The Inflammatory Mediator Leukotriene D4 Induces β-Catenin Signaling and Its Association with Antiapoptotic Bcl-2 in Intestinal Epithelial Cells*

Increased levels of the inflammatory mediator leukotriene D4 (LTD4) are present at sites of inflammatory bowel disease, and such areas also exhibit an increased risk for subsequent cancer development. It is known that LTD4 affects the expression of many proteins that influence survival and proliferation of intestinal epithelial cells. We demonstrate here that after LTD4 exposure, β-catenin translocates to the nucleus where it signals activation of the TCF/LEF family of transcription factors. These events are mediated via a phosphatidylinositol 3-kinase-dependent phosphorylation of the inhibitory Ser-9 residue of glycogen synthase kinase 3β. We also show that in the presence of LTD4, free β-catenin translocates to the mitochondria where it associates with the cell survival protein Bcl-2. We hypothesize that LTD4 may enhance cell survival via activation of β-catenin signaling, in particular, by promoting the association of β-catenin with Bcl-2 in the mitochondria. Similar to Wnt-1 signaling, LTD4 signals an increased level of free β-catenin and elevated TCF/LEF promotor activity. This work in intestinal epithelial cells further lends credence to the idea that inflammatory signaling pathways are intrinsically linked with potential oncogenic signals involved in cell survival and apoptosis.

Leukotriene D 4 (LTD 4 ) 2 is a powerful proinflammatory mediator, which is formed from arachidonic acid through the action of 5-lipoxygenase (1). Arachidonic acid is the common precursor of a group of mediators, collectively called eicosanoids (2). LTD 4 is known to mediate its effects through specific cell surface receptors belonging to the G protein-coupled receptor family. Two such receptors have been cloned, CysLT 1 (3) and CysLT 2 (4), of which the CysLT 1 receptor has been demonstrated to have a much higher affinity for LTD 4 (4). Earlier results from our group demonstrated that the CysLT 1 receptor signals through at least two different G proteins (5,6).
LTD 4 is associated with the pathogenesis of several inflammatory disorders, such as asthma and inflammatory bowel disease (IBD) (7). The IBDs, including ulcerative colitis and Crohn's disease, are charac-terized by an increased infiltration of inflammatory leukocytes into the intestinal wall where they have the capacity to cause nonspecific tissue injury. In addition, these inflammatory cells, in particular the neutrophils, are a significant source of mediators with important regulatory roles in the inflammatory process. Consequently, the IBD-effected intestine contains a significant source of eicosanoids, readily available for interaction with the CysLT receptors expressed by intestinal epithelial cells (8,9). During the last decade intense research has focused on identifying effective inhibitors of soluble eicosanoid mediators (10 -12). Most clinical trials with such inhibitors, often short term studies, have so far not revealed any advantageous effects on the inflammatory component of IBD compared with presently available treatments (11,12).
Not only is there a well established connection between IBD and increased frequency of neoplastic transformation (13), but a more general link between chronic inflammation and increased risk of developing cancer has been suggested (14). Therefore, the findings from our group that exposure to LTD 4 increases survival and proliferation of intestinal epithelial cells (15) are highly interesting. Moreover, we have shown that LTD 4 causes up-regulation of cyclooxygenase-2 (COX-2), the antiapoptotic protein Bcl-2, and the multifunctional protein ␤-catenin (16). We have also recently shown that the CysLT 1 receptor is up-regulated in colon cancer tissue and that LTD 4 signaling facilitates the survival of cancer cells (8).
␤-Catenin is a protein with many distinct roles. It is found at the cell membrane, where it forms part of the adherens-type junctions, by linking the intracellular domain of the classical type cadherins to the actin cytoskeleton (17,18). It is also an effector molecule of the Wnt signaling pathway. The presence of a Wnt signal allows ␤-catenin to translocate to the nucleus, where it activates transcription together with members of the transcription factor family TCF/LEF (19). Normally, in mature cells the adenomatous polyposis coli (APC) protein maintains intracellular ␤-catenin homeostasis by the formation of a cytosolic complex, which also includes axin and glycogen synthase kinase (GSK-3␤), (20,21). Binding of ␤-catenin to this complex facilitates its phosphorylation at serine residues 33, 37, and 41 by GSK-3␤, leading to subsequent ubiquitination and degradation (22). However, in most cases of colon cancer, the function of the APC protein is altered, the extent of which can vary from incorrect function to total loss of the protein (19,23,24). The uncontained ␤-catenin is free to enter the nucleus and activate transcription of potentially oncogenic target genes (25,26), such as c-myc (27) and cyclin D1 (28). Recently it was shown that Wnt signaling might also be involved in the regulation of cell survival by protecting cells from cytochrome c leakage (29), thus indirectly suggesting a role for ␤-catenin in this process.
