Activating Transcription Factor 3 and Early Growth Response 1 Are the Novel Targets of LY294002 in a Phosphatidylinositol 3-Kinase–Independent Pathway

Introduction

Phosphatidylinositol 3-kinase (PI3K) signaling is a well-characterized pathway known to promote cell growth and proliferation and suppresses apoptosis in many types of cancer. Akt, a crucial downstream target of PI3K, is responsible for many of the biological consequences of PI3K activation ( 1). A number of studies have shown that deregulated/inappropriate activation of the PI3K/Akt pathway, achieved through the numerous genetic and epigenetic alterations, contributes substantially to tumorigenesis. Indeed, activated PI3K signaling followed by phosphorylation of Akt leads to progression of colon adenocarcinoma tumorigenesis ( 2). Inhibition of this pathway results in the suppression of cell proliferation and/or the induction of apoptosis in several types of cancer cells. PI3K inhibitors LY294002 and wortmannin have been widely used to study the role of PI3K in cellular responses. LY294002 shows antitumorigenic activity in human colon cancer cells in vivo ( 3) and the inhibition of PI3K by LY294002 can cause induction of apoptosis and cell growth arrest in vitro ( 3, 4). LY294002 treatment induces several proapoptotic genes, including nonsteroidal anti-inflammatory drug-activated gene (NAG-1), in HCT-116 cells ( 4) and wortmannin treatment induces tumor necrosis factor–related apoptosis-inducing ligand expression in HT-29, HCT-116, and Caco-2 colon cancer cells ( 5). In addition, LY294002 enhances sodium butyrate–induced apoptosis and decreases KM20 cell viability ( 6). Thus, therapies using LY294002 may provide promising new alternatives for the treatment of colorectal cancer.

Transcription factors play a key role in controlling cell proliferation, cell cycle progression, and apoptosis and also have been known as potential targets for therapeutic drugs ( 7). Activating transcription factor 3 (ATF3) encodes a member of the ATF/cAMP-responsive element binding protein (CREB) family of transcription factors ( 8, 9). ATF3 has been known as a stress-inducible gene and induction of the ATF3 expression correlates with cellular damage. For instance, ATF3 expression is regulated by a variety of signals, including cytokines, genotoxic agents, and compounds which induce cell death ( 10). Although the physiologic functions of ATF3 are not well understood, there is some evidence that ATF3 functions as a proapoptotic gene in several cancer cells. ATF3 is involved in the stress kinase–mediated signal transduction leading to caspase protease activation ( 11). Hartman et al. ( 12) showed that ATF3 mediates stress-induced apoptosis in β cells. Overexpression of ATF3 enhances the apoptotic effect of curcumin, a polyphenol derived from the turmeric plant, in MDA-1986 cells ( 13). Epicatechin gallate, one of the major catechins in green tea, and indole-3-carbinol induce proapoptotic gene NAG-1 expression mediated by ATF3 in human colorectal cancer cells ( 14, 15). Topoisomerase inhibitor, etoposide- and camptothecin-induced apoptosis, and caspase activity are accelerated by ATF3 overexpression in human epitheloid carcinoma HeLa-S3 cells ( 16). These studies show that ATF3 transcription factor can mediate cytotoxic agent–induced apoptosis in several model systems.

Early growth response 1 (EGR-1, also known as NGFI-A, Zif268, krox24, and TIS8) is another transcription factor family that is regulated by diverse stimuli. Expression of EGR-1 and its role in cancer are complicated and EGR-1 may either inhibit or stimulate growth depending on the cellular context ( 17). It has been shown that activation of the prostaglandin E₂ (PGE₂) receptor (EP4) by PGE₂ induces EGR-1 expression and in turn leads to the expression of tumorigenic gene products such as cyclin D1 and PGE₂ synthase ( 18). EGR-1 can also induce metastasis-related factors in vitro, such as the vascular endothelial growth factor receptor Flt-1 and matrix metalloproteinase ( 19, 20), indicating that EGR-1 may act as a master protein in directing invasion and metastasis during cancer progression. On the other hand, a number of reports also indicate that EGR-1 acts as a tumor suppressor gene. EGR-1 is down-regulated in several types of neoplasia as well as in an array of tumor cell lines. EGR-1 induces cell growth arrest and apoptosis ( 21– 24) and is an important factor involved in neuronal apoptosis ( 25). EGR-1 is induced very early in the apoptotic process where it mediates the activation of downstream regulators such as p53 ( 24). EGR-1 may also activate phosphatase and tensin homologue (PTEN) tumor suppressor gene during UV irradiation ( 26) and suppresses the growth of transformed cells both in soft agar and in athymic nude mice ( 27). Whereas these results indicate that EGR-1 plays a significant role in growth suppression, the consequences of altered EGR-1 expression may be different, depending on cell context. This variety may be dependent on expression of other members of the EGR-1 binding proteins or on the EGR-1 target gene usage in different cells. Therefore, EGR-1 may regulate many genes involved in divergent functions. In this report, we have found that ATF3 is a novel target gene of EGR-1 and mediates LY294002-induced apoptosis in human colorectal cancer cells. These effects are caused by a PI3K/Akt–independent pathway.

