Targeted Salmon Gene Array (SalArray): A ... - ACS Publications

Feb 10, 2007 - The above-named findings demonstrate the interaction (cross-talk) between these two signalling systems. Inhibitory AhR−ER cross-talk ...
0 downloads 0 Views 235KB Size
474

Chem. Res. Toxicol. 2007, 20, 474-488

Targeted Salmon Gene Array (SalArray): A Toxicogenomic Tool for Gene Expression Profiling of Interactions Between Estrogen and Aryl Hydrocarbon Receptor Signalling Pathways Anne Skjetne Mortensen and Augustine Arukwe* Department of Biology, Norwegian UniVersity of Science and Technology (NTNU), Høgskoleringen 5, 7491 Trondheim, Norway ReceiVed October 6, 2006

In toxicogenomics, gene arrays are valuable tools in the identification of differentially expressed genes and potentially identify new gene biomarkers altered by exposure of organisms to xenobiotic compounds, either singly or as complex mixtures. In this study, we investigated the mechanisms of interaction between estrogen receptor (ER) and aryl hydrocarbon receptor (Ah receptor or AhR) signalling pathways using toxicogenomic approaches. First, we generated cDNA libraries using suppressive subtractive hybridization (SSH) of clones containing differentially expressed genes from Atlantic salmon (Salmo salar) separately exposed to ER and AhR agonists. Second, a targeted gene array (SalArray) was developed based on true-positive differentially expressed genes. In the experimental setup, primary cultures of salmon hepatocytes isolated by a two-step perfusion method were exposed for 48 h to nonylphenol (NP; 5 µM) and 3,3′,4,4′-tetrachlorobiphenyl (TCB; 1 µM), singly and combined, in the absence or presence of antagonists. Using a targeted SalArray, we demonstrate that exposure of salmon to NP singly or in combination with TCB produced differential gene expression patterns in salmon liver. Array analysis showed that exposure of hepatocytes to NP mainly altered genes involved in the estrogenic pathway, including genes for steroid hormone synthesis and metabolism. The anti-estrogenic properties of TCB were demonstrated in the array analysis as genes induced by NP were decreased by TCB. To study the effects of TCB on ER-mediated transcription, hepatocytes were treated for 48 h with tamoxifen (Tam; 1 µM) and ICI182,780 (ICI; 1 µM). The effect of AhR on ER-mediated transcription was investigated by blocking AhR activity with R-naphthoflavone (ANF; 0.1 and 1 µM). Quantitative real-time polymerase chain reactions confirmed the changes in expression of ERR, ERβ, vitellogenin (Vtg), zona radiata protein (Zr-protein), and vigilin for the ER pathway and AhRR, AhRβ, AhRR, ARNT, CYP1A1, UDPGT, and a 20S proteasome β-subunit for the AhR pathway. We found that exposure to NP and TCB both singly and in combination produced gene expression patterns that were negatively influenced by individual receptor antagonists. TCB caused decreased ER-mediated gene expression, and NP caused decreased AhR-mediated responses. Inhibition of AhR with ANF did not reverse the effect of TCB on ER-mediated transcription suggesting that AhRs do not have a direct role on TCB-mediated decreases of ER-mediated responses. In contrast, the inhibition of ER with Tam and ICI reversed the transcription of AhR-mediated responses (except AhRR). Taken together, the findings in the present study demonstrate a complex mode of ER-AhR interaction, possibly involving competition for common cofactors. This complex mode of interaction is further supported by the observation that the presence of ER antagonists potentiated the transcription of AhR isoforms and their mediated responses when TCB was given alone (more so for AhRβ). Thus, the inhibitory ER-AhR interactions can be used to further investigate specific genes found to be affected in our targeted SalArray chip that are important for the reproductive effects of endocrine disruptors. Introduction In the environment, contaminants typically occur as complex mixtures resulting in a situation whereby certain chemicals may have either synergistic or antagonistic effects upon the toxicological action of another contaminant. These complex interplays between environmental chemicals with different modes of action have been a subject for systematic investigations in our laboratory (1). The physiological effects of estrogens on growth and development are mediated through the estrogen receptors (ERR and ERβ). Ligand-bound ERs function as transcription factors by binding to estrogen response elements (EREs) upstream of estrogen responsive genes (2, 3). ERs belong to * To whom correspondence should be addressed. Tel: +47 73 596265. Fax: +47 73 591309. E-mail: [email protected].

the nuclear receptor (NR) superfamily that shares a common architecture of DNA-binding domain that contains a two-zinc finger structure at the center (3). A typical NR contains a variable N-terminal (A/B domain), a DNA-binding (C-domain), a linker (D-domain), and a conserved region (E/F domain) that contains the ligand-binding domain (4). In the A/B and LB domains, ERs contain, respectively, two transcriptional activation domains, namely, ligand-independent activation function (AF1) and ligand-dependent activation function (AF2), by which they are able to regulate target gene expression (4). In oviparous vertebrates, E2 is the major estrogen. Hepatic ERs are stimulated to induce the transcription of vitellogenin (Vtg), a precursor of yolk protein (5), and zona radiata protein (Zr-protein or Zrp) that forms the inner core of the eggshell (6). Natural estrogens released from the ovary stimulate Vtg and Zr-protein gene

10.1021/tx6002672 CCC: $37.00 © 2007 American Chemical Society Published on Web 02/10/2007

Toxicogenomics of ER-AhR Gene Interactions

Chem. Res. Toxicol., Vol. 20, No. 3, 2007 475

expression. Exogenous estrogenic chemicals (xenoestrogens) such as nonylphenol (NP) have been shown to induce hepatic expression of Vtg and Zr-protein genes in immature and male fish (7, 8). The aryl hydrocarbon receptor (AhR or Ah receptor) is a member of the basic-helix-loop-helix (bHLH)-Per-ARNT-Sim (PAS) family of transcription factors. AhR resides in the cytoplasmic compartment of cells as a multiprotein chaperone complex that includes the heat shock protein-90 (Hsp90; 9). Upon ligand binding, AhR undergoes conformational changes, resulting in chaperone release and subsequent translocation to the nucleus (9). In the nucleus, AhR dissociates from associated proteins and forms a heterodimer with AhR nuclear translocator (ARNT; 10, 11). Persistent chemicals such as polychlorinated biphenyls, polycyclic aromatic hydrocarbons, and dioxins have the ability to cause acute and chronic toxicity through activation of genes that encode enzymes involved in the oxidative metabolism of these compounds including cytochrome P450 enzymes (CYP1A1, CYP1A2, and CYP1B1) and uridinediphosphate glucuronosyltransferase (UDPGT). Therefore, the molecular basis for Vtg and CYP1A1 gene expression shows that the Vtg and CYP1A1 gene activations are receptor-mediated responses that are ligand structure-dependent interactions with ER and AhR, probably involving several isoforms. The antiestrogenic activities of AhR agonists have been observed (12-14). In teleost fish, both in vivo and in vitro studies have described that exposure to AhR agonists could be associated with reduced Vtg synthesis or impaired gonad development (15-17). The above-named findings demonstrate the interaction (cross-talk) between these two signalling systems. Inhibitory AhR-ER cross-talk has been demonstrated in breast cancer cells, rodent uterus, and mammary tumors (14). Recently in our laboratory, a bidirectional cross-talk between ER and AhR was demonstrated using rainbow trout primary hepatocyte cultures (1). Specifically, our data show that the AhR agonist 3,3′,4,4′tetrachlorobiphenyl (TCB) significantly inhibited NP-induced transcription of ERs and their signalling pathways and that NP inhibited TCB-induced CYP1A1 and AhR gene expressions (1). The combination of toxicology with genomics (toxicogenomics) has become a scientific subdiscipline of toxicology. In toxicological sciences, almost without exception, gene expression is altered as either a direct or an indirect result of toxicant exposure (18). Depending upon the severity and duration of the toxicant exposure, genomic analysis may be short-term toxicological responses leading to impacts on survival and reproduction (parental and offspring fitness). Therefore, gene array technology has become a powerful tool in molecular biology with potential to reveal genetic signatures in organisms that can be used to predict toxicity of these compounds (1, 19-21). Thus, the present study was designed with the objective of investigating the mechanism of interactions (cross-talk) between the ER and the AhR signalling pathways using targeted gene array technology. To establish the inhibitory effects of TCB (AhR ligand) on the expression of ER-controlled genes, we inversely evaluated if NP (ER agonists) exerts a similar control over AhRmediated gene expression patterns. In addition, we wanted to elucidate whether these effects will be dependent on cellular levels of the respective receptors or their isoforms. To study whether TCB will alter ER-mediated gene transcription independent of cellular ER content, we deployed partial (tamoxifen, Tam) and absolute (ICI182,780, ICI) ER antagonists to block the ERs, and inversely, R-naphthoflavone (ANF) was used to block the AhR.

