Expression of NAG-1, a transforming growth factor-beta superfamily member, by troglitazone requires the early growth response gene EGR-1.

Troglitazone (TGZ) and 15-deoxy-Delta(12,14)-prostaglandin J2 (PGJ2) are peroxisome proliferator-activated receptor-gamma (PPARgamma) ligands that have been shown to possess pro-apoptotic activity in human colon cancer. Although these compounds bind to PPARgamma transcription factors as agonists, emerging evidence suggests that TGZ acts independently of PPARgamma in many functions, including apoptosis. We previously reported that TGZ induces an early growth response transcription factor (EGR-1) by the ERK phosphorylation pathway rather than by the PPARgamma pathway (Baek, S. J., Wilson, L. C., Hsi, L. C., and Eling, T. E. (2003) J. Biol. Chem. 278, 5845-5853). In this report, we show that the expression of the antitumorigenic and/or pro-apoptotic gene NAG-1 (nonsteroidal anti-inflammatory drug-activated gene-1) is induced by TGZ and correlates with EGR-1 induction. In cotransfection and gel shift assays, we show that EGR-1-binding sites are located within region -73 to -51 of the NAG-1 promoter and have an important role in the transactivation of TGZ-induced NAG-1 expression. In contrast, PGJ2 induced NAG-1 protein expression, but PGJ2 may not affect the same region that TGZ does in the NAG-1 promoter. The effect of PGJ2 is probably PPARgamma-dependent because a PPARgamma antagonist inhibited the PGJ2-induced expression of NAG-1. TGZ-induced NAG-1 expression was not inhibited by the PPARgamma antagonist. The fact that TGZ-induced NAG-1 expression was accompanied by the biosynthesis of EGR-1 also suggests that EGR-1 plays a pivotal role in TGZ-induced NAG-1 expression. Our results suggest that EGR-1 induction is a unique property of TGZ, but is independent of PPARgamma activation. The up-regulation of NAG-1 may provide a novel explanation for the antitumorigenic property of TGZ.

The EGR-1 transcription factor (also known as NGFI-A, TIS8, krox-24, and zif268) is a member of the immediate-early gene family and encodes a nuclear phosphoprotein involved in the regulation of cell growth and differentiation in response to signals such as mitogens, growth factors, and stress stimuli. EGR-1 has been proposed as a tumor suppressor gene (26,27). EGR-1 activates the PTEN (phosphatase and tensin homolog) tumor suppressor gene during UV irradiation (28), and reexpression of EGR-1 suppresses the growth of transformed cells both in soft agar and in athymic nude mice (29). EGR-1 induction is both p53-dependent and p53-independent (30 -32). Moreover, EGR-1 is down-regulated in several types of neoplasia as well as in an array of tumor cell lines (33,34). These results suggest that EGR-1 has a role in growth suppression.
The nonsteroidal anti-inflammatory drug-activated gene NAG-1 was identified from an indomethacin-induced gene library (35). NAG-1 (also known as MIC-1, GDF-15, placental transforming growth factor-␤ (TGF-␤), and PLAB) represents a divergent member of the TGF-␤ superfamily. NAG-1 has antitumorigenic and pro-apoptotic activities as assessed by in vivo and in vitro assays (35)(36)(37)(38). The expression of NAG-1 in human colon tissue was seen only in the tips of the villi, where apoptosis occurs (39). Although the expression of NAG-1 is regulated by several nonsteroidal anti-inflammatory drugs independent of cyclooxygenase (40), it is also regulated by several antitumorigenic compounds, including resveratrol (38), genistein (41), and the retinoid 6-(3-(1-adamantyl)-4-hydroxyphenyl)-2-naphthalene carboxylic acid (42). We have previously reported the cloning and characterization of the 3.5-kb NAG-1 promoter (43). Although Sp1 and chicken ovalbumin upstream promoter transcription factor-1 are essential factors in the regulation of the basal level of NAG-1 expression, compoundinduced NAG-1 expression at the transcriptional level has not been fully characterized.