Materials and Methods

Cell lines and reagents. Human colorectal cancer cells (HCT-116, SW480, Caco-2, and HT-29 cells) were purchased from American Type Culture Collection (Manassas, VA). Inhibitors or antibodies and their sources were as follows: LY294002 was purchased from Promega (Madison, WI); wortmannin was from MP Biomedicals (Eschwege, Germany); 1L-6-hydroxymethyl-chiro-inositol 2(R)-2-O-methyl-3-O-octadecylcarbonate was from Calbiochem (San Diego, CA); anti-ATF3, EGR-1, and actin antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA); anti–poly(ADP-ribose) polymerase, Akt, and phospho-glycogen synthase kinase 3β antibodies were from Cell Signaling Technology (Beverly, MA); and anti-PTEN antibody was from NeoMarkers (Fremont, CA).

Construction of plasmids. The luciferase constructs containing the ATF3 promoter were previously described ( 28). The putative interleukin-6 response element-binding protein (IL-6 RE-BP) binding site or GC-rich sequence (GC box) was deleted using the QuickChange II site-directed mutagenesis kit (Stratagene, Cedar Creek, TX). To delete these sites on the −84/+34 ATF3 promoter region, the following primers were used: ΔIL-6 RE-BP sense, 5′-gtaagcttgcaacacggagtaaacgaccgc-3′; ΔIL-6 RE-BP antisense, 5′-actccgtgttgcaagcttacttagatcgca-3′; ΔGC box sense, 5′-gagtaaacgaagcctgagggctataaaagg-3′; ΔGC box antisense, 5′-ccctcaggcttcgtttactccgtgttgcca-3′. The deletions were confirmed by DNA sequencing. Plasmid for ATF3 overexpression experiment, pCG-ATF3, was kindly provided by Dr. Tsonwin Hai (Ohio State University, Columbus, OH). The full-length EGR-1, Sp1, and NGFI-A-binding protein (NAB1) cDNA in the pcDNA3 expression vector and the pEBS1⁴ luc construct were previously described ( 29– 31).

Transient transfection and luciferase reporter assays. HCT-116 cells were plated in 12-well plates at 10⁵ per well and were grown for 16 hours. Plasmid mixtures containing ATF3 promoter linked to luciferase and pRL-null (Promega) and/or indicated expression vectors were transfected by LipofectAMINE (Invitrogen, Carlsbad, CA) according to the protocol of the manufacturer. After transfection, the medium was replaced with serum-free medium and the indicated agent was added. Cells were harvested in luciferase lysis buffer and luciferase activity was determined and normalized to the pRL-null luciferase activity using a dual luciferase assay kit (Promega).

Small interfering RNA transfections. ATF3 small interfering RNA (siRNA) was purchased from Santa Cruz Biotechnology. HCT-116 cells were transfected with the ATF3 siRNA at a concentration of 100 nmol/L or with negative control siRNA using TransIT-TKO transfection reagent (Mirus, Madison, WI). After transfection for 24 hours, the medium was replaced with serum-free medium containing vehicle or LY294002. The cells were incubated for 24 hours and total protein was subjected to Western blot analysis.

DNA microarray experiment. Five micrograms of total RNA were used as a template for cDNA synthesis and the cDNA was labeled with biotin-dUTP using AmpoLabeling-LPR kit (SuperArray Bioscience Co., Frederick, MD). The cDNA probe was applied to prehybridized Human cyclic AMP (cAMP)/Ca²⁺ Signaling PathwayFinder Gene Array (HS-028) and Human Signal Transduction in Cancer Gene Array (HS-044) membranes. The hybridization was done at 60°C for 12 hours. After washing, the membranes were blocked and treated with alkaline phosphatase–conjugated streptavidin and finally exposed to CDP-Star alkaline phosphatase chemiluminescent substrate. The membrane was exposed to X-ray film. The intensity of the spots was compared using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as a positive control.

Adenoviral infections. HCT-116 cells were infected with viral vectors as previously described ( 32). Briefly, LacZ, PTEN, and dominant-negative Akt (dnAkt) were transduced at 30, 90, and 30 multiplicity of infection, respectively. After infection for 1 hour, the cells were grown for 24 hours and total cell lysates were harvested for Western blot analysis.

