A Concentration Addition Model for the Activation of ... - Oxford Journals

4 downloads 0 Views 479KB Size Report
including indomethacin, carbon tetrachloride, cocaine, digitoxin, and parathion (Robinson et al., 1976, 1979; Solomon et al.,. 1971). However, we must be ...
TOXICOLOGICAL SCIENCES 107(1), 93–105 (2009) doi:10.1093/toxsci/kfn206 Advance Access publication October 1, 2008

A Concentration Addition Model for the Activation of the Constitutive Androstane Receptor by Xenobiotic Mixtures William S. Baldwin*,1 and Jonathan A. Roling† *Institute of Environmental Toxicology, Clemson University, Pendleton, South Carolina 29670; and †Biological Sciences, Bridgewater State College, Bridgewater, 02325 Massachusetts Received July 25, 2008; accepted September 20, 2008

The effects of contaminants are typically studied in individual exposures; however, environmental exposures are rarely from a single contaminant. Therefore, the study of chemical mixtures is important in determining the effects of xenobiotics. The constitutive androstane receptor (CAR) responds to endobiotics and xenobiotics, and in turn induces detoxification enzymes involved in their elimination. First, we compared several androgens as inverse agonists, including androgens allegedly used by Bay Area Laboratory Co-operative to enhance athletic performance. CAR inverse agonists ranked in order of potency were dihydroandrosterone (DHA) > tetrahydrogestrinone (THG) > androstanol > norbolethone. Therefore, we used DHA as an inverse agonist during transactivation assays. Next, we examined the effects of several pesticides, plasticizers, steroids, and bile acids on CAR activation. Our data demonstrates that several pesticides and plasticizers, including diethylhexylphthalate, nonylphenol, cypermethrin, and chlorpyrifos activate CAR. Both full and partial CAR activators were discovered, and EC50 values and Hillslopes were determined for use in the concentration addition models. Concentration addition models with and without restraint values to account for partial activators were developed. Measured results from transactivation assays with a mixture of two to five chemicals indicate that the concentration addition model without restraints correctly predicts activity unless all of the chemicals in the mixture are partial activators, and then restraint values be considered. Overall, our data indicates that it is important to consider that we are exposed to a milieu of chemicals, and the efficacy of each individual chemical is not the sole factor in determining CAR’s activity in mixture modeling. Key Words: CAR; mixtures; androgens.

The constitutive androstane receptor (CAR; NR1I3) is an orphan nuclear receptor similar to its cousin the pregnane X recpeptor (PXR; NR1I2) that when activated acts as a master regulator of the phase I, II, and III enzymes and transporters critical for detoxification of steroids, bile acids, and xenobiotics (Chang, 2006; Chen et al., 2008; Wei et al., 2000; Zollner 1 To whom correspondence should be addressed at Clemson University, Inst of EnvironToxicology, Biological Sciences, 509 Westinghouse Rd., P.O. Box 709, Pendleton, SC 29670. Fax: 864-646-2277. E-mail: [email protected].

et al., 2006). Detoxification genes induced through CAR include several CYP family members (CYPs 1A, 2A, 2B, 2C, 3A) (Jackson et al., 2006; Wei et al., 2000; Wortham et al., 2007), sulfotransferases, uridine diphospho-glucuronyltransferases (Chen et al., 2007; Huang et al., 2004a), and multiple transporters (Assem et al., 2004; Kast et al., 2002). These genes act in concert to metabolize, detoxify, and eliminate chemical hazards (Pascussi et al., 2008; Swales and Negishi, 2004). CAR’s ligand-binding pocket is smaller and less flexible than PXR’s (Suino et al., 2004; Watkins et al., 2001), and therefore CAR is less promiscuous than PXR (Suino et al., 2004). However, ligand binding is not required for CAR activation. For example, phenobarbital (PB) activates CAR through an AMP kinase phosphorylation cascade (Rencurel et al., 2005; Shindo et al., 2007). It has been hypothesized that the majority of CAR activators work through an indirect pathway (Shindo et al., 2007) and therefore many CAR activators are not referred to as ligands or agonists. Chemicals known to activate CAR include PB (Sueyoshi et al., 1999), 5b-pregnane-3,20-dione (Moore et al., 2000), metyrapone (Honkakoski et al., 2001), 1,4-bis-(2-(3,5,-dichloropyridyloxy)) benzene (TCPOBOP) (Tzameli et al., 2000), phenytoin (Wang et al., 2004), DDE (Wyde et al., 2003), garlic (Fisher et al., 2007), and nonylphenol (Hernandez et al., 2007), but the list of chemicals known to activate CAR is much shorter than the list of chemicals known to activate PXR (Kretschmer and Baldwin, 2005). CAR received its name because of its high constitutive activity, which can be repressed by inverse agonists such as the androstanes (Forman et al., 1998). The high constitutive activity of CAR in vitro makes the use of androstanes or other inverse agonists necessary when performing transactivation assays in order to increase the sensitivity of transactivation assays. The decreased sensitivity of CAR transactivation assays coupled with the difficulties of performing CAR activation assays with human CAR (hCAR), which lacks a potent inverse agonist, makes screening for hCAR activation more difficult than screening for hPXR activation. Although it is likely that CAR is not as promiscuous as PXR because of its smaller

Ó The Author 2008. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For permissions, please email: [email protected]

94

BALDWIN AND ROLING

ligand-binding pocket; reduced screening for CAR activators may account for the significantly shorter list of xenobiotic activators of CAR (Kretschmer and Baldwin, 2005). For the most part, short-term activation of CAR is an important defense mechanism for the protection from xenobiotics and toxic endobiotics; however, CAR activation has been implicated in drug-drug interactions. Phenytoin and PB activation of CAR are associated with increased vitamin D metabolism by CYP3A4 and a subsequent reduction in bone mineral density (Xu et al., 2006). The production of N-acetyl-p-benzoquinone imine, the toxic metabolite of acetaminophen (APAP), is significantly enhanced by CAR activation, and mice lacking CAR are not as susceptible to APAP toxicity (Zhang et al., 2002). However, we are not exposed to only one chemical at a time. We are often exposed to a milieu of different chemicals due to occupational, environmental, or pharmaceutical exposures (Kolpin et al., 2002; Lang, 1995), and CAR responds to these mixtures. The Agency for Toxic Substances and Disease Registry supports the identification of environmental mixtures, joint toxicity assessment, and experimental testing in order to improve our understanding of the toxicity and health effects of chemical mixtures (Mumtaz et al., 2002). In addition, The Food Quality Protection Act of 1996 (1996) requires that the United States Environmental Protection Agency consider the cumulative effects of pesticides that have a common mechanism of action. For example, low doses of a mixture of several chemicals that activate CAR may significantly induce detoxification enzymes important in xenobiotic metabolism and cause idiosyncratic adverse drug reactions (Xu et al., 2006; Zhang et al., 2002), alter bile acid homeostasis (Wagner et al., 2005), or perturb energy metabolism (Ding et al., 2006). Mixtures of chemicals may have activities that are greater than one chemical alone (additive or synergistic effects) or less than predicted (antagonistic). Chemical mixtures are considered additive when their combined effects produce a predictable response based on the measured effects of single chemicals (Bliss, 1939; Gennings et al., 2004; Meadows et al., 2002), and antagonistic or synergistic interactions can be determined using a previously published concentration addition model equation (Berenbaum, 1985). Thus, if all the chemicals work through the same mechanism and do not interact each chemical can be considered dilutions of the same chemical for the purpose of a concentration addition model. In general, two approaches are used to predict the toxicity of chemical mixtures: concentration addition or independent joint action (Drescher and Boedeker, 1995). Concentration addition assumes that chemicals share the same mechanism of action (Bliss, 1939), and is the method we use in this study because we are studying chemical activation of CAR. The objective of our study was to determine the additive effects of multiple CAR activators and produce a model that would explain how a mixture of CAR activators activates CAR. CAR activators may act as full or partial activators, and

partial activation appears to be relatively common. Therefore, a maximal response of 100% cannot be assumed for each chemical as the model would overestimate the responses. Consequently, we must consider whether chemicals act as partial activators. This was done by compressing each normalized dose-response curve by the percent of its maximal activity compared with the full agonist TCPOBOP. This allowed us to develop a model that considers the partial agonist-type activity of each individual chemical and test whether partial activators perturb the measured activity of a chemical mixture. There has been significantly less research performed investigating the activation of CAR by environmentally toxicants than PXR. Therefore, we first examined the ability of several environmentally relevant chemicals to activate CAR using dihydroandrosterone (DHA), a new CAR inverse agonist described here. Several CAR activators were selected for further characterization including the determination of EC50 and Hillslope values. Then we measured the ability of mixtures of two to five chemicals including both partial and full activators to work in concert to induce CAR activity using a luciferase-coupled reporter assay. Results were compared with a concentration addition model without restraints and two different concentration addition models with partial activator restraint values. A model was developed using the data derived above to putatively explain the action of a mixture of CAR activators.

