Identification of Nonsteroidal Anti-inflammatory Drug-activated Gene (NAG-1) as a Novel Downstream Target of Phosphatidylinositol 3-Kinase/AKT/GSK-3β Pathway*

The signaling pathway of phosphatidylinositol 3-kinase (PI3K)/AKT, which is involved in cell survival, proliferation, and growth, has become a major focus in targeting cancer therapeutics. Nonsteroidal anti-inflammatory drug-activated gene (NAG-1) was previously identified as a gene induced by several anti-tumorigenic compounds including nonsteroidal anti-inflammatory drugs, peroxisome proliferator-activated receptor γ ligands, and dietary compounds. NAG-1 has been shown to exhibit anti-tumorigenic and/or pro-apoptotic activities in vivo and in vitro. In this report, we showed a PI3K/AKT/glycogen synthase kinase-3β (GSK-3β) pathway regulates NAG-1 expression in human colorectal cancer cells as assessed by the inhibition of PI3K, AKT, and GSK-3β. PI3K inhibition by LY294002 showed an increase in NAG-1 protein and mRNA expression, and 1l-6-hydroxymethyl-chiro-inositol 2(R)-2-O-methyl-3-O-octadecylcarbonate (AKT inhibitor) also induced NAG-1 expression. LY294002 caused increased apoptosis, cell cycle, and cell growth arrest in HCT-116 cells. Inhibition of GSK-3β, which is negatively regulated by AKT, using AR-A014418 and lithium chloride completely abolished LY294002-induced NAG-1 expression as well as the NAG-1 promoter activity. Furthermore, the down-regulation of GSK-3 gene using small interference RNA resulted in a decline of the NAG-1 expression in the presence of LY294002. These data suggest that expression of NAG-1 is regulated by PI3K/AKT/GSK-3β pathway in HCT-116 cells and may provide a further understanding of the important role of PI3K/AKT/GSK-3β pathway in tumorigenesis.

The signaling pathway of phosphatidylinositol 3-kinase (PI3K) 1 /AKT involved in the receptor signal transduction through tyrosine kinase receptors has been one of the major potential targets for novel cancer therapeutics. A number of studies have shown that activation of the PI3K/AKT pathway triggers a cascade of responses involved in cell survival, proliferation, and growth (1)(2)(3)(4)(5). PI3K catalyzes the generation of phosphatidylinositol 3,4,5-trisphosphate, which constitutes a second messenger for activating downstream components such as AKT. Subsequently, the activated AKT causes tumor cell survival/inhibition of apoptosis, induces proliferation and cell growth, and stimulates angiogenesis by phosphorylating numerous downstream targets (1,6). Because AKT delivers antiapoptotic survival signals by phosphorylating the pro-apoptotic protein BAD (7), inhibition of PI3K/AKT signaling using LY294002 can cause apoptosis in various human cancer cells in vitro (8,9). In addition, the anti-tumorigenic activity of LY294002 has been reported in colon cancer cells in vitro as well as in vivo (10). However, the molecular mechanism by which LY294002 induces apoptosis has not been studied in detail.
Glycogen synthase kinase-3 (GSK-3) is a primary target of AKT, which inactivates GSK-3 function by phosphorylation. GSK-3 was initially described as an enzyme involved in glycogen metabolism, but for now it is known to regulate a diverse array of cell functions (11,12). A number of studies have linked GSK-3 to apoptosis and cell proliferation. Overexpression of GSK-3 induces apoptosis in prostate cancer cells (13). Increased cAMP levels promote survival of neuronal cells by inactivating GSK-3␤ via a protein kinase A-dependent mechanism (14). GSK-3␤ inhibits anti-apoptotic molecules, including heat shock factor-1 and the associated expression of heat shock protein (15) which may, in turn, stimulate apoptosis. In contrast, findings in other cell types suggest that GSK-3␤ mediates cell survival. For example, disruption of the murine GSK-3␤ gene caused embryonic lethality because of increased hepatic apoptosis, which may associate with excessive production of tumor necrosis factor (16). Because it has been suggested that PI3K signaling promotes tumorigenesis in human colorectal cancer cells (17), it is of great interest to determine whether GSK-3 counteracts the anti-apoptotic effect of the PI3K pathway and promotes apoptosis in human colorectal cancer cells.