Bcl-2 is the founding member of a family of proteins that are regulators of apoptosis. Bcl-2 itself was originally found in lymphoma and is believed to protect cells from programmed cell death (30 -32). This protective action occurs largely through its effect on mitochondrial membrane integrity (33). One way by which Bcl-2 mediates its effect is through altered levels of mitochondrial and cytosolic free calcium (34). A novel functional interaction between Bcl-2 and inositol 1,4,5-triphosphate receptors, with a role in modulating survival of cells, was also recently uncovered (34). In this context it is noteworthy that we found extracellular signal-related kinase 1/2-dependent up-regulation of Bcl-2 in intestinal epithelial cells responding to LTD 4 (35).
The present study was initiated to investigate further our previous finding that LTD 4 induces an increased expression of ␤-catenin and the antiapoptotic protein Bcl-2 in epithelial cells. Our aim was to understand better the link between eicosanoid inflammatory signaling and carcinogenesis.

EXPERIMENTAL PROCEDURES
Materials-Antibodies against ␤-catenin and total GSK-3␤ were purchased from Transduction Laboratories (Lexington, KY). The COX-2 antibody was from Abcam (Cambridge, UK). The phosphospecific GSK-3␤ antibody was purchased from Cell Signaling Technology (Beverly, MA). Lamin B, actin, HA tag, Bcl-2, and Wnt-1 antibodies used for immunoprecipitation and immunoblotting were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The Bcl-2 antibody used for immunostaining and the caspase-3 fluorometric substrate peptide Ac-DEVD-AMC (Ac-Asp-Glu-Val-Asp-7-amino-4-methylcoumarin) were from Upstate Biotechnology (Lake Placid, NY). Protein G-agarose originated from Oncogene Research Products (San Diego, CA). LTD 4 and MK571 were from Cayman Chemical Company (Ann Arbor, MI). The fluorescent-labeled Alexa 488 goat-anti-mouse antibody, the Alexa 594 donkey-anti-goat antibody, Hoechst 34580, and the Mitotracker probe were purchased from Molecular Probes (Eugene, OR). Peroxidase-linked goat anti-mouse antibody and mouse IgG were from Dako A/S (Copenhagen, Denmark). Hyperfilm was from Amersham Biosciences, and Lipofectamine 2000 was purchased from Invitrogen. Pertussis toxin was purchased from Speywood Pharma Ltd. (Maidenhead, UK) and ZM198,615 was a gift from B. Andersson, Global Product Director Pulmicort and Accolate (AstraZeneca R&D, Lund, Sweden). Antibodies against voltage-dependent anion channel, LY294002, and wortmannin were acquired from Calbiochem. All other chemicals were of analytical grade and obtained from Sigma.
Constructs-pEGFP-N 1 was from Clontech (Palo Alto, CA). TOP-Flash and FOPFlash plasmids were from Upstate Biotechnology. The Bcl-2-FLAG plasmid was a kind gift from Prof. Yasuo Uchiyama (Osaka University, Osaka, Japan). The Bcl-2 sequence was further amplified by PCR using sense 5Ј-GGA ACG TCG ACA TGG CGC ACG CTG GGA GAA-3Ј and antisense 5Ј-GGA ACG TCG ACT CAC TTG TGG CTC AGA TAG-3Ј primers. The Bcl-2 sequence was cloned into the pEGFP-N 1 expression vector using SalI, and the correct orientation of the insert was checked using the PstI restriction enzyme. HA-S33Y-␤catenin and HA-wt-␤-catenin constructs were generously provided by Dr. Ben-Zeév (Weizmann Institute of Science, Rehovot, Israel). The constitutively active mutant of GSK-3␤ 9A constructs were generously provided by Dr. Birnbaum (Howard Hughes Medical Institute, University of Pennsylvania).
Cell Culture-Human intestinal epithelial nontransformed (Int 407) cells, which exhibit typical epithelial morphology and growth (36), were cultured as a monolayer to approximately 80% confluence for 5 days in Eagle's basal medium (15). The cells were tested regularly to ensure the absence of mycoplasma contamination.
Cell Lysates and Immunoprecipitation-The cells were cultured for 5 days and serum starved for 2 h before stimulation with 80 nM LTD 4 for 0.5-24 h in the absence of serum. Treatment was terminated by the addition if ice-cold buffer A (9) supplemented with 1% Triton X-100. Flasks were kept on ice for 30 min, and the remaining cell debris was scraped loose into the lysis buffer and homogenized 10 times on ice with a Dounce glass grinder. The cell suspension was cleared by centrifugation at 200 ϫ g for 10 min at 4°C to remove whole cells and cell debris. Finally, the lysates were cleared by centrifugation at 10,000 ϫ g for 15 min at 4°C. The samples were evaluated for, and compensated to, equal protein content. Before immunoprecipitation, membrane fractions were precleared with 1 g of mouse IgG and 15 l of protein G-agarose. The samples were immunoprecipitated with 2 g of ␤-catenin or Bcl-2 antibody for 90 min at 4°C. Thereafter 20 g of protein G-agarose beads were added, and the samples were rotated for an additional 45 min at 4°C. The precipitates were washed multiple times with buffer A.