Western blot analysis. Cells were washed once with PBS and lysed with radioimmunoprecipitation assay (RIPA) buffer (1× PBS, 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS) supplemented with protease (1 μg/mL aprotinin, 1 μg/mL leupeptin, and 1 mmol/L phenylmethylsulfonyl fluoride) and phosphatase inhibitors (1 mmol/L Na₃VO₄ and 1 mmol/L NaF). The soluble protein concentrations were determined by bicinchoninic acid protein assay kit (Pierce Biotechnology, Rockford, IL). Proteins were separated by SDS-PAGE and transferred onto nitrocellulose membrane (Osmonics, Minnetonka MN). The membranes were incubated first with primary antibody at 4°C overnight and then with horseradish peroxidase–conjugated secondary antibody for 1 hour. The signal was detected by enhanced chemiluminescence system (Amersham Biosciences, Piscataway, NJ) and quantified by Scion Image Software (Scion Corp. Frederick, MD).

Electrophoretic mobility shift assay. Nuclear extracts were prepared using Nuclear Extract kit (Active Motif, Carlsbad, CA). For the electrophoretic mobility shift assay (EMSA), double-stranded oligonucleotides corresponding to the ATF3 GC-rich sequence (sense, 5′-cgaccgcgccgccagccgaccgcgccgccagc-3′; antisense, 5′-gctggcggcgcggtcggctggcggcgcggtcg-3′) were synthesized and end-labeled with [γ-³²P]ATP by T4 polynucleotide kinase (Promega). Assays were done by incubating 20 μg of nuclear extracts in binding buffer (Promega) containing 2 × 10⁵ cpm of labeled probe for 20 minutes at room temperature. To ensure the specific binding of EGR-1, the recombinant EGR-1 proteins, generated by in vitro translation (TNT Quick Coupled Transcription/Translation Systems, Promega) or obtained from Sigma (St. Louis, MO), were incubated with the labeled probe and 10- or 50-fold molar excesses of unlabeled oligonucleotide. Samples were then electrophoresed on 6% nondenaturing polyacrylamide gels with 0.25× Tris borate/EDTA and gels were dried and subjected to autoradiography.

Immunoprecipitation of EGR-1. HCT-116 cells overexpressed with EGR-1 were grown to subconfluency on a 10-cm dish and then treated with LY294002 for 24 hours. After washing with ice-cold PBS, the cells were scraped with RIPA buffer containing protease and phosphatase inhibitors and mixed at 4°C for 15 minutes. The cell suspension was centrifuged and protein concentration was measured for adjusting protein amount. After preclearing with protein A/G PLUS-agarose (Santa Cruz Biotechnology), the supernatant was incubated with 2 μg of EGR-1 antibody and protein A/G PLUS-agarose overnight at 4°C with rotation. After centrifugation, the immunoprecipitate was washed four times with PBS and RIPA buffer. The pellet was resuspended in sample buffer and subjected to 4% to 12% gradient SDS-PAGE (Invitrogen) followed by Western blotting. Anti–phosphoserine/threonine antibody (BD Biosciences, San Jose, CA) was used at a dilution of 1:2,000.

Soft agar cloning assay. HCT-116 cells were resuspended at 3,000 cells in 2 mL of 0.4% agar in McCoy's 5A medium and plated on top of 1 mL of 0.8% agar in six-well plates. Plates were incubated for 3 weeks at 37°C. Cell colonies were visualized by staining with p-iodonitrotetrazolium violet (Sigma).

Detection of apoptotic cells. The DNA contents for vehicle- and LY294002-treated HCT-116 and HT-29 cells were determined by fluorescence-activated cell sorting (FACS) as previously described ( 4). For detection of apoptosis by ATF3 overexpression, HCT-116 cells were stained with FITC-labeled Annexin V and propidium iodide using Annexin V-FITC apoptosis detection kit (BD Biosciences PharMingen, San Diego, CA) according to the instruction of the manufacturer. Briefly, HCT-116 cells were plated at 3 × 10⁵ per well in six-well plates and transfected with pCG-ATF3 by FuGENE 6 (Roche, Indianapolis, IN) for 24 hours. Subsequently, the cells were grown for 24 hours and stained with Annexin V-FITC and propidium iodide. A total of 10,000 cells were examined by flow cytometry using a Beckman Coulter Epixs XL equipped with ADC and ModFit LT software. Cells were gated on side scatter and forward scatter to exclude debris. Doublets were eliminated using peak versus integral analysis. Annexin V–positive/propidium iodide–positive and Annexin V–positive/propidium iodide–negative cell populations were determined as apoptotic cells from the total gated cells.

