Glucocorticoids Stimulate p21 Gene Expression by Targeting Multiple Transcriptional Elements within a Steroid Responsive Region of the p21 waf1/cip1 Promoter in Rat Hepatoma Cells*

Glucocorticoids can induce a G1arrest in the cell cycle progression of BDS1 rat hepatoma cells. In these cells, dexamethasone, a synthetic glucocorticoid, stimulated a rapid and selective increase in expression of the p21 cyclin-dependent kinase (CDK) inhibitor mRNA and protein and virtually abolished CDK2 phosphorylation of the retinoblastoma protein. Expression of the p27 CDK inhibitor, and other G1-acting cell cycle proteins, remained unaffected. Dexamethasone stimulated p21 promoter activity in a p53-independent manner that required functional glucocorticoid receptors. Transforming growth factor-β, which also induced a G1 cell cycle arrest of the hepatoma cells, failed to elicit this response. Analysis of 5′ deletions of the p21 promoter uncovered a glucocorticoid responsive region between nucleotides −1481 and −1184, which does not contain a canonical glucocorticoid response element but which can confer dexamethasone responsiveness to a heterologous promoter. Fine mapping of this region uncovered three distinct 50–60-base pair transcriptional elements that likely function as targets of glucocorticoid receptor signaling. Finally, ectopic expression of p21 had no effect on hepatoma cell growth in the absence of glucocorticoids but facilitated the ability of dexamethasone to inhibit cell proliferation. Thus, our results have established a direct transcriptional link between glucocorticoid receptor signaling and the regulated promoter activity of a CDK inhibitor gene that is involved in the cell cycle arrest of hepatoma cells.

Glucocorticoids regulate the transcription of a network of genes that trigger characteristic responses in specific target cells (1)(2)(3). The direction and strength of the transcriptional regulation depend on steroid receptor interactions with specific DNA recognition sequences (2, 4 -6), protein-protein interactions with transcription factors and accessory proteins (1,(7)(8)(9), the overall promoter context of the regulated gene, and the availability of specific sets of tissue-specific transcriptional regulators (4,5,9,10). As a result of these interactions, glucocorticoids can either inhibit or stimulate the in vivo and in vitro growth of many types of normal and transformed cells. Normal hepatocytes and certain hepatoma cell lines display an acute sensitivity to the anti-proliferative effects of glucocorticoids and can be utilized to examine the mechanisms of steroidinduced growth suppression (11)(12)(13)(14)(15)(16)(17). The glucocorticoid growth suppression response is controlled through cellular cascades in which the receptor-mediated transcription of primary response genes regulates the subsequent expression and/or activity of a diverse set of genes including factors important for cell cycle progression (5,7,18). However, the functional connection between the steroid-regulated transcriptional events and the cell cycle arrest of hepatic-derived cells is poorly understood at a molecular level.
Most of the hormonal cues known to drive cells through critical cell cycle transitions, or inhibit cell cycle progression, target components that act within the G 1 phase or at the G 1 /S boundary (19 -21). Progression through the cell cycle is mediated by the activation of the cyclin-dependent kinases (CDKs) 1 (21,22). CDKs are regulated by complex formation with a cyclin partner and by phosphorylation at specific residues (23)(24)(25). Two families of proteins associate with specific cyclin-CDK complexes and act as CDK inhibitors (20,23,26) as follows: the p16 family (p16/Ink4, p15, p18, and p19), which competes with cyclin D for binding to CDK4 and CDK6, thereby negatively regulating kinase activity, and the p21 family (p21, p27, and p57), which forms complexes with a wider range of cyclin-CDK complexes (26,27). The p21 protein has been shown to be a potent inhibitor of CDK2 and CDK4 kinase activity in vitro (25,28,29). Binding of the CDK inhibitors prevents the CDKmediated phosphorylation of the retinoblastoma (Rb) protein, which functions to sequester the E2F-1 transcription factor in an inactive complex (30 -32). In addition, p21 and p27 have been shown to block the in vitro phosphorylation of CDK by the CDK-activating kinase (33).
Several of the CDK inhibitors are targets of signaling pathways induced by negative growth regulators. For example, in some transforming growth factor-␤ (TGF-␤) growth-inhibited cells, cell cycle progression is blocked in late G 1 as a result of the association of p27 with newly formed cyclin E-CDK2 complexes and of an increase in p15 transcript levels (20,25). Other studies have correlated the expression of CDK inhibitors with cell cycle arrest and/or differentiation in many cell types after treatment with extracellular growth regulators, exposure to particular environmental conditions, or by ectopic expression of intracellular modulators of the cell cycle (27,34). For example, p21 is under transcriptional control by the p53 tumor suppressor gene, vitamin D3 receptor (35), retinoic acid receptor (36), the AP2 (37), E2A (38) Sp1, Sp3, and STAT1 transcription factors, and TGF-␤1 through their respective response elements in the p21 promoter (38 -42). In specific osteosarcoma and fibroblast cell lines, glucocorticoids increase the level of p21, although the mechanism of transcriptional control of p21 was not characterized in these cell types (43,44). Recent studies have indicated that the transcriptional regulation of p21 is part of physiologically important proliferative, developmental, and differentiation processes. For example, strong signaling through Raf can lead to a G 1 cell cycle arrest by up-regulation of p21 (45,46). Also, the targeted expression of p21 in hepatocytes of transgenic mice resulted in a decreased number of adult hepatocytes, a smaller liver, aberrant liver tissue organization, the failure of partial hepatectomy to stimulate liver regeneration, and an observed inhibition of cyclin D-CDK4 activity with a resulting G 1 cell cycle arrest (47). This result implicates p21 as an important target in the growth regulation of normal hepatic-derived cells and perhaps their transformed counterparts.
