Suspension-mediated Induction of Hepa lclc7 Cypla-1 Expression Is Dependent on the Ah Receptor Signal Transduction Pathway*

We have recently demonstrated that release of normal human epithelial cells from cell-substratum and/or cell-cell adhesion generates cellular signals that induce the expression of CYPlAl in the absence of xenobiotic polycyclic aromatic hydrocarbons (Sadek, C. M., and Allen- B. L. J. BioZ. Chem. 169, 16067-16074). To directly test the involvement of the Ah receptor sig- nal transduction pathway in CYPlAl induction following suspension of epithelial cells, we analyzed wild-type Hepa lclc7 cells, a subclone of the Hepa-lcl mouse hepatoma line, and two mutant Hepa lclc7 lines, Class I and Class 11. Suspension of wild-type Hepa lclc7 cells for 4 h led to an induction of steady state levels of CYPlAl mRNA, similar to that obtained following treatment of adherent cells with lo-@ M 2,3,7,84etrachloro-dibenzo-p-dioxin. Mutants of the Hepa lclc7 cells defec- tive in different aspects of the Ah receptor signal transduction pathway exhibited negligible (Class I) or no (Class 11) suspension-mediated induction of CYPlAl mRNA. Gel mobility shift analysis of nuclear extracts from suspended or 2,3,7,8-tetrachlorodibenzo-p-dioxin-treated wild-type cells showed that both treatments pro- duced identical shifts in the mobility of an XRE-contain-ing probe. Antibody supershift experiments confirmed that the Ah receptor was a of DNA-pro-tein cells. data directly that cells leads to and activation of the Ah receptor to a

tive in different aspects of the Ah receptor signal transduction pathway exhibited negligible (Class I) or no (Class 11) suspension-mediated induction of CYPlAl mRNA. Gel mobility shift analysis of nuclear extracts from suspended or 2,3,7,8-tetrachlorodibenzo-p-dioxintreated wild-type cells showed that both treatments produced identical shifts in the mobility of an XRE-containing probe. Antibody supershift experiments confirmed that the Ah receptor was a component of the DNA-protein complex from suspended wild-type Hepa lclc7 cells. These data directly demonstrate that suspension of wild-type Hepa lclc7 cells leads to nuclear localization and activation of the Ah receptor to a DNA-binding form.
Ligand-activated transcription factors regulate the expression of a number of different gene products, including members of the cytochrome P450 monooxygenase superfamily. One member of this enzyme family, CYPlAl,' is substrate-inducible and has been shown to be activated by exposure to xenobiotic polycyclic aromatic hydrocarbons, such as 2,3,7&tetrachloro-Grant AR40284 (to B. L. A,-H.). This is contribution 267 from the En-* This work was supported in part by National Institutes of Health vironmental Toxicology Center, University of Wisconsin. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
f To whom correspondence should be addressed. Tel.: 608-262-2884; The abbreviations used are: CUP, cytochrome P450 as recommended by the Committee on Cytochrome P450 Nomenclature (Nelson et al., 1993); the italicized CYP indicates the human gene and cDNA, Cypla-1 indicates the murine gene and cDNA, the human and murine mRNA and protein are written as the non-italicized symbol CYP, an Arabic number designates the P450 family, the subfamily is identified by a letter when two or more subfamilies exist within a family, and an Arabic number represents the specific gene or protein; a-MEM, a-minimal essentialmedium;TCDD,2,3,7,8-tetrachlorodibenzo-p-dioxin;XRE,xenobiotic responsive element; MOPS, 4-morpholinepropanesulfonic acid. dibenzo-p-dioxin (TCDD). TCDD-activated induction of CYPlAl occurs via a signal transduction pathway involving the aromatic hydrocarbon (Ah) receptor (Burbach et al., 19921, a 90-kDa heat shock protein (hsp9O) (Perdew, 19881, and the Ah receptor nuclear translocator protein (Arnt) (Hoffman et al., 1991;Reyes et al., 1992). Studies to date indicate that TCDD enters the cytosol where it binds an Ah receptor-hsp90 complex. This complex then interacts with Arnt, concomitantly displacing hsp9O (McGuire et al., 1994). The liganded Ah receptor-Arnt heterodimer forms an active transcription factor which binds to xenobiotic responsive elements (XREs) in the 5' promoter region of the C Y P l A l gene, as well as a battery of other Ah receptor-responsive genes (reviewed in Nebert et al. (1993) andOkey et al.( 1994)). Aside from metabolism of xenobiotic polycyclic aromatic hydrocarbons, the role of the CYPlAl enzyme is currently unclear, in large part because no endogenous substrate or Ah receptor ligand has been identified (reviewed in Poellinger et al. (1992)).
