohepatic recirculation.

The rate of hepatic E2 metabolic clearance (RMET, nmol/h) was defined as the product of intrinsic clearance (Clint) and the free E2 concentration in the liver tissue (CLfree). This first approximation of clearance, represented as a first order process, is reasonable assuming that endogenous estradiol concentrations under the experimental conditions are subsaturating. Hepatic and extrahepatic clearance of E2 is predominately described as metabolic clearance rates (equivalent to intrinsic clearance; l/h). This composite term reflects the contribution of all metabolic processes (oxidation, sulfation, etc.) involved in the clearance of E2, and is well characterized for rats and humans in the literature (Ball et al., 1983Go; Hembree et al., 1969Go). Use of intrinsic clearance rates limits uncertainties associated with constructing a composite metabolic rate based on Vmax/Km data for the various (Badawi et al., 2001Go) E2 metabolizing enzymes. Fitted intrinsic clearance values of 3.0 and 1.0 l/h were used for male and female rats, respectively, while values of 500 and 150 l/h were used for men and women, respectively (Table 2Go). Under steady state conditions, the corresponding hepatic E2 extraction for rats is 0.71 (female)–0.84 (male), and for humans is 0.75 (female)–0.85 (male) based on liver blood flow and the intrinsic clearance rates.

Total metabolic clearance rates (MCRs) in rats and humans typically exceed hepatic blood flow indicating E2 is subject to extrahepatic metabolism (blood flow in rats and humans are 1.1 and ~62.5 l/h, respectively). In humans, it has been estimated that this may account for as much as 15–25% of total E2 metabolism (Longcope et al., 1968Go), which is equivalent to clearance rates of ~6–20 l/h. In view of this, an additional term for E2 metabolism was included in the central plasma compartment where the rate of extrahepatic metabolism was defined as the product of the extrahepatic clearance rate (ClEH) and the total circulating E2 concentration. Extrahepatic clearance rates of 0.1 l/h and 5–10 l/h were assigned for rats and humans, respectively (Table 2Go), corresponding to extrahepatic E2 extractions of 0.083 for the rat and 0.074–0.14 for the human.

Modeling approach.
The general procedure for evaluating the PBPK model was first to calibrate the model against 1 data set and then validate against secondary (independent) data sets when available.

Three general categories of simulations were run for both rats and humans based on different routes of administration: (1) iv infusion, (2) iv bolus, and (3) po or intraduodenal. Details of these simulations and the test data sets used for calibration and validation are described below and in Table 4Go.


View this table:
[in this window]
[in a new window]
 
TABLE 4 Summary of Pharmacokinetic Data Sets for E2 in Rats and Humans Used for Calibration and Validation of the E2 Model
 

    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Free Fraction of E2 in Plasma and Liver Plasma
Model predictions of E2 binding and fractional distribution between albumin and SHBG were compared to observed values in plasma with varying composition of the binding proteins SHBG and albumin. Plasma concentrations of albumin and SHBG were set to reflect the measured in vivo concentration of each protein in adult rat plasma, human male plasma, and human female plasma (pregnant and nonpregnant). Simulations were conducted for 25 nM estradiol.

The predicted distributions of E2 between free, albumin- and SHBG-bound states are shown in Table 5Go. In the absence of binding proteins, 100% of E2 is was free. In plasma containing albumin and low adult levels of {alpha}FP, such as rat plasma (BMAXA = 4.1 x 105 nM, KDA = 2.3 x 104 nM), 94.7% of E2 is bound to albumin and 5.3% is free. This distribution is consistent with in vitro dialysis experiments that show only 6.0% of E2 is dialyzable in Ringer's solution containing 4% albumin (Verheugen et al., 1984Go). In human plasma, which in addition to albumin (BMAXA = 5 x 105 nM) contains the high affinity E2 binding protein SHBG (BMAXG = 20–40 nM), free E2 decreases. Modeled free fractions of E2 in male serum (2.5%) was identical to the value measured by ultrafiltration dialysis (Nagel et al., 1998Go). Similarly, free fractions in nonpregnant female (1.9%) and pregnant female plasma (0.4%) generally agreed well with those measured experimentally (~2 and 0.5% for nonpregnant and pregnant women, respectively; Dunn, 1983Go).


View this table:
[in this window]
[in a new window]
 
TABLE 5 Simulation Results of Predicted Estradiol Distribution in Plasma Containing Various Protein Compositions
 
Under some conditions, tissue uptake of highly bound compounds, including hormones, is not restricted to the unbound fraction (Mendel, 1992Go). Simulations of hepatic uptake of E2 were performed to assess the impact of plasma protein binding on E2 availability to target tissues. The liver compartment was isolated (Fig. 2Go) and used to simulate single pass hepatic extraction experiments (Verheugen et al., 1984Go) that examined E2 uptake from perfusion media containing (1) 0.1% albumin, (2) 4.0% albumin, and (3) human serum (pregnant).



View larger version (15K):
[in this window]
[in a new window]
 
FIG. 2. Diagram of the isolated liver compartment used to explore E2 binding to plasma proteins and the resultant impact on tissue uptake. CA is the total E2 concentration entering the liver plasma compartment, where it remains free (CLBf), or is bound to albumin (Cba) or sex hormone binding globulin (Cbg). Movement of E2 into the tissue subcompartment is limited to free E2 and occurs at a rate proportional to diffusional clearance (PAL).

 
The model was first calibrated to yield the measured extraction of 0.85 for the control delivery conditions (Ringer's solution with 0.1% bovine serum albumin) by adjusting diffusional clearance to 5.0 l/h. Simulation of a single pass perfusion with 4% albumin containing 0.25 nM E2 predicts extraction of only 0.24 compared to in vivo estimates of 0.77 (Verheugen et al., 1984Go). This underprediction of extraction can be attributed to the model assumption that only free E2 is available for diffusion from the vasculature into the tissue. Reducing the E2 binding affinity (Kd) for albumin to 4 x 105 nM increased estradiol free fraction to approximately 40% and resulted in a first pass extraction of ~0.77. Model predicted hepatic extraction of E2 from human pregnancy serum was ~0.03, inconsistent with the reported in vivo observed extraction of 0.48 (Verheugen et al., 1984Go). Under these conditions the presence of SHBG reduces the free E2 concentration to ~0.5%. Adjusting the free E2 fraction to 11% by increasing the SHBG:E2 Kd(app) from 1.5 nM to 47 nM resulted in accurate simulations of hepatic extraction. These simulation results indicate hepatic uptake of E2 is not limited to unbound E2 alone, and that hepatic extraction will be best described in the PBPK model with protein binding disabled for the tissue plasma compartment of the liver. Accordingly, protein binding was not used in the liver plasma compartment (all E2 is available for hepatic clearance (E2Free = E2Total) and clearance was based on total E2. Slowly perfused and uterine tissue compartments included albumin binding and used a tissue specific Kd(app) that best described the tissue and plasma kinetics of E2. SHBG was not included in these compartments since the use of albumin binding alone provided adequate fits.