Experimental Procedures Chemicals and Supplies. NP (85% of p-isomers) was purchased from Fluka Chemika-Biochemika (Buchs, Switzerland). TCB (PCB77; 99.7% pure) was purchased from Dr. Ehrenstorfer GmbH (Augsburg, Germany). Dubelco’s minimum essential medium (DMEM) with nonessential amino acid and without phenol red, fetal bovine serum (FBS), L-glutamine, and a TA cloning kit were purchased from Gibco-Invitrogen Life Technologies (Carlsbad, CA). Dimethyl sulfoxide (DMSO), 100× penicillin-streptomycinneomycin solution, collagenase, BSA, N-[2-hydroxyethyl]piperazine-N′-[2-ethanesulfonic acid], ethyleneglycol-bis-(β-aminoethylether) N,N′-tetraacetic acid, sodium dodecyl sulfate (SDS), 0.4% trypan blue, 17β-estradiol (E2), ANF, β-naphthoflavone (BNF), and Tam were purchased from Sigma Chemical (St. Louis, MO). ICI was purchased from Tocris BioScience (Bristol, England), and E.Z.N.A. total RNA kit for RNA purification was from Omega Bio-Tek (Doraville, GA). An IScript cDNA synthesis kit and iTAQSYBR green supermix with ROX were purchased from BioRad Laboratories (Hercules, CA). ULTRAhyb, Strip EZ RT kit, and RNAlater were from Ambion (Austin, TX), and [R33P]dATP was purchased from Amersham (Buckinghamshire, United Kingdom). GeneRuler 100 base pairs (bp) DNA ladder and deoxynucleotide triphosphates were purchased from Fermentas GmbH (St. Leon-Rot, Germany). Generation of Suppression cDNA Library. We generated a targeted cDNA library by performing suppressive subtractive hybridization (SSH) with liver samples from juvenile salmon exposed separately to ER agonists (NP and E2) and AhR agonists (TCB and BNF) and used them against untreated samples (solvent control samples). Juvenile immature Atlantic salmon (mean weight and length, 10 g and 9 cm, respectively) were obtained from Stjørdal hatcheries (Mera˚ker, Norway) and kept in 70 L aquaria at 10 ( 0.5 °C and 12:12 h photoperiod at the Department of Biology, Norwegian University of Science and Technology (NTNU) animal holding facilities in Trondheim. The fish were exposed intraperitoneally (ip) to E2, NP, BNF, or TCB at 10, 100, 100, or 2.5 mg/kg, respectively, using corn oil:acetone (1:1) as a vehicle. At 3 days postexposure, fish were sacrificed and liver samples were collected and transferred to RNAlater. The total RNA was isolated using phenol:chloroform. Messenger RNA (mRNA) was isolated from total RNA using an Oligotex mRNA kit (Qiagen, Valencia, CA) and used for subtractive hybridization. The SSH experiment was performed in the forward and reverse directions to obtain upregulated and down-regulated genes, respectively. Subtractive hybridizations were constructed using the Clontech (Palo Alto, CA) SSH kit following the manufacturer’s protocol as outlined below and as previously by Larkin et al. (19, 22). The SSH was performed by EcoArray Inc. (Alachua, FL) under contract. Sequenced clones were analyzed, annotated, and downloaded to the GenBank EST database under the title SalArray SSH cDNA library. Array Spotting. The 300-clone salmon membrane cDNA array (SalArray) was constructed using clones with unique expression patterns that were either up- or down-regulated in the SSH. All clones were polymerase chain reaction (PCR) amplified to yield an 8 µg sample. These were verified by agarose gel electrophoresis, then purified using MultiScreen-PCR96 Filter Plates (Millipore, Billerica, MA), and quantified to accuracy using a NanoDrop spectrophotometer (NanoDrop Technologies, Wilmington, DE). Three anti-sense Arabidopsis thaliana mRNAs were added to the array as a control for loading differences, and the array was printed by EcoArray Inc. under contract. Probe Labeling and Hybridization. Hybridization of the array was performed using 33P-labeled cDNA probes. The probes were generated using 2 µg of DNase-treated total RNA from the respective exposure conditions and 1.2 µL of A. thaliana mRNA (the same spike used in array printing) as internal standard as described in Strip EZ reverse-transcriptase Kit protocol (Ambion). Probes were purified using nucleotide removal kit (Qiagen) according to the manufacturer’s protocol. The amount of probe needed for hybridization was calculated after counting 1 µL of

476 Chem. Res. Toxicol., Vol. 20, No. 3, 2007

Figure 1. Ratio intensity plots showing loess normalization. Normalization was performed to minimize the effects of systematic errors between arrays. First, the product of spot signal intensity between exposed samples (S) and the respective control sample (C) were log2 transformed, and then, the ratio between S and C was found and log2 transformed. The results are visualized in a ratio-intensity plot (RI plot). Presentation of data prior to loess normalization (A) and after loess normalization (B) showing the trendline function in the top left corner of each figure. Note that in panel B, the distribution of data is centered around y ) 0.

purified probe in a scintillation counter. For each RNA sample, one membrane containing duplicate spots was prehybridized for 2 h in ULTRAhyb (Ambion). Hybridization was performed overnight at 60 °C using ULTRAhyb for probe dilution. The membranes were washed (2 × 15 min) in 2× SSC/0.5% SDS at 60 °C and then washed again (2 × 15 min) at high stringency in 0.5× SSC/0.5% SDS at 60 °C before exposure to BAS Image plates. Radioactivity was scanned using a Phosphoimager FLA-2000 (Fuji, Japan). Normalization and Array Analysis. Spot-density data for each membrane were individually quantified by Array Gauge v2.1 (Fugi Film), exported to Microsoft Excel, and coded by gene, treatment, and spot duplicates. These hierarchical levels of identification were critical for appropriate statistical analyses, as explained below. The general background of each membrane was subtracted from the average spot-intensity values for the duplicate spots on the membrane. Thereafter, the background-normalized spot-intensity values were further normalized with the A. thaliana spikes on the same membrane. Expression levels of genes were expressed as foldchange relative to control by dividing the signal intensity of exposed samples (Si) by the signal intensity in the respective control sample (Ci). The ratio was log2 transformed, and the measured log2(Si/Ci) ratio was visualized as a function of the log2(Si*Ci) product intensities in a ratio-intensity plot (RI plot) (Figure 1A). Finally, loess normalization was performed using S-plus statistic software 6.2 (Insightful Corp.) to minimize systematic deviations (RNA quality, probe labeling, hybridization, and development of image) in the log2(Si/Ci) ratio values of spot intensity levels between exposed samples and controls (Figure 1B) (20). The loess and spike normalizations were the only data normalization performed on the array data, since other options such as normalization with a suite of reference genes did not give useful results. The normalized ratio log2(r*) was evaluated, and stringent criteria were used to filter for genes that were regulated at least 0.3-fold as compared to their

Mortensen and Arukwe respective controls. All genes were sorted and categorized based on functional pathways. Because the statistical model used in this study is conservative, variance in duplicate spot intensities in the array data was evaluated using coefficient of variation (CV; %), which was usually between 5 and 20%. Collagenase Perfusion and Isolation of Hepatocytes. Juvenile Atlantic salmon (Salmo salar) of approximately 400-500 g were supplied by Marine Harvest AS (Dyrvik, Norway) and kept at the animal holding facilities at the Biology Department, NTNU. Fish were supplied with continuously running saltwater at a constant temperature of 10 °C. Prior to liver perfusion, all glassware and instruments were autoclaved and solutions were filtration sterilized by using 0.22 µm Millipore filter (Millipore AS, Oslo, Norway). Hepatocytes were isolated from three individuals by a two-step perfusion as described by Berry and Friend (23) and modified by Andersson et al. (24). The cell suspension was filtered through a 150 µM nylon monofilament filter and centrifuged at 50g for 5 min. Cells were washed three times with serum-free medium and finally resuspended in complete medium. Following collagenase perfusion and isolation of hepatocytes, the viability of cells was determined by the trypan blue exclusion method. A cell viability value of >90% was a criterion for further use of the cells. Cells were plated on a 35 mm Primaria culture plate (Becton Dickinson Labware) at the recommended density for monolayer cells of 5 × 106 cells in 3 mL of DMEM medium (without phenol red) containing 2.5% (v/v) FBS, 0.3 g/L glutamine, and 1% (v/v) penicillin-streptomycin-neomycin solution. The cells were cultured at 10 °C in a sterile incubator without additional O2/CO2 for 48 h prior to chemical exposure. Exposure of Hepatocytes. Primary cultures of salmon hepatocytes were exposed in triplicate to NP and TCB at 5 and 1 µM, respectively. The chemicals were given singly and also in combination. Previously, we found these concentrations to be optimal in in vitro concentrations for ER-AhR interactions in salmonids (1). To study if TCB altered ER-mediated gene transcription independent of cellular ER content, we partially and completely inhibited ER activity using Tam and ICI, respectively. Both ER antagonists were administered at 1 µM concentration based on previous salmon in vitro studies (25). To establish the possible roles of AhR in ERmediated transcription, AhR activity was blocked using the synthetic flavonoid, ANF, at 0.1 and 1 µM. ANF is a known AhR antagonist in cultured hepatocytes (26). After 48 h of preculture, cells were exposed (six plates for each exposure group) to 0.01% DMSO (control), 5 µM NP, and 1 µM TCB singly and also in combination with the individual ER antagonists. The exposure time was chosen based on a previous study showing that 48 h of exposure of hepatocytes was optimal for ER- and AhR-mediated responses (Mortensen and Arukwe, unpublished data). Media was replaced with fresh media containing the respective test chemical at 24 h. Media and cells (triplicates for each exposure group) were harvested after 48 h exposure and lysed in E.Z.N.A. lysis buffer for total RNA isolation according to the manufacturer;s protocol (Omega Bio-Tek). Quantitative PCR. Total cDNA for the real-time PCR reactions was generated from DNase, and 1 µg of total RNA from all samples was treated using poly-T primers from the iScript cDNA synthesis Kit as described by the manufacturer (Bio-Rad). Quantitative (realtime) PCR was used for evaluating gene expression profiles. For each treatment, the expression of individual gene targets was analyzed using the Mx3000P Real-Time PCR System (Stratagene, La Jolla, CA). Each 25 µL DNA amplification reaction contained 12.5 µL of iTaqSYBR Green Supermix with ROX (Bio-Rad), 1 µL of cDNA, and 200 nM each forward and reverse primers. The three-step real-time PCR program included an enzyme activation step at 95 °C (5 min) and 40 cycles of 95 °C (30 s) and 55-65 °C for 30 s, depending on the primers used (see Table 1 in the Supporting Information), and 72 °C (30 s). Controls lacking cDNA template (minus RT sample) were included to determine the specificity of target cDNA amplification as described previously (1, 27). Briefly, cycle threshold (Ct) values obtained were converted into mRNA copy number using standard plots of Ct vs log copy

Toxicogenomics of ER-AhR Gene Interactions number. The criterion for using the standard curve is based on equal amplification efficiency with unknown samples, and this is usually checked prior to extrapolating unknown samples to the standard curve. The standard plots were generated for each target sequence using known amounts of plasmid containing the amplicon of interest. Data obtained from triplicate runs for target cDNA amplification were averaged and expressed as ng/µg of initial total RNA used for reverse transcriptase (cDNA) reaction. Standard errors were calculated using S-plus statistic software 6.2 (Insightful Corp.). Statistical differences among treatment groups were tested using analysis of variance (ANOVA), and comparisons of different exposure-treated and control groups were performed using Tukey’s multiple comparison test. For all of the tests, the level of significance was set at p < 0.05, unless otherwise stated.