We recently demonstrated that both CYPlAl mRNA and enzyme activity are induced in cultured human keratinocytes, but not dermal fibroblasts, in the absence of polycyclic aromatic hydrocarbon treatment by suspending adherent cells (Sadek and Allen-Hoffmann, 1994). In this previous study, normal skin keratinocytes were suspended in serum-free and growth factorfree medium made semi-solid with an inert polymer, methylcellulose, to prevent cell-cell adhesion and alter cell shape. Suspension of adherent normal human epithelial cell types, such as keratinocytes, induces CYPlAl mRNA and enzyme activity in the absence of serum, calcium, as well as a number of growth promoting agents such as epidermal growth factor, hydrocortisone, insulin, and cholera toxin. A similar induction has also been noted following suspension of cultured normal human thymic epithelial cells.' Addition of methylcellulose does not alter the osmolarity of the medium2 and replacement of methylcellulose with colloidal silica (Percoll) in medium also supports C Y P l A l induction. Furthermore, suspension in medium alone is sufficient to elicit a response in cultured normal keratinocytes, thus ruling out any potential contaminant unique to methylcellulose as the inducing agent (Sadek and Allen-Hoffmann, 1994).
As reported in our earlier work, other members of the Ah receptor gene battery responded to either suspension or TCDD treatment in a similar manner. Therefore, we hypothesized that suspension of epithelial cells may initiate cellular events which activate the Ah receptor pathway. To test this hypothesis, we have used an epithelial cell line, Hepa lclc7, derived from a mouse liver tumor. In this report we have analyzed mutants of the Hepa lclc7 cells defective in different aspects of the Ah receptor signal transduction pathway. Our findings confirm that suspension of wild-type Hepa lclc7 cells leads to an induction of Cypla-1 via an Ah receptor-dependent pathway.

EXPERIMENTAL PROCEDURES
Cell Cultures-Hepa lclc7 wild-type cells, Class I and Class I1 mutant cell lines were the kind gift of Dr. James Whitlock, Jr. (Stanford University, Stanford, CA). Cells were maintained in a-minimal essential medium (a-MEM) without nucleosides supplemented with 10% fetal calf serum, 100 units penicillin, and 100 pg/ml streptomycin. All cultures were maintained at 37 "C in a humidified, 5% CO, atmosphere.
Deatment of Cells-Cells were treated by suspension in serum-free a-MEM containing 1.68% methylcellulose (4000 centipoises; Fisher Scientific). Sterile methylcellulose-containing medium was prepared as described previously (Sadek and Allen-Hoffmann, 1994). Briefly, 3.37 g of methylcellulose and a magnetic stir bar were autoclaved in a 250-ml polyallomer centrifuge bottle (Nalge Co., Rochester, NY). 100 ml of serum-free a-MEM was heated to 60 "C, added to the methylcellulose powder, and stirred at room temperature for 20 min. An additional 100 ml of unheated serum-free a-MEM was then added and the mixture stirred a t 4 "C for 1 h. To remove undissolved methylcellulose fibers, 100-ml aliquots of methylcellulose were centrifuged at 16,000 x g for 90 min a t 4 "C.
