Differential Gene Regulation by the Human and ... - Oxford Journals

Download PDF
5 downloads 76 Views 230KB Size Report
Comment
than the mAHRb for specific AHR ligands, such as indirubin. Each receptor has also been shown to differentially recruit LXXLL coactivator motif proteins and to ...
TOXICOLOGICAL SCIENCES 114(2), 217–225 (2010) doi:10.1093/toxsci/kfp308 Advance Access publication December 31, 2009

Differential Gene Regulation by the Human and Mouse Aryl Hydrocarbon Receptor Colin A. Flaveny, Iain A. Murray, and Gary H. Perdew1 Center for Molecular Toxicology and Carcinogenesis and the Department of Veterinary and Biomedical Sciences, the Pennsylvania State University, University Park, Pennsylvania 16802 1

To whom correspondence should be addressed at Center for Molecular Toxicology and Carcinogenesis and Department of Veterinary Science, The Pennsylvania State University, 309A Life Sciences Building, University Park, PA 16802. Fax: (814) 863-1696. E-mail: [email protected]. Received August 31, 2009; accepted December 20, 2009

The human aryl hydrocarbon receptor (hAHR) and mouse aryl hydrocarbon receptor (mAHRb) share limited (58%) transactivation domain (TAD) sequence identity. Compared to the mAHRb allele, the hAHR displays 10-fold lower relative affinity for prototypical ligands, such as 2,3,7,8 tetrachlorodibenzo-p-dioxin (TCDD). However, in previous studies, we have demonstrated that the hAHR can display a higher relative ligand-binding affinity than the mAHRb for specific AHR ligands, such as indirubin. Each receptor has also been shown to differentially recruit LXXLL coactivator motif proteins and to utilize different TAD subdomains in gene transactivation. Using hepatocytes isolated from C57BL/6J mice (Ahrb/b) and AHRTtr transgenic mice, which express hAHR protein specifically in hepatocytes, we investigated whether the hAHR and mAHRb differentially regulate genes. DNA microarray and quantitative PCR analysis of Ahrb/b and AHRTtr primary mouse hepatocytes treated with 10nM TCDD revealed that a number of established AHR target genes such as Cyp1a1 and Cyp1b1 are significantly induced by both receptors. Remarkably, of the 1752 genes induced by mAHRb and 1186 genes induced by hAHR, only 265 genes (~18%) were significantly activated by both receptors in response to TCDD. Conversely, of the 1100 and 779 genes significantly repressed in mAHRb and hAHR hepatocytes, respectively, only 462 (~49%) genes were significantly repressed by both receptors in response to TCDD treatment. Genes identified as differentially expressed are known to be involved in a number of biological pathways, including cell proliferation and inflammatory response, which suggest that compared to the mAHRb, the hAHR may play contrasting roles in TCDD-induced toxicity and endogenous AHR-mediated gene regulation. Key Words: AHR; TCDD; Ah receptor; gene regulation.

The aryl hydrocarbon receptor (AHR) is a ligand-activated basic helix-loop-helix Per-Arnt-Sim protein transcription factor (Okey et al., 1979; Poland and Knutson, 1982; Poland et al., 1987; Swanson and Bradfield, 1993). The AHR can be activated by a diverse array of compounds including planar polycyclic aromatic hydrocarbons (PAHs) such as benzo[a]-

pyrene, halogenated PAHs such as 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), and a number of naturally occurring flavonols such as quercetin (2-(3,4dihydroxyphenyl)-3,5,7-trihydroxy-4H-chromen-4-one) and plant tryptophan metabolites such as indirubin ((2Z) 2,3-biindole-2,3 (1’H,1’H)-dione (Ciolino et al., 1999; Guengerich et al., 2004). In the cytoplasm, the AHR is part of a stable cytoplasmic multiprotein complex that consists of the hepatitis B virus X–associated protein 2 (XAP2/AIP/ARA 9), p23, and two molecules of heat-shock protein 90. Ligand activation of the AHR causes it to translocate into the nucleus, dissociate from its cytoplasmic complex, and heterodimerize with the aryl hydrocarbon receptor nuclear translocator (ARNT) (Ma et al., 1995). This AHR/ARNT heterodimer then binds directly to dioxin-responsive elements (DREs/XREs) of dioxinresponsive genes, such as Cyp1a1, Cyp1B1, or Cyp1a2. Activated AHR can also interact with a number of other receptors and transcription factors such as the estrogen receptor and nuclear factor-kappa-light-chain-enhancer of activated Bcells to modulate gene expression (Kharat and Saatcioglu, 1996; Kim et al., 2000; Patel et al., 2009; Vogel and Matsumura, 2009). The degree and range of symptoms accompanying TCDD exposure drastically vary from one species to another and are probably partly due to interspecies differences in AHR structure. For example, guinea pigs are 1000-fold less sensitive to TCDDinduced toxicity than the sensitive hamster (Pohjanvirta et al., 1993), both of which express AHR proteins that are structurally divergent at the N- and C-terminal domains (Henck et al., 1981). In mice, TCDD exposure can result in immunotoxicity, developmental defects, reproductive toxicity, and both skin and liver cancers. TCDD is nongenotoxic and thus acts as a liver tumor promoter possibly through prolonged and inappropriate activation of AHR-regulated genes, such as interleukins (IL), proto-oncogenes, and inflammatory factors (Birnbaum, 1994; Silbergeld and Gasiewicz, 1989; Sutter et al., 1991; Watson et al., 1995). Conversely, human exposure to TCDD has only been

 The Author 2009. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For permissions, please email: [email protected]