Plasma and Tissue E2 Concentrations in Rats
Intravenous administration studies.
Several studies describing the plasma, and in some cases uterine and liver kinetics in rats following iv bolus or infusion of E2 were available for model calibration and validation. De Hertogh et al. (1970; 1971a,b; 1973) infused intact female rats with E2 at a rate of 6, 65, or 180 ng/h for 2–7 h and measured E2 in plasma, uterus, and liver. An initial loading dose of 0.8 and 8.0 was administered to the 6 and 65 ng/h dose groups, respectively, prior to initiating the infusion. Similar data sets for ovariectomized female rats infused for 48 h (104 and 290 ng/h) were also evaluated (Ball et al., 1983Go). Data for male rats was limited to rats given an E2 loading dose (~27 ng) followed by a 27 ng/h infusion (Farrell et al., 1988Go). Eisenfeld (1967) reported E2 concentrations in tissues (plasma, uterus, and pituitary) after iv bolus doses of 2.5 and 25 µg/kg to ovariectomized rats. Data for intact female rats (1.4 µg/kg) were obtained from Larner and Hochberg (1985). E2 plasma concentrations in male rats were available following iv doses of 5, 10, and 20 µg (Bawaarshi-Hassar et al., 1989Go). This same study provided data sets for intraduodenal administration of E2 (5, 10, and 20 µg) and are the only data found for assessing gastrointestinal absorption of E2 in rats (Bawaarshi-Hassar et al., 1989Go).

Model predictions of steady-state E2 concentrations in plasma of male and female rats (intact and ovariectomized) following iv infusion of E2 were in good agreement with all experimental data sets. The best correspondence to the data for intact (Fig. 3Go) and ovariectomized (Ball et al., 1983Go; results not shown) female rats was seen using an intrinsic clearance (Clint) of 1.0 l/h and an extrahepatic clearance rate (EHC) of 0.1 l/h. Infusion data for male rats (Fig. 4Go) were best fit with an intrinsic clearance of 3.0 l/h and extrahepatic clearance rate of 0.1 l/h (Farrell et al., 1988Go). Simulations of plasma E2 concentrations (Fig. 5Go) in male rats after iv administration of 5, 10, and 20 µg corresponded well with the experimental (Bawaarshi-Hassar et al., 1989Go) data. A higher partition coefficient (PS = 2) and male-specific hepatic binding protein (Bmax = 65 nmol/liver, Kd = 35 nM) provided the best model fits for all 3 doses.



View larger version (20K):
[in this window]
[in a new window]
 
FIG. 3. Plasma, uterine, and liver concentrations of E2 in intact female rats infused with E2 at rates of 6, 65, or 180 ng/h (De Hertogh et al., 1970Go, 1971aGo,bGo, 1973Go). Model simulations (solid lines) are in good agreement with all experimental data for plasma (circle), uterus (square) and liver (triangle). If the uterus is described as a diffusion-limited organ, the model under predicts E2 concentrations (Panel A, dotted line).

 


View larger version (14K):
[in this window]
[in a new window]
 
FIG. 4. Model simulations of steady-state plasma concentrations (line) in male rats given an iv bolus loading dose of 27 ng and an iv infusion of 27 ng/h for 3.25 h were in agreement with data (open circles) from Farrell et al.(1988).

 


View larger version (16K):
[in this window]
[in a new window]
 
FIG. 5. Model predictions of plasma E2 concentrations in male rats following iv administration of either 5, 10, or 20 µg of E2. Symbols are data points from Bawaarshi-Nassar et al.(1989).

 
Simulation results of E2 concentrations in the uteri of intact rats corresponded well with experimental data from De Hertogh et al. (1971b, 1973) (Fig. 3Go). Uterine uptake behavior in these simulations can be best characterized as nondiffusion-limited (KDUA = 5 x 105 nM and PAU = 1.0). Parameterization of the uterus as a diffusion-limited organ—a condition necessary to describe E2 uptake in uteri of ovariectomized rats—grossly underpredicts uterine E2 concentrations (Fig. 3Go, dotted line). Steady-state E2 concentrations in the liver are ~10 times greater than those seen in plasma (De Hertogh et al., 1971aGo) and also correspond well with model predictions (Fig. 3Go) when an ER concentration of 0.065 nmol/liver is used, a value that is consistent with in vitro estimates (Table 3Go).

The model successfully fit data sets for E2 plasma pharmacokinetics following iv bolus dosing in both intact (Larner and Hochberg, 1985Go) and ovariectomized rats (Eisenfeld, 1967Go). The initial distribution and elimination phases fit well for both data sets (Figs. 6 and 7GoGo); however, additional binding in the liver compartment (20 nmol/liver) was needed to extend the predicted terminal elimination phase in ovariectomized rats.



View larger version (18K):
[in this window]
[in a new window]
 
FIG. 6. Model predictions of E2 concentrations in plasma, uterus, and pituitary of ovariectomized rats following iv administration of E2 (2.5 and 25 µg/kg, filled and open circles, respectively). All data are from Eisenfeld (1967). The best predictions of E2 concentrations in the uterus were described by a diffusion-limited behavior (solid lines), in contrast to nondiffusion-limited conditions where E2 concentrations are significantly overpredicted (dashed lines).

 


View larger version (13K):
[in this window]
[in a new window]
 
FIG. 7. Model predictions of plasma E2 concentrations (line) in intact female rats following iv administration of 1.4 µg/kg E2. All data (open circles) are from Larner and Hochberg (1985).

 
Unlike simulations of uterine E2 concentrations in the intact rat, uterine uptake in the ovariectomized rat was best characterized as diffusion-limited (Fig. 6Go). Predictions of E2 in both the uterus and pituitary were dependent on the presence of ER in the tissues (Fig. 6Go), without which tissue concentration-time curves parallel plasma curves (simulation not shown). Estimates of E2 concentrations in the uterus (Fig. 6Go; dashed lines) were overpredicted if parameters for the intact female rat were used. Best fits were achieved setting the apparent free concentration of E2 in uterine plasma to ~9% (KDUA = 4 x 104 nM). The best fits to E2 concentrations in the pituitary were seen with an ER content of 0.0013 nmol/pituitary, which is ~10-fold higher than measured in vitro (Table 3Go).

Intraduodenal administration.
One study was available for assessing gastrointestinal absorption of E2 in rats. Bawaarshi-Nassar reported plasma E2 concentrations following intraduodenal administration of 5, 10, and 20 µg of E2 (Bawaarshi-Hassar et al., 1989Go). The absorption rate and intrinsic clearance were adjusted manually to arrive at single values of each (KO = 1.7 h-1, CLINT = 25.0 l/h) that achieved the best fit for the 3 dose groups. Simulations of E2 uptake from the intestinal tract after intraduodenal administration of 5, 10, and 20 µg corresponded well with the data for the high and low dose, but was overpredicted at the middle dose (Fig. 8Go). Fractional absorption (FO) was set to 1.0 since E2 absorption from the gut appears complete when given at low doses by the po route (Kuhnz et al., 1993Go).



View larger version (17K):
[in this window]
[in a new window]
 
FIG. 8. Model predictions of E2 plasma concentrations (lines) in male rats following intraduodenal administration of 5, 10, or 20 µg of E2. Symbols are data points from Bawaarshi-Hassar et al.(1989).

 
Plasma and Tissue E2 Concentrations in Humans
IV administration.
Steady-state infusion experiments in humans were reported by Longcope and Tait (1971) and Hembree et al.(1969). Hembree et al.(1969) gave both men and women (premenopausal) subjects an iv loading dose of E2 (~93 ng) followed 30 min later by an E2 infusion (2 h) at a rate of 75–150 ng/h (men) or 45–70 ng/h (women). Longcope and Tait (1971) measured steady-state plasma E2 concentrations in male volunteers given an iv loading dose (~29 ng) followed by 10-h infusions at rates between 6–15 ng/h. E2 kinetics following iv bolus administration in humans were reported in experiments conducted by Kuhnz et al.(1993) in which premenopausal women were given an E2 dose of 0.3 mg. No data were available for iv bolus administration of E2 in men. The model was calibrated using the infusion data of Hembree et al.(1969), and verified using the infusion data from Longcope and Tait (1971) as well as the iv bolus kinetics of Kuhnz and coworkers (1993).