For suspension studies, cells were removed from tissue culture plates with 0.5 mM EDTA, 0.1% trypsin, washed with serum-containing medium to inactivate residual trypsin, then with serum-free medium, and finally suspended in serum-free methylcellulose-containing medium a t a density of approximately 2 x lo6 cells/ml (100-mm dish/4 ml) in sterile 50-ml polypropylene tubes. Suspended cells were incubated a t 37 "C in a humidified 5% CO, atmosphere. Cells were recovered from suspension by repeated dilution with serum-free medium followed by centrifugation at 1000 x g. Adherent cells used for controls were rinsed and 2. Northern analysis of steady state CYPlAl mRNA levels following suspension or TCDD treatment of wild-type and mutant Hepa l c l c 7 cells. Adherent wild-type Hepa lclc7 cells and Class I and Class I1 mutants were treated with serum-free medium alone, with 10." M TCDD, or were removed from adherent cultures with trypsinEDTA and suspended in serum-free semi-solid medium for 4 h. Poly(AY RNA was isolated and Northern analysis performed using the random primer-labeled I"2PlcDNA for mouse Cypla-I. For a loading control, the same blot was stripped and rchybridized to a cDNA for glyceraldehyde-3-phosphate dehydrogenase (GAPDH).
UNTR, untreated; SUSP, suspended. treated with 4 ml of a-MEM or a-MEM containing lo-' M TCDD per 100-mm dish for the same time as suspended cells.
Northern Analysis-Poly(A)' RNA was isolated from logarithmically growing cells according to the method of Badley et al. (1988). Poly(A)+ RNA was electrophoresed in a 1.2% agarose gel containing formaldehyde and electroblotted to a Zeta probe membrane (Bio-Rad). The membrane was prehybridized and then hybridized in the presence of a random primer "P-labeled cDNA probe as recommended by the supplier. The probes used were a 1269-base pair cDNA for rat glyceraldehyde-3phosphate dehydrogenase, pGPDN5 (Fortet al., 1985), and a full-length 2620-base pair cDNAfor murine Cypla-1 (Gonzalez et al., .
Nuclear Extract Preparation-Adherent Hepa lclc7 cells were treated for 90 min with a-MEM only, M TCDD or cells were suspended following trypsinization in a-MEM containing 1.68% methylcellulose. Twelve 100-mm tissue culture dishes were used for each treatment. Nuclear extracts were prepared essentially as described by Saatcioglu et al. (1990). Following treatment of adherent cells, medium was removed and cells were scraped into 30 ml of cold phosphatebuffered saline. Cells were immediately centrifuged in 15-ml centrifuge tubes for 5 min a t 1500 rpm in a Dynac I1 centrifuge. Suspended cells were removed from methylcellulose by repeated dilution with phosphate-buffered saline followed by centrifugation. After this initial rinsing step, all cells were swelled for 10 min on ice in 1 ml of Buffer A (10 mM Hepes, pH 7.5). Next, 1 ml of Buffer B (25 mM Hepes (pH 7.5), 3 mM MgCI,, 1 mM dithiothreitol) was added and swelled cells were immediately homogenized on ice by 25 strokes in a cold (4 "C) Wheaton stainless steel tissue homogenizer and centrifuged in a Beckman Microfuge a t 14,000 rpm for 30 s. The supernatant containing the cytosolic fraction was removed and the crude nuclear pellet was resuspended in 2 x volume of Buffer C (25 mM Hepes (pH 7.5), 1 mM dithiothreitol). Nuclei were lysed by the addition of 2 M KC1 to a final concentration of 0.4 M KC1 and were rocked on a nutator at 4 "C for 30 min. Samples were adjusted to 20% glycerol and centrifuged a t 150,000 x g a t 4 "C in 0.8-ml tubes in a Beckman SW 50.1 rotor. To avoid freeze-thawing of nuclear lysates, 30-pl aliquots were stored a t -80 "C until time of assay. Protein concentrations were determined by the Coomassie Plus Protein Assay (Pierce) and ranged from 3 to 4 mg/ml. Gel Mobility Shift Assays-DNA-binding reactions were camed out using a "P-labeled 29-mer oligonucleotide as a probe (a gift from Dr. Alan Poland, McArdle Laboratory, Madison, WI). The double-stranded probe was prepared by random primer extension labeling with P2P1dCTP using a 29-mer containing the XRE corresponding to the sequence located