Isolation of a Nuclear Fraction-Cells were cultured as described above. After the indicated treatments with 80 nM LTD 4 and other supplements, cells were washed twice with buffer A and kept on ice in the same buffer for 10 min. Thereafter, the cells were scraped loose and subjected to nitrogen decompression at 1,000 p.s.i. for 10 min, using a cell disruption bomb (Parr Instrument Company, Moline, IL). The nuclear fractions were collected by centrifugation at 200 ϫ g for 10 min and washed twice with buffer A prior to analysis.
Isolation of a Mitochondrial Fraction-Cells were cultured as described above and left in serum-free culture medium for 2 h prior to initiation of an experiment. After the indicated treatments with 80 nM LTD 4 , the experiments were terminated by the addition of a buffer containing 10 mM NaCl, 1.5 mM MgCl 2 , and 10 mM Tris-HCl (pH 7.5), and the flasks were left on ice for 10 min. The cells were scraped loose, homogenized 10 times (Dounce), and then lysis buffer was added (20 mM EGTA, 1 mM EDTA, 250 mM sucrose, 1 mM Na 3 VO 4 , 4 g/ml leupeptin, and 30 g/ml phenylmethylsulfonyl fluoride). The lysates were cleared by two short centrifugations at 750 ϫϫ g, and then the mitochondria were pelleted by centrifugation at 10,000 ϫ g for 15 min.
Immunofluorescence-The cells were seeded on glass coverslips and grown for 5 days and serum-starved for 2 h before stimulation with 80 nM LTD 4 for 0.5-24 h in the absence of serum. When applicable, for the last 30 min of stimulation 100 nM Mitotracker Red was added. The medium was removed and the dishes washed several times with PBS. Cells were fixed with 4% paraformaldehyde for 15 min and subsequently permeabilized with 0.1% Triton X-100 and PBS or 0.5% Triton X-100 and PBS (␤-catenin) for 5 min. Blocking was carried out in 3% bovine serum albumin in PBS for 30 min, and the primary antibody was diluted 1:50 (Bcl-2), 1:250 (␤-catenin), and 1:500 (Lamin B) in a solution containing 3% bovine serum albumin. Incubation was carried out at room temperature for 60 min. After several washes in PBS, the secondary antibody was used at a 1:500 dilution in 3% bovine serum albumin, and the incubation lasted 1 h at room temperature. This was followed by extensive washing with PBS. Finally, coverslips were mounted on glass slides with fluorescent mounting medium. Confocal images were recorded using a Bio-Rad Radiance 2000 confocal laser scanning system with a Nikon microscope (model TE300) equipped with a 60 ϫ 1.4 Plan-APOCHROMAT oil immersion objective.
Transient Transfections and Luciferase Assays-The luciferase assays were carried out employing the Dual Luciferase Reporter Assay System from Promega (Madison, WI). Each plasmid was used at a final concentration of 1 g/ml, except for the control Renilla luciferase vector, which was present at all times at 0.2 g/ml to standardize transfection efficiency. Vector DNA was allowed to form complexes with Lipofectamine (0.0001 g/ml was used). Meanwhile, 50 -60% confluent cells in 12-well plates were washed once in serum-free medium, and the DNA-Lipofectamine mixture was added. The transfection continued at 37°C for 4 h, after which the medium was changed to normal growth medium, and the cells were allowed to recover for 48 h. Before the addition of 80 nM LTD 4 , cells were left in serum-free medium for 2 h. During this time, when applicable, they were pretreated with 500 ng/ml pertussis toxin for 2 h, 50 M LY294002 for the last 30 min, 20 M ZM198,615, or 25 M MK571 for the last 15 min, or 100 nM wortmannin for the last 10 min. After stimulation with LTD 4 for the indicated periods of time, cells were washed in PBS and lysed using 100 l/well of the DRL passive lysis buffer included in the Dual Luciferase Reporter Assay System. Lysed samples were then collected and briefly centrifuged to precipitate any debris. A 20-l portion of each lysate was transferred to a luminometer test tube predispensed with 50 l of Luciferase Assay Buffer II, and the luciferase reaction was immediately recorded using a MiniLumat LB 9506 (Berthold Technologies, Germany). The control Renilla luciferase signal was recorded after the subsequent addition of 50 l of Stop & Glow buffer, and the level of expression is given as a ratio. Triplicate samples were prepared and analyzed for each condition in every set of experiment.