Results Array Gene Expression Pattern after Exposure to NP Singly or in Combination with TCB. Salmon hepatocytes exposed to NP singly and also in combination with TCB were hybridized to the targeted SalArray; this information is presented in Table 2 of the Supporting Information. Genes involved in the AhR pathway were assigned to the drug, lipid, glucosis, and retinol metabolism/homeostasis category. The array data indicated that the transcription of AhRR, AhRβ, and AhRR decreased in hepatocytes exposed to NP. On the contrary, the phase II enzyme glutathione S-transferase (GST) increased in cells exposed to NP as compared to its solvent control (see Table 1, and for a complete list of all genes, see Table 2 in the Supporting Information). The mRNA expression of AhRR, AhRβ, AhRR, ARNT, and CYP1A1 increased in cells exposed to TCB alone or in combination to NP as compared to solvent control and NP, respectively. Genes in the estrogenic pathways were assigned to a functional category by the same name. The mRNA expression patterns of most genes in this pathway were altered in response to all treatments (see Table 1, and for a complete list of all genes, see Table 2 in the Supporting Information). While NP-exposed samples showed increased transcription of Vtg precursor, Vtg receptor, and Zr-protein, the expression of vigilin mRNA showed decreased levels after NP exposure. Hepatocytes exposed to combined NP and TCB showed increased ERR levels as compared with NP exposure alone. In contrast, Vtg, Zr-protein, and vigilin mRNA levels were decreased in combined NP and TCB exposure groups, as compared to NP alone. The expression levels of targeted genes (i.e., direct ER-mediated genes) in the estrogenic pathways were not altered in cells exposed to TCB alone. Quantitative Confirmation of Differentially Expressed Genes. On the basis of the results obtained from the array analysis, we selected four genes (ERR, ERβ, Vtg, and Zrproteins) belonging to the ER pathway and six genes (AhRR, AhRβ, AhRR, ARNT, CYP1A1, and UDPGT) in the AhR pathway together with a 20S proteasome β-subunit and vigilin (potential effect mediators) for quantitative analysis using realtime PCR with gene specific primers in three different hepatocyte cultures for each exposure group. The primary criterion for selecting these genes for real-time PCR validation is because they belong to our targeted study objectives, namely, the interactions between ER and AhR signalling pathways. Several isoforms of ARNT and UDPGT have been characterized in fish, and the primer sequences used in the real-time PCR assays were designed based on conserved regions of these genes. In this report, AhR1 and AhR2 will be used synonymously with AhRR and AhRβ, respectively. Genes Altered in the ER Signalling Cascade and Effects of Antagonists. The expression pattern of the ERR after

Chem. Res. Toxicol., Vol. 20, No. 3, 2007 477 Table 1. Genes in the Metabolic and Estrogenic Pathways Whose Expression Patterns Were Either Up- or Down-regulated by Exposure of Primary Culture of Salmon Hepatocytes to NP, Singly or in Combination with TCB, Using Targeted Salmon cDNA Array (SalArray) Analysisa name

GenBank no.

NP

TCB

drug, lipid, glucosis, and retinol metabolism/homeostasis aryl hydrocarbon receptor-R (AhRR) DQ358692 0.46 2.94 aryl hydrocarbon receptor-β (AhRβ) DQ358693 0.59 1.48 aryl hydrocarbon receptor DQ372978 1.92 1.25 repressor (AhRR) aryl hydrocarbon receptor DQ367887 1.16 1.27 translocator (ARNT) β-3a-adrenergic receptor (β3AGR) DQ372979 0.92 0.96 cyclooxygenase-2 (COX2) AJ238307 1.74 1.22 cytochrome P450 1A1 (CYP1A1) AF364076 0.79 1.29 cytochrome P450 1A3 (CYP1A3) DY802103 1.64 0.87 cytochrome P450 3A (CYP3A) DQ361036 1.21 1.22 glucose-6-phosphatase AF120150 1.28 0.96 catalytic subunit glutathione S-transferase A DY802071 0.92 1.20 (GST-A) (GST class-θ) gluthathione-S-transferase (GST) DQ367889 1.25 0.75 isotocin DY802181 0.82 1.25 L-lactate dehydrogenase B DY802175 1.32 0.66 chain (LDH-B) multidrug resistance (MDR) protein AJ344049 1.89 1.24 peroxisome proliferatorDQ294237 2.91 0.80 activated receptor-R (PPARR) peroxisome proliferatorAF342945 1.39 1.68 activated receptor-β (PPARβ) peroxisome proliferatorAJ292962 1.04 1.27 activated receptor-γ (PPARγ) retinoic acid induced gene G protein DY802127 0.59 0.91 retinoic acid receptor-R (RARR) AF290610 1.08 0.74 retinoid X receptor-R (RXRR) AJ969439 0.50 1.26 retinoid X receptor-β (RXRβ) CX725791 0.87 0.64 UDP-glucuronosyltransferase DY802180 0.84 0.78 estrogenic pathways 3β-hydroxysteroiddehydrogenase S72665 (3β-HSD) androgen receptor (AR) DQ367886 apolipoprotein A-I DY802197 cytochrome P45011R DQ361039 (CYP11R; P450ssc) cytochrome P45011β (CYP11β) DQ352841 cytochrome P45019R (aromatase R) DQ361037 cytochrome P45019β (aromatase β) DQ361038 estrogen receptor-R (ERR) DQ009007 estrogen receptor-β (ERβ) AY508959 steroidogenic acute regulatory DQ415678 protein (StAR) vigilin (high-density DY802195 lipoprotein-binding protein) Vtg precursor DY802153 Vtg receptor AY049959 zona pellucida sperm-binding protein B DY802138 Zr-protein DY802149

NP + TCB 1.28 1.11 1.26 0.96 0.79 1.57 1.22 0.92 1.24 1.16 0.92 1.47 1.05 1.36 1.64 2.31 1.37 1.41 0.73 1.18 0.89 1.13 1.13

1.04 1.28

1.61

0.95 1.48 0.76 0.96 0.76 1.37

1.22 0.39 1.26

0.40 1.72 1.24 1.07 0.83 1.97

2.60 1.31 1.25 0.90 0.98 1.14

1.08 1.80 1.57 1.26 0.75 1.74

0.60 1.18

0.52

1.79 1.41 0.83 1.32

0.30 1.13 1.17 0.46

1.04 1.08 0.96 0.75

a NP, NP exposure group as compared with solvent control; TCB, TCB exposure group as compared with solvent control; NP + TCB, NP exposure alone as compared with combined NP and TCB exposure. Each data point is derived from one biological replicate.

exposure to NP singly and in combination with TCB is shown in Figure 2A. Exposure of hepatocytes to NP alone significantly increased (2.6-fold) the expression of ERR, while combined exposure with TCB did not produce a significant effect on the NP-mediated increase on ERR (Figure 2A). In hepatocytes exposed to NP singly and in combination with TCB, the partial ER antagonist, Tam, caused a significant decrease of ERR mRNA expression (Figure 2A). On the contrary, when hepatocytes were cotreated with the absolute ER antagonist ICI, the ERR expression was totally inhibited both in the NP-exposed group alone and in combination with TCB (Figure 2A).

478 Chem. Res. Toxicol., Vol. 20, No. 3, 2007

Mortensen and Arukwe

Figure 2. Expression pattern of ERR (A), ERβ (B), Vtg (C), zona radiata protein (Zr-protein; D), and vigilin (E) mRNA levels in salmon primary hepatocyte cultures exposed to NP (5 µM) and TCB (5 µM), singly or combined, in the absence or presence of tamoxifen (Tam, 1 µM) or ICI182,780 (ICI, 1 µM). Expression of mRNA levels was quantified using quantitative PCR with gene specific primer pairs. Data are given as percentage (%-age) of the solvent control ( standard error of the mean (SEM; n ) 3). Different letters denote exposure group means that are significantly different for the respective mRNA expression using ANOVA followed by Tukey’s multiple comparison test (p < 0.05).

For ERβ, a different pattern of expression was observed between the NP-exposed group and the solvent control (Figure 2B). NP exposure did not produce transcriptional change in ERβ expression, but when given in combination with TCB, a significant 3.4-fold decrease of ERβ was observed (Figure 2B). Similar to ERR, Tam and ICI produced significant decreases

of ERβ levels both in the NP-treated group alone and in combination with TCB (Figure 2B). ICI produced an almost total inhibition of ERβ gene expression in the NP exposure group alone and in combination with TCB. Interestingly, TCB exposure alone produced a significant two-fold decrease of ERβ, as compared with control (Figure 2B). The expression of Vtg