MATERIALS AND METHODS Chemicals. Table S1 (supplemental data) provides a list of the chemicals obtained for the study of CAR activation, their CAS number, source, and their purity when available. Solvents were purchased from Fisher Scientific (Houston, TX) unless otherwise noted. Transactivation assays. The mouse CAR (mCAR) expression plasmid was a generous gift from Dr David Moore (Forman et al., 1998), and the CAR luciferase-coupled reporter was a generous gift from Dr Masahiko Negishi (Wang et al., 2003). Transactivation assays were performed in HepG2 human hepatoma cells (ATCC, Rockville, MD) cultured in phenol red free Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (HyClone, Logan, UT) and 1% Penicillin/ Streptomycin (Invitrogen) under 5% CO2 at 37°C as described previously (Hernandez et al. 2007). Briefly, HepG2 cells were plated in 12-well plates at 100,000 cells per well, and transfected 24 h later with 20 ng of the mCAR expression vector, and 100 ng of the reporter plasmid using Effectene reagent according to the manufacturer’s instructions (Qiagen, Valencia, CA). Transfected cells were treated the next day with 10lM inverse agonist to reduce CAR’s constitutive activity in combination with the chemical(s) of interest. All chemicals were dissolved in dimethyl sulfoxide (DMSO) and therefore all samples, treated and untreated, received 0.2% DMSO. Firefly luciferase activity was measured 24 h after the chemical treatment with the Steady-Glo luciferase reporter assay system (Promega, Madison, WI) according to the manufacturer’s protocol. The luminescence signal was normalized to transfected, DHA-treated cells by determining the ratio of the normalized signal from chemically treated cells over that of the CAR-transfected, DHA-treated cells. This is reported as fold activation. Data are presented as the mean of triplicate assays ± standard deviation from two to three separate assays. Statistical significance was

A CONCENTRATION ADDITION MODEL FOR THE ACTIVATION OF CAR determined by ANOVA followed by Tukey’s multiple comparison test as the post hoc test using the GraphPad Prizm 4.0 software package (La Jolla, CA). Determination of EC50 and Hillslope values. Multiple concentrations that spanned the dose-response curve were used to treat transfected HepG2 cells and luciferase activity measured. The luciferase assay data determined from the chemical responses were normalized as a percent of their maximal effect, log transformed, and plotted using a sigmoid dose-response curve using GraphPad Prizm 4.0 software in order to determine the Hillslope (slope of the doseresponse curve) and EC50 of each chemical. This is the same equation described by Rider and LeBlanc (2005) (Equation 1), where R equals the response to each chemical (0 ¼ no response and 100% ¼ maximal response). However, we use a different software program with different nomenclature, as GraphPad Prizm uses the term Hillslope instead of power. R¼



1  EC50 hs

ð1Þ

C

In addition, many CAR activators are partial activators and therefore we must consider the reduced activity of these chemicals. We used the concept of a restraint value, which is the percent measured activity of a partial activator compared with the full agonist, TCPOBOP. In turn, the normalized measured activity of each partial activator is multiplied by their restraint value, which compresses each data point on the concentration-response curve. This serves to compress the normalized dose-response curve, including the EC50, without altering the Hillslope. Ultimately, the EC50 values and Hillslopes are used to estimate the cumulative effects of mixtures using the model described below. CAR concentration addition models. The EC50 and Hillslope values were used to model the additive activity of the chemical mixtures using the formula below (Rider and LeBlanc 2005). The formula below (Equation 2) provided a predictive effect of our mixture of chemicals, where R equals the response of the mixture, Ci is the concentration of chemical i that causes a 50% response, and hs is the Hillslope. R¼

1  1þ

1 n P i¼1

hs

ð2Þ

Ci EC50i

Measured EC50 and Hillslope values were placed in the Computational Approach to the Toxicity Assessment of Mixtures (CATAM) model developed by Dr Gerald LeBlanc’s laboratory at North Carolina State University (http:// www.ncsu.edu/project/toxresearch/model5/linked_files/php_files/page3.php). The CATAM model we used predicts the effect of a chemical mixture by means of concentration addition with a prescribed set of chemicals using Equation 2. Concentration addition requires concentration-response curves from each of the individual chemicals and therefore is sometimes referred to as the single chemical required approach. Predictive concentration addition models were performed without restraint values, with restraint values considered for only the most potent partial activator in the mixture, and a weighted restraint value model that considered full activators as having a restraint value of 1.0 and determining the weighted restraint value of the chemical mixture based on the potency of each chemical. This allowed us to compare each of the three models to the measured responses. Fisher’s exact test (2 3 3) was performed to compare each model to the measured values and determine the model that fit the data best using Dr Richard Lowry’s Vassar College statistics page (http:// faculty.vassar.edu/lowry/VassarStats.html).

RESULTS

CAR Inverse Agonists and Activators: In comparison to PXR, there are few known chemical activators of CAR (Kretschmer and Baldwin, 2005). However,

95

before we could start looking for novel CAR activators we needed to optimize our transactivation assay and one way to do that is reduce CAR’s constitutive activity. Therefore, we tested a couple of synthetic and natural androgens, including androstanol (And), DHA, Norbolethone (Nor), and tetrahydrogestrinone (THG), for their efficacy as inverse agonists. Androstanol, a well characterized CAR inverse agonist (Forman et al., 1998), and DHA, a 3a-hydroxysteroid dehydrogenase catalyzed metabolite of dihydrotestosterone (Biswas and Russell, 1997), are natural androgens. THG and Nor are synthetic androgens allegedly used by the Bay Area Laboratory Co-operative to enhance the performance of several famous athletes and are known as ‘‘the clear’’ and ‘‘the clean’’ (Catlin et al., 2002, 2004; Fainaru-Wada and Williams, 2007; Labrie et al., 2005). HepG2 cells cotransfected with mCAR and a luciferase reporter construct were treated with the four androgens at 10lM and their ability to repress basal transcription was measured. All four of the androgens tested repressed CAR activity. And is the least efficacious, and the synthetic androgen, THG is the most efficacious inverse agonist at 10lM; DHA and Nor had similar potencies (Fig. 1A). Treatment of cotransfected HepG2 cells with androgen concentrations ranging from 0.1 to 30lM produced concentration-response curves (Fig. 1B) with similar slopes with the exception of Nor, which had a much flatter slope. DHA is the most efficacious CAR inverse agonist at 30lM, but is the least potent based on its EC50 of 7.64lM (Table 1). At 30lM, DHA inhibited CAR activity 99%, THG 90%, Nor 75%, and And 71%. And is the least efficacious CAR inverse agonist, but has the steepest hillslope and second lowest EC50 indicating it’s relatively high potency compared with DHA. Subsequently, we compared the potency of each of the CAR inverse agonists at 10lM in the presence of TCPOBOP. The purpose of this experiment is to determine which inverse agonist has the weakest affinity for CAR in the presence of TCPOBOP. A chemical that is easily released or outcompeted in the presence of a CAR activator would be ideal because it would show the largest change following treatment with a CAR activator and thus have the most sensitivity in the assay. Interestingly, Nor, which showed a flatter slope than the other androgens, repressed the TCPOBOP activation of CAR much more than the other androgens. TCPOBOP increased THGand And-repressed CAR activity about twofold. However, DHA is the ideal inverse agonist, as TCPOBOP increased DHArepressed CAR activity approximately threefold. Thus, DHA is outcompeted by TCPOBOP more easily than the other androgens tested (p < 0.001) (Fig. 1C). Overall, our data indicates that DHA is a reasonably strong repressor of unactivated CAR, releases from CAR easily compared with other androgens following CAR activation by TCPOBOP, and provides the most sensitivity in a CAR transactivation assay. Therefore, we used DHA to repress CAR activity in our subsequent CAR transactivation assays.