We have previously characterized glucocorticoid-sensitive and -resistant hepatoma cell variants generated from Fu5 hepatoma cell populations and have demonstrated that the dexamethasone growth suppression response is a receptor-dependent process that does not affect cell viability (11,12,14). In Fu5-derived BDS1 cells, dexamethasone induces an early G 1 block in cell cycle progression (14) mediated, in part, by an increased expression of the C/EBP␣ transcription factor (48). Given the importance of p21 in controlling progression through G 1 in normal hepatic cell types, and the known transcriptional regulatory pathways operating on the p21 gene, the regulation of p21 may be part of the signal transduction pathway through which glucocorticoids mediate their tissue-specific anti-proliferative effects. In this study, we demonstrate that the glucocorticoid G 1 cell cycle arrest of hepatoma cells is accompanied by the selective stimulation in p21 expression and activation of the p21 promoter through multiple transcription factor binding sites, within a defined glucocorticoid-regulated region. The following companion paper (82) establishes a functional role for the glucocorticoid-induced expression of the C/EBP␣ transcription factor in the regulated expression and promoter activity of p21 in hepatoma cells. The transcriptional control of p21 through at least two distinct pathways provides a basis for understanding the anti-proliferative mechanism of glucocorticoids.  (3,000 Ci/mmol) were obtained from NEN Life Science Products. Anti-p21, anti-p27, anti-CDK2, anti-CDK4, anti-CDK6, anti-cyclin D1, and horseradish peroxidase (HRP) donkey antigoat antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). HRP-conjugated goat anti-rabbit and goat anti-mouse antibodies were purchased from Bio-Rad. The enhanced chemiluminescence protein detection system and the Multiprime DNA labeling kit were purchased from Amersham Corp. PCR reagents including Taq polymerase, Taq polymerase buffer, and Perfect Match enzyme were purchased from Boehringer Mannheim. The chimeric pGRE-CAT reporter plasmid containing six glucocorticoid response elements linked to the chloramphenicol acetyltransferase (CAT) reporter gene was a generous gift from Dr. Keith R. Yamamoto (Dept. of Biochemistry and Biophysics, University of California, San Francisco). The chimeric p21 promoter CAT reporter plasmids containing deletions at Ϫ2.326 kb, Ϫ1.892 kb, Ϫ1.481 kb, Ϫ1.184 kb, Ϫ0.883 kb, Ϫ0.585 kb, and Ϫ0.291 kb, and the p21-luciferase reporter plasmids were generous gifts from B. Vogelstein (Molecular Genetics Laboratory, Johns Hopkins Oncology Center, Baltimore) and have been described previously (41). Fine mapping deletion constructs were PCR-amplified from the full-length Ϫ2.326p21 CAT reporter plasmid and cloned into pBLCAT2 (Promega). All other reagents were of highest available purity.

Materials
Hepatoma Cell Lines and Methods of Culture-Glucocorticoid-sensitive BDS1 cells and glucocorticoid-resistant EDR3 (receptor-defective) cells are epithelial tumor cells derived from the rat Reuber hepatoma (12,48). All cell lines were routinely grown in Dulbecco's modified Eagle's medium/F-12/10% fetal bovine serum at 37°C in humidified air containing 5% CO 2 . Cell culture medium was routinely changed every 48 h. Dexamethasone was added to a final concentration of 1 M as indicated, and TGF-␤ was added to a final concentration of 2 ng/ml.

Assay of DNA Synthesis by [ 3 H]Thymidine
Incorporation-Triplicate samples of asynchronously growing BDS1 or EDR3 hepatoma cells were treated for the indicated times with dexamethasone, pulse-radiolabeled for 2 h with 3 Ci of [ 3 H]thymidine (84 Ci/mmol), washed three times with ice-cold 10% trichloroacetic acid, and lysed with 300 l of 0.3 N NaOH. Lysates (100 l) were transferred directly into vials containing a liquid scintillation mixture; radioactivity was quantitated by scintillation counting.
Flow Cytometric Analyses of DNA Content-Hepatoma cells (4 ϫ 10 4 ) were plated onto Corning 6-well tissue culture dishes and treated with or without 1 M dexamethasone or 2 ng/ml TGF-␤, and the medium was changed every 24 h. After 48 h, cells were hypotonically lysed in 1 ml of DNA staining solution (0.5 mg/ml propidium iodide, 0.1% sodium citrate, 0.05% Triton X-100). Nuclear emitted fluorescence was measured with a Coulter Elite instrument with laser output adjusted to deliver 15 milliwatts at 488 nm. Cell nuclei (10 4 ) were analyzed from each sample at a rate of 300 -500 nuclei/s. The percentages of cells within the G 1 , S, and G 2 /M phases of the cell cycle were determined by analysis with the Multicycle computer program provided by Phoenix Flow Systems in the Cancer Research Laboratory Microchemical Facility of the University of California, Berkeley.
Isolation of Poly(A) ϩ RNA and Northern Blot Analysis of p21 mRNA Levels-Logarithmically growing BDS1 cells were treated with 1 M dexamethasone and/or 10 g/ml cycloheximide for 8 h. Cells were then lysed, and poly(A) ϩ RNA was isolated from hepatoma cells as described previously (14). For Northern blot analysis, 2 g of poly(A) ϩ RNA was electrophoretically separated in a 6% formaldehyde, 1% agarose gel, transferred onto Nytran nylon membranes (Schleicher & Schuell), and cross-linked in a UV Stratalinker (Stratagene). Membranes were preincubated with 100 g/ml denatured salmon sperm DNA and subsequently hybridized with cDNA probes [␣-32 P]dCTP-labeled by random primer extension (Amersham Corp.). For p21 mRNA detection, membranes were washed twice with 2 ϫ SSC, 0.1% SDS at room temperature and once for 30 min and then once for 1.5 h. The same procedure was then repeated at 50°C. For detection of GAPDH transcripts, membranes were washed with 2 ϫ SSC twice at room temperature for 10 min followed by two 30-min washes at 60°C in 0.2 ϫ SSC, 0.1% SDS. The membranes were air-dried followed by autoradiographic exposure to x-ray film 1-5 days at Ϫ80°C. p21 transcript expression was detected with a purified 850-base pair (bp) EcoRI fragment 9C from mouse p21 cDNA. GAPDH mRNA was detected with a 560-bp XbaI/HinD III fragment 3A of human cDNA.