Toxicogenomics of ER-AhR Gene Interactions

was massively induced after exposure to NP alone, resulting in a 28-fold increase of mRNA levels as compared with solvent control (Figure 2C). When hepatocytes were exposed to NP and TCB in combination, the NP-induced Vtg expression was reduced to a 16-fold increase (Figure 2C). Tam decreased NPinduced Vtg gene expression both in the NP-treated group alone or in combination with TCB (Figure 2C). A total inhibition of Vtg mRNA transcription was observed when cells were exposed with NP alone or in combination with TCB, in the presence of ICI (Figure 2C). The mRNA expression of Zr-proteins increased 1.8-fold in NP-exposed hepatocytes and 1.5-fold in cells exposed to combined NP and TCB (Figure 2D). The presence of Tam significantly reduced NP-induced Zr-protein mRNA expression (Figure 2D). In contrast to Vtg, the presence of ICI did not produce a total inhibition of Zr-protein gene expression, when hepatocytes were exposed to either NP alone or in combination with TCB (Figure 2D). The expression of vigilin mRNA followed a similar pattern with Vtg and Zr-protein in NP exposure alone in the presence and absence of Tam and ICI (Figure 2E). TCB alone did not significantly alter vigilin mRNA levels, but the effects of Tam and ICI were enhanced in the presence of TCB (Figure 2E). Genes Altered in the AhR Signalling Cascade and Effect of Antagonist. The expression pattern of AhRR after exposure to TCB singly and in combination with NP is shown in Figure 3A. While AhRR expression increased 1.5-fold in response to TCB exposure, the combined TCB and NP exposure resulted in decreased AhRR expression (Figure 3A). The presence of 0.1 and 1 µM ANF produced a concentration-dependent decrease (two- and six-fold, respectively) of AhRR mRNA in TCB-exposed cells (Figure 3A). In the combined TCB and NP treatment, the presence of ANF concentrations produced a twofold decrease of AhRR expression as compared to TCB treatment alone (Figure 3A). The expression levels of AhRβ mRNA were decreased as a response to TCB exposure singly and in combination with NP (Figure 3B). Treatment with 0.1 and 1 µM ANF produced a concentration-dependent decrease (2.6- and 41-fold respectively) of AhRβ mRNA expression after exposure to TCB alone (Figure 3B). In the combined TCB and NP exposure, the presence of ANF concentrations produced a different pattern of effect on AhRβ mRNA expression as compared to TCB exposure alone, with 0.1 µM causing a higher decrease as compared to 1 µM ANF (Figure 3B). In hepatocytes exposed to TCB singly, AhRR mRNA levels increased by 15fold (Figure 3C), and when TCB was given in combination with NP, the increase in AhRR expression was reduced to sevenfold as compared to solvent control (Figure 3C). The presence of ANF resulted in a concentration-dependent decrease of AhRR expression after exposure to TCB singly and in combination with NP (Figure 3C). For ARNT, exposure to TCB alone produced a 0.6-fold significant decrease of ARNT mRNA expression, and while the presence of 0.1 µM ANF restored ARNT levels to control level, 1 µM ANF reduced it back to TCB exposure level (Figure 3D). When TCB was given in combination with NP, an attenuation of TCB-mediated decrease (2.3-fold) of ARNT levels was observed (Figure 3D). Again, while 0.1 µM ANF decreased ARNT expression levels in cells exposed to combined TCB and NP concentrations, 1 µM ANF did not significantly alter ARNT mRNA expression (although a slight increase was observed; Figure 3D). The expression pattern of CYP1A1 is shown in Figure 3E. Exposure of cells to TCB singly produced a 17-fold increase of CYP1A1, and the presence of ANF produced a concentrationdependent significant decrease in CYP1A1 expression (Figure

Chem. Res. Toxicol., Vol. 20, No. 3, 2007 479

3E). When cells were exposed to combined TCB and NP, the TCB-induced CYP1A1 mRNA expression was reduced to 11fold from 17-fold (Figure 3E). The presence of ANF produced a significant decrease of CYP1A1 mRNA expression levels in hepatocytes exposed to TCB and NP in combination, but this effect was not ANF concentration-dependent (Figure 3E). UDPGT mRNA expression levels increased 2.3-fold after exposure to TCB alone, and the presence of ANF produced concentration-dependent significant decreases in UDPGT expression levels (Figure 3F). When TCB was given in combination with NP, a significant alteration in TCB-induced UDPGT expression was observed and the presence of ANF produced concentrationdependent significant decreases in UDPGT expression (Figure 3F). The expression of a proteasome unit mRNA followed a similar pattern with AhR-mediated gene expression after exposure to TCB alone and in combination with NP in the presence and absence of ANF concentrations (Figure 3G). ER and AhR Interactions. 1. TCB Modulation of ER Signalling Pathway. The AhR antagonist, ANF, at 0.1 or 1 µM produced significant decreases of ERR expression in hepatocytes exposed to NP and TCB in combination (Figure 4A). A similar effect was also observed in cells exposed to TCB alone showing an ANF concentrationdependent decrease of ERR mRNA expression (Figure 4A). For ERβ, while cells exposed to combined NP and TCB concentrations with 0.1 µM ANF produced a seven-fold reduction of ERβ expression, 1 µM ANF did not alter the ERβ expression levels as compared to when NP and TCB were given in combination (Figure 4B). Interestingly, when TCB was given alone in the presence of 0.1 µM ANF, the expression level of ERβ was restored back to control level, as compared with 1 µM ANF that produced decreased expression level (Figure 4B). The NPinduced Vtg expression in cells exposed to NP and TCB in combination was totally inhibited in the presence of 0.1 µM ANF (Figure 4C). Treatment with 1 µM ANF restored the NPinduced Vtg expression to control levels (Figure 4C). While the expression of Vtg in TCB-exposed samples showed a twofold increase in the presence of 0.1 µM ANF, 1 µM ANF did not produce any effect (Figure 4C). For Zr-protein, hepatocytes exposed to combined NP and TCB in the presence of 0.1 and 1 µM ANF produced significant reductions (more so for 0.1 µM ANF) of Zr-protein expression (Figure 4D). Similar with ERβ and Vtg expressions, when TCB was given alone in the presence of 0.1 µM ANF, the expression level of Zr-protein was restored back to control level (TCB by itself caused a decrease), as compared with 1 µM ANF (Figure 4D). The expression of vigilin mRNA followed a similar pattern with Vtg and Zr-protein in NP exposure alone in the presence or absence of ANF concentrations (Figure 4E). TCB alone did not significantly alter vigilin levels, but the presence of ANF produced a concentration specific decrease of vigilin mRNA TCB exposure alone or combined with NP (Figure 4E). 2. NP Modulation of AhR Signalling Pathway. Exposure of salmon hepatocytes to NP did not significantly alter transcriptional changes of AhRR, and when NP was given in combination to Tam or ICI, a two-fold reduction of AhRR mRNA was observed (Figure 5A). Exposure of cells to TCB alone produced a 1.5-fold significant increase of AhRR transcript levels (Figure 5A). While exposure with TCB in the presence of ICI did not alter the TCB-induced AhRR transcription, the presence of Tam decreased the AhRR mRNA to control level (Figure 5A). In the combined TCB and NP exposure, decreased AhRR mRNA was observed, as compared to TCB exposure

480 Chem. Res. Toxicol., Vol. 20, No. 3, 2007

Mortensen and Arukwe

Figure 3. Expression pattern of AhRR (A), AhRβ (B), AhRR (C), ARNT (D), CYP1A1 (E), UDPGT (F), and 20S proteasome subunit (G) mRNA levels in salmon primary hepatocyte cultures exposed to TCB (1 µM) and NP (5 µM), singly or combined, in the absence or presence of different concentrations (0.1 and 1 µM) of ANF. Expression of mRNA levels was quantified using quantitative PCR with gene specific primer pairs. Data are given as percentage (%-age) of the solvent control ( standard error of the mean (SEM; n ) 3). Different letters denote exposure group means that are significantly different for the respective mRNA expression using ANOVA followed by Tukey’s multiple comparison test (p < 0.05).

Toxicogenomics of ER-AhR Gene Interactions

Chem. Res. Toxicol., Vol. 20, No. 3, 2007 481

Figure 4. Transcriptional changes of ERR (A), ERβ (B), Vtg (C), zona radiata protein (Zr-protein; D), and vigilin (E) mRNA levels in salmon primary hepatocyte cultures exposed to NP (5 µM) and TCB (1 µM), singly or combined, in the absence or presence of different concentrations (0.1 and 1 µM) of the Ah-receptor antagonist, ANF. Expression of mRNA levels was quantified using quantitative PCR with gene specific primer pairs. Data are given as percentage (%-age) of the solvent control ( standard error of the mean (SEM; n ) 3). Different letters denote exposure group means that are significantly different for the respective mRNA expression using ANOVA followed by Tukey’s multiple comparison test (p < 0.05).

alone (Figure 5A). While the presence of Tam did not affect this decrease, ICI attenuated the decreased in AhRR expression in cells treated with combined TCB and NP, as compared to TCB alone (Figure 5A). Exposure to NP and TCB or combination of both chemicals did not alter AhRβ mRNA expression as compared to control (Figure 5B), but when NP was given in the presence of Tam, a 3.5-fold increase of AhRβ was observed as compared to when NP was given alone (Figure 5B). Similarly, when TCB was given in the presence of Tam and ICI, respective 8- and 8.5-fold increases of AhRβ mRNA expression were observed as compared to when TCB was given alone (Figure 5B). In cells exposed to combined TCB and NP, a different pattern of Tam and ICI effect was observed. While combined TCB and NP exposure in the presence of ICI did not alter AhRβ expression, the presence of Tam produced a 5.5-fold increase in the expression level of AhRβ as compared to when TCB

and NP were given alone (Figure 5B). The expression of AhRR did not change after exposure to NP alone or in the presence of Tam, but when NP was given in the presence of ICI, a fivefold increase of AhRR mRNA was observed (Figure 5C). A 15-fold increase of AhRR expression was observed in salmon hepatocytes exposed to TCB alone, and the presence of Tam and ICI caused nonsignificant reduction of AhRR mRNA expression (Figure 5C). In the combined TCB and NP exposure, the TCB-induced expression of AhRR expression was reduced to five-fold from 15-fold, and the presence of Tam and ICI further decreased AhRR expression in the combined TCB and NP exposure (more so with ICI; Figure 5C). For ARNT, NP exposure alone did not alter ARNT expression and the presence of ICI produced a two-fold significant decrease (but not Tam) of ARNT expression level (Figure 5D). Exposure to TCB alone or in combination with NP resulted in decreased expression

482 Chem. Res. Toxicol., Vol. 20, No. 3, 2007

Mortensen and Arukwe

Figure 5. Transcriptional changes of expression pattern of AhRR (A), AhRβ (B), AhRR (C), ARNT (D), CYP1A1 (E), UDPGT (F), and 20S proteasome subunit (G) mRNA levels in salmon primary hepatocyte cultures exposed to TCB (1 µM) and NP (5 µM), singly or combined, in the absence or presence of the ER antagonists, Tam (1 µM) and ICI (1 µM). Expression of mRNA levels was quantified using quantitative PCR with gene specific primer pairs. Data are given as percentage (%-age) of the solvent control ( standard error of the mean (SEM; n ) 3). Different letters denote exposure group means that are significantly different for the respective mRNA expression using ANOVA followed by Tukey’s multiple comparison test (p < 0.05).