between -1001 and -973 base pairs upstream of the Cypla-1 cap site. The sequence of the 29-mer probe representing the noncoding strand was 5'-GAG CTC GGA GTl' GCG TGA CAA GAG CCG GA-3'. The core XRE sequence is underlined. To reduce nonspecific binding, 10 pg of nuclear extract protein was incubated for 20 min with 2.6 pg of poly(d1-dC), and 500 ng of salmon sperm DNAin 20 p1 of buffer containing 1 mM MOPS (pH 7.9), 0.02 mM EDTA, 0.05 mM dithiothreitol, 0.3 mM MgCI,, 0.4 mM spermidine, and 0.2% Ficoll. Following this incubation, 0.4 fmol of the double-stranded radiolabeled 29-mer DNA probe containing the XRE was added to the mixture in a volume of 5 p1 and the incubation was allowed to continue for an additional 20 min. Incubations were carried out a t room temperature. Samples were electrophoresed on a 1 x TBE nonreducing 6% polyacrylamide gel that had Adherent wild-type Hepa lclc7 cells were rinsed and treated with serum-free medium alone (Untreated, lunes 1-3), with M TCDD (TCDD, lunes 4-6), or suspended in serum-free, semi-solid medium (Suspended, lunes 7-9) for 90 min. Following treatment, nuclei were isolated as described and nuclear extracts were analyzed by gel mobility shift assay using a 32P-labeled 29-mer containing the core XRE sequence. XRE-binding reactions were conducted in the absence (lunes 1,4, 5, and 8) or 5000-fold (lunes 3, 6, and 9) and 7) or presence of a 1000-fold (lanes 2, molar excess of unlabeled XRE probe. The

positions of the XRE-binding complex (R)
and the free (Free) probes are indicated. Free been cooled to 4 "C and equilibrated for 2 h a t 250 volts. The gel was run at 20 mA constant current for 1 h. Autoradiography was performed on the dried gel. For the competition assay, the indicated molar excess of unlabeled 29-mer containing the XRE was incubated concurrently with the radiolabeled probe. In the antibody supershift study, extracts were incubated with 6 1. 11 containing either 2 pg of preimmune serum or 2 pg of the A-1 rabbit anti-mouse antibody to the Ah receptor (Pollenz et ul., 1994) for an additional 20 min following incubation with the radiolabeled probe.

Suspension-induced Increases in CYPlAl mRNA Are Dependent on Functional Ah
Receptor and Arnt-To study the involvement of the Ah receptor in suspension-induced Cypla-I, we used wild-type mouse hepatoma Hepa lclc7 cells (Hankinson, 1979) and two mutant Hepa lclc7 cell lines, termed Class I and I1 (Miller et al., 1983;Whitlock and Galeazzi, 1984).
These mouse hepatoma-derived epithelial cells have been used extensively to study the mechanism of TCDD-dependent induction of CYPlAl. For this reason, they are important as a model to analyze the signal transduction mechanisms triggered by suspension of certain epithelial cell types.
The wild-type Hepa lclc7 hepatoma cell line exhibits the morphological features of epithelial cells (Fig. la). They are well spread and form cell-cell contacts that produce a characteristic cobblestone morphology. Class I mutants possess a fusiform, fibroblastic morphology and are poorly spread, with little cell-cell interaction (Fig. lb). Class I1 mutants exhibit cell-cell interaction typical of foci formation in fibroblastic cell types (Fig. IC). Neither Class I nor Class I1 cells possess morphology typical of cultured epithelial cells. All three cell lines contain a positive cis-acting Ah receptor-dependent promoter/enhancer region of Cypla-1. Wild-type cells express functional Ah receptors and Arnt protein and are highly responsive to treatment with TCDD. Both Class I and Class I1 mutants contain defects in the Ah receptor signal tranduction pathway which result in either an attenuation or inability to activate transcription of Cypla-I. Class I mutants contain functional Arnt protein but contain less than 10% of normal Ah receptor levels as measured by ligand binding (Miller et al., 1983) and photoaffinity labeling (Burbach et al., 1992). Class I1 mutants contain normal levels of Ah receptor, however, the Ah receptor will not translocate to the nucleus following ligand binding (Whitlock and Galeazzi, 1984).