Model predictions of steady-state E2 plasma concentrations following iv infusions of E2 (Fig. 9Go) were in good agreement with data for both human female (Hembree et al., 1969Go) and male subjects (Longcope and Tait, 1971Go; results not shown). The Bmax for both albumin and SHBG were set to zero in the liver plasma compartment, and separate male and female intrinsic clearance rates were estimated by fitting the plasma E2 time course. Greater hepatic and extrahepatic clearance rates were required in simulations for men compared to women (Table 2Go). Using the same model parameters as for the iv infusions the simulated plasma concentration-time curve in premenopausal women following iv bolus administration of 0.3 mg E2 was in good agreement with the observed data (Kuhnz et al., 1993Go; Fig. 10Go).



View larger version (18K):
[in this window]
[in a new window]
 
FIG. 9. Simulations of intravenous infusion studies conducted in human subjects (Hembree et al., 1969Go). Women (panel A) and men (panel B) were given an iv bolus loading dose of E2 (~0.34 nmol) followed 30 min later by a 2-h infusion with E2 (0.17–0.56 nmol/h). Simulations are shown as lines and each set of symbols represents plasma E2 concentrations measured for an individual subject.

 


View larger version (15K):
[in this window]
[in a new window]
 
FIG. 10. E2 plasma concentrations following iv administration of 0.3 mg E2 to premenopausal women. The model prediction is represented by the solid line, and each data point represents the mean plasma E2 concentration (± SD) of 14 subjects (Kuhnz et al., 1993Go).

 
Oral administration.
Test data for po pharmacokinetics of E2 in postmenopausal women were available in several reports (Cassidenti et al., 1990Go; Lyrenas et al., 1981Go; Yen et al., 1975Go) and cover administered doses ranging from 1–4 mg. Kuhnz et al.(1993), studied the plasma kinetics of E2 in premenopausal women given po doses of 2, 4, and 8 mg. All studies used micronized E2 formulations.

The model did not perform as well for orally administered E2 as it did for intravascular doses. Simulations of E2 plasma kinetics in women who received 1–8 mg of micronized E2 are shown in Figure 11Go. Initially, a Clint of 150 l/h, equivalent to the value identified in the infusion studies was used, and the po absorption and movement from the GI tract compartment to the feces was manually adjusted to obtain the best fit. The rate constant for po uptake varied from 0.005/h to 0.013/h and the rate constant for fecal elimination varied from 0.0/h to 0.04/h. Peak concentrations were approximated, but the Tmax was off by several hours for all 6 data sets. Although the model gives reasonable predictions of the peak plasma concentrations, it fails to accurately predict the time at which maximum plasma concentrations are achieved and the general shape of the curve.



View larger version (25K):
[in this window]
[in a new window]
 
FIG. 11. Model predictions (lines) of plasma E2 concentrations in human female subjects given micronized E2 orally. Kinetics following E2 doses of 2 and 4 mg are shown in (A) and (B), respectively (Lyrenas et al., 1981Go). (C) and (D), plasma kinetics after E2 doses of 1 and 2 mg (Cassidenti et al., 1990Go). (E) and (F), plasma kinetics after doses of 2 mg (Yen et al., 1975Go) and 4 mg (Kuhnz et al., 1993Go), respectively. Symbols represent experimental data points.

 
Sensitivity Analysis
The sensitivity of plasma and uterine E2 concentrations to various model parameters was assessed for iv infusion experiments in rats. Time-series sensitivity coefficients (SCs) are shown for these experiments in Figure 12Go. Single-point SCs are also listed for infusion simulations and represent the time at which steady state was reached. A SC of zero or near zero indicates the response variable (concentration) is not sensitive to the model parameter. Sensitivity coefficients greater than zero indicate that the model parameter will affect the response variable—the SC equating to the change in the output relative to the change in the parameter. Sensitivity analysis identifies those model parameters with the greatest influence on E2 kinetics. If these parameters are not known with a high degree of certainty, they become the best candidates for further study in order to reduce the uncertainty in this PK model



View larger version (26K):
[in this window]
[in a new window]
 
FIG. 12. Time-series sensitivity coefficients for E2 concentrations in uterus (A) and plasma (B) with respect to various model parameters. Sensitivity analysis was conducted for simulations of 4-h infusion experiments in rats (65 nmol E2/h).

 
Sensitivity analysis of infusion simulations confirm that infusion rate, extrahepatic clearance rate, and hepatic blood flow (a surrogate for clearance under perfusion-limited conditions) are the most sensitive model parameters determining steady-state plasma concentrations (Table 6Go), whereas, time-series SC show that the shape of the plasma concentration-time curve is more sensitive to ER binding affinity (Kd) and ER content (Bmax) in tissues (Figure 12AGo). Conversely, uterine E2 concentrations at steady-state are most sensitive to ER content and E2:ER Kd, while the shape of the plasma concentration curve is sensitive to dose rate, diffusional clearance (PAU), and E2 binding in other tissues (Table 5Go and Fig. 12BGo). Uterine concentrations are insensitive to changes in uterine volume if they occur with equivalent changes in blood flow and ER content (data not shown).


View this table:
[in this window]
[in a new window]
 
TABLE 6 Normalized Sensitivity Coefficients (SCs) for Model Parameters with Respect to Steady-State E2 Concentrations in Plasma and Uterine Tissue
 
The sensitivity of steady state plasma and uterine concentrations to dose rate demonstrates the importance of adequate characterization of the rate and timing of E2 delivery, independent of route. Particularly important for future experimental work directed towards reducing model uncertainty will be advancing our understanding of the processes controlling po uptake of E2, and measurement of po uptake rates to support revision of the simple approach applied here. Parameters such as the E2:ER Kd, hepatic blood flow and uterine estrogen receptor content, are in comparison, well known. However, further characterization of the influence of OVX and E2 treatment on regulation of uterine or other response or storage tissue ER levels will also be important. Uterine E2 concentrations are moderately sensitive to the uterine plasma E2:Albumin binding constants, parameters that were fit. Overcoming uncertainty in tissue plasma compartment E2 binding in general, will involve additional experimental work directed towards the processes influencing tissue uptake of highly bound compounds and parallel efforts to develop modeling approaches representing these processes. Quantitatively less important than dose rate to plasma E2 levels at steady state, extrahepatic clearance, uncharacterized in the rat, is also a target for additional experimental work.


    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Estradiol PBPK Model
Estradiol pharmacokinetics in rats and humans are clearly dependent on many biological factors. In the process of constructing and testing the physiological models presented in this article, some of the more important determinants of E2 disposition have been evaluated. Estrogen receptor binding, metabolic clearance, plasma protein binding, and tissue uptake were of particular interest since they are thought to directly and indirectly regulate the physiologic responses to E2.