In this study, we examine the relationship between PPAR␥ ligands and NAG-1 expression. PPAR␥ ligands, including TGZ and PGJ 2 , induce NAG-1 expression in human colorectal cancer cells. We found that TGZ induces EGR-1 expression, followed by induction of NAG-1 at the transcription level, whereas PGJ 2 does not induce EGR-1. Rather, NAG-1 seems to be induced by PGJ 2 through the PPAR␥ transcription factor because a PPAR␥ antagonist inhibited NAG-1 expression. EGR-1 induction by TGZ appears to be independent of PPAR␥ because other PPAR␥ ligands did not induce EGR-1, and PPAR␥-binding sites are not located in the TGZ response element in the NAG-1 promoter. These data suggest that the expression of NAG-1 provides a novel mechanism for understanding how TGZ exerts its antitumorigenic activity.
Transfection and Luciferase Assay-HCT-116 cells were plated in 6-well plates at 2 ϫ 10 5 cells/well in McCoy's 5A medium supplemented with 10% fetal bovine serum. After growth for 16 h, plasmid mixtures containing 1 g of NAG-1 promoter linked to luciferase and 0.1 g of pRL-null (Promega, Madison, WI) were transfected with Lipo-fectAMINE (Invitrogen) according to the manufacturer's protocol. For the cotransfection experiment, plasmid mixtures containing 0.5 g of promoter linked to luciferase, 0.5 g of expression vector, and 0.1 g of pRL-null were transfected with LipofectAMINE according to the man-ufacturer's protocol. After 48 h of transfection, the cells were harvested in 1ϫ luciferase lysis buffer, and luciferase activity was determined and normalized to the pRL-null luciferase activity with a dual luciferase assay kit (Promega). For PPAR␥ ligand treatments, the cells were treated with the ligand in the absence of serum for 24 h and then assayed for luciferase activity.
Western Blot Analysis-The level of protein expression was evaluated by Western blot analysis with anti-EGR-1 and anti-NAG-1 antibodies. Cells were grown to 60 -80% confluency in 10-cm plates, followed by 16 h of additional growing in the absence of serum. After treatment with the indicated compounds, total cell lysates were isolated using precipitation assay buffer (1ϫ phosphate-buffered saline, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS). After sonication of samples, lysate proteins (30 g) were separated by SDS-PAGE and transferred for 1 h onto nitrocellulose membrane (Schleicher & Schü ll). The blots were blocked for 1 h with 5% skim milk in Trisbuffered saline and Tween 0.05% and probed with each antibody for 2 h at room temperature. After washing with Tris-buffered saline and Tween 0.05%, the blots were treated with horseradish peroxidaseconjugated secondary antibody for 1 h and washed several times. Proteins were detected by the enhanced chemiluminescence system (Amersham Biosciences).
Preparation of Nuclear Extracts and Electrophoretic Mobility Shift Assay-Nuclear extracts were prepared as described previously (43). For the gel shift assay, double-stranded oligonucleotides (Invitrogen) were end-labeled with [␥-32 P]ATP by T4 polynucleotide kinase (New England Biolabs Inc., Beverly, MA). Assays were performed by incubating 10 g of nuclear extracts in binding buffer (Geneka Biotechnology) containing 2 ϫ 10 5 cpm of labeled probe for 20 min at room temperature. To assure the specific binding of transcription factors, the probe was chased with 1-, 10-, and 50-fold molar excesses of unlabeled wild-type oligonucleotide. For the supershift experiments, anti-EGR-1 antibody (Geneka Biotechnology) was incubated with nuclear extracts on ice for 30 min before addition to the binding reaction. Samples were then electrophoresed on 5% nondenaturing polyacrylamide gels with 0.5ϫ Tris borate/EDTA, and gels were dried and subjected to autoradiography.