96 8 7

B

6

Relative Activity

A Luciferase Activity

BALDWIN AND ROLING

*

5

*a

4

*b

3

*cd

2

140 120 100 80 60

1

20

Nor And

0

DHA

0 UT

And

DHA

THG

THG

40

Nor

-7.5

-7.0

Androgens

Relative Activity

C

-6.5

-6.0

-5.5

-5.0

-4.5

Chemical Concentration (M)

*e

3.5 3.0 2.5

*

*

2.0 1.5 1.0 0.5 0.0 And

And + TC

DHA

DHA + TC

THG

THG + TC

Nor

Nor + TC

Androgens FIG. 1. Several natural and synthetic androgens are CAR inverse agonists. HepG2 cells cotransfected with an expression plasmid for mCAR and the luciferase reporter plasmid, containing the CYP2B6 PBREM, were treated with androgens and their ability to repress basal transcription was measured. (A) CAR’s activity is significantly repressed by 10lM concentrations of each of the androgens, And, DHA, THG, and Nor. (B) Sigmoidal dose-response curves (GraphPad Prizm 4.0) from transactivation assays following treatment with androgens as inverse agonists. (C) Androgens repress CAR activation by 250nM TCPOBOP (TC). Data are presented as mean ± SD. Statistical analysis was performed by ANOVA followed by Tukey’s multiple comparison test as the post hoc test using the GraphPad Prizm 4.0 software package (n ¼ 3). An asterisk indicates statistical significance from the untreated control (A) or correspondingly treated androgens (C) (p < 0.001). The letters [a, b, or c], respectively, indicate a significant difference (p < 0.05, 0.01, 0.001), from And-treated cells, [d] indicates a significant difference (p < 0.01) from DHA-treated cells, and an [e] indicates a significant difference (p < 0.001) from cells treated with other androgens þ TC.

CAR transactivation assays were performed with a variety of environmental chemicals, steroid hormones, and bile acids at 10lM concentrations unless otherwise noted. TCPOBOP activated DHA-repressed CAR activity nearly fourfold. The only other chemicals to increase CAR activity greater than 3.4fold were mono-ethylhexylphthalate (MEHP) at 100lM (3.963), corticosterone (3.423), and three organophosphate insecticides, parathion (3.833), SSS-Tributylphosphorotritioate (3.763), and chlorpyrifos (4.823). None of the bile acids or cholesterol activated CAR at the concentration tested (10lM), and progesterone acted as an inverse agonist. The remaining CAR activators are most likely partial activators and include but are not limited to chemicals such as nonylphenol, methoxychlor, imazalil, bisphenol A, cypermethrin, and 50lM arsenite (Fig. 2). We choose seven of these chemicals with diverse structures (Fig. 3) for our mixtures studies. In order to use the concentration addition approach (Casey et al., 2004) as described in ‘‘Materials and Methods,’’ the potency and Hillslope for each chemical must be determined. In addition, efficacy is necessary for partial activators so that restraint values can be used in our concentration addition model. Therefore, we measured concentration-response curves for seven chemicals (Fig. 4), and determined their efficacy (%

activity relative to the full agonist TCPOBOP), EC50, and Hillslopes (Table 2). The seven chemicals tested were TCPOBOP and PB as positive controls, nonylphenol, bisphenol A, cypermethrin, chlorpyrifos, and phthalic acid (also known as diethylhexylphthalate). We hypothesized phthalic acid was a partial activator based on data from Figure 2 but found it was a full activator at concentrations over 100lM. Nonylphenol, cypermethrin, bisphenol A, and PB are partial activators. TCPOBOP is considerably more potent than the

TABLE 1 Potency and Efficacy of CAR Inverse Agonists Chemical And DHA Nor THG a

EC50 (lM)a

95% CIb

Hillslopec

2.531 7.636 2.683 1.542

0.313–15.23 4.930–11.86 1.580–4.604 0.254–9.385

–5.439 3.453 0.912 2.742

EC50: Effective concentration at which activity is reduced 50%. 95% confidence interval (CI). c Hillslope is the slope of the dose-response curve as defined by Equation 1. EC50, 95% CI, and Hillslope were determined using the Sigmoidal doseresponse curve program in Graphpad 4.0. b

A CONCENTRATION ADDITION MODEL FOR THE ACTIVATION OF CAR

97

FIG. 2. Several endo- and xenobiotics activate mCAR. HepG2 cells cotransfected with an expression plasmid for mCAR and a luciferase reporter plasmid containing the CYP2B6 PBREM, were treated with environmental and endobiotic chemicals at 10lM þ DHA unless otherwise noted. All data is normalized to the DHA-repressed luciferase activity. Data are presented as mean ± SD. Statistical analysis was performed by ANOVA followed by Tukey’s multiple comparison test as the post hoc test (n ¼ 3). An asterisk (p < 0.05) or two asterisks (p < 0.01) indicates statistical significance from the DHA control.

other chemicals tested and chlorpyrifos is the most potent of the environmentally relevant chemicals tested (Fig. 4; Table 2). Concentration Addition Model for CAR Mixtures of two chemicals were made from four of the chemicals in which EC50, Hillslope, and relative efficacy compared with TCPOBOP were determined (Table 2). Each of the chemicals was added in equal concentrations and three different mixture concentrations were tested in our HepG2 CAR transactivation assay. Of the four chemicals tested only chlorpyrifos is a full activator. Chemical concentrations were 0.25, 1.0, and 5.0lM each for a total of 0.5, 2.0, and 10lM combined. The measured values from the transactivation assays shown as relative activity were compared with the concentration addition model values obtained from entering the chemical concentration, EC50, and Hillslope into CATAM (Equation 2). The concentration addition model without restraints mimicked the measured CAR activity remarkably well if chlorpyrifos was one of the chemicals in the mixture (Figs. 5A–C). However, if both chemicals were partial activators than the concentration addition models with restraints were better at predicting CAR activity (Figs. 5D, F; Table 3). The concentration addition model with weighted restraints closely mimicked the model with the highest restraint values of the strongest activator in the mixture because in many of the two-chemical mixtures one of

the chemicals was dominant, especially if bisphenol A was the partial agonist in the mixture. Fisher exact tests indicated that the concentration addition model without restraints was the best model if no partial activators were present, suggesting that in the presence of a full CAR activator, partial CAR activators act like a full activator (Table 3). However the proximity of the models and the simplistic nature of the mixture indicated that more study was necessary. Therefore, we needed to test the hypothesis that partial CAR activators act as full activators in the presence of a full activator under slightly different conditions and treated CAR-transfected HepG2 cells with mixtures composed of three chemicals. HepG2 cells were exposed to four different treatments. In three of the treatment regimens, cells were treated with two partial activators and a full activator (chlorpyrifos þ two different partial activators), and in one of the treatment regimens cells were treated with three partial activators. Chemical concentrations were 0.166, 0.666, and 3.33lM each for a total of 0.5, 2.0, and 10lM combined. The measured values from the transactivation assays shown as relative activity were compared with the concentration addition CATAM values. The concentration addition model without restraints mimicked the measured CAR activity remarkably well if chlorpyrifos was one of the chemicals in the mixture (Figs. 6A–C; Table 3). In this slightly more complex mixture with more partial