The extensive binding of E2 by both albumin and SHBG, allows only a small percentage (2–5%) of E2 to circulate unbound (Dunn, 1983Go; Pardridge, 1986Go). The early work of Brodie et al.(1960) introduced the idea that only unbound drug in plasma is available to interact with membrane receptors or diffuse into tissues to elicit pharmacologic effects. Single pass extraction experiments in rats designed to study E2 uptake into target organs indicate binding to either albumin or SHBG can differentially affect tissue availability of E2 (Pardridge and Mietus, 1979aGo,bGo; Verheugen et al., 1984Go). In most cases E2 extraction is significantly greater than would be expected for uptake of unbound E2 alone. Simulations of hepatic extraction predicting limited uptake of E2 by the liver in the presence of albumin (0.24) and even less extraction with both albumin and SHBG (0.03) are inconsistent with in vivo experiments, which report a single pass extraction of ~0.85 (Verheugen et al., 1984Go). Pardridge et al. has demonstrated that in vitro protein binding constants for E2 are not predictive of in vivo uptake in the liver, brain, or uterus (Pardridge, 1981Go, 1986Go; Pardridge and Landaw, 1985Go). Instead, an apparent Kd (Kd(app)) for individual organs can be used to reflect better the bioavailable E2 fraction in the plasma passing through each organ, where Kd(app) is related to the in vitro Kd by the E2 flux rate through the capillary membrane and capillary transit time (Pardridge and Landaw, 1985Go). Where the capillary transit time is large relative to the E2 disassociation constant and the plasma tissue concentrations gradient is also large (for highly cleared tissues like the liver), protein binding restricts tissue uptake less than would be predicted based on free concentrations (Mendel, 1992Go). In the PBPK model presented here, a Kd(app) was used only for the uterus and slowly perfused tissue compartments. The Kd(app) in the liver plasma compartment was set high enough to effectively disable protein binding; E2 was free for hepatic uptake and clearance. It appears that capillary transit time is sufficiently long relative to dissociation rate of E2 from the plasma binding proteins that protein binding does not significantly restrict hepatic extraction, but rather acts as a reservoir for E2. This suggests that hepatic uptake of E2 is dependent on the total E2 concentration in the blood and not simply free E2, a concept consistent with the "free hormone transport hypothesis" proposed by Mendel (1992).

Because tissue E2 data are only available for adult rats, where only albumin binding is significant, the potential for SHBG to restrict E2 uptake in most human tissues cannot be evaluated. Model simulations of E2 kinetics in humans were only predictive when hepatic plasma protein binding was eliminated, suggesting that at least E2 uptake by the liver is not significantly restricted by SHBG. This has been confirmed experimentally by Pardridge and Mietus (1979a) who found >82% of E2 in human plasma is extracted on a single pass through the rat liver. The intrinsic clearance rates used in the model led to similar, extensive, extraction of E2 by the liver. Extraction will increase with transit time, which may be different in humans.

Metabolic clearance of E2 was restricted to the liver and plasma compartments of the model. In the liver, the rate of elimination was defined as the product of intrinsic clearance (Clint) and the free E2 concentration in liver tissue (Clfree). The total plasma metabolic clearance rate (MCRP), 0.45 l/h for the female rat, falls well within the reported range of values (MCRP = 0.36–0.97 l/h (Ball et al., 1983Go; De Hertogh et al., 1970Go; Kono et al., 1981Go; Larner and Hochberg, 1985Go; Tapper and Brown-Grant, 1975Go). The same analysis of Clint for men and women yield MCRPs consistent with values reported for humans (Hembree et al., 1969Go; Longcope et al., 1968Go; Longcope and Tait, 1971Go). Hepatic extraction was between 0.70 and 0.85 for rats and humans, consistent with the observed first pass extraction rate of 0.85 in rats (Verheugen et al., 1984Go). When expressed as a percentage of the total MCR, extrahepatic clearance represent ~22% of total clearance, and is consistent with the range of 15–25% reported by Longcope et al.(1968). The model required higher Clint values for male rats and humans as compared to females. This is consistent with the MCR differences reported between sexes and may be due to inherently greater activity of some enzymes involved in E2 metabolism. For example, 17{alpha}-2- and 4-hydroxylase activities are ~5-fold higher in microsomes from male Sprague-Dawley rats as compared to females (Dannan et al., 1986Go).

Hepatic blood flow appears to be one of the most important sources of experimental variability in determining the overall metabolic clearance of E2 (Longcope and Tait, 1971Go). This is confirmed by this PBPK model by estimated hepatic clearance rates and a sensitivity analysis that indicated near perfusion limited hepatic clearance.

As confirmed by the sensitivity analysis, the E2 concentration achieved in plasma under steady state infusion conditions is dependent on only the infusion and clearance rates (Gibaldi and Perrier, 1982Go), and is independent of other parameters such as partition coefficients, receptor binding, blood flows, diffusional clearance rates, and organ volumes. However, these parameters do have an impact on the time required to reach steady state and the shape of the concentration-time curve, as well as steady-state concentrations in tissues. Altering Kd not only changes the shape of the plasma concentration-time curve, but also impacts the steady-state concentration of E2 in the uterus. This is an important consideration in light of the recent finding of two distinct ER subtypes (ER{alpha} and ERß) possessing different dissociation constants and which have differential tissue distribution (Brandenberger et al., 1997Go; Kuiper et al., 1997Go). Uncertainty around ER content and binding affinities will greatly impact model predictions of uterine tissue dose and ultimately estimates of receptor occupation and tissue response characteristics.

The uterine ER content used in the model provided reasonable fits to all of the data sets, however, the uptake of E2 by the uterus appeared dependent on hormonal status. Uptake behavior of E2 in the ovariectomized rat, but not intact rats, was best characterized as diffusion-limited. E2 stimulation of uterine hyperemia, edema, albumin (Peterson and Spaziani, 1971Go), and trypan blue uptake (Cecil et al., 1966Go) in ovariectomized rats are all indications of enhanced uterine vascular permeability (Cullinan-Bove and Koos, 1993Go). The diffusion-limited behavior needed to describe E2 uptake in the ovariectomized rat may reflect a reduced state of vascular permeability compared to the intact state where the uterus is continuously exposed to endogenous E2. Because estrogen receptor (ER) binding is included in the uterus compartment of the model, estimates of receptor occupancy as a tissue dose metric are easily obtained. Assuming the E2 production rate in the rat ovary is ~8 ng/h, model simulations predict approximately 2% occupancy of ER in the uterus, which is identical to the estimate made by Eisenfeld (1967). Sensitivity of uterine E2 concentrations to both ER Bmax and Kd indicates uncertainty around these parameters would introduce significant error into tissue dosimetry estimates such as receptor occupancy. For this reason, ER values that are relevant to species, strain, age, and hormonal status should be used in the model when possible.

E2 distribution in the rat and human were clearly dependent on model parameters such as tissue partitioning and binding and diffusional clearance rates. Improved fits to the experimental iv data were achieved by increasing receptor binding in the slowly perfused compartment or increasing binding to a nonreceptor binding site in the liver. For male rats, this may represent the male-specific hepatic binding protein described by Rogerson and Eagon (1986), especially since the reported (Kd = 31–43 nM and Bmax = ~20 nmol/liver) and model estimated binding constants (Kd = 35 nM and Bmax = 65 nmol/liver) are similar. The need for additional binding in the lumped tissue compartment and in the female liver simply imply there is additional binding (specific or nonspecific) that is not accounted for at the model's current level of detail.

Oral absorption kinetics of E2 in rats and humans presented some difficulty. Only intraduodenal administration data were available for the male rat, while the dosage form for humans was micronized E2 (Estrace®). Although the model consistently provided reasonable estimates for the maximum plasma E2 concentrations reached, the shape of the curves did not correspond to the data. Likely, this is due to the complex nature of E2 dissolution and gut transit rates of micronized E2 that are not addressed in the model. It was necessary to increase hepatic extraction from 0.75 to 0.91 in the human and from 0.84 to 0.98 in the rate in order to fit the rat and human data; otherwise, plasma concentrations were over predicted. To accomplish these increases, intrinsic clearance (Clint) was increased from 3.0 l/h to 25.0 l/h and from 150 l/h to 500 l/h in humans and rats, respectively. Oral bioavailability of E2 is only 2–5% in both rats and humans (Bawaarshi-Hassar et al., 1989Go; Kuhnz et al., 1993Go) and is the result of first-pass metabolism by the intestinal mucosa and liver (Longcope et al., 1985Go; Meli et al., 1968Go). Using a higher intrinsic clearance value for the oral route of administration may account for additional metabolism in the gut wall. Alternative approaches—slower uptake and lower first pass metabolism—were explored and improved simulations of oral route pharmacokinetics. These more speculative analyses only suggest that additional experimental work is called for to improve our understanding of oral uptake of E2, and were therefore not presented here.