RNA Isolation and Northern Blot Analysis-After reaching 60 -80% confluency in 10-cm plates, the cells were treated at the indicated concentrations with PPAR␥ ligands in the absence of serum. For the cycloheximide experiment, the cells were treated with 5 g/ml compound for 30 min prior to TGZ treatment. Total RNAs were isolated with TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Ten g of total RNA was denatured at 55°C for 15 min, separated on a 1.2% agarose gel containing 2.2 M formaldehyde, and then transferred to Hybond-N membrane (Amersham Biosciences). After fixing the membrane by UV, blots were prehybridized in hybridization solution (Rapid-Hyb buffer, Amersham Biosciences) for 1 h at 65°C, followed by hybridization with cDNA labeled with [␣-32 P]dCTP by random primer extension (DECAprimeII kit, Ambion Inc., Austin, TX). The probes used were full-length NAG-1 fragments. After 4 h of incubation at 65°C, the blots were washed once with 2ϫ SSC and 0.1% SDS at room temperature and twice with 0.1ϫ SSC and 0.1% SDS at 65°C. mRNA abundance was estimated from the intensities of the hybridization bands of autoradiographs with a Scion Image (Scion Corp.). Equivalent loading of RNA samples was confirmed by hybridizing the same blot with a 32 P-labeled ␤-actin probe, which recognizes an mRNA of ϳ2 kb.

PPAR␥ Ligands PGJ 2 and TGZ Induce NAG-1 Expression by
Different Pathways-PPAR␥ ligands have an antitumorigenic activity that is either dependent on binding to the ligand to PPRE or independent of PPAR␥ transcriptional binding (25, 44 -46). One mechanism by which PPAR␥ ligands exert antitumorigenesis may involve the transcriptional up-regulation of antitumorigenic proteins. We measured PTEN and p53 tumor suppressor gene expression. The PTEN protein is only marginally induced, whereas the level of p53 is not altered by TGZ in HCT-116 cells (25). Interestingly, NAG-1, which has antitumorigenic activity, was significantly induced by the PPAR␥ ligands. As shown in Fig. 1A, PGJ 2 and TGZ, which are both PPAR␥ ligands, induced NAG-1 mRNA in a concentration-dependent manner (3-fold at 1 and 5 M, respectively). HCT-116 cells were also treated with 1 M PGJ 2 or 5 M TGZ for different times. Both PGJ 2 and TGZ induced NAG-1 protein expression as early as 6 h (Fig. 1B), and a marked increase in NAG-1 was observed at 24 and 48 h, indicating that PGJ 2 and TGZ induce NAG-1 expression in a dose-and time-dependent manner. In addition, the PPAR␣ ligand WY-14643 did not induce NAG-1 expression at concentrations up to 100 M (data not shown), indicating that induction of NAG-1 is specific for this PPAR␥ ligand. We then examined whether NAG-1 induction by PPAR␥ ligands is dependent on the PPAR␥ transcription factor in HCT-116 cells expressing intact PPAR␥ (25). HCT-116 cells were treated with a combination of PPAR␥ ligands and/or GW9662, a selective PPAR␥ inhibitor. Western analyses suggest that the PPAR␥ antagonist suppressed the PGJ 2 -induced NAG-1 expression, but did not suppress TGZ-induced NAG-1 expression. These findings suggest that TGZ-induced NAG-1 expression may be PPAR␥-independent, whereas PGJ 2 increased NAG-1 expression through activation of PPAR␥ (Fig. 2).
NAG-1 Promoter Activity and PPAR␥ Ligands-To evaluate the importance of cis-acting elements in conferring PPAR␥inducible NAG-1 expression, the 3.5-kb NAG-1 promoter and other deletion constructs were transfected into HCT-116 cells and then treated with either PGJ 2 or TGZ. As an internal control, the plasmid pRL-null was used to determine the transfection efficiency. As shown in Fig. 3A, a large increase in luciferase activity was observed after TGZ treatment for all NAG-1 promoter constructs. However, in contrast, an increase in luciferase activity was not observed with PGJ 2 treatment. In fact, the response of the different constructs to PGJ 2 appeared   (38). Total protein (30 g) was subjected to Western analysis as described under "Materials and Methods." to be the same as that to the vehicle. The pGLBasic3 promoterless vector was transfected into HCT-116 cells as a negative control, and no significant luciferase activity was observed with either PGJ 2 or TGZ treatment. These data suggest the presence of a positive TGZ response element in the 3.5-kb NAG-1 promoter, but the absence of a PGJ 2 response element. These results also indicate that TGZ and PGJ 2 induce NAG-1 expression by different mechanisms, which is consistent with the finding from the Western analysis experiment in which we used a PPAR␥ antagonist (Fig. 2). To investigate whether this promoter region is responsive to other ligands, pNAG133/LUCtransfected cells were treated with the ligands all-trans-RA, 9-cis-RA, retinol, rosiglitazone, 13-hydroxyoctadecadienoic acid, PGJ 2 , WY-14643, and TGF-␤1. As shown in Fig. 3B, TGZ increased luciferase activity, but the other ligands did not. This result supports the notion that TGZ increases the transcriptional activity of NAG-1 by a mechanism that does not utilize the nuclear receptor PPAR␥.