218

FLAVENY, MURRAY, AND PERDEW

shown to cause chloracne (Denison et al., 1986). Nonetheless, the Environmental Protection Agency classifies TCDD as a human carcinogen (Travis and Hattemer-Frey, 1991). Among inbred mouse strains that express the low ligandbinding affinity mouse aryl hydrocarbon receptor (mAHR)d (e.g., DBA) and the high ligand-binding affinity mAHRb (e.g., C57BL/6J) alleles, there is a marked difference in TCDD sensitivity with C57BL/6J mice displaying a relatively higher sensitivity to TCDD treatment (Ema et al., 1994; Poland and Glover, 1990; Poland et al., 1994; Weber et al., 1995). The resistance to TCDD toxicity displayed by DBA mice is apparently due to expression of the mAHRd allele, which contains an alanine to valine (A375V) amino acid substitution in the ligand-binding domain of the mAHRd that sterically impinges on the binding of typical AHR ligands, such as TCDD (Ema et al., 1994; Poland et al., 1994). Interestingly, this low-affinity mAHRd, V375 corresponding to V381, is conserved in the ligand-binding domain of the human aryl hydrocarbon receptor (hAHR), giving the hAHR a 10-fold lower relative affinity for TCDD when compared to the mAHRb (Ema et al., 1994; Ramadoss and Perdew, 2004). However, recent investigations have demonstrated that the hAHR may actually display selectivity for structurally atypical ligands, including indirubin and quercetin (Flaveny et al., 2009). This discovery suggested that the hAHR may be regulated by endogenous and exogenous ligand(s) that may be structurally different from typical highaffinity AHR ligands. This also may lend an evolutionary basis for the apparent resistance that humans display to dioxin toxicity. There is a high degree of divergence that exists between the hAHR and mAHRb C-terminal domain structures. The C-terminal domains of both receptors, which encompass the receptor transactivation domains (TADs), share only a 58% amino acid sequence identity. This amino acid sequence disparity was shown to result in differential recruitment of LXXLL coactivator–binding motifs (Flaveny et al., 2008). This finding suggests that the receptors may also differentially recruit coactivator/corepressor complexes and thus may differentially regulate gene expression following ligand activation. We decided to investigate whether the differences between the mAHRb and hAHR result in differences in receptor-regulated gene modulation in response to TCDD treatment. Toward this end, hepatocytes isolated from the hepatocyte-specific transgenic hAHR and C57BL/6J mice, which express mAHRb, were treated with TCDD or vehicle control. Total RNA from these cells was subjected to DNA microarray analysis to determine if each receptor was capable of differentially regulating gene expression. This study demonstrates that, in response to TCDD, the hAHR regulates a functionally different subset of genes compared to the mAHRb. Consequently, the accuracy of assessing human risk based on data obtained in rodent models may be confounded.

MATERIALS AND METHODS Transgenic Mice Ahrfx/fx and Ahrfx/fx/CreAlb mice generated previously (Walisser et al., 2005) were backcrossed for at least seven generations before experimental use. These backcrossed Ahrfx/fx/CreAlb mice were used to generate transgenic AHRTtr (strain name [B6.Cg-Ahrtm3.1Bra Tg (Alb-cre, Ttr-AHR)1Ghp]) by crossing Ahrfx/fx/CreAlb with transgenic hAHR expression mice derived on a fully inbred C57BL/6J background and so the AHRTtr mice were therefore backcrossed to an eighth generation. These mice expressed hAHR protein specifically in hepatocytes under the control of the transthyretin promoter as described previously (Flaveny et al., 2009). All mice were genotyped using the relevant primers listed in Supplementary table 1 and as described previously (Flaveny et al., 2009; Walisser et al., 2005). Cytosol Preparation Mouse livers were homogenized in MENG buffer (25mM 3-(Nmorpholino) propanesulfonic acid, 2mM ethylenediamenetetraacetic acid, 0.02% NaN3, 10% glycerol, pH 7.4) containing 20mM sodium molybdate and protease inhibitors (Sigma) and centrifuged at 100,000 3 g for 1 h. Immunoblotting Whole mouse liver cytosol was resolved using 8% SDS-tricine polyacrylamide gels. Proteins were transferred to polyvinyldene fluoride membrane, and AHR or b-actin proteins were detected using the mouse monoclonal antibodies RPT1 (Affinity BioReagents) or sc47778 (Santa Cruz Biotechnology Inc.) and visualized using autoradiography. Primary Hepatocyte Isolation Hepatocytes were isolated as described previously (Madden et al., 2000), with a few modifications. Briefly, mice were anaesthetized with 0.1–0.3 ml 2.5% Avertin, administered via ip injection. Hepatic perfusion was performed with Buffer-I (5mM dextrose/116mM NaCl/760lM NaH2PO4/5.3mM KCl/26mM NaHCO3/10mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid/500lM ethylene glycol tetraacetic acid, pH 7.2) for 1 min, followed by Buffer-II (0.2 mg/ml type-I collagenase [Worthington]/5.3mM KCl/116mM NaCl/5mM dextrose/26mM NaHCO3/1.6mM MgSO4/900lM CaCl2/48 lg/ml trypsin inhibitor, pH 7.2) for a further 5–10 min. Hepatic tissue was excised, transferred, and dissociated in a 100-mm plate containing 9-ml short-term media (Dulbecco’s Modified Eagles Medium/10% fetal bovine serum/2.5% dimethyl sulfoxide (DMSO)/10nM dexamethasone/100 IU/ml penicillin/100 lg/ ml streptomycin). Cells were filtered, centrifuged (500 3 g for 1 min), and resuspended in short-term media. Cell viability was assessed via trypan blue staining, and cells were seeded into type-I collagen–coated six-well plates (BD Bioscience) at a density of 1 3 106 cells/ml. After 4 h incubation at 37C, nonadherent cells were aspirated and fresh short-term media added. After overnight incubation at 37C, cells were washed with PBS and short-term media replaced with long-term hepatocyte culture media (Hepatozyme-SFM [Invitrogen]/2.5% DMSO/10nM dexamethasone/100 IU/ml penicillin/100 lg/ ml streptomycin). Hepatocytes were cultured in long-term media under standard conditions in a 37C incubator at 5% CO2, 95% air, and cell culture media was changed every 48 h. DNA microarray analysis: isolation and labeling of RNA. Primary hepatocytes were isolated from six AHRTtr mice and six Ahrb/b mice, respectively. After being cultured for 48 h, hAHR- and mAHRb-expressing hepatocytes were treated with 10nM TCDD or vehicle for 6 h. Total RNA isolated from primary hepatocytes was converted to complementary DNA (cDNA) and subjected to real-time reverse transcriptase (RT)-PCR to assess the degree of induction of AHR-responsive genes. RNA isolated using Tri-reagent (as described before) was additionally purified using RNeasy minicolumns (Qiagen). The quality of the RNA was assessed using formaldehyde agarose gels and a Bioanalyzer and RNA LabChip (Agilent Technologies) at the Penn State DNA microarray facility. Poly-A