The PBPK model developed here provides a flexible tool for integrating existing biological knowledge of the processes governing E2 pharmacokinetics, evaluating their significance and impact, as well as providing predictions of in vivo binding potency and tissue dosimetry. Overall, the simulations from our PBPK model are consistent with E2 pharmacokinetics in rats and in humans for a variety of dosing regimens. The model will be central to ongoing efforts to characterize the toxicity and dose-response relationships for the prototypical EAC, E2, and by extension, to other EACs.

Quantitative Framework for Evaluating Responses to E2 and Other EAC Compounds
The distinction between adverse and normal responses to endocrine-active compounds is a function of the timing, magnitude, and duration of target tissue exposure (PK), as well as the affinity for the receptor (e.g., ER) and the potency of the compound. Efforts to quantify the potential for EACs to cause toxicity initially focused on in vitro measures of receptor affinity. Without adequate treatment of the processes controlling pharmacokinetics—absorption, tissue distribution, metabolic and other clearances and restriction of free concentrations by plasma protein binding—the results of these experiments can be misleading.

The E2 model presented here is a biologically motivated quantitative framework that simultaneously treats both pharmacokinetic processes and receptor binding, and is readily adaptable to other EACs. Integration of this model with response models for specific tissues will ultimately create more biologically realistic models for studying the pharmacodynamics of E2 as well as other EACs.

The model is best used as an integrated part of an experimental program, used to evaluate experimental data and guide the development and evolution of and effective experimental agenda. The initial step, population of the model with in vitro derived parameter estimates of clearance, tissue partitioning, plasma protein binding, and receptor binding, facilitates evaluation of processes controlling tissue dosimetry, allowing informed decisions regarding the priority of additional experimental work. After necessary revisions to initial parameter estimates based on additional experimental work have been made and satisfactory representations of pharmacokinetics are achieved, the model can be linked to pharmacodynamic models of common response assays such as the uterotrophic response to complete the construction of a biologically motivated dose-response model. Models such as this are the best available tool for integrating quantitative knowledge regarding the processes which ultimately determine the potency of endocrine-active compounds. As such they should be seen as a useful, flexible tool for evaluating and interpreting experimental data, making informed decisions regarding research priorities directed at evaluation of the overall potency of these compounds and well as making well supported predictions.

Uncertainties, Data Gaps, and Recommendations
The nature of constructing PBPK models requires a simplification and reduction of the complexities found in biological systems, and thus by default introduces a degree of model uncertainty. In addition, the test data sets and model parameters, which were obtained from numerous sources, are potential sources of parameter uncertainty derived from measurement and sampling errors. The following discussion points address some of these uncertainties and recommend potential research to fill data gaps.

The model assumes E2 pharmacokinetics in plasma and tissues are independent of the pharmacologic activity of E2. In ovariectomized and immature animals, E2 alters uterine blood flow, vascular permeability, uterine size, and receptor levels. This list of physiologic responses implies E2 may affect its own kinetics. This could be evaluated by linking the current PBPK model to a pharmacodynamic response model for E2 effects on uterine physiology. Furthermore, an integrated model of this type will help elucidate the relationships between E2 pharmacokinetics and pharmacodynamics. Clearly, this model can be extended to compliment other work on endocrine-active compounds. Sensitivity analyses clearly indicated that ER content and binding affinity were important determinants of tissue E2 concentrations. Although ER levels have been reported for many tissues, uses of these data are confounded by nonstandardized assay methodology, inconsistent presentation of the results, and the use of surgically or hormonally manipulated test animals. All of these issues make use of published tissue receptor levels problematic. Ideally, ER tissue measurements and pharmacokinetic studies should be matched as closely as possible in terms of strain, age, and hormonal status.

Simulation results of E2 pharmacokinetics in intact and ovariectomized rats suggest tissue uptake and plasma pharmacokinetics are dependent on hormonal status. Not enough data are currently available to confirm this, pointing to a need for additional plasma and tissue kinetic studies designed to examine E2 kinetics in mature intact and ovariectomized rats.

ER binding in tissues was described using equilibrium binding constants. In order to describe the long terminal elimination phase seen following iv bolus dosing, additional binding in the liver and poorly perfused tissue compartment was required beyond the expected ER tissue content. An alternate approach worth exploring is to use a kinetic description for receptor binding in tissues that explicitly describe ligand on- and off-rates for the receptor and ligand-receptor interactions with DNA binding sites. This approach may better describe the tissue retention of E2 and not require additional binding sites in the model. A further advantage of this approach is that the measurements of E2-ER to DNA binding regions would provide a dosimeter closer to the final tissue response. It is not clear whether the data to validate such a model are available.

As mentioned above, tissue ER content was assumed to be time invariant. This is clearly an oversimplification of true receptor dynamics, which include the processes of synthesis, degradation, recycling, and replenishment (Clark et al., 1977Go; Katzenellenbogen, 1980Go). These processes may become very important in explaining E2 tissue kinetics during multiple or prolonged exposures to E2 and ultimately influence tissue response. Exploration of this would clearly be of value for specific tissue response models such as those under investigation for the uterus at ICF Kaiser.

ER subtypes were not distinguished in the model. Differences in binding affinity for E2 and tissue distribution patterns may impact tissue dose estimates. A useful extension of the current model would be to parameterize the model so each target tissue contained appropriate ER subtype binding constants.

Estimates of intrinsic hepatic clearance could be refined by studies designed to determine the actual kinetic constants for the biotransformation of E2 to estrone. Estrone is formed by the further oxidation of estradiol at the C17 position and is the primary metabolite of estradiol seen upon iv and oral administration (Kuhnz et al., 1993Go). Furthermore, the kinetics of estrone may be worth including in future models since there is some speculation that estrone is converted to E2 in target tissues (Lobo and Cassidenti, 1992Go).

The absence of time course data for target organs such as the brain, gonads, and storage compartments such as those comprising the richly and poorly perfused tissues (muscle, fat, kidney) also limit the current description of E2 pharmacokinetics. Regulatory control of serum estrogenicity is expected to provide some buffering against exposure to E2, but the range over which this compensation is effective and the timing and degree of control are not known. Adequate understanding of these processes may be particularly important for simulation of chronic exposures, where the compensation is more likely than for short infusions or bolus dose experiments.

Although the E2 response data is plentiful for rats, direct comparison between studies or with E2 pharmacokinetic studies are extremely difficult due to the diversity of experimental designs employed. Typically, an effort is made to maximize response to E2 by surgical or pharmacological manipulation of the test animals. Unfortunately, corresponding tissue and plasma E2 concentration data are not usually available for comparison. A standard set of pharmacokinetic studies for the most commonly used response testing paradigms would help address this problem.


    ACKNOWLEDGMENTS
 
We thank Dr. Melvin Andersen of Colorado State University for his characteristically insightful, leading, and important comments during the development of this article. We also acknowledge the American Chemistry Council's Endocrine Toxicology Implementation Panel for financial support.