TGZ Response Element Is Located in Region Ϫ73 to Ϫ51 of the NAG-1 Promoter-A p53 site that is controlled by several dietary antitumorigenic compounds is present in the NAG-1 promoter at position ϩ43 (38). TGZ is known to induce apoptosis by either a p53-dependent or p53-independent pathway (47,48). To examine the importance of p53 in TGZ-induced NAG-1 expression and to further define the TGZ response element, we generated two constructs containing the p53 site at position ϩ43, pNAG133/ϩ70 and pNAG41/ϩ70. These constructs were transfected into HCT-116 cells, which were treated with vehicle or TGZ. As shown in Fig. 4A, the p53 site is not responsible for TGZ-induced NAG-1 expression because pNAG41/ϩ70transfected cells did not exhibit an increase in luciferase activity in the presence of TGZ. Thus, the TGZ response element is located between positions Ϫ133 and Ϫ41 in the NAG-1 promoter. There are two Sp1 sites (Sp1-B and Sp1-C) in this region of the NAG-1 promoter, which play a pivotal role in basal level expression (43). In addition, these Sp1-B and Sp1-C sites overlap with putative EGR-1-binding sites (Fig. 4B). To demonstrate a functional role for the Sp-1 or EGR-1 site in TGZ- induced NAG-1 expression, we generated point/deletion mutation clones in the Sp1 and EGR-1 sites (Fig. 4B). The analysis of all mutant constructs revealed a dramatic reduction of luciferase activity compared with the wild-type construct, indicating that the TGZ response element may be located in region Ϫ73 to Ϫ51 region of the NAG-1 promoter. Furthermore, both Sp1 and EGR-1 may be involved in TGZ-induced NAG-1 expression.
Sp1 and TGZ Induce NAG-1 Expression-Because region Ϫ73 to Ϫ51 of the promoter contains two Sp1 sites and two EGR-1 sites, the corresponding transcription factors might bind and transactivate the TGZ-induced NAG-1 expression. To evaluate the importance of these sites, Sp1 and EGR-1 expression vectors were generated and cotransfected along with the pNAG133/LUC reporter vector into HCT-116 cells. As shown in Fig. 5A, TGZ treatment of the Sp1-transfected cells did not enhance the induction of luciferase activity compared with the vector-transfected cells. Actually, the TGZ-induced increase in luciferase activity was less in the Sp1-transfected cells than in the vector-transfected cell (3.6-fold versus 2.3-fold, respectively). These data suggest that Sp1 may negatively interfere with TGZ-induced NAG-1 expression in HCT-116 cells. Because Sp1 proteins do not appear to be involved in TGZ-induced NAG-1 expression, we examined the effect of EGR-1 on TGZinduced NAG-1 expression. EGR-1 contains a zinc finger motif that shares a DNA-binding site with Sp1. Indeed, the promoters of many genes contain a GC box, which may interact with Sp1 and EGR-1. To determine whether EGR-1 plays a pivotal role in TGZ-induced NAG-1 expression, we cotransfected an expression vector containing EGR-1. Interestingly, EGR-1 expression with the NAG-1 promoter resulted in an increase in luciferase activity after TGZ treatment (3.6-versus 6.7-fold, respectively) (Fig. 5A). Taken together, these data suggest that Sp1 and EGR-1 compete with each other in the NAG-1 promoter region and that the expression of EGR-1 is critical for TGZ-induced NAG-1 expression. To explore the functional influence of EGR-1 on the NAG-1 promoter and its interplay with Sp1, HCT-116 cells were transiently transfected with expression vectors encoding Sp1 and EGR-1 cDNAs, along with the reporter construct pNAG133/LUC. When increasing amounts of the EGR-1 expression vector were introduced into the HCT-116 cells, a dose-dependent increase was observed in the expression of NAG-1 promoter activity (Fig. 5B). Although both Sp1 and EGR-1 induced NAG-1 promoter activity, the overexpression of EGR-1 resulted in more luciferase activity than observed with Sp1. This result prompted us to clone NAB1 cDNA, a known EGR-1 repressor, and to construct an expression vector. As shown in Fig. 5C, EGR-1-mediated NAG-1 expression was inhibited by NAB1 expression. Indeed, NAB1 expression markedly reduced EGR-1-induced NAG-1 promoter activity, whereas Sp1-induced NAG-1 expression was not affected by NAB1 expression (Fig. 5C).