DIFFERENTIAL GENE REGULATION BY THE HUMAN AND MOUSE AHR (Affymetrix, Santa Clara, CA) controls were added to the RNA samples before they were labeled using GeneChip One-Cycle Target Labeling and Control Reagents (Affymetrix). Labeled samples were subsequently assessed for quality using RNA LabChip kit and Bioanalyzer to determine if minimal fragmentation was obtained. The quality of the samples was further tested using GeneChip Test 3 arrays that possess known mouse and human housekeeping gene sets. The labeled RNA used in each array was representative of hepatocytes isolated from each individual mouse. Samples that satisfied the quality control assessments were then used for full-scale hybridization and scanning using Affymetrix mouse genomewide expression set 430 2.0 arrays, which has 45,000 probe sets that analyze the expression levels of 39,000 transcripts over 34,000 well-characterized genes. DNA microarray data analysis. GeneChip Operating Software (Affymetrix) was utilized to preprocess CAB/CEL files generated from the 12 scanned DNA microarrays, which represented hepatocytes isolated from one mouse each. Data quality was initially assessed by checking the array image, B2 oligo performance, average background to noise ratios, poly-A controls, hybridization controls, and the 3# to 5# probe set ratios for control genes (e.g., b-actin or GAPDH). DNA microarray data were normalized using Probe Logarithmic Intensity Error (PLIER-MM) approximation algorithm (Affymetrix Expression Console Software 1.1). Normalized DNA microarray data outputs from TCDD- and control-treated Ahrb/b and AHRTtr hepatocytes were compared for differential expression using Significance Analysis of Microarrays (SAM, version 2.23A [Pan, 2002; Tusher et al., 2001]) with 100 permutations, KNN-10. The total number of genes induced and repressed by the hAHR or mAHRb in response to TCDD was then calculated from the SAM output gene lists. For comparing TCDDand vehicle-treated sample arrays from either Ahrb/b or AHRTtr hepatocyte expression value, comparisons were conducted at a value of 0.44 and a false discovery rate of 5%. Across the Ahrb/b or AHRTtr hepatocyte genotypes, genes were considered significantly differentially induced or repressed in TCDD-treated Ahrb/b compared to TCDD-treated AHRTtr primary hepatocytes based on a q value < 0.05. The ‘‘q value’’ is similar to a ‘‘p value’’ but is adapted to the analysis of a large number of genes and is a measure of significance in terms of the false discovery rate. Normalized array data files are available online at http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc¼ GSE17925. Differential expression of selected genes was validated using real-time RT-PCR.

219

RESULTS

Classic TCDD-Responsive Genes Cyp1a1 and Cyp1b1 Are Equally Induced in Both AHRTtr and Ahrb/b Mouse Hepatocytes AHRTtr mouse hepatocytes express hAHR protein at comparable levels to that of the mAHRb in C57BL/6J mouse hepatocytes (Fig. 1A). Conversely, Ahrfx/fx mice exhibited a lower level of expression (Fig. 1A). Ahrfx/fx, AHRTtr, and Ahrfx/fx/CreAlb mice were injected with 100 lg/kg TCDD or corn oil vehicle and assessed for expression of the AHRresponsive genes Cyp1a1, Cyp1a2, and Cyp1b1 using real-time RT-PCR. Ahrfx/fx mice express the low-affinity mAHRd allele, which has an affinity for TCDD that is comparable to that of the hAHR (Ramadoss and Perdew, 2004). Both TCDD-treated AHRTtr and Ahrfx/fx mice displayed similar levels of TCDDinduced expression of AHR-regulated genes (Fig. 1B).

Functional annotation cluster analysis of DNA microarray data. In order to identify the biological roles of the genes shown to be differentially regulated by the hAHR and mAHRb, the SAM output of differentially regulated genes was subjected to cluster analysis using the DAVID Functional Annotation Clustering Tool (Dennis et al., 2003; Huang da et al., 2009) (http://david.abcc.ncifcrf.gov/summary.jsp). Real-time RT-PCR. Primary hepatocytes isolated from 8- to 10-week-old AHRTtr, Ahrfx/fxCreAlb, and Ahrb/b male mice were treated with 10nM TCDD or vehicle control for 6 h. For in vivo experiments, two groups of three 8-week-old male AHRTtr, Ahrfx/fx/CreAlb, and Ahrfx/fx mice were each treated via ip injection with 100 lg/kg TCDD or corn oil vehicle controls for 6 h. Total messenger RNA (mRNA) from whole-liver sections or primary hepatocytes was isolated using Tri-reagent (Invitrogen). RNA was then converted to cDNA using ABI cDNA archive synthesis kit (ABI), and mRNA expression was quantified using real-time RT-PCR (BioRad) and the relevant primers and normalized using control b-actin primers. All primer sequences are listed in Supplementary table 1. For all RT-PCR reactions, primer efficiencies were 100% (±10%) and melting curves were assessed to ensure that no nonspecific PCR products were formed or stable primer dimerization occurred. SD is represented by y-errors bars in all graphs shown. Statistical analysis. Real-time RT-PCR data were analyzed using two-way ANOVA and Bonferroni posttests, and p values < 0.05 were considered statistically significant.