Immunoprecipitation and CDK2 Kinase Assay-Hepatoma cells were cultured for 0, 6, 12, 24, or 30 h in growth media with or without 1 M dexamethasone and then rinsed twice with PBS, harvested, and stored as dry pellets at Ϫ70°C. For the immunoprecipitations, cells were lysed for 15 min in immunoprecipitation buffer (50 mM Tris-HCl, pH 7.4, 200 mM NaCl, 0.1% Triton X-100) containing protease and phosphatase inhibitors (50 g/ml phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 5 g/ml leupeptin, 0.1 g/ml NaF, 10 g/ml ␤-glycerophosphate, and 0.1 mM sodium orthovanadate). After centrifugation at 1.4 ϫ 10 4 ϫ g for 10 min at 4°C, the total protein in each sample was determined by Bradford assay. Samples were diluted to 400 g of protein/ml of immunoprecipitation buffer. Samples were pre-cleared for 2 h at 4°C with 40 l of protein A-Sepharose beads (Pharmacia Biotech, Sweden) and 1 g of rabbit IgG. After a brief centrifugation to remove pre-cleared beads, 0.3 g of anti-CDK2 antibody was added to each sample and incubated on a rocking platform at 4°C for 2 h. Subsequently, 10 l of protein A-Sepharose beads were added to each sample, and the slurries were incubated on the rocking platform at 4°C for 30 min. The beads were then washed five times with ice-cold IP buffer and twice with ice-cold kinase buffer (50 mM HEPES, 10 mM MgCl 2 , 5 mM MnCl 2 , 0.1 g/ml NaF, 10 g/ml ␤-glycerophosphate, and 0.1 mM sodium orthovanadate).
For the kinase assay, each immunoprecipitated sample was resuspended in 25 l of kinase buffer containing 20 mM ATP, 5 mM dithiothreitol, 0.18 g of Rb carboxyl-terminal domain protein substrate (Santa Cruz Biotechnology), and 10 Ci of [␥-32 P]dATP (6000 Ci/mmol). Reactions were incubated for 15 min at 30°C and stopped by adding an equal volume of 2 ϫ loading buffer (10% glycerol, 5% ␤-mercaptoethanol, 3% SDS, 62.5 mM Tris-HCl, pH 6.8, and bromphenol blue). Reaction products were boiled for 10 min and then electrophoretically fractionated in SDS-10% polyacrylamide gels. Gels were stained with Coomassie Blue to monitor loading and de-stained overnight with 3% glycerol. Subsequently, gels were dried and visualized by autoradiography.
Transfection Procedures-Logarithmically growing hepatoma cells were transfected by electroporation as described previously (48). Single cell suspensions were washed twice with sterile PBS and resuspended in electroporation buffer (270 mM sucrose, 7 mM sodium phosphate buffer, pH 7.4, 1 mM MgCl 2 ). Cells (400 l; 1-2 ϫ 10 7 cells/sample) were dispensed into sterile cuvettes. In all transfection experiments, the cells and 14 g of pBLCAT2 empty vector and 16 g of p21 promoter-CAT (or p21 promoter-luciferase) reporter construct or 30 g of expression vector DNA were gently mixed, electrically pulsed five times (400 V square wave pulse for 99 s) using a BTX 800 Transfector apparatus (BTX Inc., San Diego, CA), and incubated on ice for 10 min. Transfected cells were plated into pre-warmed Dulbecco's modified Eagle's medium/F12, 10% fetal bovine serum in 100-mm tissue culture dishes and propagated at 37°C. Twenty-four-hour intervals after transfection, cells were re-incubated with fresh medium with or without 1 M dexamethasone or 2 ng/ml TGF-␤ and harvested after 48 h.
CAT Assays-For CAT assays, cells were harvested, washed twice in PBS, resuspended in 100 mM Tris-HCl, pH 7.8, and lysed by alternating between an ethanol/dry ice bath and a 37°C water bath, 5 min per cycle. Cell lysates were heated at 68°C for 15 min, centrifuged at 1.4 ϫ 10 4 ϫ g for 10 min, and the supernatant fractions were recovered. CAT activity in the cell extracts containing 20 -50 g of lysate protein was measured by a quantitative non-chromatographic assay (49). The enzyme assay was carried out in 100 mM Tris-HCl, pH 7.8, 1 mM aqueous chloramphenicol, and 1 Ci of [ 3 H]acetyl coenzyme A (final reaction volume of 250 l). The reaction mixture was gently overlaid with 4 ml of Econofluor water immiscible scintillation fluorochrome (NEN Life Science Products). CAT activity was monitored by direct measurement of radioactivity by liquid scintillation counting. The enzyme activity was expressed as a function of 3 H-acetylated chloramphenicol produced per g of protein present in corresponding cell lysates. For each assay procedure, reaction mixtures were incubated at 37°C for 3-8 h. Mock transfected cells were used to establish basal level activity for both assays.