NAG-1 Induction by TGZ Is Mediated by EGR-1-Because
EGR-1 has an important role in TGZ-induced NAG-1 expression in HCT-116 cells at the transcriptional level, we sought to determine whether EGR-1 expression is altered during TGZ treatment in HCT-116 cells. HCT-116 cells were treated with 5 M TGZ at the indicated times, and cell extracts were analyzed for EGR-1 by Western analysis. EGR-1 expression was increased after 2 h of treatment with TGZ and then decreased after 12 h of treatment (Fig. 6A), whereas Sp1 was not changed during the TGZ treatment (data not shown). NAG-1 expression

FIG. 4. GC box located in the NAG-1 promoter has an important role in TGZ-induced NAG-1 expression.
A, the constructs pNAG133/ϩ70 and pNAG41/ϩ70 were transfected with the pRL-null control vector. Both constructs contain the p53-binding site, but the pNAG41/ϩ70 construct does not contain the two Sp1-binding sites (Sp1-B and Sp1-C). The constructs pNAG133/ϩ70 and pNAG41/ϩ70 (1 g each) were transfected into HCT-116 cells and treated with either vehicle or 5 M TGZ. Transfection efficiency for luciferase activity was normalized to Renilla luciferase (pRL-null vector) activity. B, mutations in the Sp1/EGR-1-binding sites affect the activity of the NAG-1 promoter in the presence of TGZ. The construction of NAG-1 promoter vectors with point or deletion mutations has been described previously (43). Wild-type (pNAG133/ LUC) or Sp1/EGR-1 site mutant reporters (1 g each) and pRL-null (0.1 g) were cotransfected into HCT-116 cells. After 24 h of treatment with either vehicle or TGZ, the cell lysates were isolated, and luciferase (LUC) activity was measured. The x axis shows relative luciferase units (RLU; firefly luciferase/Renilla luciferase). The results are the means Ϯ S.D. of four independent transfections. The point mutations are underlined.
was also increased by TGZ treatment, with the increase in EGR-1 expression by TGZ always preceding that in NAG-1 expression. EGR-1 can be phosphorylated by ERK kinases, but the importance of the phosphorylation in the biological function of EGR-1 is not understood (49). TGZ induces ERK1/2 phosphorylation and activity in these cells, which appear to regulate the expression of EGR-1 by post-transcriptional and posttranslational mechanisms (25). Therefore, the effect of ERK1/2 inhibition on TGZ-induced NAG-1 expression was examined. As shown in Fig. 6B, TGZ did not increase the expression of NAG-1 in the presence of the MAPK inhibitor PD98059. The inhibition of NAG-1 expression was dependent on the concentration of the MAPK inhibitor (Fig. 6B). This result is consistent with a previous report (25) and further indicates that MAPK has a pivotal role in TGZ-stimulated induction of NAG-1. In addition, the pNAG133/LUC construct was cotransfected with the EGR-1 expression vector, and the cells were incubated with TGZ, PD98059, or both. Treatment with the MAPK inhibitor PD98059 resulted in the reduction of NAG-1 promoter activity in the presence of EGR-1 expression (Fig.  6C). These findings suggest that the ERK1/2 kinase may play an important role in the transactivation of EGR-1-induced NAG-1 expression. Finally, to determine whether TGZ-induced NAG-1 expression requires de novo synthesis, HCT-116 cells were pretreated with or without cycloheximide for 30 min, followed by treatment with 5 M TGZ. NAG-1 mRNA was induced by TGZ treatment (Fig. 6D). However, in the presence of cycloheximide, TGZ did not increase the level of NAG-1 mRNA, suggesting that TGZ-induced NAG-1 expression requires de novo protein synthesis. These data are compatible with the notion that the increase in NAG-1 biosynthesis dependent on TGZ requires the de novo synthesis of EGR-1.