Results

LY294002 has antitumorigenic activity and induces ATF3 expression. 2-(4-Morpholinyl)-8-phenyl-4H-1-benzopyran-4-one (LY294002) has been shown to have antitumorigenic activity in human colorectal cancer both in vivo and in vitro ( 3). We have also reported that LY294002 causes cell growth arrest and induction of apoptosis in HCT-116 cells ( 4). To confirm the effect of LY294002 on antitumorigenesis, we did a soft agar cloning assay. As shown in Fig. 1A , a significant reduction in the number of colonies was observed with LY294002 treatment compared with vehicle treatment, suggesting that LY294002 inhibits an anchorage-independent proliferation of HCT-116 cells. Based on our observations and previous reports, LY294002 has an antitumorigenic activity and affects Ca²⁺ signaling ( 3, 4, 33). We conducted a gene expression profiling study to identify novel mediators of LY294002 using DNA microarray membranes designed for signal transduction in cancer and the cAMP/Ca²⁺ signaling pathway. Enhanced expression of several mRNAs was detected in LY294002-treated HCT-116 cells; ATF3, arginine vasopressin, proliferating cell nuclear antigen, and tyrosine hydroxylase with the Human cAMP/Ca²⁺ Signaling PathwayFinder Gene Array membrane ( Fig. 1B); and β-casein and p53 with the Human Signal Transduction in Cancer Gene Array membrane (data not shown). We focused on ATF3 transcription factor because it is one of the genes most strongly induced by LY294002 in HCT-116 cells. The induction of ATF3 transcripts was confirmed by reverse transcription-PCR, showing more than 2-fold induction (data not shown). LY294002-induced ATF3 expression was also confirmed by Western blot analysis. Induction of ATF3 expression was observed at 30 μmol/L LY294002 ( Fig. 1C) and was detected in a time-dependent manner ( Fig. 1D). ATF3 induction by LY294002 (50 μmol/L) was also seen in SW480 and Caco-2 cells but not in HT-29 cells ( Fig. 1E). Consistent with Western blot analysis, LY294002-induced apoptosis in HT-29 cells was very weak compared with that in HCT-116 cells ( Fig. 1F) although LY294002 caused G₁ cell cycle arrest in HT-29 cells (vehicle: 61.2 ± 1.8% versus LY294002: 83.6 ± 0.7%; P < 0.001). These results indicate that ATF3 expression may play an important role in LY294002-induced apoptosis in several types of human colorectal cancer cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Antitumorigenic compound LY294002 induces ATF3 expression in several colorectal cancer cell lines. A, HCT-116 cells were grown in 0.4% soft agar for 3 weeks and stained with p-iodonitrotetrazolium. The colonies were counted from multiple, randomly selected fields. Columns, mean from three replicate experiments; bars, SD. B, HCT-116 cells were incubated for 24 hours with LY294002 at the concentration of 50 μmol/L or vehicle. After isolation of RNA, membrane microarray analysis was done using the GEArray Q Series Human cAMP/Ca²⁺ Signaling PathwayFinder Gene Array. Levels of gene transcripts were normalized to the levels of GAPDH. C to E, HCT-116, SW480, Caco-2, or HT-29 cells were treated with vehicle or LY294002 (10-50 μmol/L). At the indicated times, the cell lysates were harvested for Western blot analysis. Equal loading was confirmed by determining actin immunoreactivity. F, apoptosis was analyzed by FACS using HCT-116 and HT-29 cells treated with LY294002 (50 μmol/L) for 24 hours. Columns, mean fold increase over apoptotic percentage of vehicle-treated cells from three replicate experiments; bars, SD. *, P < 0.05; ***, P < 0.001, versus vehicle treatment (Student's t test).

ATF3 contributes to LY294002-induced apoptosis in HCT-116 cells. Although there is ample evidence that induction of ATF3 expression correlates with cellular damage and/or apoptosis ( 11– 14, 16), ATF3 also has antiapoptotic effects in cardiac myocytes ( 34) and superior nerve ganglion neurons ( 35). These studies indicate that the role of ATF3 may depend on cell context. We used Annexin V staining to assess whether the overexpression of ATF3 results in the induction of apoptosis in human colorectal cancer cells. HCT-116 cells were transfected with ATF3 expression vector and apoptosis was analyzed. Increased ATF3 expression and a significant increase in cells that are undergoing apoptosis were observed in ATF3 overexpressed HCT-116 cells ( Fig. 2A ). To assess the involvement of ATF3 in LY294002-induced apoptosis, we did knockdown of endogenous ATF3 gene by RNA interference and analyzed apoptosis-related protein poly(ADP-ribose) polymerase expression. As shown in Fig. 2B, treatment of control siRNA–transfected cells with LY294002 increased cleaved poly(ADP-ribose) polymerase expression by 4-fold whereas treatment of ATF3 siRNA–transfected cells with LY294002 increased cleaved poly(ADP-ribose) polymerase expression by 2-fold compared with vehicle treatment. These results suggest a linkage between ATF3 expression and LY294002-induced apoptosis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