    NOTES
 
1 To whom correspondence should be sent at present address: Pfizer Inc., Groton Laboratories, Mail slot 4098, Eastern Point Road, Groton, CT 06340. E-mail: david_r_plowchalk{at}groton.pfizer.com. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Andersen, M. E. (1981). A physiologically based toxicokinetic description of the metabolism of inhaled gases and vapors: Analysis at steady state. Toxicol. Appl. Pharmacol. 60, 509–526.[ISI][Medline]

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, 38–50.[Abstract/Free Full Text]

Andersen, M. E., Clewell, J. H., III, Gearhart, J., and Allen, B. C. (1997). A pharmacodynamic model of the rat estrus cycle in relation to endocrine disruptors. J. Toxicol. Environ. Health 51, 189–209.[ISI][Medline]

Aten, R. F., Dickson, R. B., and Eisenfeld, A. J. (1978). Estrogen receptor in adult male rat liver. Endocrinology 103, 1629–1635.[Abstract]

Attardi, B., and Palumbo, L. (1981). Effects of nafoxidine on the luteinizing hormone surge: Temporal distribution of estrogen receptors and induction of cytoplasmic progestin receptors in the hypothalamus-preoptic area, pituitary, and uterus of the immature rat. Endocrinology 109, 1365–1374.[Abstract]

Badawi, A., Cavalieri, E., and Rogan, E. (2001). Role of human cytochrome P450 1A1, 1A2, 1B1, and #A4 in the 2-, 4-, and 16-{alpha}-hydroxylation of 17ß-estradiol. Metab. Clin. Exp. 50, 1001–1003.[ISI][Medline]

Baird, D. (Ed.) (1973). The Secretion of Estrogen from the Ovary in Normal and Abnormal Menstrual Cycles. American Elsevier, New York.

Ball, P., Emons, G., Kayser, H., and Teichmann, J. (1983). Metabolic clearance rates of catechol estrogens in rats. Endocrinology 113, 1781–1783.[Abstract]

Bawaarshi-Hassar, R. N., Hussain, A. A., and Crooks, P. A. (1989). Nasal absorption and metabolism of progresterone and 17beta-estradiol in the rat. Drug Metab. Dispos. 17, 248.[Abstract]

Biegel, L. B., Flaws, J. A., Hirshfield, A. N., O'Connor, J. C., Elliott, G. S., Ladics, G. S., Silbergeld, E. K., Van Pelt, C. S., Hurtt, M. E., Cook, J. C., and Frame, S. R. (1998). 90-day feeding and one-generation reproduction study in Crl:CD BR rats with 17ß-estradiol. Toxicol. Sci. 44, 116–142.[Abstract]

Brandenberger, A. W., Tee, M. K., Lee, J. Y., Chao, V., and Jaffe, R. B. (1997). Tissue distribution of estrogen receptors-{alpha} (ER-{alpha}) and -ß (ER-ß) mRNA in the midgestational human fetus. J. Clin. Endocrinol. Metab. 82, 3509–3512.[Abstract/Free Full Text]

Brodie, B., Kurtz, H., and Schanker, L. (1960). The importance of dissociation constant and lipid solubility on influencing the passage of drugs into the CSF. J. Pharmacol. Exp. Ther. 130, 20–25.[ISI][Medline]

Bruce, N. W. (1976). The distribution of blood flow to the reproductive organs of rats near term. J. Reprod. Fertil. 46, 359–362.[Abstract]

Cassidenti, D. L., Vijod, A. G., Vijod, M. A., Stanczyk, F. Z., and Lobo, R. A. (1990). Short-term effects of smoking on the pharmacokinetic profiles of micronized estradiol in postmenopausal women. Am. J. Obstet. Gynecol. 163, 1953–1960.[ISI][Medline]

Cecil, H. C., Hannum, J. A., Jr., and Bitman, J. (1966). Quantitative characterization of uterine vascular permeability changes with estrogen. Am. J. Physiol. 211, 1099–1102.[Free Full Text]

Clark, J., Paszko, Z., and Peck, E., Jr. (1977). Nuclear binding and retention of the receptor estrogen complex: relation to the agonistic and antagonistic properties of estriol. Endocrinology 100, 91–96.[Abstract]

Clark, J. H., Campbell, P. S., and Peck, E. J., Jr. (1972). Receptor-estrogen complex in the nuclear fraction of the pituitary and hypothalamus of male and female immature rats. Neuroendocrinology 10, 218–228.[ISI][Medline]

Clark, J. H., and Mani, S. K. (1994). Actions of ovarian steroid hormones. In The Physiology of Reproduction (E. Knobil and J. D. Neill, Eds.), pp. 1011–1059. Raven Press, New York.

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.), pp. 1–245. Springer-Verlag, New York.

Cullinan-Bove, K., and Koos, R. D. (1993). Vascular endothelial growth factor/vascular permeability factor expression in the rat uterus: Rapid stimulation by estrogen correlates with estrogen-induced increases in uterine capillary permeability and growth. Endocrinology 133, 829–837.[Abstract]

Dahlberg, E. (1982). Characterization of the cytosolic estrogen receptor in rat skeletal muscle. Biochim. Biophys. Acta 717, 65–75.[ISI][Medline]

Dannan, G. A., Porubek, D. J., Nelson, S. D., Waxman, D. J., and Guengerich, F. P. (1986). 17ß-estradiol 2- and 4-hydroxylation catalyzed by rat hepatic cytochrome P-450: Roles of individual forms, inductive effects, developmental patterns, and alterations by gonadectomy and hormone replacement. Endocrinology 118, 1952–1960.[Abstract]

Danzo, B. (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.[ISI][Medline]

De Hertogh, R., Ekka, E., and Vanderheyden, I. (1971a). Equilibrium concentration of (6,7-3H)oestrone and (6,7-3H)17ß-oestradiol metabolites in female rat tissues, during long-term intravenous infusions. Acta Endocrinol. 66, 401–416.[ISI][Medline]

De Hertogh, R., Ekka, E., Vanderheyden, I., and Hoet, J. J. (1970). Metabolic clearance rates and the interconversion factors of estrone and estradiol-17ß in the immature and adult female rat. Endocrinology 87, 874–880.[ISI][Medline]

De Hertogh, R., Ekka, E., Vanderheyden, I., and Hoet, J. J. (1971b). In vivo observations on reversible binding of estradiol-17ß,6,7-3H in rat uterus. Measurement of a "biological exchange constant" and of the "tissue capacity." Endocrinology 88, 165–174.[ISI][Medline]

De Hertogh, R., Ekka, E., Vanderheyden, I., and Hoet, J. J. (1973). Subcellular distribution and binding of (6,7-3H) estradiol in rat uterus, at equilibrium, after long-term intravenous infusion. J. Steroid Biochem. 4, 289–299.[ISI][Medline]

Dickson, R. B., and Eisenfeld, A. J. (1979). Estrogen receptor in liver of male and female rats: Endocrine regulation and molecular properties. Biol. Reprod. 21, 1105–1114.[ISI][Medline]

Dionne, F. T., Lesage, R. L., Dube, J. Y., and Tremblay, R. R. (1979). Estrogen binding proteins in rat skeletal and perineal muscles: In vitro and in vivo studies. J. Steroid Biochem. 11, 1073–1080.[ISI][Medline]

Duffy, M. J., and Duffy, G. J. (1978). Estradiol receptors in human liver. J. Steroid Biochem. 9, 233–235.[ISI][Medline]

Dunn, J. F. (1983). Transport of estrogens in human plasma. In Catechol Estrogens (G. R. Merriam, Ed.), pp. 167–176. Raven, New York.

Dunn, J. F. (1984). Computer simulation of steroid transport in human plasma. Serono Symp. Publ. Raven Press 14, 277–285.