FIG. 1. AHRTtr mice express functional hAHR protein that is inducible with TCDD treatment. (A) Western blot using anti-AHR and control b-actin antibodies of total protein isolated from AHRTtr, Ahrfx/fx, Ahrfx/fx/CreAlb, and Ahrb/b primary mouse hepatocytes. (B) Real-time RT-PCR: mRNA from whole livers of TCDD-treated (n ¼ 3) AHRTtr, Ahrfx/fx, and Ahrfx/fx/CreAlb mice were subjected to RT-PCR using primers for Cyp1a1, Cyp1b1, and Cyp1a2. Mice were treated with 100 lg/kg TCDD or vehicle for 6 h via ip injection. (C) Realtime RT-PCR of Cyp1a1 mRNA expression in TCDD-treated primary hepatocytes isolated from AHRTtr, Ahrb/b, and Ahrfx/fx/CreAlb mice. Primary hepatocytes isolated from AHRTtr, Ahrb/b, and Ahrfx/fx/CreAlb mice were treated with 10nM TCDD for 6 h. Total RNA, isolated from liver and hepatocytes, was converted to cDNA, and gene expression was quantified using real-time RTPCR and primers for Cyp1a1, Cyp1a2, and Cyp1b1 (Supplementary table 1). *p < 0.05, **p < 0.01, ***p < 0.001.

220

FLAVENY, MURRAY, AND PERDEW

Interestingly, a small degree of constitutive expression of the AHR-regulated genes Cyp1a1 and Cyp1a2 was observed in liver samples from the control AHRTtr mice, which was not seen in relevant samples from Ahrfx/fx mice (Fig. 1B). Also, TCDDtreated Ahrfx/fx/CreAlb mouse livers exhibited a small degree of Cyp1a1, Cyp1a2, and Cyp1b1 activation, possibly due to the presence of other liver cell types that express mAHR. The degree of AHR induction as measured by Cyp1a1 activity was also similar between AHRTtr and Ahrb/b hepatocytes treated with 10nM TCDD (Fig. 1C). A dose of 10nM TCDD yielded a maximal level of Cyp1a1 mRNA induction as was previously determined in hepatocyte-based TCDD-induced dose-response experiments (Flaveny et al., 2009). However, control AHRTtr and Ahrb/b hepatocytes did not exhibit the constitutive activity observed in the control-treated whole livers from AHRTtr mice (Figs. 1B and 1C). In addition, Ahrfx/fx/CreAlb hepatocytes did not display any increase in Cyp1a1 mRNA expression in response to TCDD treatment. Since AHRd expression levels in Ahrfx/fx were low relative to hAHR protein expression, and given that the hAHR displayed reduced constitutive activity in cultured AHRTtr primary hepatocytes, we decided to perform DNA microarray experiments utilizing primary hepatocytes from AHRTtr and Ahrb/b mice. In addition, the use of primary hepatocytes circumvented the complications that may have arisen due to the multiple cell types present in the liver. Therefore, total RNA isolated from TCDD- and control-treated AHRTtr and Ahrb/b primary hepatocytes was subjected to DNA microarray analysis. The mAHRb and hAHR Mutually Modulate a Limited Number of Genes in Primary Hepatocytes The transcriptional profiles of vehicle- and TCDD-treated hepatocytes from AHRTtr and Ahrb/b mice were assessed using Affymetrix mouse genome-wide expression set 430 2.0. Normalized array data from control and treated samples were then assessed for differential expression using SAM. The numbers of genes significantly induced or repressed were obtained from the SAM gene list outputs (Fig. 2). The hAHR and mAHRb mutually induced 265 genes, which was only ~15% of the total number of genes significantly induced by the mAHRb and ~22% of the total number of genes significantly induced by the hAHR in response to TCDD (Fig. 2). Interestingly, the hAHR and mAHRb also mutually repressed 462 genes that was ~59% of the total number of genes repressed by the hAHR, while only ~42% of the total number of genes were repressed by the mAHRb in response to TCDD (Fig. 2). hAHR and mAHRb Differentially Regulate a Number of Genes Involved in Distinct Cellular Pathways TCDD-treated AHRTtr and Ahrb/b were assessed for differential expression using SAM to compare and identify genes that are repressed or induced in AHRTtr compared to the

FIG. 2. Number of genes induced or repressed in hepatocytes isolated from hAHR- and mAHR-expressing mouse primary hepatocytes in response to 10nM TCDD treatment for 6 h. DNA microarray data were normalized using the Probe Logarithmic Intensity Error (PLIER-MM) algorithm. Normalized DNA microarray data outputs from TCDD- and control-treated Ahrb/band AHRTtr hepatocytes were compared for differential expression using SAM (Pan, 2002; Tusher et al., 2001) with 100 permutations, KNN-10. The total number of genes induced and repressed by the hAHR or mAHRb in response to TCDD was then calculated from the SAM output gene lists and plotted graphically.

Ahrb/b. Genes that were identified as differentially regulated were involved in a number of pathways. The pathways identified include immune function: IL10 and complement component 1, q subcomponent, alpha polypeptide (C1qa); cellular metabolism: acireductone dioxygenase 1 (Adi1) and insulin-degrading enzyme (Ide); and cell proliferation: epidermal growth factor (Egf) and gelsolin (Gsn) (Table 1, Fig. 4). Differentially expressed genes were subjected to cluster analysis using DAVID Functional Annotation Clustering Tool (Dennis et al., 2003; Huang da et al., 2009). Compared to the hAHR, the mAHRb was found to disproportionally regulate genes involved in metabolism and membrane transport in response to TCDD treatment. Conversely, the hAHR appeared to disproportionally regulate genes involved in immune response and cell proliferation compared to the mAHRb (Fig. 3). hAHR and mAHRb Differentially Regulate the Genes Oxtr, Adi1, Egf, and Gsn in Primary Hepatocytes A number of genes identified as differentially expressed through DNA microarray analysis were selected and confirmed for differential regulation using real-time RT-PCR. Interestingly, oxytocin receptor (Oxtr), involved in the regulation of parturition and late-onset obesity, and acireductone dioxygenase 1 (Adi1), important to cellular recycling of methionine and mRNA processing, were differentially upregulated by the mAHRb in primary hepatocytes in response to TCDD treatment (Fig. 4A, Table 2). In contrast, epidermal growth factor (Egf), involved in the regulation of cellular differentiation and growth, and gelsolin (Gel), important to metabolism, the maintenance of cellular structure and motility as well as platelet development, are differentially upregulated in hAHR-expressing primary mouse hepatocytes (Fig. 4B, Table 2). Interestingly, there was a notable difference in fold induction observed between the genes identified in the arrays and the real-time RT-PCR data quantitatively; however, qualitatively, the data are consistent. Also the basal levels of Oxtr and Adi1 in particular