ATF3 in part mediates LY294002-induced apoptosis. A, HCT-116 cells were transfected with pcDNA3 or pCG-ATF3. The cells were grown for 24 hours and subsequently stained with Annexin V-FITC and propidium iodide and analyzed by flow cytometry. Columns, mean from three independent transfections; bars, SD. Significance for the apoptotic cell populations after transfection with pCG-ATF3 was calculated by Student's t test; **, P < 0.01, versus pcDNA3-transfected cells. B, HCT-116 cells were transfected with control or ATF3 siRNA, followed by LY294002 (50 μmol/L) treatment for 24 hours. The cell lysates were harvested for Western blot analysis. Apoptosis was assessed by cleaved poly(ADP-ribose) polymerase expression. Columns, mean from three independent transfections; bars, SE. ***, P < 0.001, versus control siRNA–transfected cells in the presence of LY294002 (Paired t test).

Evidence for PI3K-independent pathway in LY294002-induced ATF3 expression. Inhibition of PI3K activity by LY294002 and Akt inactivation is likely responsible for the antitumorigenic and apoptotic action of this compound. However, LY294002 affects signaling pathways other than PI3K ( 33, 36, 37). To investigate the molecular mechanism by which LY294002 induces ATF3 expression in a PI3K-independent manner, we treated HCT-116 cells with the PI3K inhibitor wortmannin. As shown in Fig. 3A , wortmannin did not affect ATF3 expression, implying independence of PI3K inhibition in LY294002-induced ATF3 expression. A significant inhibition of PI3K activity by LY294002 and wortmannin was confirmed by dephosphorylation of glycogen synthase kinase 3β, which is a downstream target of the PI3K/Akt pathway. In addition, a specific Akt inhibitor, 1L-6-hydroxymethyl-chiro-inositol 2(R)-2-O-methyl-3-O-octadecylcarbonate, was used to assess the contribution of Akt in LY294002-induced ATF3 expression. Treatment with Akt inhibitor did not increase ATF3 expression ( Fig. 3B). To obtain further evidence, PTEN, which antagonizes PI3K activity ( 38), or dnAkt was overexpressed using adenoviral vectors. Consistently, neither PTEN nor dnAkt affected ATF3 expression compared with LacZ adenovirus control infection ( Fig. 3C), supporting the conclusion that LY294002 causes ATF3 induction in a PI3K/Akt–independent pathway.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

ATF3 induction by LY294002 is independent of PI3K/Akt pathway. A, HCT-116 cells were treated with vehicle, wortmannin (0.5 and 1 μmol/L), or LY294002 (50 μmol/L) for 24 hours. B, HCT-116 cells were treated with vehicle, 1L-6-hydroxymethyl-chiro-inositol 2(R)-2-O-methyl-3-O-octadecylcarbonate, a selective Akt inhibitor (10 μmol/L), or LY294002 (50 μmol/L) for 24 hours. C, adenoviral vectors were used to overexpress PTEN, dnAkt, and LacZ (control). Twenty-four hours after treatment with the indicated inhibitor or infection, the cell lysates were harvested to perform Western blot analysis.