To directly test the involvement of the Ah receptor signal transduction pathway in suspension-mediated induction of Cypla-1, we analyzed wild-type Hepa lclc7 and the two mutant Hepa lclc7 cell lines, Class I and Class 11. We first tested the responsiveness of adherent wild-type, Class I, and Class I1 Hepa lclc7 cells to either TCDD treatment or suspension by Northern analysis (Fig. 2). All cells were treated for 4 h in serum-free medium with or without M TCDD or by suspension in serum-free, semi-solid medium. We found wild-type cells to be responsive to both TCDD and suspension treatments to a similar degree. Suspension-mediated induction of wildtype Hepa lclc7 CYPlAl mRNA also occurred in the absence of the pH indicator phenol red (data not shown). Class I cells have lower levels of Ah receptors and thus were less responsive to TCDD treatment when compared to wild-type Hepa lclc7 cells. did not occur in cells defective in the Ah receptorIArnt pathway, our findings strongly indicate a role for both the Ah receptor and Arnt protein in suspension-induction of the Cypla-1 gene. The reason for the exceedingly modest induction of CYPlAl mRNA following suspension of Class I mutants is currently unclear. However, this response may reflect the suspensionmediated release or production of an endogenous ligand(s) in lower quantity or with less ability, compared to TCDD, to activate the fewer number of Ah receptors present in these cells. For example, suspension may produce a lower concentration of endogenous ligand than lo-' M TCDD and therefore ligandreceptor interaction is less likely to occur. Regardless of the specific mechanism of suspension-mediated Ah receptor activation, our studies suggest that cell shape or cell-cell and cellsubstratum interactions dramatically influence this signal transduction pathway.
The DNA Binding Activity of the Ah Receptor Is Induced by Suspension-During TCDD treatment, the Ah receptor translocates to the nucleus and forms a heterodimeric transcription factor with Arnt. This complex then binds XRE sequences in the DNAto initiate transcription. Because we observed that the suspension-induction of CYPlAl mRNA is dependent on Ah receptor levels and nuclear translocation, we hypothesized that suspension of wild-type Hepa lclc7 cells would lead to Ah receptor activation similar to that observed following exposure of adherent cells to TCDD. We next performed gel mobility shift assays to determine if suspension of the wild-type Hepa lclc7 cells leads to formation of an active XRE-binding transcription factor. Adherent wild-type cells were treated with serum-free medium alone, with TCDD, or were suspended for 90 min. Nuclear extracts were incubated with a radiolabeled 29-mer probe containing the core XRE sequence. Both TCDD treatment and suspension of the wild-type cells led to identical shifts in probe mobility (Fig. 3, lunes 4 and 7). As expected, the 29-mer probe was not bound by nuclear extracts of untreated adherent control cells (Fig. 3, lune 1 ). To ensure that the XRE containing 29-mer was properly labeled and that the shifted DNA was specifically recognized by the Ah receptor complex, we also performed competition experiments using unlabeled oligomer containing the XRE. Nuclear extracts from adherent, TCDD-treated, and suspended cells were incubated with a 1000or 5000-fold molar excess of unlabeled oligomer in addition to the radiolabeled probe. As expected, the intensity of the shifted band was reduced to background levels upon addition of cold competitor (Fig. 3, lanes 5, 6, 8, and 9). These results demonstrate that suspension of wild-type Hepa lclc7 cells produces a nuclear DNA-binding factor of the same size and in a similar quantity as TCDD treatment.