EGR-1 Protein Binds to the NAG-1 Promoter-Because TGZ dramatically induces EGR-1 expression prior to NAG-1 induction, the EGR-1 proteins should bind to the NAG-1 promoter. To confirm that the EGR-1 protein binds to such promoter sites, a gel shift assay was performed with [␥-32 P]ATP-radiolabeled oligonucleotide probes corresponding to region Ϫ73 to Ϫ44 of the human NAG-1 promoter (Fig. 7), and nuclear extracts were prepared from either vehicle-or TGZ-treated HCT-116 cells. In addition, unlabeled wild-type oligonucleotides were incubated at 10 and 50 times the concentration to compete FIG. 5. EGR-1 transactivates NAG-1 promoter activity. A, the coexpression of EGR-1 enhances TGZ-induced NAG-1 promoter activity. HCT-116 cells were cotransfected with 0.5 g of the pNAG133/LUC construct and with 0.5 g of empty vector pcDNA3, EGR-1, or Sp1. Cells were treated with either vehicle or TGZ after transfection, and luciferase activity was assayed after 24 h of treatment. As an internal control, the pRL-null vector (0.1 g) was used to correct for transfection efficiency. The results shown are the means Ϯ S.D. of three independent transfections. B, shown is the transactivation of NAG-1 promoter activity by Sp1 and EGR-1. HCT-116 cells were transfected with the pNAG133/LUC construct (1 g) and a combination of EGR-1 and Sp1 at the indicated amounts. The empty pcDNA3.1 vector was used to keep total plasmid DNA constant. Luciferase assay was performed after 48 h, and activity is reported as relative luciferase units (RLU). As an internal control, the pRL-null vector (0.1 g) was used to adjust transfection efficiency. The results shown are the means Ϯ S.D. of three independent transfections. C, NAB1 expression suppresses EGR-1-induced NAG-1 activity. HCT-116 cells were cotransfected with 1 g of the pNAG133/LUC construct and with the indicated amounts of EGR-1, NAB1, Sp1, or empty pcDNA3 vector alone or in combination. The empty pcDNA3.1 vector was used to keep total plasmid DNA constant. Luciferase activity was assayed after 48 h as described under "Materials and Methods." As an internal control, the pRL-null vector (0.1 g) was used to adjust transfection efficiency. The results shown are the means Ϯ S.D. of three independent transfections. with the labeled probe to confirm the specificity of protein binding to this region of the promoter (Fig. 7, lanes 3-5 and 9 -11). The shifted complexes observed resulted from specific binding (Fig. 7, lane 2), which is similar to results shown previously (43). When nuclear extracts from vehicle-treated cells were incubated with the unlabeled oligonucleotide before the addition of the radiolabeled oligonucleotide probe, binding of the radiolabeled probe was reduced with increasing concentrations of the unlabeled wild-type oligonucleotides. Similarly, when nuclear extracts from TGZ-treated cells were incubated with the unlabeled oligonucleotide, shifted bands were competed out with unlabeled oligonucleotides (Fig. 7, lanes 9 -11). These data indicate that the shifted bands represent a specific protein binding to the NAG-1 promoter sequence. In addition, we also performed a gel shift assay in the presence of anti-EGR-1 antibody to demonstrate supershifting. Shifted bands from nuclear extracts from vehicle-treated cells represent only Sp1 family proteins (and not EGR-1 proteins) because anti-EGR-1 antibody did not supershift these bands (Fig. 7, lane 6). However, when nuclear extracts from TGZ-treated cells were mixed with the labeled oligonucleotide, the addition of anti-EGR-1 antibody caused a supershift (SS). Thus, we observed EGR-1 binding to the NAG-1 promoter region only with nuclear extracts from TGZ-treated cells. DISCUSSION In this study, we report for the first time that EGR-1 induction by TGZ results in the increased expression of the nonsteroidal anti-inflammatory drug-activated gene NAG-1 (also known as MIC-1, GDF-15, PLAB, and placental TGF-␤). NAG-1 is associated with pro-apoptosis (36 -38, 50), anti-inflammatory activity (51), and antitumorigenesis (35) in several model systems. Although other PPAR␥ ligands induce NAG-1 expression in HCT-116 cells via PPAR␥ activation, TGZ uses a unique mechanism that requires EGR-1 to mediate NAG-1 expression.