EGR-1 mediates LY294002-induced ATF3 expression at the transcriptional level. To investigate the effects of LY294002 on transcriptional activity of ATF3, ATF3 promoter constructs containing −1850/+34 (pATF3−1850/+34) or −84/+34 (pATF3−84/+34) were transfected into HCT-116 cells and luciferase activity was measured. As shown in Fig. 4A , LY294002 transactivated ATF3 promoter (pATF3−84/+34) in a concentration-dependent manner, up to 50 μmol/L. Moreover, a significant increase of reporter activity in both pATF3−1850/+34 and pATF3−84/+34 constructs was observed in LY294002-treated cells but not in Akt inhibitor–treated cells ( Fig. 4B), consistent with previous data showing that ATF3 is induced by LY294002 but not by Akt inhibitor ( Fig. 3B). Dramatic activation by LY294002 was observed in both ATF3 promoter constructs, indicating that there are positive regulators within −84/+34 bp of ATF3 promoter. The promoter region within −84 bp was further investigated to find any specific cis-acting elements conferring LY294002-induced ATF3 expression. Two different putative binding sites, IL-6 RE-BP and GC-rich sequence, were identified in the −84/+34 promoter region using Transcription Element Search System (http://www.cbil.upenn.edu/tess/). To determine the functional role for these sites in LY294002-induced ATF3 expression, we generated ATF3 promoter internal deletion clones and the reporter activity in the presence of LY294002 was compared with the wild-type construct. As shown in Fig. 4C, the transfection of constructs with deleted putative GC-rich sequence results in a significant reduction in the LY294002-induced promoter activity compared with wild-type pATF3−84/+34 construct whereas deletion of the putative IL-6 RE-BP site did not affect the promoter activity induced by LY294002. Because transcription factors Sp1 and EGR-1 bind to GC-rich sequence and regulate gene expression, we speculate that Sp1 and/or EGR-1 may be involved in LY294002-induced ATF3. Sp1 or EGR-1 expression vector was cotransfected along with the pATF3−84/+34 reporter vector and luciferase activity was measured. Overexpression of EGR-1 in the absence of LY294002 did not affect ATF3 promoter activity whereas Sp1 overexpression showed ∼3.1-fold increase compared with control (pcDNA3), suggesting that Sp1 expression up-regulates ATF3 promoter at the basal level (data not shown). However, LY294002 treatment of the EGR-1-overexpressing cells enhanced the ATF3 promoter activity compared with vector-transfected cells (16.5-fold versus 6.7-fold, respectively). The dramatic increase of luciferase activity in LY294002-treated EGR-1-overexpressing cells was markedly reduced by deletion of GC-rich sequence on pATF3−84/+34 ( Fig. 4D), suggesting that GC-rich sequence is required, at least in part, to mediate EGR-1 action on ATF3 promoter. To confirm the functional influence of EGR-1 on the ATF3 promoter, expression vectors encoding EGR-1 and/or NAB1 cDNAs were transfected with the reporter construct pATF3−84/+34 into HCT-116 cells. NAB1 has been known to block the biological activity of EGR-1 ( 39, 40). As shown in Fig. 4E, NAB1 alone did not affect ATF3 promoter activity; however, NAB1 markedly down-regulated the ATF3 promoter activation induced by EGR-1 in the presence of LY294002. Taken together, these data suggest that EGR-1 contributes to LY294002-induced ATF3 transactivation in human colorectal cancer cells. Finally, we have examined whether LY294002 enhances a reporter gene containing four copies of EGR-1 binding sites. As shown in Fig. 4F, LY294002 increases a luciferase activity in pEBS1⁴ luc–transfected cells. This indicates that LY294002 may not only induce ATF3 promoter but also may induce other promoters containing EGR-1 binding sites.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Overexpression of EGR-1 augments LY294002-induced ATF3 promoter activity. A, pATF3−84/+34 (0.5 μg) was transiently transfected with 0.05 μg of pRL-null vector into HCT-116 cells. Subsequently, the cells were treated with vehicle or LY294002 (10-50 μmol/L). B, pATF3−1850/+34 or pATF3−84/+34 (0.5 μg each) was transfected and the cells were treated with vehicle, Akt inhibitor (10 μmol/L), or LY294002 (50 μmol/L) for 24 hours. C, pATF3−84/+34 or its internal deletion clones (ΔIL-6 RE-BP and ΔGC box; 0.5 μg each) were transfected and the cells were treated with vehicle or LY294002 (50 μmol/L) for 24 hours. D, pATF3−84/+34 or pATF3−84/+34ΔGC box (0.2 μg each) was transfected with pcDNA3 (0.3 μg), Sp1, or EGR-1 (0.15 μg each) into HCT-116 cells and the cells were treated with vehicle or LY294002 (50 μmol/L) for 24 hours. E, pATF3−84/+34 (0.2 μg) were transfected with pcDNA3 (0.3 μg), EGR-1, or NAB1 (0.15 μg each) into HCT-116 cells and the cells were treated with vehicle or LY294002 (50 μmol/L) for 24 hours. F, the pEBS1⁴ luc construct was transfected with pRL-null vector into HCT-116 cells. The cells were incubated for 24 hours with vehicle or LY294002 (50 μmol/L). The promoter activities were measured by luciferase activity. Transfection efficiency for luciferase activity was normalized to the Renilla luciferase (pRL-null vector) activity. Columns, mean from three to six independent transfections; bars, SD. Fold induction over vehicle treatment per each transfection.