In the present study, we identify NAG-1 as a novel downstream target of the PI3K/AKT/GSK-3␤ pathway. The inhibition of PI3K or AKT, which results in the activation of GSK-3␤, enhanced NAG-1 expression. The down-regulation of GSK-3␤ gene by small interference RNA (siRNA) and GSK-3␤ inhibition using a specific GSK-3␤ inhibitor abolished LY294002 (PI3K inhibitor)-induced NAG-1 expression. Therefore, these data demonstrated that NAG-1 is one of the downstream proteins of the PI3K/AKT/GSK-3␤ pathway, which may explain LY294002-induced apoptosis in human colorectal cancer cells.
Transfection and Luciferase Assay-HCT-116 cells were plated in 12-well plates at 10 5 cells/well in McCoy's 5A medium supplemented with 10% fetal bovine serum. After growth for 16 h, plasmid mixtures containing 0.5 g of NAG-1 linked to luciferase and 0.05 g of pRL-null (Promega, Madison, WI) were transfected by LipofectAMINE (Invitrogen) according to the manufacturer's protocol. After transfection, the media were replaced with serum-free media, and the inhibitors were added. Cells were harvested in 1ϫ luciferase lysis buffer, and luciferase activity was determined and normalized to the pRL-null luciferase activity using a dual luciferase assay kit (Promega, Madison, WI).
Western Blot Analysis-Cells were grown to 60 -80% confluency in 6-cm plates followed by 0 -48 h of treatment in the presence of the indicated inhibitors. Total cell lysates were isolated using RIPA buffer (1ϫ PBS, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS) containing protease inhibitors, and the soluble protein concentrations were determined by BCA protein assay kit (Pierce). Proteins (30 g) were separated by SDS-PAGE and transferred for 1 h onto nitrocellulose membrane. The blots were blocked for 1 h with 5% skim milk in TBS/Tween 0.05% (TBS-T) and probed with each antibody (1:1000 dilution, 5% skim milk in TBS-T) at 4°C overnight. After washing with TBS-T, the blots were treated with horseradish peroxidase-conjugated secondary antibody for 1 h and washed several times. The signal was detected by the enhanced chemiluminescence system (Amersham Biosciences). The contour length of signal on images was measured using the program Scion Image.
Northern Blot Analysis-Total RNA was isolated with Trizol Reagent (Invitrogen), according to the manufacturer's instructions. NAG-1 cDNA was labeled with biotin-N 4 -dCTP using Biotin Random Primer Kit (Pierce). Total RNA (10 g) was separated on 1.2% agarose gels containing formaldehyde and transferred to nylon membranes. Hybrid-ization and chemiluminescent signal detection was performed using North2South Chemiluminescent Hybridization and Detection Kit (Pierce).
RNA Interference-GSK-3␣/␤ siRNA was purchased from Cell Signaling Technology. HCT-116 cells were transfected with the GSK-3␣/␤ siRNA at a concentration of 50 nM or negative control siRNA (Ambion, Austin, TX), using TransIT-TKO transfection reagent (Mirus, Madison, WI). Twenty-four hours after transfection, the medium was replaced with serum-free media containing vehicle or LY294002. The cells were incubated for 24 h and harvested to perform Western blot analysis.
Cell Proliferation Assay-The cell proliferation assay was performed using the CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega, Madison, WI). The assay was carried out according to the manufacturer's protocol. In 96-well plates, cells were plated at 1,000 cells/well in 100 l of media. After 16 h, the cells were treated with LY294002 in the presence of serum and incubated for different time points. Methanethiosulfonate/phenazine methosulfate solution (20 l/ well) was added and incubated for 1 h at 37°C, 5% CO 2 . Absorbance was read at 490 nm using a microplate reader (Universal Microplate Reader, ELX 800, Bio-Tek Instruments, Winooski, VT).