Luciferase Assays-For luciferase assays, cells were harvested by washing twice in PBS and lysed in 1 ml of Promega lysis buffer. Construction of p21 Promoter-CAT Reporter Genes-A p21-CAT reporter plasmid containing 2.326 kb of the human p21 promoter sequence upstream of the RNA start site was a gift from B. Vogelstein, and has been previously described (41). For fine mapping, wild type internal promoter fragments were cloned from the 2.326-kb p21 promoter using a PCR cloning strategy, into a BamHI/HinD III site on the pBLCAT2 vector which contains a thymidine kinase (tk) minimal promoter. Each PCR reaction contained 1 unit of Taq polymerase, Taq polymerase buffer (10 mM Tris-HCl, 50 mM KCl, and 2.5 mM MgCl 2 , pH 8.3), 0.1 unit of Perfect Match enzyme, and 20 M primers specifically designed to amplify each internal fragment. The 297-bp fragment Ϫ1481/Ϫ1184 was amplified using the following: 5ЈTAATTAAGCTTC-TGTGTCCTCCCACCCCT3Ј and 5ЈAATTTGGATCCATCTACCTCACA-CCCCTGAC3Ј; the 199-bp fragment Ϫ1383/Ϫ1184 with 5ЈACTGGAA-GCTTGCATGTCTGGGCAGAGATTT3Ј and 5ЈAATTTGGATCCATCT-ACCTCACACCCCTGAC3Ј; the 100-bp fragment Ϫ1481/Ϫ1381 with 5ЈTAATTAAGCTTCTGTGTCCTCCCACCCCT3Ј and 5ЈTTATTGGATC-CCCCAGTCTTCTTCCTCTAAC3Ј; the 60-bp fragment Ϫ1441/Ϫ1381 with 5ЈTTAGGAAGCTTCAGAGGAGAAAGAAGCCT3Ј and 5ЈTTATT-GGATCCCCCAGTCTTCTTCCTCTAAC3Ј; the 50-bp fragment Ϫ1481/Ϫ1431 with 5ЈTAATTAAGCTTCTGTGTCCTCCCACCCCT3Ј and 5ЈGCAAAGGATCCTTCTCCTCTGCTGTGGGGAT3Ј; and the 50-bp fragment Ϫ1383/Ϫ1333 with 5ЈACTGGAAGCTTGCATGTCTGG-GCAGAGATTT3Ј and 5ЈGCAAAGGATCCTACTGACATCTCAGGCTG-3Ј. Fragments were then cloned into the plasmid pTK-CAT by ligation with the Takara DNA ligation kit (Panvera, Madison, WI). To verify each step in the cloning procedure, BamHI/HindIII fragments larger than 60 bp were resolved on 1% agarose gels and visualized by ethidium bromide staining. BamHI/HindIII fragments between 50 and 60 bp in length were end-labeled using the Random Primed Labeling Kit (Boehringer Mannheim) for 30 min at room temperature and then separated on a 10% native polyacrylamide gel in 0.15 M NaCl, 0.017 M sodium citrate and exposed to autoradiography film for 30 min. All constructs were subsequently confirmed by DNA sequencing.
Cell Foci Assay-To determine the effects of p21 overexpression on hepatoma cell growth, a cell foci assay was utilized as described previously (48). BDS1 cells were cotransfected with 29 g of the p21 expression vector, along with 2.7 g of pCNot which carries the neomycin resistance gene. A second set of hepatoma cells was also cotransfected with a control vector and the pCNot neomycin resistance plasmid. Forty-eight hours after transfection, cells were washed twice with PBS and propagated with fresh medium containing G418 (400 g/ml). Transfected cells were grown under G418 selection for 2 weeks. Single cell suspensions were made by trypsinization, and 10 4 cells were replated on 100-mm tissue culture dishes. The transfected cells were cultured for 2 weeks in medium supplemented with G418 (200 g/ml) in the presence or absence of dexamethasone. Cells were washed twice with PBS, fixed, and stained with 10% formalin, 0.5% crystal violet. The foci area within each plate was quantitated using the NIH image program after scanning each culture of fixed hepatoma cells with a UMAX UC630 scanner.

Dexamethasone Stimulates p21 CDK Inhibitor Protein Levels and Reduces CDK Activity in Growth Suppressible BDS1
Hepatoma Cells-We have previously shown that glucocorticoids induce a G 1 block in cell cycle progression in BDS1 rat hepatoma cells (14,48) which suggested that this steroid may selectively regulate the expression of G 1 -acting cell cycle components. As an initial test of the kinetic relationship between the expression of cell cycle components and the glucocorticoid growth arrest of an asynchronously growing population of BDS1 hepatoma cells, the rate of DNA synthesis and the production of cell cycle proteins were monitored over a 24-h time course of dexamethasone treatment. Western blot analysis of total hepatoma cell extracts revealed that dexamethasone strongly stimulated the level of the p21 CDK inhibitor protein but had no effect on p27 protein levels, a member of the same family of CDK inhibitors (Fig. 1, upper panels). Dexamethasone induced a relatively rapid increase in p21 protein production that peaked between 10 and 12 h. After 24 h of steroid treatment, the level of p21 protein still remained 2-fold above basal levels (Fig. 1, lower panel). In contrast to p21 production, the protein levels of p27, CDK2, CDK4, CDK6, and cyclin D1, other G 1 -acting cell cycle components, remained unchanged after glucocorticoid treatment (Fig. 1). Time course analysis of [ 3 H]thymidine incorporation in the asynchronously growing population of hepatoma cells revealed that the rise in p21 protein approximately coincided with the dexamethasone-mediated inhibition of DNA synthesis during the first 12 h of steroid treatment (Fig. 1, lower panel). The peak of the dexamethasone induction of p21 protein levels occurred prior to the observed maximal inhibition of [ 3 H]thymidine incorporation which suggests a causal relationship between p21 production and the glucocorticoid-regulated growth arrest of hepatoma cells.
The p21 protein is known to interact with and inhibit each of the G 1 -acting cyclin-CDK complexes (23,29). To functionally test if the glucocorticoid-mediated increase in p21 protein had an effect on CDK activity, the ability of immunoprecipitated CDK2 to phosphorylate the retinoblastoma protein (Rb) in vitro was examined throughout a 24-h time course of dexamethasone treatment. As shown in Fig. 2, the level of CDK2-mediated Rb phosphorylation activity was strongly inhibited by dexamethasone in a manner consistent with the stimulation in p21 protein levels. Moreover, a greater amount of p21 protein co-immunoprecipitated with CDK2 from dexamethasone-treated cells as compared with untreated cells (data not shown). After 24 h in dexamethasone, virtually no CDK2 activity was observed. Thus, the glucocorticoid stimulation of p21 protein expression results in an inhibition of the enzymatic activity of CDK2, which is likely to play a role in the G 1 cell cycle arrest observed in hepatoma cells.