EGR-1 binds to ATF3 promoter. To determine whether EGR-1 can interact with the ATF3 promoter, EMSA was done using the oligonucleotides containing two copies of GC-rich sequence on the ATF3 promoter region. Nuclear extracts were purified from either vehicle- or LY294002-treated HCT-116 cells and used for EMSA. First, to identify whether EGR-1 binds to the ATF3 promoter, in vitro synthesized EGR-1 protein was incubated with the probe. As shown in Fig. 5A , EGR-1 can interact with the ATF3 promoter and complex formation was diminished by the addition of 10- or 50-fold molar excess of cold oligonucleotides. Second, to evaluate the effect of LY294002 treatment on the binding affinity of EGR-1 for its DNA recognition element, nuclear extracts from vehicle- or LY294002-treated cells were used. The intensities of the corresponding EGR-1-probe complex were not different between vehicle- and LY294002-treated cells ( Fig. 5A, right) although the protein complexes were competing out with addition of cold oligonucleotide ( Fig. 5A, middle). This suggests that LY294002 may not enhance the EGR-1 binding affinity to the ATF3 promoter. It has been reported that conformational changes such as phosphorylation lead to transactivation of the target gene without increasing its DNA binding affinity ( 41, 42). To analyze the phosphorylation status of EGR-1 in LY294002-treated HCT-116 cells, EGR-1-overexpressing cells were treated with vehicle or LY294002 and immunoprecipitated with EGR-1 antibody, followed by Western blotting with antibody against phosphoserine/threonine or EGR-1. Initially, the same transfection efficiency of EGR-1 was confirmed by Western blot analysis when the cells were overexpressed with EGR-1 (data not shown). As shown in Fig. 5B, highly phosphorylated serine/threonine residues on EGR-1 were detected in the LY294002-treated HCT-116 cells compared with vehicle-treated cells. These results suggest that EGR-1 phosphorylation in LY294002-treated cells at serine/threonine residues may be important for EGR-1-mediated ATF3 expression. Consistent with EMSA data, there is no induction of EGR-1 expression observed in LY294002-treated cells ( Fig. 5C). Therefore, once EGR-1 binds to DNA, the EGR-1 phosphorylation may facilitate transcriptional activation of ATF3 in the presence of LY294002.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Possible role of EGR-1 phosphorylation for LY294002-induced ATF3 expression. A, gel shift assays were done with in vitro translated (IVT) EGR-1 protein or nuclear extracts (NE) from either vehicle (V) or LY294002(L)-treated HCT-116 cells for 24 hours. The nuclear extracts (20 μg) were incubated with a ³²P-labeled double-stranded oligonucleotide corresponding to the GC-rich sequence of ATF3 promoter. Competitions were done in the presence of 10- and 50-fold molar excesses of nonradiolabeled probe. The binding reactions were resolved by 6% nondenaturing acrylamide electrophoresis. B, HCT-116 cells were transfected with EGR-1 expression vector and grown to subconfluency. Twenty-four hours after incubation with either vehicle or LY294002 (50 μmol/L), the cell lysates were harvested. The cell extracts were immunoprecipitated (IP) with antibody against EGR-1 and immunoblotted (IB) with antibodies against phosphoserine/threonine and EGR-1. C, HCT-116 cells were treated with LY294002 (50 μmol/L). At the indicated time points, the cell lysates were prepared for Western blot analysis. HCT-116 cell lysates treated with 30 μmol/L of sulindac sulfide were used as a positive control (C; ref. 54).

Discussion

LY294002 has been widely used as a specific PI3K inhibitor and recognized to have antitumorigenic and proapoptotic properties in several types of cancer cell lines based on the inhibitory effect of PI3K. In this study, we report that ATF3 is up-regulated by LY294002 ( Fig. 1) and mediates apoptosis induced by LY294002 in human colorectal cancer cells ( Fig. 2). Further, LY294002-induced ATF3 expression is partially mediated by EGR-1 transcription factor at the ATF3 promoter region ( Fig. 4). Interestingly, ATF3 expression induced by LY294002 is independent of the PI3K pathway because another PI3K inhibitor (wortmannin) and an Akt inhibitor do not enhance ATF3 expression ( Fig. 3) nor do biological inhibitors of the PI3K/Akt pathway. Although both LY294002 and wortmannin inhibit PI3K activity, they are structurally distinct compounds. A synthetic compound, LY294002, was designed as a PI3K inhibitor based on the flavonoid quercetin ( 43). Wortmannin was originally isolated from Penicillium wortmannii and subsequently identified as a specific PI3K inhibitor ( 44). Unlike wortmannin, LY294002 has been shown to affect other targets. For instance, LY294002, but not wortmannin, causes a decrease of peak Ca²⁺ responses to serotonin in rat airway smooth muscle cells ( 33). LY294002 directly binds to the estrogen receptor and subsequently inhibits 17β-estradiol-induced transcriptional activity of an estrogen-sensitive reporter gene ( 37). LY294002 inhibits monocyte chemoattractant protein-1 expression in a PI3K inhibition–independent manner ( 36). These data support our results, showing that LY294002 has additional targets besides PI3K in human colorectal cancer cells. The physiologic function of ATF3 is likely dependent on cellular context ( 10, 12, 13, 16, 34, 35). We have also suggested that ATF3 plays an important role in mediating apoptotic signals in colorectal cancer cells ( 14). Along with previous reports, we found that LY294002 does not cause a functional induction of apoptosis in HT-29 cells where ATF3 expression is not induced by LY294002 ( Fig. 1F). Furthermore, knockdown of ATF3 gene by siRNA suppressed LY294002-induced apoptosis compared with vehicle ( Fig. 2B). These results indicate that ATF3 expression mediates, at least in part, LY294002-induced apoptosis in colorectal cancer.