Apoptosis and Cell Cycle Analysis-The DNA contents for vehicleand LY294002-treated HCT-116 cells were determined by fluorescenceactivated cell sorter (FACS). HCT-116 cells were plated at 3 ϫ 10 5 cells/well in 6-well plates, incubated for 16 h, and then treated with LY294002 in the presence of serum. The cells (attached and floating cells) were then harvested, washed with PBS, fixed by the slow addition of cold 70% ethanol to a total of 1 ml, and stored at Ϫ20°C overnight. The fixed cells were pelleted, washed with PBS, and stained in 0.5 ml of 20 g/ml propidium iodide solution. A total of 10,000 cells was examined by flow cytometry using Beckman Coulter Epixs XL equipped with ModFit LT software by gating on an area versus width dot plot to exclude cell debris and cell aggregates. Apoptosis was measured by the level of sub-diploid DNA content. induced NAG-1 expression, and the signal intensities increased more than 2-fold compared with that from vehicle (Fig. 1). Because PI3K is considered to regulate various cellular processes such as apoptosis, proliferation, growth, and cytoskeletal rearrangement (1-5) and LY294002 treatment induces apoptosis, we have further focused on the PI3K pathway and NAG-1 expression.

Effect of Kinase Inhibitors on NAG-1 Expression-In
Enhanced NAG-1 Expression by PI3K and AKT Inhibition-NAG-1 expression was measured after treatment with various concentrations of LY294002 (1-50 M). Expression of NAG-1 was induced by LY294002 in a concentration-dependent manner, with a high increase in expression observed at 50 M ( Fig. 2A). We have also examined time-dependent expression of NAG-1. As shown in Fig. 2B, NAG-1 was induced by LY294002 treatment at 6 -12 h. In addition, GSK-3␤ was de-phosphorylated by LY294002 treatment, which is consistent with previous evidence that PI3K inhibition leads to a decrease in the phosphorylation and hence the activation of GSK-3␤. Although there is strong evidence that PI3K activates AKT by phosphorylation, which is likely responsible for many biological consequences of PI3K activation, PI3K has also been shown to regulate the activation of other cellular targets such as the serum-and glucocorticoidinducible kinase and the small GTP-binding proteins RAC1 and CDC42 in an AKT-independent manner (1). Therefore, we determined whether AKT is the downstream target of PI3K regarding LY294002-induced NAG-1 expression. A selective AKT inhibitor, 1L-6-hydroxymethyl-chiro-inositol 2(R)-2-O-methyl-3-O-octadecylcarbonate was incubated with HCT-116 cells, and NAG-1 expression was measured. As shown in Fig. 2C, NAG-1 was induced in the presence of the AKT inhibitor as well as PI3K inhibitor, indicating that NAG-1 induction by LY294002 may be directly linked to the PI3K/AKT pathway.
Effects of PI3K Inhibition on Cell Growth and Apoptosis in HCT-116 Cells-To investigate the effects of PI3K inhibition on the cell growth of HCT-116 cells, the cells were treated at several concentrations of LY294002. A concentration-dependent inhibition of cell growth was observed, and 50 M LY294002 completely arrested cell growth (Fig. 3A). Subsequently, FACS analysis was performed to determine whether LY294002 affects apoptosis as measured by sub-G 1 population of the cells. FACS analysis revealed that 50 M LY294002 caused the significant increased apoptosis (vehicle, 1.3 Ϯ 0.3%; LY294002, 6.6 Ϯ 1.4%) (Fig. 3B) and G 1 cell cycle arrest (Table I). In addition, the LY294002-induced apoptosis was confirmed by annexin V staining, which can detect early apoptotic cells, and the result showed that LY294002 caused a 3-fold increase compared with vehicle-treated HCT-116 cells (data not shown). These data are consistent with previous reports, which show that LY294002 induced cell growth arrest and apoptosis in the other cell lines (29 -33). Taken together with previous reports, our data indicate the inhibition of PI3K results in growth arrest, apoptosis induction, and G 1 cell cycle arrest in human colorectal cancer cells.