Dexamethasone Stimulation of p21 Transcript Levels Does Not Require de Novo Protein Synthesis-The relatively rapid increase in p21 protein production observed after dexamethasone treatment suggested that this glucocorticoid response is due to the increased expression of p21 mRNA. To determine if glucocorticoids stimulate p21 transcript levels, poly(A) ϩ RNA was isolated from hepatoma cells treated with or without dexamethasone for 4 h, and Northern blots were hybridized with a p21-specific cDNA probe. As shown in Fig. 3, glucocorticoidtreated cells express significantly higher levels of p21 transcripts than untreated cells. GAPDH transcripts remained unchanged by dexamethasone and represent a gel loading control. Thus, the glucocorticoid stimulation of p21 protein production can be accounted for by a corresponding induction of p21 transcripts. Poly(A) ϩ RNA was also isolated from a parallel set of glucocorticoid-treated and untreated cells incubated with the protein synthesis inhibitor, cycloheximide. As shown in Fig. 3, the glucocorticoid induction of p21 transcripts occurred in the absence of de novo protein synthesis, suggesting that the regulation of p21 gene expression is mediated, in part, by one or more pre-existing cellular components that are likely to be involved in the transcriptional control of this gene.
Dexamethasone Stimulates p21 Promoter Activity in a p53 Independent Manner That Requires the Presence of the Glucocorticoid Receptor-The kinetics of dexamethasone-induced p21 transcripts and protein in BDS1 hepatoma cells suggested that this response is mediated by the transcriptional activation of the p21 gene. However, the p21 promoter does not contain a canonical glucocorticoid response element (44), indicating that the receptor signaling pathway likely regulates the activity and/or expression of one or more transcription factors that control expression of this CDK inhibitor gene. It is well established that one such transcriptional regulator of the p21 promoter is the p53 tumor suppressor protein which can cause a

FIG. 2. Dexamethasone inhibition of CDK2 kinase activity.
Cells were cultured with or without 1 M dexamethasone, and at the indicated times, CDK2 was immunoprecipitated from cell lysates and assayed for in vitro kinase activity using the carboxyl terminus of the Rb protein as a substrate. One control immunoprecipitation contained rabbit anti-IgG with no added anti-CDK2 antibodies (No IP). The kinase reaction mixtures were electrophoretically fractionated, and the level of [ 32 P]Rb analyzed by autoradiography. G 1 cell cycle arrest of certain tumor cells (21,39,50). Moreover, we have recently uncovered a functional interaction between p53 and glucocorticoid receptor signaling in certain epithelial cell types (51). Thus, it was important to determine if glucocorticoids can stimulate p21 promoter activity and if this response can occur independently of the key p53 DNA element. To examine directly these possibilities, reporter plasmids were utilized that contain fragments of the p21 promoter linked to the firefly luciferase reporter gene. The Ϫ2.4-kb p21 promoter fragment contains the functional p53 DNA-binding site at Ϫ2.28 kb, whereas the other p21 promoter fragment is a 5Ј deletion ending at Ϫ2.256 kb (Fig. 4, diagram) and is not responsive to the transcriptional effects of p53 (40). BDS1 hepatoma cells were transiently transfected with each reporter plasmid, treated with or without dexamethasone for 48 h, and assayed for luciferase specific activity. As shown in Fig. 4, each p21 promoter fragment conferred glucocorticoid responsiveness to the luciferase reporter gene following dexamethasone stimulation. This result demonstrates that the p21 promoter is a transcriptional target of glucocorticoid receptor signaling and that regulation occurs independently of the critical p53 consensus sequence.
To test if the dexamethasone stimulation of p21 promoter activity required the presence of functioning glucocorticoid receptors, the wild type Ϫ2.4-kb p21-luciferase reporter plasmid was transfected into glucocorticoid receptor-containing (GRϩ) BDS1 hepatoma cells or into glucocorticoid receptor-deficient (GRϪ) EDR3 hepatoma cells originally selected for their resistance to the glucocorticoid growth arrest (12). Treatment with dexamethasone failed to induce p21 promoter activity in GRdeficient EDR3 hepatoma cells under conditions in which this promoter was inducible in GR-positive BDS1 cells (Fig. 4, top  panels), thereby demonstrating the receptor dependence of this response. As controls, parallel cultures of either GR-containing BDS1 cells or GR-deficient EDR3 cells were either transiently transfected with the pGRE-CAT reporter plasmid, which contains six copies of the consensus glucocorticoid response element linked to the bacterial chloramphenicol acetyltransferase (CAT) reporter gene, or examined for the dexamethasone inhibition of [ 3 H]thymidine incorporation. As expected, dexamethasone strongly stimulated CAT activity 50 -100-fold and inhibited DNA synthesis of GR-containing BDS1 cells but had no effect on either response in the GR-deficient EDR3 cells (Fig. 5,  middle and lower panels). This observation implies that the glucocorticoid induction of p21 promoter is a biologically significant receptor-dependent response associated with the growth regulation of hepatoma cells.