Toxicogenomics of ER-AhR Gene Interactions

levels of ARNT (Figure 5D). While the presence of Tam produced a slight increase of ARNT expression levels, the presence of ICI produced significant increases of ARNT as compared to when TCB was given alone or in combination with NP (Figure 5D). The ER antagonists (Tam and ICI) did not produce alterations in TCB-induced CYP1A1 expression (Figure 5E). Exposure of hepatocytes to TCB alone produced a 10-fold increase of CYP1A1 mRNA expression, and combined TCB and NP exposure significantly decreased CYP1A1 expression (Figure 5E). The presence of Tam did not alter the expression of CYP1A1 in cells exposed to TCB singly or in combination with NP (Figure 5E). On the contrary, the presence of ICI produced a two-fold significant decrease of CYP1A1 expression when TCB was given in combination with NP (Figure 5E). Exposure of cells to TCB produced a 1.5-fold increase of UDPGT expression, while combined exposure with NP reduced TCBinduced UDPGT expression (Figure 5F). UDPGT mRNA expression was reduced to control levels in the presence of Tam when TCB was given alone or in combination with NP (Figure 5F). The presence of ICI had no effect under these exposure conditions (Figure 5F). NP exposure alone did not alter UDPGT expression, but treatment with both Tam and ICI produced a two-fold significant reduction of UDPGT transcript level (Figure 5F). The expression of a proteasome subunit mRNA only followed a similar pattern with CYP1A1 gene expression after exposure to TCB alone in the presence or absence of ER antagonists (Tam and ICI; Figure 5G). In the combined TCB and NP exposure, the presence of Tam produced a complete inhibition of the proteasome subunit (Figure 5G).

Discussion Evaluation of SalArray Data. The molecular basis for ERand AhR-mediated gene expressions shows that these gene activations are receptor-mediated, at least in part, and responses that are ligand structure-dependent interactions with ER and AhR probably involve several isoforms and cofactors. Previously, we reported that TCB, an AhR agonist with known antiestrogenic activity, potentiated and inhibited in vivo ERmediated NP-induced Vtg and Zr-protein gene and protein expression patterns in Atlantic salmon, with respective responses being dependent on TCB and NP dose ratios and sequential order of exposure and influenced by seasonal changes (16). To further characterize the molecular mechanism(s) behind these responses, the combined analytical power of SSH, quantitative (real-time) PCR, and primary salmon hepatocyte culture was used with single concentration of NP and TCB, given singly and also in combination, in studying the differential gene expression patterns of chemically responsive genes. In addition, exposure with NP and TCB was also performed in the presence and absence of ER and AhR antagonists to study the underlying mechanisms. First, we generated cDNA libraries of clones containing differentially expressed genes with combined liver samples from salmon exposed separately to ER agonists (NP and E2) and AhR agonists (TCB and BNF) and controls (carrier solvent exposed). mRNA differences between the combined chemically exposed liver samples and the control were assessed using SSH to enrich for genes that were responsive to the respective receptor agonist in the liver. Because the SSH technique favors the enrichment of high abundance transcripts and is therefore very susceptible to a high false-positive rate (19, 22), we performed the hybridization in both forward (up-regulated) and reverse (downregulated) directions to maximize the detection and identification

Chem. Res. Toxicol., Vol. 20, No. 3, 2007 483

of agonists responsive genes. Second, a targeted gene array was developed based on apparent true-positive differentially expressed genes. Therefore, our targeted SalArray represents a unique suite of differentially expressed genes that were either up- or down-regulated in response to ER and AhR agonist exposure. After sequencing, BLAST identification, genes were annotated and downloaded to EST GenBank. Because the SSH technique did not identify genes in the transcription machinery in either the ER (ER isoforms), the AhR (AhR isoforms, ARNT, and AhRR), or the steroidogenesis (aromatase gene isoforms, StAR, and 3β-HSD) pathways in the liver subtraction libraries, these genes were generated by normal PCR, cloned, sequenced, and subsequently measured by array analysis in this study. We also measured the expression of ER and AhR isoforms, ARNT, and AhRR expression by real-time PCR in hepatocyte samples used for mechanistic studies (see later). Using a targeted salmon gene array, we demonstrate that exposure of salmon to NP and TCB singly or in combination produced differential gene expression patterns in salmon liver. Array analysis showed that exposure of hepatocytes to NP mainly altered expression of genes involved in the estrogenic pathway, including steroid hormone synthesis and metabolism and many genes involved in the AhR signalling pathway. In addition, our array analysis also demonstrated that genes involved in steroid hormone synthesis and metabolism such as aromatase gene isoforms, StAR, CYP11β, and 3β-HSD were affected by all three exposure conditions. Endocrine toxicology research has mainly focused on estrogenicity that involves direct ER-mediated effects. Several studies have reported the expression of steroidogenic enzymes and proteins in fish hepatic tissues (28-30). To our knowledge, the present study represents the first report on the parallel expression of aromatase gene isoforms, StAR, CYP11β, and 3β-HSD in fish liver, and the subsequent modulation by an estrogen mimic. The teleost liver is not a typical steroid-producing organ but rather a steroidmetabolizing organ. Therefore, the physiological role or consequences of NP-mediated expression of aromatase isoforms, StAR, and 3β-HSD in salmon liver is yet to be determined. Nevertheless, the induction of steroidogenic enzymes and proteins is highly tissue and cell type specific and is controlled by different promoters and second messenger pathways. These pathways may provide various targets for interaction with xenobiotics, and studies are ongoing in our laboratory to evaluate the quantitative expression patterns of hepatic steroidogenic genes and proteins after exposure to environmental contaminants. In our array analysis, we compared cells exposed to NP singly and with TCB in combination. The antiestrogenic properties of TCB were illustrated in the array analysis as genes induced by NP were decreased by TCB. This antiestrogenic effect of TCB has previously been investigated by single gene analysis in vivo and in vitro fish studies (1, 16), but this effect has not, to our knowledge, previously been described using a targeted gene array. Exposure to TCB is known to affect genes in the AhR signalling pathway, with secondary effects that may alter the expression of genes in other pathways. When hepatocytes were exposed to combined NP and TCB, the androgen receptor, CYP11β, aromatase-β, ovomucoid, StAR, and VtgR gene expressions were up-regulated, while the aromatase-R, ERβ, Vtg, and Zr-protein were down-regulated. These findings suggest an interaction between the ER and the AhR signalling pathways. The mechanisms behind these interactions were investigated using antagonists for both receptor pathways and will be discussed below.

484 Chem. Res. Toxicol., Vol. 20, No. 3, 2007

Genes Altered in the ER Signalling Cascade and Effects of Antagonists. The ERs belong to the large family of NRs and mediate their actions through binding to estrogens or estrogen mimics (2). We investigated the modulatory effects of TCB on NP-induced ER signalling and the possible effects of ER antagonists (Tam and ICI). The mRNA transcription of ERR and ERβ and two E2 responsive genes (Vtg and Zr-protein) were studied using real-time PCR. We found that exposure of hepatocytes to NP and TCB singly or in combination produced distinct expression patterns of each ER isotype. ERR (but not ERβ) expression was altered in response to NP treatment, and combined NP and TCB treatment produced decreased and increased expression of ERR and ERβ, respectively, as compared with NP alone. Combined NP and TCB exposure produced a slight increase, as compared to NP alone, of ERR on the array hybridization. A unique aspect of both ER isoforms is that the partial and absolute ER antagonists (Tam and ICI, respectively) caused significant decreases of the expression pattern after exposure to NP singly or in combination with TCB. Interestingly, the expression pattern of ERR and ERβ differed when TCB was given alone in the presence of Tam and ICI. Here, while both antagonists caused a two-fold decrease of ERR, only Tam (and not ICI) decreased the expression of ERβ when TCB was given alone. These findings indicate a differential effect of TCB on these receptors or a different TCB-mediated regulatory mechanism. In mammals, the tissue and cell specific roles of ER isoforms have been described (3). In teleost species, the expression profile of these ERs shows that both isoforms are expressed in fish liver (31) with different binding capacity and ability to induce transcription of E2-mediated genes (32). The current study supports previous findings on the relationship between ER isoforms, Vtg and Zr-protein, transcription levels in salmon liver exposed to NP (1, 7, 25). Furthermore, we showed that the presence of Tam and ICI in NP exposure singly and also in combination with TCB produced a differential effect on ERR and ERβ expression patterns. This finding is interesting because the ER has two domains required for transcriptional activation (AF). The AF1 domain is located in the N-terminal and is ligand-independent, while the AF2 domain is ligand-dependent and lies within the ligand-binding domain (2). When an agonist binds to the ER, conformational change in the protein that is followed by activation of AF1 and AF2 is produced (4). Tam is a partial ER antagonist that mediates its effects through inactivation of the AF2 domain. Thus, treatment with Tam leaves the AF1 domain active while cofactor recruitment for transcriptional activation is still possible, showing both potential for estrogenic and antiestrogenic effect. On the other hand, binding of ICI to ER inactivates both the AF1 and the AF2 domains together with access to coactivator that result to rapid proteasome-dependent induced degradation of the receptor (13). Therefore, our findings show a possible Tam and ICI-mediated proteasome degradation of ERR and ERβ only in the presence of the ER agonist (NP). This was further supported by the findings that in our array system, a proteasome-like subunit was significantly downregulated in the combined NP and TCB exposure and was later confirmed with real-time PCR (see below). It should also be noted that TCB alone decreased ERβ expression (but not ERR) to below control level, and this effect is consistent with the expression pattern of the proteasome subunit. Because it has been established that AhR agonists such as TCB do not competitively bind to the steroid hormone receptors nor do steroid hormone receptors bind to the AhR (14), we speculate that this effect is most likely a result of functional and structural