To confirm that the nuclear complex binding the 29-mer probe contains the Ah receptor, we conducted an antibody supershift experiment (Fig. 4). Nuclear extracts were generated from untreated (lunes 1-31, TCDD-treated (lanes 4-61, or SUSpended (lunes 7-9) wild-type cells and incubated with the radiolabeled XRE-containing 29-mer in the presence or absence of an antibody against the Hepa lclc7 Ah receptor. The A-1 antibody used in our experiments was generated against a recombinant protein (amino acids 61-419 of the AhR) encoded by a cDNA to the mouse Ah"" allele (Pollenz et al., 1994). This antibody recognizes authentic Ah receptor protein in both wildtype Hepa lclc7 cells and in both Class I and 11 mutants. Both TCDD treatment (lane 4 ) and suspension (lane 7) induced identical shifts in gel mobility, similar to the results shown in Fig. 3. Addition of the A-1 antibody to XRE-binding reactions containing the nuclear extracts from adherent TCDD-treated cells (lane 6) or suspended cells (lane 9) resulted in further retardation in gel mobility (supershift) of the radiolabeled complex relative to preimmune serum (lanes 5 and 8). As anticipated, nuclear extracts from untreated adherent wild-type cells bound little radiolabeled XRE-containing probe under any condition (lanes 1 3 ) . These results demonstrate that TCDD treatment and suspension not only lead to a similar shift in band mobility but that activated Ah receptor is a component of the nuclear extract from both treatments. Based on our results in Fig. 2 and the identical gel shifts and supershifts produced by these two treatments, we suspect that Arnt is a component of the XRE-binding complex as well. These data directly demonstrate that suspension of wild-type Hepa lclc7 cells leads to activation of the Ah receptor to a DNA-binding form in the absence of xenobiotics.

DISCUSSION
In this report, we demonstrate that the suspension-mediated induction of Cypla-1 requires an intact Ah receptor signaling pathway. As with human keratinocytes, suspension of adherent wild-type Hepa lclc7 cells causes induction of Cypla-1 gene expression in the absence of polycyclic aromatic hydrocarbon treatment (Sadek and Allen-Hoffmann, 1994). The cellular signals responsible for suspension-mediated signal transduction through the Ah receptor pathway are currently unknown. It is possible that suspension promotes production or release of a n endogenous ligand(s) thereby leading to activation of an Ah receptor to a DNA-binding form via a mechanism similar to polycyclic aromatic hydrocarbon induction. It is also reasonable to postulate that the cell shape and adhesion changes that result from suspension may disrupt the Ah receptor-hsp9O complex sufficiently to activate non-liganded Ah receptor to a DNAbinding form. The molecular chaperone hsp90 is necessary to maintain the Ah receptor in a polycyclic aromatic hydrocarbon ligand-binding conformation (Whitelaw et al., 1993). However, release of hsp90 is required for activation of the Ah receptor to a DNA-binding form. Following disruption of hsp90-Ah receptor complexes, the Ah receptor exhibits a loss of affinity for its cognate ligands (Pongratz et al., 1992). In vitro studies have shown that disruption of the Ah receptor-hsp90 complex in the absence of ligand is sufficient to promote Ah receptor binding to the XRE (Pongratz et al., 1992). Recent studies by McGuire et al. (1994) suggest that release of hsp90 from the Ah receptor is dependent not only on ligand binding but also on interaction with Arnt. That is, in vitro activation of the Ah receptor by ligand-induced release of hsp90 appears to require interaction with functional Arnt. By analogy, suspension of epithelial cells may disrupt the association of hsp9O and the Ah receptor directly, thus promoting interaction with Arnt and consequently promoting XRE binding and Cypla-1 induction. Alternatively, suspension may alter cellular pools ofAh receptor and Arnt and in this manner facilitate Amt-mediated release of hsp9O from the Ah receptor. Therefore, the presence of an endogenous ligand may or may not be necessary for suspensionmediated induction of CYPlAl in the epithelial models we have investigated.
Taken together, our results suggest that an endogenous ligand(s) or other cellular change resulting from suspension of adherent wild-type Hepa lclc7 cells activates the Ah receptor to a DNA-binding form and promotes its nuclear localization. The biological role ofAh receptors in normal tissue homeostasis is a relatively unexplored area. Further studies will focus on confirming the presence of endogenous ligands as well as determining the role of Arnt in suspension-mediated induction of CYPlAl in epithelial cell types. Isolation and characterization of putative endogenous ligands or other cellular mechanisms of Ah receptor activation will increase our understanding of this signal transduction pathway in normal developing and adult tissues as well as pathological conditions.