Activation of the p21 Promoter Is Specific for the Glucocorticoid Cell Cycle Arrest of Hepatoma Cells-BDS1 hepatoma cells can be growth-inhibited by either glucocorticoids or by TGF-␤. This phenotype allowed us to test if the regulation of p21 promoter activity is a specific glucocorticoid receptor response or generally reflects the growth-suppressed state of the cells. Parallel BDS1 cell cultures were tested for the stimulation of p21 promoter activity and regulation of DNA content by flow cytometry after treatment with either dexamethasone or TGF-␤. To monitor p21 promoter activity, cells were transiently transfected with the Ϫ2.326-kb p21-CAT reporter plasmid that contains a fragment of the p21 promoter (Ϫ2326 to ϩ55) linked to the CAT reporter gene. Cells were treated with dexamethasone or TGF-␤ for 48 h, and cell lysates were assayed for CAT specific activity. As shown in Fig. 6, treatment with dexamethasone, but not with TGF-␤, stimulated p21 promoter activity. Importantly, flow analysis of nuclear DNA content after propidium iodide staining confirmed that both dexamethasone and TGF-␤ induced a G 1 block in cell cycle progression under the conditions of our assay (Fig. 6, right  panels). Thus, even though both glucocorticoids and TGF-␤ can induce a G 1 cell cycle arrest, the regulation of p21 gene expression is specific for the glucocorticoid growth suppression pathway in BDS1 hepatoma cells.
Identification of the Glucocorticoid Responsive Region of the p21 Promoter by Deletion Analysis-Sequence analysis of the p21 promoter revealed no obvious canonical glucocorticoid response elements. To determine which region within the p21 promoter is responsible for the glucocorticoid-mediated transcriptional activation, BDS1 hepatoma cells were transfected with a series of p21-CAT reporter genes containing a series of 5Ј deletions starting at Ϫ2326 bp upstream of the p21 gene and terminating at ϩ55 within the p21 gene. BDS1 cells were treated with or without 1 M dexamethasone for 48 h and cell FIG. 3. Effects of dexamethasone on the expression of p21 transcripts in hepatoma cells treated with or without cycloheximide. BDS1 hepatoma cells were treated with or without 1 M dexamethasone (Dex) in the presence or in the absence of 10 g/ml cycloheximide (Chx) for 8 h. The isolated poly(A) ϩ RNA was electrophoretically fractionated, and Northern blots were probed for p21 transcripts (p21) as described under "Experimental Procedures." As a loading control, the Northern blots were reprobed for GAPDH which is a constitutively expressed transcript. lysates assayed for CAT activity. As shown in Fig. 7, dexamethasone stimulated transcriptional activity of p21-CAT constructs containing the three largest p21 promoter regions with deletions ending at Ϫ2326, Ϫ1892, and Ϫ1481 bp. Maximum stimulation (2-5-fold) of p21-CAT activity by dexamethasone was observed with these constructs. Glucocorticoid inducibility of the p21 promoter was lost in deletions beyond Ϫ1184 bp which indicates that promoter sequences between nucleotides Ϫ1481 and Ϫ1184 of the p21 promoter is the glucocorticoid responsive region. The minor inductions observed with these more extensive deletions were not statistically significant or reproducible between experiments. In addition, the observed variation in basal level of CAT activity was likely to due to the selective elimination of functional transcription factor binding sites depending on the deletion construct or the relative efficiency of transfection.
The Glucocorticoid Responsive Region of the p21 Promoter Alone Can Confer Responsiveness to Dexamethasone and Contains Multiple DNA Elements That Are under Glucocorticoid Control-To functionally test if the 297-bp glucocorticoid responsive region defined by the deletion analysis can confer dexamethasone responsiveness to a heterologous promoter, the Ϫ1.481and Ϫ1.184-bp fragment was linked immediately upstream of the tk minimal promoter sequences driving the bacterial CAT gene (1481/Ϫ1184p21-tkCAT). BDS1 hepatoma cells were transiently transfected with Ϫ1481/Ϫ1184p21-tkCAT, and CAT activity was monitored in cell extracts isolated from dexamethasone-treated and untreated cells. As shown in Fig. 8, dexamethasone induced CAT expression by approximately 3-fold compared with hepatoma cells not treated with dexamethasone. Reporter gene activity in cells transfected with the minimal promoter ptkCAT alone was low and unaffected by dexamethasone treatment (data not shown). These results establish that the glucocorticoid responsive region of the p21 promoter in transfected hepatoma cells is located between Ϫ1481-bp and Ϫ1184-bp upstream of the RNA start site.
To define the minimum number of transcriptional elements under steroid control within the glucocorticoid responsive region of the p21 promoter, smaller promoter fragments from within the Ϫ1481 to Ϫ1184-bp region were generated either through convenient restriction site deletions, PCR cloning, or by oligonucleotide synthesis. These promoter fragments were each cloned into the ptkCAT reporter plasmid and transfected into BDS1 hepatoma cells. CAT activity was subsequently measured in dexamethasone-treated and untreated cells. Initially, two p21 promoter fragments (Ϫ1481 to Ϫ1381 and Ϫ1383 to Ϫ1184) were generated that constitute the entire 297-bp glucocorticoid responsive region. Both reporter plasmids (Ϫ1481/Ϫ1381p21-tkCAT and Ϫ1383/Ϫ1184p21-tkCAT) were glucocorticoid inducible to approximately the same extent as the full 297-bp glucocorticoid responsive region (Fig. 8). The Ϫ1481 to Ϫ1381 p21 promoter fragment on the 5Ј side of the glucocorticoid responsive region was further subdivided into two fragments (Ϫ1481 to Ϫ1431 and Ϫ1441 to Ϫ1381), and the activity of both corresponding reporter plasmids (Ϫ1481/ Ϫ1431p21-tkCAT and Ϫ1441/Ϫ1381p21-tkCAT) was also dexamethasone inducible in transfected hepatoma cells (Fig. 8).
These results indicate the existence of at least two functioning elements within the Ϫ1481 to Ϫ1381-bp fragment of the p21 promoter that are direct or indirect targets of the glucocorticoid receptor signaling pathway which may account for the dexamethasone induction of p21 transcripts in the absence de novo protein synthesis.