Mortensen and Arukwe

differences between ER isoforms that should be further investigated. Genes Altered in the AhR Signalling Cascade and Effects of Antagonists. We investigated the modulatory effects of NP on TCB-induced AhR signalling and the effects of AhR antagonist (ANF). We observed that TCB induced AhR signalling by transcriptional changes of AhR isoforms (AhRR and AhRβ), ARNT, AhRR, CYP1A1, and UDPGT. In accordance with the present study, the induced transcription of phase I and II biotransformation enzymes by TCB has previously been reported (1, 16). While the expression of AhRR and AhRR was induced together with CYP1A1 and UDPGT by TCB exposure, the AhRβ and the AhR nuclear dimerization partner, ARNT, were reduced. Although the biochemical and molecular properties of AhR have been characterized in mammalian cells, there are still uncertainties concerning the regulation, interactions with other proteins, and transcriptional properties of AhRs. The synthetic flavonoid 7,8-benzoflavone (ANF) binds to AhR with moderate affinity (26). In mammalian cell lines, ANF is found to compete with TCDD for cytosolic AhR binding resulting in the inhibition of TCDD-induced CYP1A1 gene expression and decreased formation of the nuclear AhR complex (33). ANF has previously been used successfully to inhibit the AhR in studies investigating the role of this receptor with cell cultures system (4, 33-35). In the present study, we treated hepatocytes with two concentrations of ANF (0.1 and 1 µM) and found a concentration-dependent decrease of TCB-induced AhRR, AhRβ, AhRR, CYP1A1, and UDPGT gene expressions but not ARNT. A similar effect was also observed when TCB was given in combination with NP, except for ARNT and AhRβ. In the combined TCB and NP exposure group, a 0.1 µM ANF concentration caused a higher decrease of ARNT and AhRβ than 1 µM (see Figure 3B,D). The consistency in the pattern of effect between ER and AhR signalling pathways in the combined TCB and NP exposure with the AhR antagonists (ANF) should also be noted (compare Figures 3B,D and 4AD). It is known that TCDD and related compounds induce AhR expression in several teleost species. In zebrafish (Danio rerio) embryo and liver cell line, TCDD induced a dose-dependent increase of AhR2 mRNA expression (36). A similar effect was also observed in rainbow trout where the AhR2 and AhR2β were elevated in gonadal cell lines and kidney tissue (37). In addition, these authors did not observe increases in mRNA expression of neither AhR2 nor AhR2β mRNA after TCDD exposure in rainbow trout liver or spleen (37). Elsewhere, TCDD or TCB doses did not affect transcriptional changes of AhR2 mRNA expression in Atlantic tomcod (Microgadus tomcod) liver (38). For ARNT expression, we observed that the presence of ANF caused significant effects on TCB-exposed hepatocytes, with 0.1 µM ANF elevating ARNT to control levels and 1 µM reducing ARNT to TCB-exposed levels (see Figure 3D). The overall function of ARNT is not fully understood in teleost, while in mammalian cells, this protein appears to be constitutively active (39). In view of the present study and others, a comparison of the in vivo endogenous response with in vitro reporter assays that have utilized different AhR isoforms from rainbow trout suggests that AhRR may account for the CYP1A1 induction by TCB in our system (37). It has been shown by Karchner and co-workers (40) that the amino acid sequence of AhR1 is most closely related to mammalian AhRs, which mediate the molecular response after exposure to halogenated aromatic hydrocarbons. The AhR1 (or AhRR) mRNA is nearly undetectable in many tissues that exhibit TCDD (and related

Toxicogenomics of ER-AhR Gene Interactions

compounds)-inducible CYP1A1 expression, implying that AhR2 (or AhRβ) is capable of mediating this response (41). The transcriptional capability of bHLH-PAS family of transcription factors is not well understood, and their individual in vivo functions are still the subject of current discussions. For example, by showing that the AhR may be involved in regulating the development of the vascular system in liver and other organs of AhR-null mice, the developmental, metabolic, and cardiovascular phenotypes of AhR-null mice provide important clues to the numerous functions of the AhRs (42-44). ER and AhR Cross-Talk. An important part of molecular toxicology research is to examine and recognize how complex chemical mixture exposures affect biological systems. To understand mechanisms, knowledge of protein-protein interactions between various NRs and transcription factors is required. Several reports have shown that AhR ligands possess antiestrogenic properties (16, 45, 46). Klinge and co-workers reported in vitro that AhR interacts directly with ERR in a ligand specific manner (47). We have previously reported bidirectional interactions between the ER and the AhR in the salmon in vitro system (1). In the present study, we observed that TCB decreased the expression of NP-induced transcription of ERR, ERβ, Vtg, and Zr-protein. The antiestrogenic action of TCB was less effective than Tam and ICI. However, we also observed that inhibition of AhR by ANF did not restore TCB-decreased gene expression of NP-induced ER signalling. Likewise, studies of the TCDD ability to bind to ER demonstrated that this strong AhR agonist did not compete with E2 for binding to the ER (48). Four possible mechanisms have been suggested for the antiestrogenic actions of AhR agonists: (i) increased rate of E2 metabolism, (ii) decreased cellular ER isoform levels, (iii) suppression of E2-induced transcription, and (iv) ER-AhR competition for transcriptional cofactors (2, 12). When these potential mechanisms are put into the context of the present study, degradation of endogenous E2 (or E2 mimics) by metabolizing enzymes induced by AhR leads to reduced ERmediated transcription. The involvement of CYP1A1 in E2 metabolism was previously investigated in female carp by Smeets and co-workers (49) and reported that the antiestrogenicity of different AhR ligands in female carp was found to be mediated through the AhR, not involving the CYP1A1. It was shown in a recent study that 3-methylcholanthrene (3-MC) exerts its effects by activating the AhR/ARNT heterodimer, which is able to interact with the unliganded ER, leading to induction of estrogenic pathway (50). This finding provides relevance and mechanistic explanation with the findings in the present study, showing no clear pattern of increased ER gene expression in response to decreased CYP1A1 after treatment with the AhR antagonist, ANF. The ER degradation by proteasomes induced by AhR has been explained as a possible mechanism of the antiestrogenicity (13, 51). In addition to activating AhR, TCDD is found to rapidly reduce the level of AhR protein in cells and mechanistic studies have established that the turnover is mediated through the 26S proteasome, involving ubiquitination of AhR, and requires the transcription activation domain of AhR (52, 53). Despite the fact that ANF is a partial AhR antagonist, our data do not support this mechanism since inhibition of AhR by ANF did not restore ER-mediated gene transcription and no significant proteasome transcript changes were observed (see Figure 3G). The proteasome involvement on TCB-mediated antiestrogenicity is not conclusive since it is possible that we may have quantified the wrong proteasome subunit. The choice of our proteasome subunit was based on its differential expres-

Chem. Res. Toxicol., Vol. 20, No. 3, 2007 485

sion pattern on our subtractive cDNA library after exposure to ER and AhR agonists. Interestingly, the presence of low AhR antagonist concentration (0.1 µM ANF) produced significant increase of ERβ, Vtg, and Zr-protein expression levels in TCBexposed samples but not ERR. Previously in an AhR defective mouse cell line, it was shown that while E2 increased ER-ERE complex formation, TCDD alone did not have any effect and the binding of ER to ERE was completely lost in cells simultaneously treated with both E2 and TCDD, leading the authors to conclude that TCDD was no longer antiestrogenic in the mutated cell line since AhR was required for ER transactivation of the ERE (54). Another possible target for AhR-mediated antiestrogenicity is the mRNA stability of ER and its transcriptional downstream products (Vtg and Zr-proteins). Dodson and Shapiro performed RNA gel mobility shift assays and found an estrogen-inducible mRNA stabilizing protein that bound specifically to Vtg mRNA in an area previously implicated in estrogen-mediated stabilization of Vtg mRNA. The stability of mRNA is determined by site specific mRNA endonuclease activities. Reviewed by Ostareck-Lederer (55), the endonuclease-catalyzed mRNA decay is regulated through the binding of RNA-binding proteins to target mRNAs that prevent their cleavage by endonucleases. Vigilin or high-density lipoprotein-binding protein is an ubiquitous protein in vertebrate cells (56). For example, the stability of liver Vtg mRNA in Xenopus laeVis is regulated by an E2induced vigilin that binds specifically to a 3′-UTR segment of the Vtg mRNA and protects it from degradation (56). This is in accordance with the present study, showing that the expression of vigilin mRNA followed a parallel pattern with Vtg and explains the decreased Vtg (and Zr-protein) mRNA levels by Tam and ICI that was enhanced in the presence of TCB. However, the vigilin pathway does not seem to explain the mechanism for TCB-mediated down-regulation of Vtg and Zrprotein levels, probably because TCB is a weak AhR agonist and ER antagonist, as compared to Tam and ICI. It should also be noted that the expression levels of vigilin were enriched by the SSH technique and therefore present on our targeted SalArray chip. On the AhR signalling pathway, we observed that the NP decreased the transcription of AhR gene battery to below TCBexposed levels, indicating that NP has anti-AhR signalling effects. Interestingly, the expression of AhRβ and ARNT was enhanced by Tam and ICI (i.e., ER antagonists) in TCB exposure alone and in combination with NP (note that these exposure conditions had no effect). This suggests the involvement of ERs in the repression of AhR gene signalling that probably involves ER-mediated gene products and other signalling pathways or responses. We observed minor exposure specific changes in the gene expression of ARNT. Considering that ARNT functions as a dimerization partner for several proteins in the bHLH-PAS protein superfamily (39) and it is found to be constitutively expressed in human tissues (39, 46) only minor alterations in ARNT gene expression can be expected in response to xenobiotic exposures. However, on the basis of sequence homology with an ER transcription factors p160, Brunnberg and co-workers investigated the role of ARNT as a coactivator of ER-dependent transcription and found that ARNT functions as a coactivator of ER and that this effect was due to the C-terminal domain and not the conserved bHLH or PAS domains (46). In addition, although the ARNT contains a less complex activation domain as compared to AhR, the activation domains of AhR and ARNT are located in the carboxy-terminal of both genes. During CYP1A1 (and other genes) activation,