Dunn, J. F., Merriam, G. R., Eil, C., Kono, S., Loriaux, D. L., and Nisula, B. C. (1980). Testosterone-estradiol binding globulin binds to 2-methoxyestradiol with greater affinity than to testosterone. J. Clin. Endocrinol. Metab. 51, 404–406.[Abstract]

Eisenfeld, A. J. (1967). Computer analysis of the distribution of [3H]estradiol. Biochim. Biophys. Acta 136, 498–507.[ISI][Medline]

Eisenfeld, A. J., and Aten, R. F. (1979). Estrogen receptor in the mammalian liver. Adv. Steroid Biochem. Pharmacol. 7, 91–117.[Medline]

Farrell, G. C., Koltai, A., and Murray, M. (1988). Source of raised serum estrogens in male rats with portal bypass. J. Clin. Invest. 81, 221–228.[ISI][Medline]

Germain, B. J., Campbell, P. S., and Anderson, J. N. (1978). Role of the serum estrogen-binding protein in the control of tissue estradiol levels during postnatal development of the female rat. Endocrinology 103, 1401–1410.[Abstract]

Gibaldi, M., and Perrier, D. (1982). Pharmacokinetics. Marcel Dekker, New York.

Ginsburg, M., MacLusky, N. J., Morris, I. D., and Thomas, P. J. (1975). Physiological variation in abundance of oestrogen specific high-affinity binding sites in hypothalamus, pituitary and uterus of the rat. J. Endocrinol. 64, 443–449.[Abstract]

Guyton, A. C. (1991). Textbook of Medical Physiology. W.B. Saunders, Philadelphia.

Hagenfeldt, Y., and Eriksson, H. A. (1988). The estrogen receptor in the rat kidney. Ontogeny, properties and effects of gonadectomy on its concentration. J. Steroid Biochem. 31, 49–56.[ISI][Medline]

Hembree, W. C., Bardin, C. W., and Lipsett, M. B. (1969). Estrogen metabolic clearance rates and transfer factors. J. Clin. Invest. 48, 1809–1819.[ISI][Medline]

Hruska, R. E., and Olson, J. R. (1989). Species differences in estrogen receptors and in the response to 2,3,7,8-tetrachlorodibenzo-p-dioxin exposure. Toxicol. Lett. 48, 289–299.[ISI][Medline]

ILSI (1994). Physiological Parameter Values for PBPK Models. International Life Sciences Institute, Risk Sciences Institute.

Jehan, Q., Srivasta, S., Akhlaq, M., Ahmad, A., and Setty, B. S. (1982). Kinetics of distribution and retention of 3H-oestradiol-17ß in rat tissues: A comparative study with free oestradiol and after its incorporation into liposomes. Endokrinologie 80, 8–12.[ISI][Medline]

Jobling, S., Reynolds, T., White, R., Parker, M., and Sumpter, J. (1995). A variety of environmentally persistent chemicals, including some phthalate plasticizers, are weakly estrogenic. Environ. Health Perspect. 103, 582–587.[ISI][Medline]

Katzenellenbogen, B. S. (1980). Dynamics of steroid hormone receptor action. Annu. Rev. Physiol. 42, 17–35.[ISI][Medline]

Kedderis, L. B., Mills, J. J., Andersen, M. E., and Birnbaum, L. S. (1993). A physiologically based pharmacokinetic model for 2,3,7,8-tetrabromodibenzo-p-dioxin (TBDD) in the Rat: Tissue distribution and CYP1A induction. Toxicol. Appl. Pharmacol. 121, 87–98.[ISI][Medline]

Kerr, J. B., Maddocks, S., and Sharpe, R. M. (1992). Testosterone and FSH have independent, synergistic and state-dependent effects upon spermatogenesis in the rat testis. Cell Tissue Res. 268, 179–189.[ISI][Medline]

Khor, S. P., and Mayersohn, M. (1991). Potential error in the measurement of tissue to blood distribution coefficients in physiological pharmacokinetic modeling. Residual tissue blood. I. Theoretical considerations. Drug Metab. Dispos. 19, 478–485.[Abstract]

Klotz, D., Beckman, B., Hill, S., McLachlan, J., Walters, M., and Arnold, S. (1996). Identification of environmental chemicals with estrogenic activity using a combination of in vitro assays. Environ. Health Perspect. 104, 1084–1089.[ISI][Medline]

Kono, S., Brandon, D. D., Merriam, G. R., Loriaux, D. L., and Lipsett, M. B. (1981). Metabolic clearance rate and uterotropic activity of 2-hydroxyestrone in rats. Endocrinology 108, 40–43.[Abstract]

Kuhnz, W., Gansau, C., and Mahler, M. (1993). Pharmacokinetics of estradiol, free and total estrone, in young women following single intravenous and oral administration of 17ß-estradiol. Arzneim-Forsch/Drug Res. 43, 966–973.

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 a and b. Endocrinology 138, 863–870.[Abstract/Free Full Text]

Langlois, P. L. (1970). The size of the normal uterus. J. Reprod. Med. 4, 220–228.[ISI][Medline]

Larner, J. M., and Hochberg, R. B. (1985). The clearance and metabolism of estradiol and estradiol-17-esters in the rat. Endocrinology 117, 1209–1214.[Abstract]

Lees, P. D., Lynch, D. T., Richards, H. K., Lovick, A. H., Perry, S., and Pickard, J. D. (1992). Blood flow in portal systems with special reference to the rat pituitary gland. J. Cereb. Blood Flow Metab. 12, 128–138.[ISI][Medline]

Lobo, R. A., and Cassidenti, D. L. (1992). Pharmacokinetics of oral 17ß-estradiol. J. Reprod. Med. 37, 77–84.[ISI][Medline]

Longcope, C., Gorbach, S., Goldin, B., Woods, M., Dwyer, J., and Warram, J. (1985). The metabolism of estradiol; oral compared to intravenous administration. J. Steroid Biochem. 23, 1065–1070.[ISI][Medline]

Longcope, C., Layne, D. S., and Tarr, J. F. (1968). Metabolic clearance rates and interconversions of estrone and 17b-estradiol in normal males and females. J. Clin. Invest. 47, 93–106.[ISI]

Longcope, C., and Tait, J. F. (1971). Validity of metabolic clearance and interconversion rates of estrone and 17ß-estradiol in normal adults. J. Clin. Endocrinol. Metab. 32, 481–490.[ISI][Medline]

Lyrenas, S., Carlstrom, K., Backstrom, T., and von Schoultz, B. (1981). A comparison of serum oestrogen levels after percutaneous and oral administration of oestradiol-17ß. Br. J. Obstet. Gynecol. 88, 181–187.[ISI][Medline]

Manni, A., Pardridge, W. M., Cefalu, W., Nisula, B. C., Bardin, C. W., Santner, S. J., and Santen, R. J. (1985). Bioavailability of albumin-bound testosterone. J. Clin. Endocrinol. Metab. 61, 705–710.[Abstract]

Martucci, C. P., and Fishman, J. (1993). P450 enzymes of estrogen metabolism. Pharmacol. Ther. 57, 237–257.[ISI][Medline]

Masseyeff, R., Gilli, J., Krebs, B., Calluaud, A., and Bonet, C. (1975). Evolution of {alpha}-fetoprotein serum levels throughout life in humans and rats, and during pregnancy in the rat. Ann. N.Y. Acad. Sci. 259, 17–28.[ISI][Medline]

Meli, A., Cargill, D. I., Giannina, T., and Steinetz, B. G. (1968). Studies on the transport of estrogens by the rat small intestine in vivo. Proc. Soc. Exp. Biol. Med. 129, 937–944.