221

DIFFERENTIAL GENE REGULATION BY THE HUMAN AND MOUSE AHR

TABLE 1 Differentially Expressed Genes mAHRb Relative to hAHR. Gene List Was Generated Using SAM. Relative Intensities of Probe Logarithmic Intensity Error (PLIER-MM) Normalized Signals from Selected TCDD-Induced Genes in mAHRb Arrays were Compared to that of hAHR Arrays. Genes Were Clustered into Functional Groups Using DAVID Functional Annotation Clustering Tool

ENTREZ gene ID Cell proliferation 18667 12043 13645 227753 Metabolism 15925 104923 18430 Immune response 16153 12259 12262

GENPEPT accession

Gene name

Gene symbol

hAHR signal*

mAHRb signal*

Relative fold change** mAHRb/hAHR

BAC35662 AAA37282 AAH60741 BAE4129

Progesterone receptor B-cell leukemia/lymphoma 2 Epidermal growth factor Gelsolin

Pgr Bcl2 Egf Gsn

2.537 2.288 4.328 1.748

0.405 0.582 2.513 0.905

0.16 0.28 0.58 0.62

BAC26288 CAD10349 BAE25333

Insulin-degrading enzyme Acireductone dioxygenase 1 Oxytocin receptor

Ide Adi1 Oxtr

1.532 0.394 0.743

0.766 9.437 3.427

0.50 23.90 4.63

AAA39274 BAE34901

Interleukin 10 Complement component 1, q subcomponent, alpha polypeptide Complement component 1, q subcomponent, C chain

Il10 C1qa

2.717 6.860

6.195 0.027

2.28 0.004

C1qc

7.657

0.056

0.008

AAH43945

*PLIER-MM normalized array signal intensities corresponding to the listed genes from TCDD-treated hepatocytes from hAHR and mAHRb. **Ratio of PLIER-MM normalized array signal intensities corresponding to the same genes in mAHRb and hAHR TCDD-treated hepatocytes.

(Fig. 4) between the hAHR- and mAHRb-expressing hepatocytes differed.

DISCUSSION

The hAHR and mAHRb differ in amino acid sequence identity at residue A375V within the N-terminal ligand-binding domains of both receptors. This small difference, however, has

FIG. 3. Functional annotation cluster analysis of the genes that are differentially induced or repressed by the mAHR and hAHR. hAHR and mAHRb signal: Probe Logarithmic Intensity Error (PLIER-MM) normalized signal for TCDD-induced samples. Change fold induction: relative PLIER-MM normalized signals for TCDD-treated mAHRb and hAHR samples. Gene clusters of differentially regulated genes were generated using DAVID Functional Annotation Clustering Tool (Dennis et al., 2003; Huang da et al., 2009). The number of genes for each functional cluster differentially regulated by each receptor was then calculated from the gene cluster lists and represented graphically.

previously been shown to be responsible for contrasts in the relative affinities of the mAHRb and hAHR for the ligand TCDD. This amino acid residue is found in the middle of the ligand-binding pocket in models of the AHR ligand-binding domain (Denison et al., 2002; Poland et al., 1994). The mAHRb has a higher relative binding affinity for TCDD compared to the hAHR, whereas the hAHR has a greater relative binding affinity for indirubin compared to the mAHR (Flaveny et al., 2009; Ramadoss and Perdew, 2004). Indeed, hAHR and mAHRb differences in ligand selectivity may be indicative of contrasting functional roles each receptor may have evolved to perform in response to endogenous or exogenous AHR-ligand activation. The contribution of receptor structural divergence to interspecies differences in receptor activity is not without precedence. Structural divergence between human and mouse peroxisome proliferator activated receptor (PPAR)-a was shown to result in differences in susceptibility to PPAR-a ligand– induced toxicity. In contrast to inbred mice, transgenic mice expressing human PPAR-a were shown to be resistant to fibrateinduced hepatomegaly and liver cancer. This appears to be at least in part due to the differential modulation of genes involved in cellular growth regulation (Cheung et al., 2004). Also, the constitutive androstane receptor (CAR) ligand, meclizine, was shown to be an agonist and an inverse agonist for the structurally divergent mouse CAR and human CAR, respectively (Huang et al., 2004). Even the subtle two amino acid residue divergence between human and mouse farnesoid-X-receptor amino acid

222

FLAVENY, MURRAY, AND PERDEW

FIG. 4. Real-time RT-PCR of differentially expressed genes in primary hepatocytes derived from AHRTtr, Ahrfx/fx, and Ahrfx/fx/CreAlb mice. (A) Adi1 and Oxtr mRNA expression assessed using the cognate primers (Supplementary table 1) and total RNA isolated from AHRTtr, Ahrfx/fx, and Ahrfx/fx/CreAlb mice. (B) Real-time RT-PCR: Egf and Gsn mRNA expression was assayed using RNA isolated from AHRTtr, Ahrfx/fx, and Ahrfx/fx/CreAlb mice. (A–B) Primary hepatocytes isolated from AHRTtr, Ahrb/b, and Ahrfx/fx/CreAlb mice were treated with 10nM TCDD or DMSO vehicle control for 6 h. All experiments were repeated three times. Total RNA samples were generated in triplicate, with samples represented by the three genotypes (AHRTtr, Ahrfx/fx, and Ahrfx/fx/CreAlb) originating from hepatocytes isolated from one mouse each. RNA samples were generated using a new batch of control- and TCDD-treated hepatocytes that were generated independently of the hepatocytes RNA samples subjected to microarray analysis. For all experiments, total RNA, isolated from hepatocytes, was converted to cDNA and gene expression was quantified using the cognate primers (Supplementary table 1). Real-time RT-PCR data were analyzed using two-way ANOVA and Bonferroni posttests, and p values < 0.05 were considered statistically significant. *p < 0.05, **p < 0.01, and ***p < 0.001.