486 Chem. Res. Toxicol., Vol. 20, No. 3, 2007

the ARNT activation domain does not contribute to the activation of AhR complex (57). It has been reported by Whitelaw and co-workers (58) that the activation of the AhR is repressed by the central ligand-binding segment of the receptor. The presence of ER antagonists in NP and TCB exposure singly increased the expression of several genes in the AhR signalling pathway. Hence, we suggest a direct interaction between the ER and the AhR transcriptional machinery. In general, the present data are consistent with previous studies showing that NP (i.e., estrogen mimic) and E2 significantly suppressed hepatic CYP1A1 mRNA levels, EROD activity, and CYP1A1 protein in in vivo and in vitro experiments using several teleost species (7, 17). On the basis of the possible mechanisms explained above, we hypothesize that NP can bind the CYP1A1 protein, and through this binding, NP or its metabolites may inhibit the CYP1A1 expression (59). Alternatively, the inhibitory action of NP could be mediated, at least in part, through the ERs where the ER-NP complex interfered with the AhR transcriptional machinery directly or with the CYP1A1 and indirectly regulate AhR-induced gene expression through binding the XRE. In addition, NP may control the recruitment of other coactivators that are not studied in the present study. Furthermore, the consistency between AhRR, CYP1A1, and UDPGT expression pattern suggests that this repressor singly may have caused the decrease in CYP1A1 and UDPGT levels. The AhRR-ARNT heterodimerization may negatively regulate AhR-driven gene expression through transcriptional repression (60). In accordance with our data, the modulation of CYP1A1 by NP, E2, and BNF was recently shown to parallel the AhRR gene expression (61). Any of the above-mentioned mechanisms may have caused the NP effect on AhR signaling. This is supported by the fact that the BHLHPAS (Per-AhR/ARNT-Sim homology sequence) of transcription factor usually acts in association with each other through the formation of heterodimers (AhR/ARNT or AhRR/ARNT) that subsequently bind to the XRE sequences in the promoter regions of the target genes to regulate their expression. In summary, the use of targeted SalArray gene chip proved to be a sensitive screening and hypothesis-generating tool for differentially expressed genes after chemical exposure, singly or as complex mixtures. Interestingly, the unique hepatic expression patterns of steroidogenic genes affected by NP and TCB provide a new biological pathway that is not well-studied. While supporting the tested hypothesis, our SalArray chip has generated a new hypothesis that needs to be further investigated. The AhR agonist (TCB) functioned as anti-NP-induced effect, and NP produced anti-AhR-induced effect or as inhibitor of AhRR, AhRR, ARNT, CYP1A1, and UDPGT expression. Taken together, the findings in the present study demonstrate a complex mode of ER-AhR interaction, indicating that TCB-mediated antiestrogenicity may involve a possible ER-AhR competition for common cofactors. The complex mode of interaction is further supported by the observation that the presence of ER antagonists potentiated the transcription of AhR isoforms and their mediated responses when TCB was given alone (more so for AhRβ). In our laboratory, we are still performing studies on cross-talk between the ER and the AhR signal transduction systems and underlying mechanism(s) by which xenobiotics and xenoestrogens interact with each other. This complex interaction between two different classes of ligand-activated receptors provides novel mechanistic insights on signalling pathways. In addition, the inhibitory AhR-ER interactions can be used to further investigate specific genes found to be affected in our

Mortensen and Arukwe

targeted SalArray chip that are important for the reproductive effects of endocrine disruptors. Acknowledgment. The NTNU doctoral fellowship grant to A.S.M. financed this study. We thank Solveig Gaasø at Marine Harvest Norway AS for supplying the experimental fish. We are grateful to Marte Braathen for assistance during sampling and Tommy Jørstad for helpful discussions during array analysis. Supporting Information Available: Primer pair sequences, their GenBank numbers, amplicon size, and annealing temperature conditions for target mRNA of interest investigated with real-time PCR. Complete list of sequences whose expression patterns were either up- or down-regulated by exposure of primary culture of salmon hepatocytes to NP, singly or in combination with TCB using targeted salmon cDNA array (SalArray) analysis. This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) Mortensen, A. S., Tolfsen, C. C., and Arukwe, A. (2006) Gene expression patterns in estrogen (nonylphenol) and aryl hydrocarbon receptor agonists (PCB-77) interaction using rainbow trout (Oncorhynchus mykiss) primary hepatocyte culture. J. Toxicol. EnViron. Health A 69, 1-19. (2) Nilsson, S., Makela, S., Treuter, E., Tujague, M., Thomsen, J., Andersson, G., Enmark, E., Pettersson, K., Warner, M., and Gustafsson, J. A. (2001) Mechanisms of estrogen action. Physiol. ReV. 81, 1535-1565. (3) Matthews, J., and Gustafsson, J.-Å. (2003) Estrogen signaling: A subtle balance between ER alpha and ER beta. Mol. InterVentions 3, 281-292. (4) Aranda, A., and Pascual, A. (2001) Nuclear hormone receptors and gene expression. Physiol. ReV. 81, 1269-1304. (5) Mommsen, P. T., and Walsh, P. J. (1988) Vitellogenesis and oocyte assembly. In Fish Physiology (Hoar, W. S., Randall, D. J., and Donaldson, E. M., Eds.) pp 347-406, Academic Press, New York. (6) Oppen-Berntsen, D. O., Gram-Jensen, E., and Walther, B. T. (1992) Zona radiata proteins are synthesized by rainbow trout (Oncorhynchus mykiss) hepatocytes in response to oestradiol-17 beta. J. Endocrinol. 135, 293-302. (7) Arukwe, A., Celius, T., Walther, B. T., and Goksoyr, A. (2000) Effects of xenoestrogen treatment on zona radiata protein and vitellogenin expression in Atlantic salmon (Salmo salar). Aquat. Toxicol. 49, 159170. (8) Arukwe, A., Knudsen, F. R., and Goksoyr, A. (1997) Fish zona radiata (eggshell) protein: A sensitive biomarker for environmental estrogens. EnViron. Health Perspect. 105, 418-422. (9) Kazlauskas, A., Sundstrom, S., Poellinger, L., and Pongratz, I. (2001) The hsp90 chaperone complex regulates intracellular localization of the dioxin receptor. Mol. Cell. Biol. 21, 2594-2607. (10) Whitlock, J. P., Jr. (1999) Induction of cytochrome P4501A1. Annu. ReV. Pharmacol. Toxicol. 39, 103-125. (11) Hoffman, E. C., Reyes, H., Chu, F. F., Sander, F., Conley, L. H., Brooks, B. A., and Hankinson, O. (1991) Cloning of a factor required for activity of the Ah (dioxin) receptor. Science 252, 954-958. (12) Safe, S., and Wormke, M. (2003) Inhibitory aryl hydrocarbon receptorestrogen receptor R-cross-talk and mechanisms of action. Chem. Res. Toxicol. 16, 807-816. (13) Wormke, M., Stoner, M., Saville, B., Walker, K., Abdelrahim, M., Burghardt, R., and Safe, S. (2003) The aryl hydrocarbon receptor mediates degradation of estrogen receptor alpha through activation of proteasomes. Mol. Cell. Biol. 23, 1843-1855. (14) Safe, S., Astroff, B., Harris, M., Zacharewski, T., Dickerson, R., Romkes, M., and Biegel, L. (1991) 2,3,7,8-Tetrachlorodibenzo-pdioxin (TCDD) and related compounds as antioestrogens: Characterization and mechanism of action. Pharmacol. Toxicol. 69, 400409. (15) Anderson, M. J., Olsen, H., Matsumura, F., and Hinton, D. E. (1996) In vivo modulation of 17 beta-estradiol-induced vitellogenin synthesis and estrogen receptor in rainbow trout (Oncorhynchus mykiss) liver cells by beta-naphthoflavone. Toxicol. Applied. Pharmacol. 137, 210218. (16) Arukwe, A., Yadetie, F., Male, R., and Goksoyr, A. (2001) In vivo modulation of nonylphenol-induced zonagenesis and vitellogenesis by the antiestrogen, 3,3′4,4′-tetrachlorobiphenyl (PCB-77) in juvenile fish. EnViron. Toxicol. Pharmacol. 10, 5-15.

Toxicogenomics of ER-AhR Gene Interactions (17) Navas, J. M., and Segner, H. (2000) Antiestrogenicity of betanaphthoflavone and PAHs in cultured rainbow trout hepatocytes: Evidence for a role of the arylhydrocarbon receptor. Aquat. Toxicol. 51, 79-92. (18) Volz, D. C., Bencic, D. C., Hinton, D. E., Law, J. M., and Kullman, S. W. (2005) 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) induces organ- specific differential gene expression in male Japanese medaka (Oryzias latipes). Toxicol. Sci. 85, 572-584. (19) Larkin, P., Knoebl, I., and Denslow, N. D. (2003) Differential gene expression analysis in fish exposed to endocrine disrupting compounds. Comp. Biochem. Physiol., Part B: Biochem. Mol. Biol. 136, 149161. (20) Brown, M., Robinson, C., Davies, I. M., Moffat, C. F., Redshaw, J., and Craft, J. A. (2004) Temporal changes in gene expression in the liver of male plaice (Pleuronectes platessa) in response to exposure to ethynyl oestradiol analysed by macroarray and Real-Time PCR. Mutat. Res. 552, 35-49. (21) Mazurais, D., Montfort, J., Delalande, C., and Gac, F. L. (2005) Transcriptional analysis of testis maturation using trout cDNA macroarrays. Gen. Comp. Endocrinol. 142, 143-154. (22) Ji, W., Wright, M. B., Cai, L., Flament, A., and Lindpaintner, K. (2002) Efficacy of SSH PCR in isolating differentially expressed genes. BMC Genomics 3, 12. (23) Berry, M. N., and Friend, D. S. (1969) High-yield preparations of isolated rat liver parenchymal cells. A biocemical and fine structural study. J. Cell Biol. 43, 506-520. (24) Andersson, T., Fo¨rlin, L., and Hansson, T. (1983) Biotransformation of 7- ethoxycoumarin in isolated perfused rainbow trout liver. Drug. Metab. Dispos. 11, 494-498. (25) Celius, T., Haugen, T. B., Grotmol, T., and Walther, B. T. (1999) A sensitive zonagenetic assay for rapid in vitro assessment of estrogenic potency of xenobiotics and mycotoxins. EnViron. Health. Perspect. 107, 63-68. (26) Merchant, M., and Safe, S. (1995) In vitro inhibition of 2,3,7,8tetrachlorodibenzo-p-dioxin-induced activity by alpha-naphthoflavone and 6-methyl-1,3,8-trichlorodibenzofuran using an aryl hydrocarbon (Ah)-responsive construct. Biochem. Pharmacol. 50, 663-668. (27) Arukwe, A. (2005) Modulation of brain steroidogenesis by affecting transcriptional changes of steroidogenic acute regulatory (StAR) protein and cholesterol side chain cleavage (P450scc) in juvenile Atlantic salmon (Salmo salar) is a novel aspect of nonylphenol toxicity. EnViron. Sci. Technol. 39, 9791-9798. (28) von Hofsten, J., Karlsson, J., Jones, I., and Olsson, P. E. (2002) Expression and regulation of fushi tarazu factor-1 and steroidogenic genes during reproduction in Arctic char (SalVelinus alpinus). Biol. Reprod. 67, 1297-1304. (29) Wang, Y., and Ge, W. (2004) Cloning of zebrafish ovarian P450c17 (CYP17, 17alpha-hydroxylase/17,20-lyase) and characterization of its expression in gonadal and extra-gonadal tissues. Gen. Comp. Endocrinol. 135, 241-249. (30) Zhou, L. Y., Wang, D. S., Senthilkumaran, B., Yoshikuni, M., Shibata, Y., Kobayashi, T., Sudhakumari, C. C., and Nagahama, Y. (2005) Cloning, expression and characterization of three types of 17betahydroxysteroid dehydrogenases from the Nile tilapia, Oreochromis niloticus. J. Mol. Endocrinol. 35, 103-116. (31) Sabo-Attwood, T., Kroll, K. J., and Denslow, N. D. (2004) Differential expression of largemouth bass (Micropterus salmoides) estrogen receptor isoforms alpha, beta, and gamma by estradiol. Mol. Cell. Endocrinol. 218, 107-118. (32) Menuet, A., Le Page, Y., Torres, O., Kern, L., Kah, O., and Pakdel, F. (2004) Analysis of the estrogen regulation of the zebrafish estrogen receptor (ER) reveals distinct effects of ERalpha, ERbeta1 and ERbeta2. J. Mol. Endocrinol. 32, 975-986. (33) Merchant, M., Arellano, L., and Safe, S. (1990) The mechanism of action of alpha-naphthoflavone as an inhibitor of 2,3,7,8-tetrachlorodibenzo-p-dioxin-induced CYP1A1 gene expression. Arch. Biochem. Biophys. 281, 84-89. (34) Ferraris, M., Flora, A., Chiesara, E., Fornasari, D., Lucchetti, H., Marabini, L., Frigerio, S., and Radice, S. (2005) Molecular mechanism of the aryl hydrocarbon receptor activation by the fungicide iprodione in rainbow trout (Oncorhynchus mykiss) hepatocytes. Aquat. Toxicol. 72, 209-220. (35) Bemanian, V., Male, R., and Goksoyr, A. (2004) The aryl hydrocarbon receptor-mediated disruption of vitellogenin synthesis in the fish liver: Cross-talk between AHR- and ERalpha-signalling pathways. Comp. Hepatol. 3, 2. (36) Tanguay, R. L., Abnet, C. C., Heideman, W., and Peterson, R. E. (1999) Cloning and characterization of the zebrafish (Danio rerio) aryl hydrocarbon receptor. Biochim. Biophys. Acta 1444, 35-48. (37) Abnet, C. C., Tanguay, R. L., Hahn, M. E., Heideman, W., and Peterson, R. E. (1999) Two forms of aryl hydrocarbon receptor type 2 in rainbow trout (Oncorhynchus mykiss). Evidence for differential expression and enhancer specificity. J. Biol. Chem. 274, 15159-15166.