The Ϫ1383 to Ϫ1184-bp fragment of the p21 promoter also appears to contain two distinct transcriptional elements under glucocorticoid control. The following paper (82) demonstrates the functional requirement for the C/EBP␣ transcription factor and its corresponding DNA-binding site at Ϫ1270 (see diagram in Fig. 8) in the glucocorticoid stimulation of p21 gene expression. A second regulated element in the Ϫ1383to Ϫ1184-bp fragment of the p21 promoter was identified by transfection of a CAT reporter plasmid containing nucleotides Ϫ1383 to Ϫ1333 bp (forming Ϫ1383/Ϫ1333p21-tkCAT). As also shown in Fig. 8, Ϫ1383/Ϫ1333p21-tkCAT was dexamethasone inducible in transfected BDS1 cells indicating the presence of a regulated element within this 50-bp promoter fragment. Some differences were observed in the transcriptional efficiency of each of the tested p21 promoter fragments which was likely due to the presence or absence of DNA sites for regulatory factors that modulate basal promoter activity. Taken together, these results indicated the presence of at least three distinct glucocorticoid responsive transcriptional elements within the p21 promoter (see map in Fig. 8).
Ectopic Expression of p21 Facilitates the Glucocorticoid Growth Suppression Response-As a functional test for the FIG. 5. Glucocorticoid stimulation of p21 promoter activity requires expression of functional glucocorticoid receptors. Glucocorticoid receptor-positive BDS1 hepatoma cells and receptor-negative EDR3 hepatoma cells were transiently transfected with the Ϫ2.4p21-luc reporter plasmid, treated with or without 1 M dexamethasone for 48 h, and luciferase-specific activity determined as described in Fig. 4. One parallel set of hepatoma cells was transfected with the pGRE-CAT reporter plasmid to assess general glucocorticoid responsiveness. Transfected cells were treated with 1 M dexamethasone for 48 h and assayed for CAT-specific activity as described under "Experimental Procedures." Another set of dexamethasone-treated and untreated cells was examined for DNA synthesis by the incorporation of [ 3 H]thymidine as described under "Experimental Procedures." Given values were calculated as an average of triplicate experiments, and the error bars indicate the standard deviation. potential role of the p21 CDK inhibitor in the glucocorticoid growth suppression response, BDS1 hepatoma cells were cotransfected with either a p21 expression vector or with a vector control and the neomycin resistance expression vector, pBC-MGneo. After selection of positively transfected cells by propagation for 2 weeks in 400 g/ml G418, equal numbers of cells were re-plated in the presence or absence of dexamethasone and assayed for cell growth by a transient cell foci assay. The efficiency of cell foci formation was determined by calculating the average integrated density of cells observed in each condition. Consistent with our previous results (48), treatment with dexamethasone caused a significant reduction in the number of cell foci (Fig. 9). Transfection of the p21 expression vector in the absence of steroid had no apparent effect on the number of hepatoma cell foci. However, expression of p21 significantly reduced the formation of cell foci observed in the presence of glucocorticoids (Fig. 9). Thus, ectopic expression of p21 facilitates the ability of glucocorticoids to inhibit the proliferation of hepatoma cells and suggests that transcriptional control of this CDK inhibitor plays a role in the G 1 cell cycle arrest but that its regulation is not sufficient to mediate the overall growth suppression response. DISCUSSION Our results have established a direct mechanistic link between the glucocorticoid receptor-mediated G 1 cell cycle arrest of hepatoma cells and the regulated promoter activity of the p21 CDK inhibitor gene. The stimulation of p21 promoter activity is a specific glucocorticoid receptor response that accounts for the selective increase in p21 gene expression and inhibition of CDK kinase activity in rat hepatoma cells. This observation indicates that glucocorticoids induce an early G 1 block in cell cycle progression (14,48), in part through the regulation of p21 gene expression. Although the glucocorticoid responsive region of the p21 promoter does not contain a canonical glucocorticoid response element, glucocorticoid responsiveness can be conferred to a heterologous promoter by several closely linked sub-regions of the p21 promoter. We propose that glucocorticoids stimulate the p21 promoter through a combination of two distinct transcriptional mechanisms involving di- rect glucocorticoid receptor-protein interactions with pre-existing promoter-bound transcription factors and indirect glucocorticoid-induced expression of the C/EBP␣ transcription factor that targets the p21 gene (see model in Ref. 82). This complex regulation of p21 transcription in hepatoma cells may reflect the fine tuning of p21 gene transcription that is needed to respond to the multiple regulatory pathways that influence the proliferation and differentiated functions of liver-derived cells (52)(53)(54). Our results suggest that the cell type, environmental and hormonal conditions, and the availability of particular sets of tissue-specific transcriptional regulators and accessory factors influence the precise region of the p21 promoter that is targeted by the glucocorticoid receptor signaling pathway.
Fine mapping studies uncovered several distinct promoter fragments of 50 -60 bp within the glucocorticoid responsive region that can confer steroid responsiveness to a heterologous promoter in transfected hepatoma cells. Consistent with our observation that dexamethasone can induce the level of p21 transcripts in the absence of de novo protein synthesis, each of these glucocorticoid response fragments encodes putative consensus DNA-binding sites for transcription factors that are members of gene families known to interact directly or functionally with the glucocorticoid receptor. These transcription factors include Ets 2, PEA 3, and GATA1 and GATA2 (55)(56)(57). Although mutation of the Ets 2 site located at Ϫ1365 did not abrogate the dexamethasone induction of the promoter fragment, the mutation decreased the basal activity of the promoter 4-fold, 2 which indicates that the Ets 2 transcription factor may be involved in some aspect of the glucocorticoid activation of the p21 promoter. For example, the glucocorticoid receptor could interact with pre-existing factors on this region of the promoter or recruit factors through protein-protein interactions to activate transcription through assembly of a transcriptionally competent complex. We are currently attempting to define each of the transcriptional regulators and to characterize the potential combinatorial effects between these transcription factors and the glucocorticoid receptor that mediate the steroid stimulation of p21 promoter activity and the G 1 cell cycle arrest of hepatoma cells. Our results suggest that glucocorticoids control the cell cycle in these growth-suppressed cells, in part, by multifactorial control of p21 transcription.