98

BALDWIN AND ROLING OH Cl

Cl

O

Cl

O N

O

Cl H N

O H3C N

N

H

O

CH3

H3C

O

NH

O O

Cl

Cl

H3C

Nonylphenol MW=220.39 Log Kow=4.48

Phenobarbital MW=232.235

TCPOBOP MW=402.1

Cypermethrin MW=416.35 Log Kow=6.60

OH

Cl O H3C

CH3

OH O

O

S

H 3C OH

Cl

N

P O

O Cl

H3C OH

Bisphenol A MW=228.3 Log Kow=3.32

Phthalic acid MW=166.1 Log Kow=0.79

Chlorpyrifos MW=350.5836 Log Kow=4.699

FIG. 3. Structures of some of the different chemicals that activate CAR, and that we used to perform concentration-response curves.

activators, the concentration addition model with weighted restraints underestimated the activity of the mixture (Figs. 6A–C; Table 3) because partial activators appear to act like full activators in the presence of a full activator within the mixture. If all three chemicals were partial activators than the concentration addition models with restraint values better predicted CAR activity (Fig. 6D). Overall, the data suggests that the activity of partial activators could be underestimated in situations in which a full activator is present in an environmental mixture. Lastly, we considered a more environmentally relevant situation in which cells were exposed to lower concentrations of several chemicals. We treated HepG2 cells with five different chemicals (phthalic acid, chlorpyrifos, nonylphenol, cypermethrin, and bisphenol A) at multiple concentrations varying from 0.042 to 2lM of each chemical at equal concentrations for a total of 0.21–10lM concentrations of the five chemicals combined. The chemicals significantly activated CAR at a combined concentration of 0.85 or 0.17lM of each chemical (Fig. 7), demonstrating that low concentrations of several chemicals in the nanomolar range can activate CAR

and this may be environmentally relevant especially for occupations in the agriculture or plastics industries. Second, the concentration addition model without restraint values reasonably mimicked the measured CAR activity when using two full activators (phthalic acid, chlorpyrifos) and three partial activators (nonylphenol, cypermethrin, bisphenol A) (Fig. 7) significantly better than either of the restraint models (Table 3). This further demonstrates that partial activators act like full activators in the presence of full activators.

DISCUSSION

CAR is an often mysterious nuclear receptor because it has high constitutive activity in vitro (Baes et al., 1994) and is activated through both direct (ligand activation, i.e., TCPOBOP; Tzameli et al. 2000) and indirect means of nuclear translocation induced by protein phosphatase 2A mediated dephosphorylation of serine 202 (i.e., PB) (Hosseinpour et al., 2006; Kawamoto et al., 1999). These properties can make CAR difficult to study, provide at least two pathways of

A CONCENTRATION ADDITION MODEL FOR THE ACTIVATION OF CAR TCPOBOP nonylphenol phenobarbital cypermethrin bisphenol A chlorpyrifos phthalic acid

390

Relative Activity

340 290 240 190 140 90

-9

-8

-7

-6

-5

-4

-3

Concentration (M) FIG. 4. Dose-response curve of selective CAR agonists. Dose-response curves of seven chemicals were determined using GraphPad Prism 4.0 for the purpose of calculating the efficacy, EC50, and Hillslope of each chemical. Data points are shown as mean ± SD (n ¼ 3).

activation, and add additional experiments to a study such as the determination of an appropriate inverse agonist. In our case, we investigated chemicals that can activate and inactivate CAR in addition to the effect of a mixture of chemicals on CAR activity. CAR activators increased CAR activity in a concentration addition model in a predictable manner with the exception of partial activators as they acted as full activators in the presence of an experimentally verified full activator. The loss of partial activator type effects by chemicals in the presence of a full activator is another interesting CAR peculiarity. Humans are exposed to a mixture of different chemicals through occupational, dietary, environmental, and pharmaceutical exposures that CAR and other xenosensors respond to.

TABLE 2 Potency and Efficacy of CAR Agonists Chemical TCPOBOP Chlorpyrifos Cypermethrin Nonylphenol Phthalic acid Bisphenol A PB

EC50 (lM)a

95% CI (lM)b

Hillslopec

Restraintd

0.0234 0.7258 1.967 2.581 4.976 12.24 314.6

0.0138–0.0397 0.3724–1.415 1.295–2.988 1.716–3.881 0.0624–312.3 0.0068–2722 12.62–494.2

3.066 1.335 1.988 1.597 0.718 1.443 1.202

NA NA 0.67 0.60 NA 0.50 0.47

Note. NA ¼ not applicable. EC50, 95% CI, and Hillslope were determined using the Sigmoidal dose-response curve program in Graphpad 4.0. a EC50: effective concentration at which activity is reduced 50%. b 95% confidence Interval (CI). c Hillslope is the slope of the dose-response curve as defined by Equation 1. d Restraint or restraint value refers to the efficacy of the chemical as a partial CAR agonist and indicates the relative activity compared with the full agonist TCPOBOP.

99

The Food Quality Protection Act of 1996 (1996) requires that the United States Environmental Protection Agency consider the cumulative effects of pesticides that have a common mechanism of action for this reason. We hypothesized that a mixture of CAR activators would work through a simple concentration addition type mechanism because a group of chemicals that work through the same mechanism typically act additively but sometimes antagonistically in the case of partial agonists that inhibit the activity of other agonists (Lu and Serrero, 1999; Miller et al., 2008). To prove this hypothesis, we tested full and partial activators on CAR using concentration addition for estimating the activity of the mixture. All three models worked to varying degrees with the two chemical mixtures. The concentration addition model without restraints mimicked the measured values if a full agonist was present, the concentration addition model with the largest partial restraint value worked best if all the chemicals were partial agonists, and the concentration addition model with weighted restraints worked well in both conditions. However, the weighted restraint model also suggested that one chemical may dominate a simple mixture (chlorpyrifos). Furthermore, if all the chemicals were partial agonists, the concentration addition model with the highest partial restraint value from the mixture did not show a statistically significant difference from the concentration addition model with weighted restraints. Taken together, more robust mixtures with more partial agonists were needed to determine the best model and therefore three and five chemicals were used in following studies. When three chemicals or more were used, neither of the concentration addition models with restraint values correctly predicted the action of a mixture if a full activator was present (Table 3). This strengthened our theory that partial activators can act like a full activator in the presence of a full activator. Furthermore, partial activators were neither antagonistic nor partially additive in these cases, which would have reduced the activity of a full activator (Figs. 5–7) (Lu and Serrero, 1999; Miller et al., 2008). Therefore, the role or activity of partial agonists did not need to be considered in our models unless all of the CAR activators were partial agonists (Figs. 5, 6), suggesting the activity of partial CAR activators could be underestimated. Furthermore, if all of the chemicals in the mixture are partial activators, the largest restraint value should be considered based on our limited sampling to date. Mathematically, this essentially is the same as a mixture containing a full activator because the restraint value equals one (i.e. no restraint value) in a mixture containing a full activator. Thus the largest restraint value is always the one that should be used as the multiplier to Equation 2. Our current studies indicate that both concentration addition models (with and without restraint values) are the same if the restraint value used is the largest of the chemicals in the mixture. We hypothesize that partial agonists are acting as stronger partial activators or full activators in mixtures because CAR