Chem. Res. Toxicol., Vol. 20, No. 3, 2007 487 (38) Roy, N. K., and Wirgin, I. (1997) Characterization of the aromatic hydrocarbon receptor gene and its expression in Atlantic tomcod. Arch. Biochem. Biophys. 344, 373-386. (39) Gu, Y. Z., Hogenesch, J. B., and Bradfield, C. A. (2000) The PAS superfamily: Sensors of environmental and developmental signals. Annu. ReV. Pharmacol. Toxicol. 40, 519-561. (40) Karchner, S. I., Powell, W. H., and Hahn, M. E. (1999) Identification and functional characterization of two highly divergent aryl hydrocarbon receptors (AHR1 and AHR2) in the teleost Fundulus heteroclitus. Evidence for a novel subfamily of ligand-binding basic helix loop helix-Per-ARNT-Sim (bHLH-PAS) factors. J. Biol. Chem. 274, 33814-33824. (41) Powell, W. H., Bright, R., Bello, S. M., and Hahn, M. E. (2000) Developmental and tissue-specific expression of AHR1, AHR2, and ARNT2 in dioxin-sensitive and -resistant populations of the marine fish Fundulus heteroclitus. Toxicol. Sci. 57, 229-239. (42) Gonzalez, F. J., and Fernandez-Salguero, P. (1998) The aryl hydrocarbon receptor: Studies using the AHR-null mice. Drug. Metab. Dispos. 26, 1194-1198. (43) Fernandez-Salguero, P., Pineau, T., Hilbert, D. M., McPhail, T., Lee, S. S., Kimura, S., Nebert, D. W., Rudikoff, S., Ward, J. M., and Gonzalez, F. J. (1995) Immune system impairment and hepatic fibrosis in mice lacking the dioxin-binding Ah receptor. Science 268, 722726. (44) Thurmond, L. M., Lauer, L. D., House, R. V., Cook, J. C., and Dean, J. H. (1987) Immunosuppression following exposure to 7,12-dimethylbenz[a]anthracene (DMBA) in Ah-responsive and Ah-nonresponsive mice. Toxicol. Appl. Pharmacol. 91, 450-460. (45) Matthews, J., Wihlen, B., Thomsen, J., and Gustafsson, J. A. (2005) Aryl hydrocarbon receptor-mediated transcription: Ligand-dependent recruitment of estrogen receptor alpha to 2,3,7,8-tetrachlorodibenzop-dioxin-responsive promoters. Mol. Cell. Biol. 25, 5317-5328. (46) Brunnberg, S., Pettersson, K., Rydin, E., Matthews, J., Hanberg, A., and Pongratz, I. (2003) The basic helix-loop-helix-PAS protein ARNT functions as a potent coactivator of estrogen receptor-dependent transcription. Proc. Natl. Acad. Sci. U.S.A. 100, 6517-6522. (47) Klinge, C. M., Bowers, J. L., Kulakosky, P. C., Kamboj, K. K., and Swanson, H. I. (1999) The aryl hydrocarbon receptor (AHR)/AHR nuclear translocator (ARNT) heterodimer interacts with naturally occurring estrogen response elements. Mol. Cell. Endocrinol. 157, 105-119. (48) Harris, M., Zacharewski, T., and Safe, S. (1990) Effects of 2,3,7,8tetrachlorodibenzo-p-dioxin and related compounds on the occupied nuclear estrogen receptor in MCF-7 human breast cancer cells. Cancer Res. 50, 3579-3584. (49) Smeets, J. M., van Holsteijn, I., Giesy, J. P., and van den Berg, M. (1999) The anti-estrogenicity of Ah receptor agonists in carp (Cyprinus carpio) hepatocytes. Toxicol. Sci. 52, 178-188. (50) Ohtake, F., Takeyama, K., Matsumoto, T., Kitagawa, H., Yamamoto, Y., Nohara, K., Tohyama, C., Krust, A., Mimura, J., Chambon, P., Yanagisawa, J., Fujii-Kuriyama, Y., and Kato, S. (2003) Modulation of oestrogen receptor signalling by association with the activated dioxin receptor. Nature 423, 545-550. (51) Callige, M., and Richard-Foy, H. (2006) Ligand-induced estrogen receptor alpha degradation by the proteasome: new actors? Nucl. Recept. Signal 4, e004. (52) Ma, Q., and Baldwin, K. T. (2000) 2,3,7,8-tetrachlorodibenzo-p-dioxininduced degradation of aryl hydrocarbon receptor (AhR) by the ubiquitin-proteasome pathway. Role of the transcription activaton and DNA binding of AhR. J. Biol. Chem. 275, 8432-8438. (53) Pollenz, R. S. (1996) The aryl-hydrocarbon receptor, but not the arylhydrocarbon receptor nuclear translocator protein, is rapidly depleted in hepatic and nonhepatic culture cells exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Mol. Pharmacol. 49, 391-398. (54) Kharat, I., and Saatcioglu, F. (1996) Antiestrogenic effects of 2,3,7,8tetrachlorodibenzo-p-dioxin are mediated by direct transcriptional interference with the liganded estrogen receptor. Cross-talk between aryl hydrocarbon- and estrogen-mediated signaling. J. Biol. Chem. 271, 10533-10537. (55) Ostareck-Lederer, A., Ostareck, D. H., Standart, N., and Thiele, B. J. (1994) Translation of 15-lipoxygenase mRNA is inhibited by a protein that binds to a repeated sequence in the 3′ untranslated region. EMBO J. 13, 1476-1481. (56) Dodson, R. E., and Shapiro, D. J. (1997) Vigilin, a ubiquitous protein with 14 K homology domains, is the estrogen-inducible vitellogenin mRNA 3′-untranslated region-binding protein. J. Biol. Chem. 272, 12249-12252. (57) Ko, H. P., Okino, S. T., Ma, Q., and Whitlock, J. P., Jr. (1996) Dioxininduced CYP1A1 transcription in vivo: The aromatic hydrocarbon receptor mediates transactivation, enhancer-promoter communication, and changes in chromatin structure. Mol. Cell. Biol. 16, 430-436. (58) Whitelaw, M. L., Gustafsson, J. A., and Poellinger, L. (1994) Identification of transactivation and repression functions of the dioxin

488 Chem. Res. Toxicol., Vol. 20, No. 3, 2007 receptor and its basic helix-loop-helix/PAS partner factor Arnt: Inducible versus constitutive modes of regulation. Mol. Cell. Biol. 14, 8343-8355. (59) Arukwe, A., and Goksoyr, A. (1997) Changes in three hepatic cytochrome P450 subfamilies during a reproductive cycle in Turbot (Scophthalmus maximus L.). J. Exp. Zool. 277, 313-325. (60) Karchner, S. I., Franks, D. G., Powell, W. H., and Hahn, M. E. (2002) Regulatory interactions among three members of the vertebrate aryl

Mortensen and Arukwe hydrocarbon receptor family: AHR repressor, AHR1, and AHR2. J. Biol. Chem. 277, 6949-6959. (61) Maradonna, F., Polzonetti, V., Bandiera, S. M., Migliarini, B., and Carnevali, O. (2004) Modulation of the hepatic CYP1A1 system in the marine fish Gobius niger, exposed to xenobiotic compounds. EnViron. Sci. Technol. 38, 6277-6282.

TX6002672