To clarify the molecular mechanism by which LY294002 induces ATF3 expression, the promoter activity of ATF3 was assessed by luciferase assay in response to LY294002. We found that LY294002 transactivates ATF3 and the proximal promoter region spanning positions −84 to +34 located LY294002 response elements ( Fig. 4B). Interestingly, deletion of GC-rich sequence on the proximal promoter region resulted in a significant reduction in LY294002-induced ATF3 promoter activity ( Fig. 4C). However, the result of deletion of GC-rich sequence in the ATF3 promoter also implies involvement of additional factors in response to LY294002 although we have shown that EGR-1 plays an important role in LY294002-activated ATF3 promoter. IL-6 RE-BP is not involved in LY294002-induced ATF3 but other cis-acting elements, including activator protein 2α and GATA-1 within −84 bp of ATF3 promoter, have not been considered in this report. Therefore, further experimentation is required to exclude the possibility that other cis-acting elements may be involved in the LY294002-induced ATF3 expression.

It has been reported that functional interplay between Sp1 and EGR-1 contributes to either activation or suppression of their target genes because GC-rich sequence overlaps the DNA sequence for binding of Sp1 and EGR-1 ( 30, 45– 47). Indeed, EGR-1 overexpression enhances the LY294002-induced promoter activity (16.5-fold increase) whereas Sp1 expression prevents the induction (2.6-fold increase; Fig. 4D). In addition, cotransfection with both EGR-1 and Sp1 suppressed the promoting effect of EGR-1 on ATF3 promoter (data not show). Thus, Sp1 transcription factor negatively regulates ATF3 promoter in the presence of LY294002 although Sp1 expression can activate the promoter at the basal levels (data not shown). This is consistent with a previous report that Sp1 plays a pivotal role in the basal transcription machinery ( 29).

EGR-1 contains several serine/threonine phosphorylation sites in its sequences. There is conflicting evidence about a role for phosphorylation of EGR-1 in influencing its DNA binding and transcriptional activities ( 48, 49). In this report, we have determined the EGR-1 phosphorylation status and alteration of its DNA binding affinity by LY294002 because LY294002 does not change EGR-1 protein levels ( Fig. 5C). Highly phosphorylated serine/threonine residues of EGR-1 were observed in the LY294002-treated HCT-116 cells compared with vehicle-treated cells ( Fig. 5B). However, LY294002 treatment did not change the EGR-1 binding affinity to the ATF3 promoter as assessed by EMSA ( Fig. 5A). Phosphorylation has been known to create a high-affinity binding site for other proteins and may allow the high-affinity binding of EGR-1 to the coactivator. In fact, the transcription factor CREB is activated by protein kinase A phosphorylation, and interaction of phosphorylated CREB with coactivators such as CREB-binding protein and p300 modulates cAMP-regulated gene expression ( 50, 51). Richards et al. ( 41) reported that phosphorylation does not change the DNA binding affinity of CREB. As shown in our study, LY294002 alters EGR-1 phosphorylation and this alteration could play an important role for interaction with coactivators rather than increased DNA binding affinity. Indeed, CREB-binding protein and p300 have been reported to interact with EGR-1 to modulate gene transcription ( 52, 53). Casein kinase II has been suggested to phosphorylate EGR-1 although its molecular mechanism has not been elucidated in detail ( 49). We examined whether inhibition of casein kinase II prevents LY294002-induced ATF3 expression. Casein kinase II inhibitor (4,5,6,7-tetrabromo-2-azabenzimidazole) did not affect LY294002-induced ATF3 expression although staurosporine (a broad range kinase inhibitor) did (data not shown). It remains to be elucidated what kinase was activated by LY294002, leading to the phosphorylation of EGR-1.

In conclusion, we have shown that transcription factor ATF3 expression is increased by LY294002 in a PI3K-independent manner. Our results suggest that EGR-1 plays an important role in mediating LY294002-induced ATF3 expression ( Fig. 6 ). Up-regulation of ATF3 may provide a novel explanation for the antitumorigenic properties of LY294002 in colorectal cancer cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Schematic diagram of ATF3 up-regulation by LY294002. In addition to the inhibitory effect on PI3K, LY294002 up-regulates ATF3 in human colorectal cancer cells. Unlike LY294002, wortmannin does not induce ATF3 expression. EGR-1 in part mediates LY294002-induced ATF3 transactivation. ATF3 and EGR-1 are identified as novel target genes for LY294002 and play an important role for the antitumorigenic effect of LY294002 in colorectal cancer cells.