100

BALDWIN AND ROLING

100

100

A

75

80 60

50 chlorpyrifos + cypermethrin model model + restraint model + wt restraint

40 20 0

Relative Activity

150 125

0.0

2.5

5.0

7.5

10.0

B

nonylphenol + cypermethrin model model + restraint model + wt restraint

25 0

12.5

0.0

80

100

2.5

5.0

7.5

10.0

12.5

E

60

75

40

50

nonylphenol + chlorpyrifos model model + restraint model + wt restraint

25 0 0.0

100

D

2.5

5.0

7.5

10.0

0

12.5

0.0 100

C

nonylphenol + bisphenol A model model + restraint model + wt restraint

20

2.5

5.0

7.5

10.0

12.5

F

75

75

50

50 chlorpyrifos+ bisphenol A model model + restraint model + wt restraint

25 0 0.0

2.5

5.0

7.5

10.0

12.5

cypermethrin + bisphenol A model model + restraint model + wt restraint

25 0 0.0

2.5

5.0

7.5

10.0

12.5

Chemicals (uM) FIG. 5. Concentration addition models for CAR activity following activation by a two-chemical mixture. Equal concentrations of two chemicals were combined and CAR activity measured. Concentrations shown in the x-axis are the total combined concentrations of the two chemicals. Measured CAR activity was compared with a concentration addition model without restraint values, a concentration addition model with restraint values for the most potent partial agonist in the mixture, and a concentration addition model that weighted each chemicals restraint value based on potency (Equation 2). EC50 values were multiplied by the partial agonist value to determine the concentration addition model with restraint values. (A) chlorpyrifos and cypermethrin, (B) chlorpyrifos and nonylphenol, (C) chlorpyrifos and bisphenol A, (D) nonylphenol and cypermethrin, (E) nonylphenol and bisphenol A, (F) bisphenol A, and cypermethrin. Measured values (mean ± 95% confidence intervals; n ¼ 3) are shown in black with filled boxes, the concentration addition model is shown with triangles and dashed lines in red, the concentration addition model with restraint values is shown with upside down triangles with dotted lines in blue, and the concentration addition model with weighted restraints is shown with diamonds with alternating dashed and dotted lines in green.

can be activated through at least two distinct pathways. For example, full activators may be ligands or true agonists that recruit a full complement of cofactors necessary for a full response, whereas partial activators are acting through one or more indirect pathways that do not recruit the full complement of cofactors and therefore cannot induce a full response by themselves (i.e., PB). However in combination the full complement of cofactors is available at the response element. Then again, ligand binding to CAR has not been tested for the partial and full activators measured so this hypothesis is currently speculative. Conversely, it is possible that indirect activators potentiate the effects of ligand activators. An example of this would be estrogen receptor alpha (ERa). ERa is activated by insulin-like growth factor-1 or epidermal growth factor (Ignar-Trowbridge et al., 1993, 1995) through a mitogen activated protein kinasedependent pathway in addition to its ligand-dependent pathway

(Kato et al., 1995). Activation through the two distinct pathways combined increased ERa activity significantly above estrogen alone (Ignar-Trowbridge et al., 1995, 1996). Thus, the increase in measured CAR activity from the predicted concentration addition model with restraint values to a model without restraint values may actually approximate a maximum activity under a specific set of uncharacterized conditions. However, this is speculative and does not consider why the concentration addition model without restraints predicts the measured values at concentrations well below saturation of CAR activity. Further studies are necessary to determine the mechanisms of cooperation between full and partial agonists. Future models may need to consider the different slopes of concentration-response curves at lower chemical concentrations. In several experiments the models underestimated the measured activity at the lowest concentrations where the individual chemical concentrations were below their observed

101

A CONCENTRATION ADDITION MODEL FOR THE ACTIVATION OF CAR

TABLE 3 Comparison of the Three Concentration Addition Models and Determination of the Models that Fit within the 95% Confidence Intervals of Measured Values and the Models that Mimic Measured Values Most Closely Model closest to measuredb Points with 95% CIa

2 chems

Full activator is in the chemical mixture Without restraint 5/9 With restraint 1/9 With weighted restraint 4/9 p Value 0.0001 Only partial activators are in the chemical mixture Without restraint 2/9 With restraint 3/9 With weighted restraint 2/9 p Value 0.281

Modelc

3 chems

5 chems

2 chems

3 chems

5 chems

5/9 2/9 0/9 0.0001

3/5 2/5 0/5 0.001

8/9* 1/9 0/9

8/9* 0/9 1/9

5/5* 0/5 0/5

0/3 0/3 0/3 0.679

3/5 2/5 0/5 NA

4/9 4/9 1/9

0/3 2/3 1/3

NA NA NA

Note. Determination of the models that fit closest to measured values and their Fisher’s exact test values. *An asterisk indicates statistical significance by Fisher’s exact test (2 3 3) to compare the models that most closely mimicked the measured values. Fisher’s exact tests were performed as described in ‘‘Materials and Methods.’’ p Values for the data comparing the three different models to the measured values are provided in the table. NA ¼ not applicable. a Number of points that the model is within the 95% confidence interval of measured values over the total number of data points. b Number of times that the model is closest to the measured values over the total number of data points. c The three models compared with the measured values are the concentration addition model without restraints, the concentration addition model with restraints (largest restraint value among the partial agonists), and the concentration addition model with weighted restraints (a weighted average of the restraint values from each partial agonist in the chemical mixture).

effect concentrations or at the bottom of the sigmoidal dose-response curve. Previous experiments with multiple environmental estrogens (8–11 chemicals) used at levels below their no observable effect concentrations also showed greater measured effects than concentration addition models predicted (Rajapakse et al., 2002; Silva et al., 2002). With CAR, even measured activity from a mixture of partial activators most closely resembled the concentration addition model without restraint at lower concentrations (Table 3; Figs. 5, 6); however, measured activity with higher chemical concentrations of only partial activators clearly mimicked concentration addition models with restraint values better. Therefore, the data appears as if synergism occurred, but we predict that low dose models combined with restraint factors, and potentially other mitigating factors determined in future experiments may accurately predict the combined effects of low dose exposure to CAR activators. Overall, CAR activators acted in an additive fashion that could be modeled using concentration addition demonstrating that several environmental chemicals at low concentrations may cause an unexpected idiosyncratic drug reaction. This was further demonstrated using five chemicals at low concentrations that significantly induced CAR activity (Fig. 7). In a world in which we are exposed to an increasing number of chemicals through either occupational or erratic dietary, environmental or pharmaceutical exposures; the potential for drug-toxicant interactions may increase. CAR is also important in the regulation of bile acid homeostasis and energy metabolism, thus exposure to several low dose CAR activators

may affect metabolic diseases, alter bile acid homeostasis, and increase liver disease (Ding et al., 2006; Guo et al., 2003; Wagner et al., 2005; Zhang et al., 2004). In addition, repression of CAR activity by androgens was investigated to optimize the CAR transactivation assay and determine the potential effects of synthetic androgens used to enhance performance. All of the androgens tested repressed CAR activity, including the often used inverse agonist And. However, DHA was the most efficacious inverse agonist in the CAR transactivation assay. DHA is a strong inhibitor at high concentrations, except when TCPOBOP was present and then DHA released easily allowing sensitive measurement of CAR activation. Therefore we used DHA in our transactivation assays (Figs. 1, 2). However, the synthetic androgen THG, known as ‘‘the clear’’, is the most potent inverse agonist (Table 1; Fig. 1). THG, which has a similar structure to gestrinone and trenbolone, has potential as an inverse agonist, but its use as a synthetic androgen by young athletes may cause unexpected adverse drug reactions. Nor, the synthetic androgen known as ‘‘the clean,’’ is also a CAR inverse agonist and inhibitor of TCPOBOP activation of CAR. This makes Nor an extremely poor inverse agonist for our purpose, and furthermore its flat slope increases its variability within each set of transactivation assays. However, Nor is an interesting chemical because of its potential to cause adverse drug interactions as a CAR inhibitor and synthetic androgen used by young athletes even in the presence of CAR activators. Interestingly, Nor is known to interact with and alter the toxicity of a number of chemicals,