sequences were shown to be responsible for the contrasting differences in human and mouse bile acid metabolism (Cui et al., 2002). We also previously showed that the hAHR and mAHRb can differentially recruit LXXLL coactivator–binding motifs, possibly due to differences in the amino acid sequences

of the C-terminal TADs of each receptor (Flaveny et al., 2008). Other previously published studies involving humanized AHR mice have shown that the hAHR responds differently to TCDDinduced activation in vivo (Moriguchi et al., 2003). Therefore, we postulated that the degree of variation exhibited at the Cterminal domain between the mAHRb and hAHR may lead to differential regulation of gene expression by the human and mouse AHR. Recent research involving interspecies comparison of DNA microarray data obtained from rat and human hepatocytes treated with TCDD and PCB 126 has highlighted the toxicogenomic differences between rat and human AHR activity (Carlson et al., 2009). However in this and other studies aimed at comparing of AHR activity across species, the results obtained are limited by the need to identify human gene orthologs. The process of finding human gene orthologs is complicated by the fact that only 37% of rat/mouse genes that possess putative DREs have a known human ortholog (Sun et al., 2004). The need to search for human/rodent orthologs, which is required to compare TCDD-induced expression profiles across different species, is avoided through the experimental approach employed here. Also, in most cases, rodent and human primary cells require, and thus are exposed to different culture conditions or are derived using distinct processes, both factors that may unpredictably affect interspecies comparisons of DNA microarray–based expression profiles. Also, interspecies comparison of AHR-mediated gene regulation often utilizes transformed human and mouse cell types of mixed or dissimilar origins and thus may be difficult to accurately compare. In this study, we aimed to identify toxicogenomic differences between the mAHR and hAHR activity in response to TCDD treatment. Therefore, in order to more accurately test this hypothesis, we deduced that the hAHR and mAHRb activity can be more precisely compared when expressed in the same cell types that are cultured under identical conditions. Thus, we chose to specifically express the human AHR in hepatocytes on an Ahrfx/fx/Crealb conditional knockout background (Flaveny et al., 2009). Within these mice, mAHRd expression is replaced by the hAHR only in hepatocytes. In this study, the activity of both receptors within hepatocytes derived from mice with similar genetic backgrounds (C57BL/6J) and exposed to the same hepatocyte

TABLE 2 Functional Significance of Genes Confirmed to be Differentially Induced by the mAHRb and hAHR ENTREZ gene ID

GENPEPT accession

Gene name

Gene symbol

DRE position

Differentially regulated by

104923 13645 227753

CAD10349 AAH60741 BAE41291

Acireductone dioxygenase 1 Epidermal growth factor Gelsolin

Adi1 Egf Gsn

652 796 47

mAHRb hAHR hAHR

18430

BAE25333

Oxytocin receptor

Oxtr

493

mAHRb

Biological role mRNA processing, methionine recycling Cell proliferation, growth, and differentiation Platelet development, cell structure, motility, apoptosis, and cancer Parturition, breast-feeding, late-onset obesity, metabolism

DIFFERENTIAL GENE REGULATION BY THE HUMAN AND MOUSE AHR

microenvironments was compared. Therefore, this experimental approach circumvented a number of confounding factors associated with comparison of human and rodent DNA microarray data obtained from TCDD-treated human and rodent tissues or cells. Since the AHRTtr, Ahrfx/fx, and Ahrfx/fx/CreAlb mice used in this study were not sufficiently backcrossed to be considered fully congenic, the differences in the genetic background of the mice used may have influenced the results obtained. However, it is unlikely that the differential regulation observed herein is due primarily to disparities in the genetic background of the mice used in this study. In addition, the difference in mAHRb and hAHR affinity for TCDD is one potential limitation to the methodology used here to compare the human and mouse AHR. We first considered attempting to address this in part through comparison of the hAHR with the low-affinity mAHRd that is expressed in Ahrfx/fx mice. The difference in relative ligand affinity between the mAHRd and hAHR for TCDD has been estimated to be twofold (Ramadoss and Perdew, 2004). However, AHRTtr and Ahrfx/fx mice did not display comparable levels of AHR expression in the liver. Therefore, to avoid introducing further uncertainty associated with differences in receptor protein levels, the mAHRb allele, which was expressed at similar levels to the hAHR, was utilized. Also to overcome the difference between mAHRb and hAHR receptor affinity, a saturating dose of TCDD was used that yielded a similar level of Cyp1a1 induction in hepatocytes for both the mAHRb and the hAHR. However, human and mouse hepatocytes may also differentially express coregulators, which may also contribute to observable differences in hAHR- and mAHRb-mediated gene regulation as well. Additionally, mouse and human coregulators may be structurally divergent, and as a result, the hAHR may interact differentially with mouse coregulators compared to the mAHRb. Moreover, the mAHRb and hAHR differentially interact with LXXLL coactivator motifs and therefore may potentially bind selectively to different coactivators (Flaveny et al., 2008). As a result, the observed differential gene regulation displayed by the hAHR and mAHRb may have been influenced by interspecies differences in coregulator expression, recruitment, or receptor/coregulator interactions. Nonetheless, in this study, our findings demonstrate that the mAHRb and hAHR are capable of regulating distinct genes in hepatocytes when both receptors are maximally activated by TCDD. Interestingly, the differences in fold induction observed between the genes identified in the arrays and the real-time RT-PCR data quantitatively differ. Differences between microarray and RT-PCR analysis in terms of the mechanism of detection of expression and disparities in sensitivity that exist between the two techniques may be partially responsible for the differences in fold induction observed. Nonetheless, the data are intended to emphasize distinct qualitative interspecies differences in receptor activity on an identical cellular background, which is consistent across both the real-