GSK-3␤ Mediates LY294002-induced NAG-1 Expression-Several downstream targets of AKT have been identified including GSK-3␤, mammalian target of rapamycin (mTOR), BAD, IB kinase, and MDM2 (1). To clarify the responsible downstream components of AKT regarding LY294002-induced NAG-1 expression, lithium chloride, a GSK-3 inhibitor, or ra-pamycin, an mTOR inhibitor, was used. HCT-116 cells were treated with lithium chloride in the presence of LY294002 or rapamycin alone because AKT inactivates GSK-3 or activates mTOR by phosphorylation. As shown in Fig. 4A, rapamycin treatment had no effect on the NAG-1 expression; however, lithium chloride suppressed the expression of NAG-1 induced by LY294002 in a concentration-dependent manner. Because lithium chloride is a nonspecific inhibitor for GSK-3␣ and GSK-3␤, isomer-specific inhibitors were used. Additional AR-A014418 (GSK-3␤ specific inhibitor) (34) treatment with LY294002 potently suppressed the LY294002-induced NAG-1 expression, and AR-A014418, at the concentration of 20 -50 M, completely abolished not only the LY294002 induction but also the basal levels of NAG-1 expression (Fig. 4B). Treatment with 50 M AR-A014418 alone slightly decreased the NAG-1 expression (data not shown).
GSK-3␤ Plays an Important Role for LY294002-induced NAG-1 Promoter Activity and Apoptosis-To obtain further evidence that PI3K inhibition induces NAG-1 expression through GSK-3␤, we have examined the NAG-1 promoter activity in the presence of the PI3K inhibitor. The 3.5-kb NAG-1 promoter and the deletion clones were transfected into HCT-116 cells and were then treated with LY294002 (50 M) with or without AR-A014418 (50 M). As shown in Fig. 5A, a significant increase in luciferase activity was observed in treatment with LY294002 in all the NAG-1 promoter constructions (pNAG3500/LUC, pNAG1086/LUC, and pNAG133/LUC). However, the inhibition of GSK-3␤ by AR-A014418 suppressed the luciferase activity to the levels of vehicle-treated cells. These results are consistent with Western blot analysis, shown in Fig. 4B. Based on the result showing that GSK-3␤ inhibition suppresses LY294002induced NAG-1 expression and the promoter activity, we investigated the effects of AR-A014418 on apoptosis. HCT-116 cells were treated with LY294002 with or without AR-A014418, and expression of PARP, which is target of caspases, was analyzed. An increased cleavage of PARP was detected in the cells treated with LY294002, compared with that with vehicle. However, AR-A014418 was able to restore the LY294002-increased cleavage of PARP, which suggests that AR-A014418 treatment could block the LY294002-induced apoptosis in HCT-116 cells (Fig. 5B).  supported by the inhibition of GSK-3␤ expression with siRNA GSK-3␤. This approach permits direct assessment of the GSK-3␤ involvement in the LY294002-induced NAG-1 expression. HCT-116 cells were transiently transfected with a control siRNA and GSK-3␣/␤ siRNA, followed by the treatment with LY294002. As shown in Fig. 6A, transfection with GSK-3␣/␤ siRNA completely knocked down GSK-3␤ expression and decreased the LY294002-induced NAG-1 expression, suggesting that GSK-3␤ plays a pivotal role for LY294002-induced NAG-1 expression. We have shown that GSK-3␤ is an important protein for the LY294002-induced NAG-1 expression as well as apoptosis. To see if there are other mediators involved between GSK-3␤ and the NAG-1 promoter, we first performed the cycloheximide experiment. HCT-116 cells were pretreated with or without 10 g/ml cycloheximide for 30 min, followed by treatment with vehicle or 50 M LY294002. As shown in Fig.  6B, NAG-1 mRNA was induced by LY294002 treatment; however, in the presence of cycloheximide, LY294002 did not increase the levels of NAG-1 mRNA, suggesting that LY294002induced NAG-1 expression requires de novo protein synthesis. Cycloheximide treatment resulted in the induction of NAG-1 mRNA, possibly via accumulation of NAG-1 mRNA.