In a previous study, we identified a proximal promoter region spanning positions Ϫ133 to ϩ41 that functions in the basal expression of NAG-1 in HCT-116 cells (43). Binding of the transcription factors Sp1, Sp2, Sp3, and chicken ovalbumin upstream promoter transcription factor-1 to this region is crucial for the regulation of basal level expression (43). To extend our initial studies, we performed, in the present investigation, a detailed functional analysis of the NAG-1 promoter region in the presence of PPAR␥ ligands. A TGZ response element is located between positions Ϫ73 and Ϫ51 in the NAG-1 promoter region. In addition to Sp family proteins, the EGR-1 transcription factor bound to the same region and transactivated the TGZ-induced NAG-1 expression (Figs. 5A and 7). Furthermore, interplay between EGR-1 and Sp1 was required for TGZ responsiveness of NAG-1 promoter activity. As shown in Fig. 7, Sp1 controlled the basal level of NAG-1 expression, whereas EGR-1 fully controlled TGZ-induced NAG-1 expression. Because Sp1 and EGR-1 sites are located in the same region of the NAG-1 promoter, expression of Sp1 might compete with EGR-1 and result in the reduction of TGZ-induced NAG-1 expression. As shown in Fig. 5A, cotransfection of EGR-1 increased NAG-1 promoter activity after TGZ treatment, whereas Sp1 expression resulted in the reduction of NAG-1 promoter activity compared with empty vector transfection. Functional interplay between Sp1 and EGR-1 has been described for a number of human gene promoters (49,(52)(53)(54). The cis-acting element required for Sp1/EGR-1 binding is usually represented by a GC-rich sequence (GC box). In general, EGR-1 does not bind to Sp1 consensus sites, and conversely, Sp1 does not compete with EGR-1 at its recognition motifs; but the presence of overlapping Sp1/EGR-1 sites allows binding and functional interplay of both factors (55,56). For example, Sp1 and EGR-1 interplay has been reported for the human platelet-derived growth factor ␣-chain gene (55). Upon stimulation by phorbol 12-myristate 13-acetate, EGR-1 displaces constitutively bound Sp1 in the promoter and stimulates the transcriptional activity of the platelet-derived growth factor ␣-chain gene. Fig. 8 illustrates a model for EGR-1 regulation of the NAG-1 promoter after TGZ treatment. The increase in ERK1/2 activity that occurs following treatment with TGZ increases the promoter activity and mRNA stability of EGR-1 (25). The increase in EGR-1 protein results in the specific association of this protein with cognate sites in the NAG-1 promoter, which then increases NAG-1 expression. The Sp1 transcription factor competes with EGR-1 at the same site in the promoter. In addition, the NAB1 repressor can also inhibit EGR-1-induced NAG-1 expression by binding to the EGR-1 protein. Overexpression of NAB1 blocks transcription mediated by EGR-1; and, furthermore, NAB1 does not act by blocking DNA binding or nuclear localization of EGR-1 (57). Thus, TGZ-induced NAG-1 expression is regulated in HCT-116 cells by at least three proteins, EGR-1, Sp1, and NAB1 (Fig. 8).