223

time and the microarray data presented. In this key aspect, this report provides data that are qualitatively consistent. Also the basal levels of the Oxtr and Adi1, identified to be differentially regulated (Fig. 4), differed between mAHRband hAHR-expressing hepatocytes. This observed difference in basal expression may be due to a myriad of underlying reasons that may possibly be AHR dependent. In particular, this observation does hint at possible differences in the regulatory activity of the unliganded hAHR and mAHRb. Interestingly, comparisons of DNA microarray expression data from vehicle-treated Ahrb/b and AHRTtr hepatocytes revealed no significant differences in the basal levels of genes expressed in both hepatocyte genotypes (data not shown). Therefore, this study provides supporting evidence that the structural differences that the hAHR and mAHR exhibit may contribute to distinct biological hAHR- and mAHR-regulated gene expression in hepatocytes. It is possible that this study may be highlighting differences in gene regulation that potentially underpin interspecies differences in receptor physiological roles. But identifying definite differences in mouse and human endogenous or toxicological function is beyond the scope of this study and should be the subject of future research. Future studies therefore should focus on potential differences in the endogenous and toxicologically relevant roles of the hAHR and mAHR in development, immune response, cancer, and reproduction by generating transgenic mice that express the hAHR in various tissues. The estimation of human toxicological risk for xenobiotic AHR ligands depends on the accuracy of inbred rodents as model systems from which to extrapolate the possible consequences of human exposure to toxic AHR activators. It is cogent and necessary to continually improve on the model systems used to estimate human risk in order to develop a scientifically sound understanding of the caveats and limitations of these models. This investigation identifies an additional challenge to the accuracy of the current approach of extrapolating toxicological data from high-affinity/sensitive mAHRb expressing mouse strains employed to estimate human health risk associated with exposure to AHR ligands, such as TCDD. In this study, we showed that the hAHR and mAHRb can differentially regulate gene expression and thus highlighted another limitation of using data from nontransgenic model systems to predict toxicological consequences of human exposure. These findings also illustrate the importance of transgenic humanized mouse model systems such as the AHRTtr mouse utilized in this investigation to the scientific endeavor of accurately understanding the possible threats that AHR ligand exposure poses to humans.

SUPPLEMENTARY DATA

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

224

FLAVENY, MURRAY, AND PERDEW

FUNDING

National Institutes of Health (ES04869); grant from the Dow Chemical Company.

ACKNOWLEDGMENTS

We thank Dr Christopher Bradfield for kindly providing the Ahrfx/fx/CreAlb mice and helpful discussions and Dr Naomi Altman for providing her advice on microarray data analysis.

REFERENCES Birnbaum, L. S. (1994). Evidence of the role of the Ah receptor in responses to dioxin. Prog. Clin. Biol. Res. 387, 139–154. Carlson, E. A., McCulloch, C., Koganti, A., Goodwin, S. B., Sutter, T. R., and Silkworth, J. B. (2009). Divergent transcriptomic responses to aryl hydrocarbon receptor agonists between rat and human primary hepatocytes. Toxicol. Sci. 112, 257–272. Cheung, C., Akiyama, T. E., Ward, J. M., Nicol, C. J., Feigenbaum, L., Vinson, C., and Gonzalez, F. J. (2004). Diminished hepatocellular proliferation in mice humanized for the nuclear receptor peroxisome proliferator-activated receptor alpha. Cancer Res. 64, 3849–3854. Ciolino, H. P., Daschner, P. J., and Yeh, G. C. (1999). Dietary flavonols quercetin and kaempferol are ligands of the aryl hydrocarbon receptor that affect CYP1A1 transcription differentially. Biochem. J. 340(Pt 3), 715–722. Cui, J., Heard, T. S., Yu, J., Lo, J. L., Huang, L., Li, Y., Schaeffer, J. M., and Wright, S. D. (2002). The amino acid residues asparagine 354 and isoleucine 372 of human farnesoid X receptor confer the receptor with high sensitivity to chenodeoxycholate. J. Biol. Chem. 277, 25963–25969. Denison, M. S., Pandini, A., Nagy, S. R., Baldwin, E. P., and Bonati, L. (2002). Ligand binding and activation of the Ah receptor. Chem. Biol. Interact. 141, 3–24. Denison, M. S., Vella, L. M., and Okey, A. B. (1986). Structure and function of the Ah receptor for 2,3,7,8-tetrachlorodibenzo-p-dioxin. Species difference in molecular properties of the receptors from mouse and rat hepatic cytosols. J. Biol. Chem. 261, 3987–3995. Dennis, G., Jr, Sherman, B. T., Hosack, D. A., Yang, J., Gao, W., Lane, H. C., and Lempicki, R. A. (2003). DAVID: Database for Annotation, Visualization, and Integrated Discovery. Genome Biol. 4, P3. Ema, M., Ohe, N., Suzuki, M., Mimura, J., Sogawa, K., Ikawa, S., and FujiiKuriyama, Y. (1994). Dioxin binding activities of polymorphic forms of mouse and human arylhydrocarbon receptors. J. Biol. Chem. 269, 27337–27343. Flaveny, C., Reen, R. K., Kusnadi, A., and Perdew, G. H. (2008). The mouse and human Ah receptor differ in recognition of LXXLL motifs. Arch. Biochem. Biophys. 471, 215–223. Flaveny, C. A., Murray, I. A., Chiaro, C. R., and Perdew, G. H. (2009). Ligand selectivity and gene regulation by the human aryl hydrocarbon receptor in transgenic mice. Mol. Pharmacol. 75, 1412–1420. Guengerich, P. F., Martin, M. V., McCormick, W. A., Nguyen, L. P., Glover, E., and Bradfield, C. A. (2004). Aryl hydrocarbon receptor response to indigoids in vitro and in vivo. Arch. Biochem. Biophys. 423, 309–316. Henck, J. M., New, M. A., Kociba, R. J., and Rao, K. S. (1981). 2,3,7,8tetrachlorodibenzo-p-dioxin: acute oral toxicity in hamsters. Toxicol. Appl. Pharmacol. 59, 405–408.