102

BALDWIN AND ROLING 120

125

A

100

C

100

80

75

60 50

40

4NP, Chlor, Cyper model model + restraint model + wt restraint

Relative Activity

20 0 0.0

100

2.5

5.0

7.5

10.0

0

12.5

0.0

100

B

4NP, Chlor, Bis A model model + restraint model + wt restraint

25

2.5

5.0

7.5

10.0

12.5

D

75 75 50

50 Chlor, Cyper, BisA model model + restraint model + wt restraint

25 0 0.0

2.5

5.0

7.5

10.0

12.5

4NP, Cyper, BisA model model + restraint model + wtrestraint

25 0 0.0

2.5

5.0

7.5

10.0

12.5

Chemicals (uM) FIG. 6. Concentration addition models for CAR activity following activation by a three-chemical mixture. Equal concentrations of three chemicals were combined and CAR activity measured. Concentrations shown in the x-axis are the total combined concentrations of the three chemicals. Measured CAR activity was compared with a concentration addition model without restraint values, a concentration addition model with restraint values for the most potent partial agonist in the mixture, and a concentration addition model that weighted each chemicals restraint value based on potency (Equation 2). EC50 values were multiplied by the partial agonist value to determine form the concentration addition model with restraint values and (A) nonylphenol, chlorpyrifos, and cypermethrin, (B) chlorpyrifos, cypermethrin, and bisphenol A (C) nonylphenol, chlorpyrifos, and bisphenol A, (D) nonylphenol, cypermethrin, bisphenol A. Measured values (mean ± 95% confidence intervals; n ¼ 3) are shown in black with filled boxes, the concentration addition model is shown with triangles and dashed lines in red, the concentration addition model with restraint values is shown with upside down triangles with dotted lines in blue, and the concentration addition model with weighted restraints is shown with diamonds with alternating dashed and dotted lines in green.

including indomethacin, carbon tetrachloride, cocaine, digitoxin, and parathion (Robinson et al., 1976, 1979; Solomon et al., 1971). However, we must be careful when extrapolating our results with natural and synthetic androgens to humans because the androgens tested to date have proven to be much weaker hCAR inverse agonists than mCAR inverse agonists (Forman et al., 1998; Huang et al., 2004b; Moore et al., 2002). Further work needs to be done with primary cells or humanized mice to determine the efficacy of these chemicals as hCAR inverse agonists. Lastly, we used our CAR transactivation assay to screen for new chemical activators of CAR and found several environmentally relevant chemicals and steroid hormones act as full and partial activators (Fig. 2). None of the bile acids tested acted as CAR activators at 10lM concentrations. Several chemicals activated CAR such as the pesticides parathion, chlorpyrifos, tributylphosphorotritioate, fenitrothion, cypermethrin, alachlor, metalachlor, endosulfan, and others. Plasticizers such as MEHP, phthalic acid, nonylphenol, and bisphenol A also increased CAR activity as did the steroid corticosterone. Progesterone is an inverse agonist. In addition, both arsenic and monosodium acid methane arsenate activated

CAR (Fig. 2). It has been noted that the list of chemicals known to activate CAR is much shorter than the list of chemicals that activate PXR (Kretschmer and Baldwin, 2005), probably due to PXR’s larger and flexible binding pocket (Watkins et al., 2001). However, there has also been significantly less screening for CAR activators and Figure 2 suggests that CAR’s ability to be activated through two distinct pathways probably allows it to show high promiscuity in a fashion similar to PXR. Furthermore, because there are so many CAR activators, the potential for them to work in combination and induce hepatic drug-metabolizing enzymes is relatively high. Overall, the purpose of our study was to screen for CAR activators and inactivators and then use this information to determine the effect of a mixture of chemicals on CAR activity. Several different chemicals with distinct structures, including steroids, pesticides, and plasticizers alter CAR activity. These chemicals work together in an additive fashion to increase CAR activity in a predictable manner. The Food Quality Protection Act of 1996 requires that the United States Environmental Protection Agency consider the cumulative effects of pesticides that have a common mechanism of action;

A CONCENTRATION ADDITION MODEL FOR THE ACTIVATION OF CAR

103

REFERENCES Relative Activity

100

Assem, M., Schuetz, E. G., Leggas, M., Sun, D., Yasuda, K., Reid, G., Zelcer, N., Adachi, M., Strom, S., Evans, R. M., et al. (2004). Interactions between hepatic Mrp4 and Sult2a as revealed by the constitutive androstane receptor and Mrp4 knockout mice. J. Biol. Chem. 279, 22250–22257.

80 60 40 measured model model + restraint model+ wt restraint

20 0 0.0

2.5

5.0

7.5

10.0

12.5

(uM) Chemical Mixtures FIG. 7. Concentration addition models for CAR activity following activation by a five-chemical mixture. Equal concentrations of five chemicals were combined and CAR activity measured as described in ‘‘Materials and Methods.’’ Equivalent concentrations of nonylphenol, cypermethrin, bisphenol A, phthalic acid, and chlorpyrifos were added to each well. Concentrations shown in the x-axis are the total combined concentrations of the five chemicals. Each chemical was supplied at concentrations of 0.042, 0.085, 0.170, 0.60, and 2.0lM, respectively for final concentrations of 0.21, 0.42, 0.85, 3, and 10lM. Activity of the mixture was compared with the full agonist activity of 0.250lM TCPOBOP to determine relative activity. Measured CAR activity was compared with a concentration addition model without restraint values, a concentration addition model with restraint values for partial agonists, and a concentration addition model with weighted restraints. Data are presented as mean ± 95% CI.

however, industrial chemicals and pharmaceuticals can further enhance the effects of chemicals on CAR activity. The increased use of chemicals and pharmaceuticals in modern societies improves the likelihood of unexpected adverse consequences of chemical use because of exposure to several similar acting chemicals at low doses. Our data indicate that it is important to consider that we are exposed to a milieu of chemicals, and these chemicals can act additively to increase CAR activity.

SUPPLEMENTARY DATA

Supplementary data are available online at http://toxsci. oxfordjournals.org/.

FUNDING

EPA (grants U-91620701, NIH S06 GM008012); start-up funds from Clemson University; and EPA-STAR-MAI fellowship (U-91620701) supported J.R.

ACKNOWLEDGMENTS

We would like to thank Yim Chan and Laura Chapman for their help with the transactivation assays, and Dr Don Catlin of the UCLA Olympic Laboratory for the THG and Nor.