DISCUSSION
The NSAID-activated gene (NAG-1) has anti-tumorigenic and pro-apoptotic activities, and its expression is induced by many anti-tumorigenic compounds (22)(23)(24)(25)(26)(27)35), suggesting that NAG-1 mediates actions of many anti-cancer compounds. Although direct transcription factors of NAG-1 promoter activity including EGR-1, SP1, p53, and COUP-TF1 have been identified (23, 28, 35), the detailed signaling mechanism affecting NAG-1 expression in cancer cells has not been elucidated. In this report, we have shown, for the first time, that NAG-1 expression can be regulated by GSK-3␤, which provides a novel mechanism of NAG-1 regulation and links the apoptotic activity to NAG-1 expression on circumstantial conditions in human colorectal cancer cells.
The PI3K pathway has been implicated in the apoptosis and cell growth of many cell lines, and inhibition of PI3K exhibited apoptosis induction (2,4,5). Based on previous reports, we have investigated the mechanism of NAG-1 induction by PI3K inhibition. In addition to the concentration-and time-dependent NAG-1 induction by PI3K inhibition, AKT inhibition was also able to induce NAG-1 expression (Fig. 2). These results provide further evidence that AKT is responsible for the action of PI3K on NAG-1 regulation.
LY294002 is well known as a specific PI3K inhibitor, and inactivation of PI3K using LY294002 results in G 1 arrest of glioblastoma cell (29) and choroidal melanoma cell (30) and increased apoptosis in small cell lung cancer cell (31), malignant glioma cell (32), and ovarian cancer cell (33). LY294002 treatment in several colorectal cancer cells resulted in the induction of apoptosis in vitro as well as in vivo (10). As shown in Fig. 3B, FACS analysis demonstrated that LY294002 treatment significantly caused induction of apoptosis in HCT-116 cells. Cell growth arrest was also observed in HCT-116 cells treated with LY294002, suggesting that this growth arrest may contribute to the enhanced apoptosis. These data confirm that the PI3K pathway is closely related in cell survival and proliferation, and its inhibition leads to apoptosis and cell growth arrest in human colorectal cancer.

FIG. 5. GSK-3␤ mediates LY294002-induced NAG-1 promoter activity and apoptosis.
A, the NAG-1 promoter constructs were described previously (28). Each construct (0.5 g) was cotransfected with 0.05 g of pRL-null vector into HCT-116 cells using LipofectAMINE, and then the cells were treated with vehicle, or LY294002 (50 M), and/or AR-A014418 (50 M). After 24 h, the promoter activity was measured by luciferase activity. Transfection efficiency for luciferase activity was normalized to Renilla luciferase (pRL-null vector) activity. Relative luciferase units (RLU) indicate firefly luciferase activity/Renilla luciferase activity. Each value represents mean Ϯ S.D. from six independent transfections. B, HCT-116 cells were treated with LY294002 (50 M) and/or AR-A014418 (50 M) for 24 h. The cell lysates were harvested to perform Western blot analysis. Intact PARP and cleaved PARP appeared around 116 and 89 kDa, respectively. Equal amount of loading proteins was estimated by actin probing.
AKT inhibition was also able to induce NAG-1 expression. AKT is phosphorylated by PI3K, and this event triggers a cascade of responses. Activation of AKT causes tumor cell survival, inhibition of apoptosis, induction of proliferation, cell growth, and stimulation of angiogenesis through downstream targets such as the pro-apoptotic proteins BAD, procaspase-9, and the transcription factors Forkhead, cyclic AMP elementbinding protein, and IB kinase (1,6). In fact, prostaglandin E 2 stimulates proliferation, migration, and invasion of colorectal carcinoma cells by an epidermal growth factor receptor-dependent activation of AKT (36,37). NAG-1 is induced by some cyclooxygenase inhibitors, thereby inhibiting prostaglandin synthesis, suggesting that the expression of NAG-1 regulation by PI3K/AKT pathway may in part be dependent on prostaglandins. The activated AKT negatively regulates biological effects of GSK-3, which contributes to apoptosis and proliferation. Another downstream target of AKT, mTOR, is also known to mediate some of the transforming effects of AKT. Whereas an mTOR inhibitor, rapamycin, did not affect the NAG-1 expression, a specific GSK-3 inhibitor, lithium chloride, abolished the LY294002-enhanced expression of NAG-1 (Fig. 4A). Because apoptosis-related protein tumor necrosis factor-related apoptosis-inducing ligand (38) and p53 (39) are also known to be regulated by GSK-3, GSK-3 may be an important protein in apoptosis.