Mendel, C. M. (1992). The free hormone hypothesis: Distinction from the free hormone transport hypothesis. J. Androl. 13, 107–116.[Abstract]

Milligan, S., Balasubramanian, A., and Kalita, J. (1998). Relative potency of xenobiotic estrogens in an acute in vivo mammalian assay. Environ. Health Perspect. 106, 23–26.[ISI][Medline]

Murphy, J. E., Janzen, D. B., and Gargas, M. L. (1995). An in vitro method for determination of tissue partition coefficients of non-volatile chemicals such as 2,3,7,8-tetrachlorodibenzo-p-dioxin and estradiol. J. Appl. Toxicol. 15, 147–152.[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. Soc. Exp. Biol. Med. 217, 300–309.[Abstract]

Nishikawa, Y., Ikegami, H., Sakata, M., Mizutani, T., Morishige, K., Kurachi, H., Hirota, K., Miyake, A., and Tanizawa, O. (1993). Ovariectomy increases the level of estrogen receptor mRNA and estrogen receptor binding sites in female rat adipose tissue. J. Endocr. Invest. 16, 579–583.[ISI][Medline]

Nisula, B. C., and Dunn, J. F. (1979). Measurement of the testosterone binding parameters for both testosterone-estradiol binding globulin and albumin in individual serum samples. Steroids 34, 771–791.[ISI][Medline]

Notides, A. C. (1970). Binding affinity and specificity of the estrogen receptor of the rat uterus and anterior pituitary. Endocrinology 87, 987–992.[ISI][Medline]

Oyesanya, O. A., Parsons, J. H., Collins, W. P., and Campbell, S. (1995). Total ovarian volume before human chorionic gonadotrophin administration for ovulation induction may predict the hyperstimulation syndrome. Hum. Reprod. 10, 3211–3212.[Abstract]

Pardridge, W. M. (1981). Transport of protein-bound hormones into tissues in vivo. Endocr. Rev. 2, 103–123.[ISI][Medline]

Pardridge, W. M. (1986). Serum bioavailability of sex steroid hormones. Clin. Endocrinol. Metab. 15, 259–278.[ISI][Medline]

Pardridge, W. M., and Landaw, E. M. (1984). Tracer kinetic model of blood-brain barrier transport of plasma protein-bound ligands. Empiric testing of the free hormone hypothesis. J. Clin. Invest. 74, 745–752.[ISI][Medline]

Pardridge, W. M., and Landaw, E. M. (1985). Testosterone transport in brain: Primary role of plasma protein-bound hormone. Am. J. Physiol. 249, E534–E542.[Abstract/Free Full Text]

Pardridge, W. M., and Mietus, L. J. (1979a). Transport of protein-bound steroid hormones into liver in vivo. Am. J. Physiol. 237, E367–E372.[ISI][Medline]

Pardridge, W. M., and Mietus, L. J. (1979b). Transport of steroid hormones through the rat blood-brain barrier. Primary role of albumin-bound hormone. J. Clin. Invest. 64, 145–154.[ISI][Medline]

Pedersen, S. B., Fuglsig, S., Sjogren, P., and Richelsen, B. (1996a). Identification of steroid receptors in human adipose tissue. Eur. J. Clin. Invest. 26, 1051–1056.[ISI][Medline]

Pedersen, S. B., Hansen, P. S., Lund, S., Andersen, P. H., Odgaard, A., and Richelsen, B. (1996b). Identification of oestrogen receptors and oestrogen receptor mRNA in human adipose tissue [see comments]. Eur. J. Clin. Invest. 26, 262–269.[ISI][Medline]

Peterson, R. P., and Spaziani, E. (1971). Autoradiographic localization of albumin-131I accumulation in the estrogen-stimulated rat uterus. Endocrinology 89, 1280–1286.[ISI][Medline]

Renoir, J. M., Mercier-Bodard, C., and Baulieu, E. E. (1980). Hormonal and immunological aspects of the phylogeny of sex steroid binding plasma protein. Proc. Natl. Acad. Sci. U.S.A. 77, 4578–4582.[Abstract]

Robbins, S., Cotran, R., and Kumar, U. (1989). Pathologic Basis of Disease. Saunders, Philadelphia.

Rogerson, B. J., and Eagon, P. K. (1986). A male specific hepatic estrogen binding protein: characteristics and binding properties. Arch. Biochem. Biophys. 250, 70–85.[ISI][Medline]

Romkes, M., Piskorska-Pliszczynska, J., and Safe, S. (1987). Effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin on hepatic and uterine estrogen receptor levels in rats. Toxicol. Appl. Pharmacol. 87, 306–314.[ISI][Medline]

Rosner, W. (1991). Plasma steroid-binding proteins. Endocrinol. Metab. Clin. North Am. 20, 697–720.[ISI][Medline]

Rosner, W., and Smith, R. N. (1975). Isolation and characterization of the testosterone-estradiol-binding globulin from human plasma. Use of a novel affinity column. Biochemistry 14, 4813–4820.[ISI][Medline]

Setchell, B. P. (1978). The Mammalian Testis. Cornell University Press, Ithaca, NY.

Shelby, M. D., Newbold, R. R., Tully, D. B., and Davis, V. L. (1996). Assessing environmental chemicals for estrogenicity using a combination of in vitro and in vivo assays. Environ. Health Perspect. 104, 1296–1300.[ISI][Medline]

Shouwei, H., Zhilan, P., Qianping, T., Hu, Z., and Lingling, W. (1993). Analysis of estrogen receptors in normal bone and bone tumor tissues. J. West. China Univ. Med. Sci. 4, 160–162.

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, 113–122.[ISI][Medline]

Stahl, S., Chun, T., and Gray, W. (1998). Phytoestrogens act as estrogen agonists in an estrogen-responsive pituitary cell line. Toxicol. Appl. Pharmacol. 152, 41–48.[ISI][Medline]

Stock, M. K., and Metcalfe, J. (1994). Maternal physiology during gestation. In The Physiology of Reporduction (E. Knobil and J. D. Neill, Eds.), pp. 947–983. Raven Press, New York.

Sudo, K., Monsma, F. J., Jr., and Katzenellenbogen, B. S. (1983). Antiestrogen-binding sites distinct from the estrogen receptor: Subecellular localization, ligand specificity, and distribution in tissues of the rat. Endocrinology 112, 425–434.[ISI][Medline]

Tapper, C. M., and Brown-Grant, K. (1975). The secretion and metabolic clearance rates of oestradiol in the rat. J. Endocrinol. 64, 215–227.[Abstract]

van der Walt, L. A., Sanfilippo, J. S., Siegel, J. E., and Wittliff, J. L. (1986). Estrogen and progestin receptors in human uterus: Reference ranges of clinical conditions. Clin. Physiol. Biochem. 4, 217–228.[ISI][Medline]

Verheugen, C., Pardridge, W. M., Judd, H. L., and Chaudhuri, G. (1984). Differential permeability of uterine and liver vascular beds to estrogens and estrogen conjugates. J. Clin. Endocrinol. Metab. 59, 1128–1132.[Abstract]

Waynford, H. B., and Flecknell, P. A. (1992). Experimental and Surgical Technique in the Rat. Academic Press, London.

Yen, S. S., Martin, P. L., Burnier, A. M., Czekala, N. M., Greaney, M. O., Jr., and Callantine, M. R. (1975). Circulating estradiol, estrone and gonadotropin levels following the administration of orally active 17ß-estradiol in postmenopausal women. J. Clin. Endocrinol. Metab. 40, 518–521.[Abstract]

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, R569–R575.[Abstract/Free Full Text]