Baes, M., Gulick, T., Choi, H.-S., Martinoli, M. G., Simha, D., and Moore, D. D. (1994). A new orphan member of the nuclear hormone receptor superfamily that interacts with a subset of retinoic acid response elements. Mol. Cell. Biol. 14, 1544–1552. Berenbaum, M. C. (1985). The expected effect of a combination of agents: The general solution. J. Theor. Biol. 114, 413–431. Biswas, M. G., and Russell, D. W. (1997). Expression cloning and characterization of oxidative 17b- and 3a-hydroxysteroid dehydrogenases from rat and human prostate. J. Biol. Chem. 272, 15959–15966. Bliss, C. I. (1939). The toxicity of poisons applied jointly. Ann. Appl. Biol. 26, 585–615. Casey, M., Gennings, C., Carter, W. H. J., Moser, V. C., and Simmons, J. E. (2004). Detecting interaction(s) and assessing the impact of component subsets in a chemical mixture using fixed-ratio mixture ray designs. J. Agric. Biol. Environ. Stat. 9, 339–361. Catlin, D. H., Abrens, B. D., and Kucherova, Y. (2002). Detection of norbolethone, an anabolic steroid never marketed, in athletes’ urine. Rapid Commun. Mass Spectrom. 16, 1273–1275. Catlin, D. H., Sekera, M. H., Ahrens, B. D., Starcevic, B., Chang, Y. C., and Hatton, C. K. (2004). Tetrahydrogestrinone: Discovery, synthesis, and detection in urine. Rapid Commun. Mass Spectrom. 18, 1245–1249. Chang, T. K. H. (2006). Synthetic drugs and natural products as modulators of constitutive androstane receptor (CAR) and pregnane X receptor (PXR). Drug Metab. Rev. 38, 51–73. Chen, H. L., Liu, Y. J., Chen, H. L., Wu, S. H., Ni, Y. H., Ho, M. C., Lai, H. S., Hsu, W. M., Hsu, H. Y., Tseng, H. C., et al. (2008). Expression of hepatocyte transporters and nuclear receptors in children with early and latestage biliary atresia. Pediatr. Res. 63, 667–673. Chen, X., Zhang, J., Baker, S. M., and Chen, G. (2007). Human constitutive androstane receptor mediated methotrexate induction of human dehydroepiandrosterone sulfotransferase (hSULT2A1). Toxicology 231, 224–233. Ding, X., Lichti, K., Kim, I., Gonzalez, F. J., and Staudinger, J. L. (2006). Regulation of constitutive androstane receptor and its target genes by fasting, cyclic AMP, HNF4a and the coactivator PGC-1a. J. Biol. Chem. 281, 26540–26551. Drescher, K., and Boedeker, W. (1995). Assessment of the combined effects of substances—The relationship between concentration addition and independent action. Biometrics 51, 716–730. Fainaru-Wada, M., and Williams, L. (2007). In Game of Shadows: Barry Bonds, BALCO, and the Steroids Scandal that Rocked Professional Sports. Penguin Group, New York, NY. Fisher, C. D., Augustine, L. M., Maher, J. M., Nelson, D. M., Slitt, A., Klaassen, C. D., Lehman-McKeeman, L. D., and Cherrington, N. J. (2007). Induction of drug-metabolizing enzymes by garlic and allyl sulfide compounds via activation of the CAR and NRF2. Drug Metab. Dispos. 35, 995–1000. Public Law 104-170, 104th Congress. August 3, 1996. Forman, B. M., Tzameli, I., Choi, H. S., Chen, J., Simha, D., Seol, W., Evans, R. M., and Moore, D. D. (1998). Androstane metabolites bind to and deactivate the nuclear receptor CAR-beta. Nature 395, 612–615. Gennings, C., Carter, W. H., Jr, Carney, E. W., Charles, G. D., Bollapudi, B. B., and Carchman, R. A. (2004). A novel flexible approach for evaluating fixed ratio mixtures of full and partial agonists. Toxicol. Sci. 80, 134–150. Guo, G. L., Lambert, G., Negishi, M., Ward, J. M., Brewer, H. B., Jr, Kliewer, S. A., Gonzalez, F. J., and Sinal, C. J. (2003). Complementary roles

104

BALDWIN AND ROLING

of farnesoid X receptor, pregnane X receptor, and constitutive androstane receptor in protection against bile acid toxicity. J. Biol. Chem. 278, 45062–45071.

Meadows, S. L., Gennings, C., Carter, W. H., Jr, and Bae, D.-S. (2002). Experimental designs for mixtures of chemicals along fixed ratio rays. Environ. Health Perspect. 110(Suppl. 6), 979–983.

Hernandez, J. P., Huang, W., Chapman, L. M., Chua, S., Moore, D. D., and Baldwin, W. S. (2007). The environmental estrogen, nonylphenol, activates the constitutive androstane receptor (CAR). Toxicol. Sci. 98, 416–426.

Miller, J. R., Siripurkpong, P., Hawes, J., Majdalawieh, A., Ro, H. S., and McLeod, R. S. (2008). The trans-10, cis-12 isomer of conjugated linoleic acid decreases adiponectin assembly by PPARgamma-dependent and PPARgamma-independent mechanisms. J. Lipid Res. 49, 550–562.

Honkakoski, P., Jaaskelainen, I., Kortelahti, M., and Urtti, A. (2001). A novel drug-regulated gene expression system based on the nuclear receptor constitutive androstane receptor (CAR). Pharm. Res. 18, 146–150. Hosseinpour, F., Moore, R., Negishi, M., and Sueyoshi, T. (2006). Serine 202 regulates the nuclear translocation of constitutive active/androstane receptor. Mol. Pharmacol. 69, 1095–1102. Huang, W., Zhang, J., and Moore, D. D. (2004a). A traditional herbal medicine enhances bilirubin clearance by activating the nuclear receptor CAR. J. Clin. Invest. 113, 137–143. Huang, W., Zhang, J., Wei, P., Schrader, W. T., and Moore, D. D. (2004b). Meclizine is an agonist ligand for mouse constitutive androstane receptor (CAR) and an inverse agonist for human CAR. Mol. Endocrinol. 18, 2402–2408. Ignar-Trowbridge, D. M., Pimentel, M., Parker, M. G., McLachlan, J. A., and Korach, K. S. (1996). Peptide growth factor cross-talk with the estrogen receptor requires the A/B domain and occurs independently of protein kinase C or estradiol. Endocrinology 137, 1735–1744. Ignar-Trowbridge, D. M., Pimentel, M., Teng, C. T., Korach, K. S., and McLachlan, J. A. (1995). Cross talk between peptide growth factor and estrogen receptor signaling systems. Environ. Health Perspect. 103, 35–38. Ignar-Trowbridge, D. M., Teng, C. T., Ross, K. A., Parker, M. G., Korach, K. S., and McLachlan, J. A. (1993). Peptide growth factors elicit estrogen receptor-dependent transcriptional activation of an estrogenresponsive element. Mol. Endocrinol. 7, 992–998. Jackson, J. P., Ferguson, S. S., Negishi, M., and Goldstein, J. A. (2006). Phenytoin induction of the cyp2c37 gene is mediated by the constitutive androstane receptor. Drug Metab. Dispos. 34, 2003–2010. Kast, H. R., Goodwin, B., Tarr, P. T., Jones, S. A., Anisfeld, A. M., Stoltz, C. M., Tontonoz, P., Kliewer, S., Willson, T. M., and Edwards, P. A. (2002). Regulation of multidrug resistance-associated protein 2 (ABCC2) by the nuclear receptors pregnane X receptor, farnesoid X-activated receptor, and constitutive androstane receptor. J. Biol. Chem. 277, 2908–2915. Kato, S., Endoh, H., Masuhiro, Y., Kitamoto, T., Uchiyama, S., Sasaki, H., Masushige, S., Gotoh, Y., Nishida, E., Kawashima, H., et al. (1995). Activation of the estrogen receptor through phosphorylation by mitogenactivated protein kinase. Science 270, 1491–1494. Kawamoto, T., Sueyoshi, T., Zelko, I., Moore, R., Washburn, K., and Negishi, M. (1999). Phenobarbital-responsive nuclear translocation of the receptor CAR in induction of the CYP2B gene. Mol. Cell. Biol. 19, 6318–6322. Kolpin, D. W., Furlong, E. T., Meyer, M. T., Thurman, E. M., Zaugg, S. D., Barber, L. B., and Buxton, H. T. (2002). Pharmaceuticals, Hormones, and Other Organic Wastewater Contaminants in U.S. Streams, 1999-2000: A National Reconnaissance. Environ. Sci. Technol. 36, 1202–1211. Kretschmer, X. C., and Baldwin, W. S. (2005). CAR and PXR: Xenosensors of endocrine disrupters? Chem. Biol. Interact. 155, 111–128. Labrie, F., Luu-The, V., Calvo, E., Martel, C., Cloutier, J., Gauthier, S., and Belle, P. (2005). Tetrahydrogestrinone induces a genomic signature typical of a potent anabolic steroid. J. Endocrinol. 184, 427–433. Lang, L. (1995). Strange brew: Assessing risk of chemical mixtures. Environ. Health Perspect. 103, 142–145. Lu, R., and Serrero, G. (1999). Resveratrol, a natural product derived from grape, exhibits antiestrogenic activity and inhibits the growth of human breast cancer cells. J. Cell. Physiol. 179, 297–304.