Huang, W., Zhang, J., Wei, P., Schrader, W. T., and Moore, D. D. (2004). Meclizine is an agonist ligand for mouse constitutive androstane receptor (CAR) and an inverse agonist for human CAR. Mol. Endocrinol. 18, 2402–2408. Huang da, W., Sherman, B. T., and Lempicki, R. A. (2009). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat. Protoc. 4, 44–57. 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. Kim, D. W., Gazourian, L., Quadri, S. A., Romieu-Mourez, R., Sherr, D. H., and Sonenshein, G. E. (2000). The RelA NF-kappaB subunit and the aryl hydrocarbon receptor (AhR) cooperate to transactivate the c-myc promoter in mammary cells. Oncogene 19, 5498–5506. Ma, Q., Dong, L., and Whitlock, J. P., Jr. (1995). Transcriptional activation by the mouse Ah receptor. J. Biol. Chem. 270, 12697–12703. Madden, C. R., Finegold, M. J., and Slagle, B. L. (2000). Expression of hepatitis B virus X protein does not alter the accumulation of spontaneous mutations in transgenic mice. J. Virol. 74, 5266–5272. Moriguchi, T., Motohashi, H., Hosoya, T., Nakajima, O., Takahashi, S., Ohsako, S., Aoki, Y., Nishimura, N., Tohyama, C., Fujii-Kuriyama, Y., et al. (2003). Distinct response to dioxin in an arylhydrocarbon receptor (AHR)humanized mouse. Proc. Natl. Acad. Sci. U.S.A. 100, 5652–5657. Okey, A. B., Bondy, G. P., Mason, M. E., Kahl, G. F., Eisen, H. J., Guenthner, T. M., and Nebert, D. W. (1979). Regulatory gene product of the Ah locus. Characterization of the cytosolic inducer-receptor complex and evidence for its nuclear translocation. J. Biol. Chem. 254, 11636–11648. Pan, W. (2002). A comparative review of statistical methods for discovering differentially expressed genes in replicated microarray experiments. Bioinformatics 18, 546–554. Patel, R. D., Murray, I. A., Flaveny, C. A., Kusnadi, A., and Perdew, G. H. (2009). Ah receptor represses acute-phase response gene expression without binding to its cognate response element. Lab. Invest. 89, 695–707. Pohjanvirta, R., Unkila, M., and Tuomisto, J. (1993). Comparative acute lethality of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), 1,2,3,7,8pentachlorodibenzo-p-dioxin and 1,2,3,4,7,8-hexachlorodibenzo-p-dioxin in the most TCDD-susceptible and the most TCDD-resistant rat strain. Pharmacol. Toxicol. 73, 52–56. Poland, A., and Glover, E. (1990). Characterization and strain distribution pattern of the murine Ah receptor specified by the Ahd and Ahb-3 alleles. Mol. Pharmacol. 38, 306–312. Poland, A., Glover, E., and Taylor, B. A. (1987). The murine Ah locus: a new allele and mapping to chromosome 12. Mol. Pharmacol. 32, 471–478. Poland, A., and Knutson, J. C. (1982). 2,3,7,8-tetrachlorodibenzo-p-dioxin and related halogenated aromatic hydrocarbons: examination of the mechanism of toxicity. Annu. Rev. Pharmacol. Toxicol. 22, 517–554. Poland, A., Palen, D., and Glover, E. (1994). Analysis of the four alleles of the murine aryl hydrocarbon receptor. Mol. Pharmacol. 46, 915–921. Ramadoss, P., and Perdew, G. H. (2004). Use of 2-azido-3-[125I]iodo-7,8dibromodibenzo-p-dioxin as a probe to determine the relative ligand affinity of human versus mouse aryl hydrocarbon receptor in cultured cells. Mol. Pharmacol. 66, 129–136. Silbergeld, E. K., and Gasiewicz, T. A. (1989). Dioxins and the Ah receptor. Am. J. Ind. Med. 16, 455–474. Sun, Y. V., Boverhof, D. R., Burgoon, L. D., Fielden, M. R., and Zacharewski, T. R. (2004). Comparative analysis of dioxin response elements in human, mouse and rat genomic sequences. Nucleic Acids Res. 32, 4512–4523.

DIFFERENTIAL GENE REGULATION BY THE HUMAN AND MOUSE AHR Sutter, T. R., Guzman, K., Dold, K. M., and Greenlee, W. F. (1991). Targets for dioxin: genes for plasminogen activator-2 and interleukin-1. Science 254, 811–828. Swanson, H. I., and Bradfield, C. A. (1993). The AH-receptor: genetics, structure and function. Pharmacogenetics 3, 213. Travis, C. C., and Hattemer-Frey, H. A. (1991). Human exposure to dioxin. Sci. Total Environ. 104, 97–127. Tusher, V. G., Tibshirani, R., and Chu, G. (2001). Significance analysis of microarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. U.S.A. 98, 5116–5121. Vogel, C. F., and Matsumura, F. (2009). A new cross-talk between the aryl hydrocarbon receptor and RelB, a member of the NF-kappaB family. Biochem. Pharmacol. 77, 734–745.

225

Walisser, J. A., Glover, E., Pande, K., Liss, A. L., and Bradfield, C. A. (2005). Aryl hydrocarbon receptor-dependent liver development and hepatotoxicity are mediated by different cell types. Proc. Natl. Acad. Sci. U.S.A. 102, 17858–17863. Watson, M. A., Devereux, T. R., Malarkey, D. E., Anderson, M. A., and Maronport, R. R. (1995). H-ras oncogene mutation spectra in B6C3F1 and C57BL/6 mouse liver tumors provide evidence for TCDD promotion of spontaneous and vinyl carbamate-initiated liver cells. Carcinogenesis 16, 1705–1710. Weber, L. W. D., Lebofsky, M., Stahl, B. U., Smith, S., and Rozman, K. K. (1995). Correlation between toxicity and effects on intermediary metabolism in 2,3,7,8-tetrachlorodibenzo-p-dioxin-treated male C57BL/6J and DBA/2J mice. Toxicol. Appl. Pharmacol. 131, 155–162.