Sp1 protein expression does not change during TGZ treatment in HCT-116 cells, and there is no obvious PPRE with a direct repeat within the Ϫ133 NAG-1 promoter, supporting the conclusion that NAG-1 induction by TGZ is not dependent on PPAR␥ activation or an Sp1 transcription factor. In contrast, PGJ 2 may induce NAG-1 expression by the PPAR␥ transcription factor. The PPAR␥ antagonist GW9662 inhibited PGJ 2induced NAG-1 expression, and cycloheximide treatment of HCT-116 cells did not suppress PGJ 2 -induced NAG-1 expression (data not shown). Taken together, these data suggest that PGJ 2 elicits its effect by a PPAR␥ transcription factor, but a functional PPRE site in the NAG-1 promoter region has not been confirmed.
The EGR-1 transcription factor (also know as NGFI-A, TIS8, krox-24, and zif268) is a member of a transcription factor family that contains three zinc fingers and preferentially binds to the GC-rich DNA core sequence 5Ј-GCGGGGGCG-3Ј; each finger contacts 3 bases within this sequence. However, the sequence 5Ј-TGCGT(G/A)GGCGGT-3Ј has been determined as a high affinity consensus site for EGR-1 (58), indicating that there is a variation in the core sequence. Indeed, the EGR-1 sites in the NAG-1 promoter contain 3 mismatch base pairs compared with the EGR-1 core sequence. These are apparently functional sites because they can bind the EGR-1 protein (Fig.  7). Furthermore, two EGR-1-binding sites may work together to transactivate NAG-1 expression because mutation of one EGR-1 site showed dramatic reduction of NAG-1 promoter activity (Fig. 4B). The exact mechanism of how two EGR-1 sites play a role in TGZ-induced NAG-1 expression is currently under investigation.
TGZ-induced ERK1/2 activity appears to be critical in regulating EGR-1 expression (25). In addition, EGR-1 is a nuclear phosphoprotein (59), but the biological significance of this modification is unknown. Some reports suggest that the phosphorylated forms of EGR-1 are bound to DNA more efficiently than the non-phosphorylated forms (60). Our data suggest that the EGR-1 protein is not only induced after TGZ treatment, but also subsequently highly phosphorylated (25) by the ERK1/2 FIG. 7. EGR-1 preferentially binds to the NAG-1 promoter after TGZ treatment of HCT-116 cells. Gel shift assays were performed with nuclear extracts from either vehicle-or TGZ-treated HCT-116 cells. Ten g of nuclear extracts was incubated with a 32 P-labeled doublestranded oligonucleotide corresponding to region Ϫ73 to Ϫ44 of the human NAG-1 promoter (shown at the top). Competitions were done in the presence of 1-, 10-, and 50-fold molar excesses of non-radiolabeled oligonucleotide corresponding to the wild-type oligonucleotide shown at the top. The binding reactions were resolved by 5% nondenaturing acrylamide electrophoresis. Supershift assays were performed by a 30-min preincubation of the reaction mixture with 2 g of anti-EGR-1 antibody prior to the addition of the radiolabeled probe. Nuclear extracts were obtained from either vehicle-treated (lanes 2-6) or 5 M TGZ-treated (lanes 8 -12) HCT-116 cells. Cold Oligo, unlabeled oligonucleotide; S, shifted bands; SS, supershifted bands.
pathway. The MAPK inhibitors, including PD98059 (MEK inhibitor), attenuated TGZ-induced NAG-1 protein expression and promoter activity (Fig. 6, B and C). These data indicate that MAPK is responsible for the expression and phosphorylation of EGR-1. On the other hand, TGZ-induced ERK1/2 phosphorylation may result in the phosphorylation of PPAR␥, and it has been well documented that the phosphorylation of PPAR␥ results in the inactivation of PPAR␥ activity (61). Other PPAR␥ ligands such as PGJ 2 up-regulate NAG-1 expression by a mechanism dependent on the PPRE. However, this additional activity of the PPAR␥ ligand TGZ to stimulate ERK activity inhibits any increase in NAG-1 expression dependent on PPAR␥. Thus, TGZ-induced expression of NAG-1 is essentially regulated by the increased expression of EGR-1 under these experimental conditions.