Ectopic expression of p21 facilitated the glucocorticoid suppression of cell foci formation in transfected hepatoma cells but had no apparent anti-proliferative effect in the absence of added steroid. This result suggests that the glucocorticoid regulation of p21 gene expression may be important for the cell cycle arrest of hepatoma cells but that p21 alone may not be sufficient to cause the growth arrest. Conceivably, other gene targets of the same glucocorticoid-regulated transcription factors that control p21 gene expression may play an important role in the cell cycle arrest. Our results are consistent with the correlation of the negative regulation of CDK activity by p21 with the growth arrest of many different types of tumorigenic cells (58 -62). Expression of p21 may be important to control the growth and differentiation of certain tumors. For example, the tumor suppressor, BRCA1, requires p21 expression to inhibit S phase progression in human colon cancer cells (63). In addition, the level of p21 protein is reduced in some primary melanomas (64) and is up-regulated during the growth arrest observed during p53-deficient murine erythroleukemia cell differentiation (65).
An emerging concept from numerous studies is that a key cellular strategy for regulation of cell cycle progression, DNA replication, and p53-dependent apoptosis is to alter the expression of p21 (39,40,50). It is well established that the p53 tumor suppressor protein stimulates transcription of p21 through functional p53 response elements in the p21 promoter (40), which may mediate the anti-apoptotic and anti-oncogenic actions of this tumor suppressor protein (40,66,67). Mice homozygously null for p21 develop normally, but the embryonic 2 E. C. Wang, E. J. Cram, and G. L. Firestone, unpublished results.  9. Analysis of the effects of ectopic expression of p21 on the glucocorticoid growth suppression of hepatoma cells by a cell foci assay. BDS1 hepatoma cells were cotransfected with either a p21 expression vector or with vector control DNA and with the pCNot which carries the neomycin resistance gene. Transfected cells were cultured for 2 weeks in G418 and then approximately 10 4 cells were re-plated on 100-mm tissue culture dishes and cultured for 2 weeks in the presence or absence of 1 M dexamethasone (Dex) and G418. Cells were then washed and stained with formalin/crystal violet to visualize the cell foci (A). The foci area within each plate was quantitated using NIH Image as described under "Experimental Procedures." The average integrated densities were determined and illustrated graphically in B. Error bars reflect the standard deviation of three independent integrated density measurements.
fibroblasts of these mice were not able to arrest in G 1 in response to DNA damage, a cellular response mediated by p53 (66). Many transformed cell lines that lack p53 also lack p21 in CDK complexes (68), suggesting that the basal level expression of p21 depends upon p53 function. In contrast, p21 gene expression can be regulated in a p53-independent manner by a variety of regulators in normal and transformed cells (35, 42, 46, 69, 70 -74). Similarly, we found that the p21 promoter from which the primary functional p53 DNA-binding site had been deleted remained dexamethasone responsive in transfected hepatoma cells, which demonstrated that the glucocorticoid stimulation of p21 promoter activity occurs in a p53-independent manner. In addition, the functionally defined glucocorticoid responsive region of the promoter does not include p53-regulated elements.
Given the diversity of signaling pathways influencing p21 transcription (75), control of p21 gene expression may represent a general cellular strategy for regulating p21 function. Our results have firmly established that one such cellular cascade is the glucocorticoid growth suppression response in hepatoma cells. The stimulation of p21 gene expression and promoter activity are relatively specific for this CDK inhibitor because the level of the related p27 protein did not change after dexamethasone treatment. Several other studies have shown the potential involvement of p21 in the receptor-mediated steroid hormone control of cell growth, implicating a transcriptional mechanism. For example, the anti-estrogen (76) or antiprogestin (77) inhibition of cell cycle progression was accompanied by an increase in p21 expression in T47D human breast cancer cells. Moreover, in certain osteosarcoma and fibroblast cell lines, glucocorticoids increased the level of p21 gene products (43,44). In these studies, neither the mechanism of transcriptional control nor the effects of ectopic expression of p21 were evaluated, although a nuclear run-on analysis using fibroblasts showed that glucocorticoids stimulate the rate of transcription of the p21 gene (44). Several reports have established that the p21 promoter is specifically regulated during cellular differentiation. For example, Sp3 is involved in the induction of p21 promoter activity during keratinocyte differentiation (78). Both vitamin D3 and retinoic acid, which act through members of the steroid/thyroid hormone receptor family, induce myeloid cell differentiation and directly stimulate p21 transcription through their corresponding DNA-binding sites in the p21 gene promoter (35,36).
Numerous studies have shown that glucocorticoids regulate gene expression, cellular function, and proliferation of normal and transformed hepatic derived tissue (10,14,15,17,48,79). Our study suggests a role for p21 gene expression in the glucocorticoid-regulated cell cycle control of liver epithelial tumor cells and has provided further evidence for the existence of a steroid-regulated G 1 restriction point in these transformed cells. Conceivably, members of the p21 CDK inhibitor gene family may also play a key tumor suppression role in hepatic derived tissues (80) and during normal liver development (81). We have uncovered a direct mechanistic link between glucocorticoid receptor signaling and the transcriptional control of the p21 CDK inhibitor gene. The following paper (82) describes the functional relationship between the glucocorticoid stimulation of p21 promoter activity and the C/EBP␣ transcription factor, which plays a key role in regulating the tissue-specific functions of the liver (52). Our future efforts will focus on determining the precise receptor-protein interactions and functional role for the p21 promoter-bound factors controlling the steroid responsiveness of this gene and the glucocorticoid-mediated cell cycle arrest of hepatoma cells and perhaps other glucocorticoid-sensitive tumor cells.