There are two mammalian GSK-3 isoforms encoded by distinct genes, GSK-3␣ and GSK-3␤ (40). Recent studies have characterized GSK-3␤ as being closely associated with apoptosis. Sanchez et al. (41) reported GSK-3␤ as a downstream target of PI3K-mediated apoptosis through the inhibition of the NF-B pathway in astrocytes. Inhibition of GSK-3␤, which is also involved in Wnt signaling pathway, led to activated ␤-catenin-associated transcription and enhanced survival of neoplastic cells in B cell chronic lymphocytic leukemia (42). Thus, GSK-3␤ promotes apoptosis in different cell types. Indeed, our results showed that GSK-3␤ has priority on NAG-1 regulation, because a specific GSK-3␤ inhibitor, AR-A014418, was completely able to block the induction of NAG-1 by PI3K inhibition. Although AKT phosphorylates a number of downstream targets, this inhibition implies that GSK-3␤ is responsible for NAG-1 regulation through the PI3K/AKT pathway. To confirm this pharmacological approach, the GSK-3 gene was suppressed by GSK-3-specific siRNA. The GSK-3␤ silencing demonstrated that GSK-3␤ is required for LY294002-induced  7. Biochemical pathway leading to induction of NAG-1 expression following PI3K inhibition. Inhibition of PI3K induces NAG-1 expression through the following process. LY294002 exhibits the inhibitory action of PI3K. Subsequently, the suppressed signal inactivates AKT. Because AKT phosphorylates GSK-3␤ and inactivates the ability of its biological effects, the suppressed AKT leads to an increase in GSK-3␤ activity. GSK-3␤ affects NAG-1 promoter through an unknown protein, which is required to be synthesized by LY294002. The enhanced NAG-1 expression may mediate apoptosis and/or anti-tumorigenesis in the HCT-116 cells. NAG-1 expression (Fig. 6A). To evaluate whether inhibition of GSK-3␤ suppresses LY294002-mediated apoptosis, cleavage of PARP was analyzed. AR-A014418 restored the increased cleavage of PARP by LY294002. Therefore, this result indicates a characteristic apoptotic pathway of PI3K/AKT/GSK-3␤ and suggests that GSK-3␤ plays an important role in apoptosis regulated by the PI3K/AKT pathway (Fig. 7).
The NAG-1 promoter constructs were used to evaluate the transcriptional effects of PI3K inhibition. LY294002 showed an increase in luciferase activity, and AR-A014418 restored the upregulation of this promoter. Furthermore, by using deletion clones pNAG1086/LUC and pNAG133/LUC, LY294002 still showed an increased luciferase activity, suggesting that the NAG-1 promoter within Ϫ133 may be important for the LY294002-induced transcriptional activity. However, these effects were abolished when GSK-3␤ was inhibited. As shown in Fig. 2B, the significant induction of NAG-1 appeared at 12 h after LY294002 treatment. This slow response implicates that there must be downstream targets of GSK-3␤, because de-phosphorylation/activation of GSK-3␤ was seen at 1 h after LY294002 treatment (Fig. 2B). Indeed, we found that de novo synthesis is required for LY294002-induced NAG-1 expression as assessed by cycloheximide experiments (Fig. 6B). Results from the promoter assay showed that known transcription factors to bind the NAG-1 promoter, including Sp1, COUP-TF1, and EGR-1, are not responsible for the LY294002-induced NAG-1 expression (data not shown). The molecular mechanisms of GSK-3␤ on NAG-1 expression need further investigation.
In summary, we demonstrated that NAG-1 is regulated by the PI3K/AKT/GSK-3␤ pathway in human colorectal cancer cells. Our finding contributes to the further understanding of the PI3K/AKT/GSK-3␤ pathway in human cancer and identifies NAG-1 as a novel target of GSK-3␤ to facilitate its antitumorigenic activity.