Moore, L. B., Maglich, J. M., McKee, D. D., Wisely, B., Willson, T. M., Kliewer, S. A., Lambert, M. H., and Moore, J. T. (2002). Pregnane X receptor (PXR), constitutive androstane receptor (CAR), and benzoate X receptor (BXR) define three pharmacologically distinct classes of nuclear receptors. Mol. Endocrinol. 16, 977–986. Moore, L. B., Parks, D. J., Jones, S. A., Bledsoe, R. K., Consler, T. G., Stimmel, J. B., Goodwin, B., Liddle, C., Blanchard, S. G., Willson, T. M., et al. (2000). Orphan nuclear receptors constitutive androstane receptor and pregnane X receptor share xenobiotic and steroid ligands. J. Biol. Chem. 275, 15122–15127. Mumtaz, M. M., Tully, D. B., El-Masri, H. A., and De Rosa, C. T. (2002). Gene induction studies and toxicity of chemical mixtures. Environ. Health Perspect. 110(Suppl. 6), 947–956. Pascussi, J. M., Gerbal-Chaloin, S., Duret, C., Daujat-Chavanieu, M., Vilarem, M. J., and Maurel, P. (2008). The tangle of nuclear receptors that controls xenobiotic metabolism and transport: Crosstalk and consequences. Annu. Rev. Pharmacol. Toxicol. 48, 1.1–1.31. Rajapakse, N., Silva, E., and Kortenkamp, A. (2002). Combining xenoestrogens at levels below individual no-observed-effect-concentrations dramatically enhances steroid hormone action. Environ. Health Perspect. 110, 917–921. Rencurel, F., Stenhouse, A., Hawley, S. A., Friedberg, T., Hardie, D. G., Sutherland, C., and Roland, C. (2005). AMP-activated protein kinase mediates phenobarbital induction of CYP2B gene expression in hepatocytes and a newly derived human hepatoma cell line. J. Biol. Chem. 280, 4367–4373. Rider, C. V., and LeBlanc, G. A. (2005). An integrated addition and interaction model for assessing toxicity of chemical mixtures. Toxicol. Sci. 87, 520–528. Robinson, C. P., Smith, P. W., McConnell, J. K., and Endecott, B. R. (1976). Comparison of protective effects of ethylestrenol, norbolethone, and spironolactone against lethality from acute doses of parathion and paraoxon in female rats. J. Pharm. Sci. 65, 595–596. Robinson, C. P., Smith, P. W., McConnell, J. K., and Endecott, B. R. (1979). Effect of treatment with norbolethone on parathion toxicity in male rats. Proc. Okla. Acad. Sci. 59, 74–75. Shindo, S., Numazawa, S., and Yoshida, T. (2007). A physiological role of AMP-activated protein kinase in phenobarbital-mediated constitutive androstane receptor activation and CYP2B induction. Biochem. J. 401, 735–741. Silva, E., Rajapakse, N., and Kortenkamp, A. (2002). Something from ‘‘nothing’’—Eight weak estrogenic chemicals combined at concentrations below NOECs produce significant mixture effects. Environ. Health Perspect. 36, 1751–1756. Solomon, H. M., Reich, S., Spirt, N., and Abrams, W. B. (1971). Interactions between digitoxin and other drugs in vitro and in vivo. Ann. N. Y. Acad. Sci. 179, 362–369. Sueyoshi, T., Kawamoto, T., Zelko, I., Honkakoshi, P., and Negishi, M. (1999). The repressed nuclear receptor CAR responds to phenobarbital in activating the human CYP2B6 gene. J. Biol. Chem. 274, 6043–6046. Suino, K., Peng, L., Reynolds, R., Li, Y., Cha, J. Y., Repa, J. J., Kliewer, S. A., and Xu, H. E. (2004). The nuclear xenobiotic receptor CAR: Structural determinants of constitutive activation and heterodimerization. Mol. Cell 16, 893–905. Swales, K., and Negishi, M. (2004). Car, driving into the future. Mol. Endocrinol. 18, 1589–1598.

A CONCENTRATION ADDITION MODEL FOR THE ACTIVATION OF CAR

105

Tzameli, I., Pissios, P., Schuetz, E. G., and Moore, D. D. (2000). The xenobiotic compound 1,4-bis[2-(3,5-dichloropyridyloxy)]benzene is an agonist ligand for the nuclear receptor CAR. Mol. Cell. Biol. 20, 2951–2958.

P450 oxidoreductase genes determines interindividual variability in basal expression and activity of a broad scope of xenobiotic metabolism genes in the human liver. Drug Metab. Dispos. 35, 1700–1710.

Wagner, M., Halilbasic, E., Marschall, H. U., Zollner, G., Fickert, P., Langner, C., Zatloukal, K., Denk, H., and Trauner, M. (2005). CAR and PXR agonists stimulate hepatic bile acid and bilirubin detoxification and elimination pathways in mice. Hepatology 42, 420–430.

Wyde, M. E., Bartolucci, E., Ueda, A., Zhang, H., Yan, B., Negishi, M., and You, L. (2003). The environmental pollutant 1,1-dichloro-2,2-bis (p-chlorophenyl)ethylene induces rat hepatic cytochrome P450 2B and 3A expression through the constitutive androstane receptor and pregnane X receptor, constitutive androstane receptor and pregnane X receptor. Mol. Pharmacol. 64, 474–481.

Wang, H., Faucette, S., Moore, R., Sueyoshi, T., Negishi, M., and LeCluyse, E. (2004). Human constitutive androstane receptor mediates induction of CYP2B6 gene expression by phenytoin. J. Biol. Chem. 279, 29295–29301. Wang, H., Faucette, S., Sueyoshi, T., Moore, R., Ferguson, S., Negishi, M., and LeCluyse, E. L. (2003). A novel distal enhancer module regulated by pregnane X receptor/constitutive androstane receptor is essential for the maximal induction of CYP2B6 gene expression. J. Biol. Chem. 278, 14146–14152. Watkins, R. E., Wisely, G. B., Moore, L. B., Collins, J. L., Lambert, M. H., Williams, S. P., Willson, T. M., Kliewer, S. A., and Redinbo, M. R. (2001). The human nuclear xenobiotic receptor PXR: Structural determinants of directed promiscuity. Science 292, 2329–2333. Wei, P., Zhang, J., Egan-Hafley, M., Liang, S., and Moore, D. D. (2000). The nuclear receptor CAR mediates specific xenobiotic induction of drug metabolism. Nature 407, 920–923. Wortham, M., Czerwinski, M., He, L., Parkinson, A., and Wan, Y. J. (2007). Expression of constitutive androstane receptor, hepatic nuclear factor 4a, and

Xu, Y., Hashizume, T., Shuhart, M. C., Davis, C. L., Nelson, W. L., Sakaki, T., Kalhorn, T. F., Watkins, P. B., Schuetz, E. G., and Thummel, K. E. (2006). Intestinal and hepatic CYP3A4 catalyze hydroxylation of 1 alpha, 25dihydroxyvitamin D(3): Implications for drug-induced osteomalacia. Mol. Pharmacol. 69, 56–65. Zhang, J., Huang, W., Chua, S. S., Wei, P., and Moore, D. D. (2002). Modulation of acetaminophen-induced hepatotoxicity by the xenobiotic receptor CAR. Science 298, 422–424. Zhang, J., Huang, W., Qatanani, M., Evans, R. M., and Moore, D. D. (2004). The constitutive androstane receptor and pregnane X receptor function coordinately to prevent bile acid-induced hepatotoxicity. J. Biol. Chem. 279, 49517–49522. Zollner, G., Marschall, H. U., Wagner, M., and Trauner, M. (2006). Role of nuclear receptors in the adaptive response to bile acids and cholestasis: Pathogenetic and therapeutic considerations. Mol. Pharmacol. 3, 231–251.