Assessment of Caspase-3 Activity-Cells, transfected or not with wild type or S33Y mutated ␤-catenin, were cultured in 12-well plates, and during the last 4 h they were exposed or not to 100 M NS-398 in the presence or absence of 80 nM LTD 4 . Thereafter, the cells were lysed on ice for 30 min in 300 l of buffer containing 1% (v/v) Triton X-100, 20 mM Tris-HCl (pH 7.5), and 150 mM NaCl. 50-l samples from such lysates were suspended in 200 l of reaction buffer containing 20 mM HEPES, 2 mM dithiothreitol, and 10% glycerol and added to Nunc Polysorb 96-well white fluorescence measurement plates. 5 l of caspase-3 fluorometric substrate Ac-Asp-Glu-Val-Asp-AMC was subsequently added. The plates were incubated at 37°C for 1 h, and the fluorescence of each well was then measured at 390 and 460 nm (excitation and emission wavelengths) using a BMG plate reader (Offenburg, Germany). Triplicate samples were prepared and analyzed for each condition in every set of experiments.
Hoechst Staining and Trypan Blue Exclusion-The cells were cultured for 2 days on glass coverslips for Hoechst staining and directly in 35 ϫ 10-mm Petri dishes for trypan blue exclusion analysis. Cell transfected or not with S33Y-␤-catenin or GSK-3␤ 9A were exposed for 3 days to 100 M NS-398 in the absence or presence of 80 nM LTD 4 . The experiments were terminated by washing the cells on the coverslips (Hoechst staining) three times with PBS, fixing them with 4% (w/v) paraformaldehyde for 15 min at 37°C, and then staining them for 15 min in 10 M/ml Hoechst 34580 suspended in PBS. The coverslips were then washed again and mounted on glass slides. The cells were examined, and representative images were taken. To determine viability, the cells growing in the Petri dishes were harvested and stained with 0.2% (w/v) trypan blue suspended in PBS. After washing the percentages of viable cells were calculated.
Statistical Analysis-Results are expressed as the means Ϯ S.E. All comparisons between mean values were performed by use of analysis of variance statistics with subsequent post hoc testing according to Tukey-Kramer at ␣ 0.05; p values of Ͻ0.05 were considered significant.

LTD 4 -induced Phosphorylation of GSK-3␤-GSK-3␤
is known to phosphorylate ␤-catenin directly and induce its ubiquitination and degradation by proteasomes. Because GSK-3␤ can be inactivated through phosphorylation at its Ser-9 residue via a PI 3-kinase-dependent activation of Akt (37), such events have the potential of increasing the amount of free ␤-catenin (38). Earlier results from our group show that LTD 4 stimulation induces an increase in ␤-catenin protein levels (16) and Akt activation (39). We therefore investigated whether LTD 4 -mediated ␤-catenin up-regulation involves changes in GSK-3␤ phosphorylation status. Int 407 cells were stimulated with LTD 4 for different periods of time and the level of phosphorylated GSK-3␤ evaluated by immunoblotting. The level of Ser-9 phosphorylation of GSK-3␤ was increased significantly (5-fold that of basal) after 30 min of LTD 4 treatment and still remained elevated after 24 h (Fig. 1A, representative blot and accumulated densitometric data). The membranes were always reprobed with an antibody against total GSK-3␤ to ensure equal loading (Fig. 1, A-C). Subsequent experiments revealed that LTD 4 stimulation (1 h) caused a parallel increase in the cellular level of ␤-catenin (Fig. 1B, representative blot and accumulated densitometric data). These LTD 4induced changes in GSK-3␤ phosphorylation status and in the level of ␤-catenin were both reduced by the CysLT 1 receptor antagonist ZM198,615 and the PI 3-kinase inhibitor LY294002 (Fig. 1B). As a control, we inhibited GSK-3␤ with 10 mM LiCl, a procedure known to be mediated via phosphorylation of Ser-9 (40). Our data clearly show that treatment with LiCl results in phosphorylation of GSK-3␤ on Ser-9, and importantly in an increase in the level of ␤-catenin (Fig. 1C). Furthermore, we show that both LiCl and LTD 4 increased the level of COX-2 expression (Fig. 1C). Because ␤-catenin-mediated TCF/LEF activity can regulate transcription of the COX-2 gene, these results lend further support to the notion that LTD 4 -induced phosphorylation of GSK-3␤ (Ser-9) and the accumulation of free ␤-catenin could be interrelated. To test this possibility further, we transfected cells with the constitutively active GSK-3␤ 9A. In the GSK-3␤ 9A-transfected cells we found a clear reduction of the LTD 4 -induced increase in ␤-catenin expression compared with cells transfected with an empty vector (Fig. 1D). In parallel we observed a reduction in the level of GSK-3␤ (Ser-9) phosphorylation in the GSK-3␤ 9A-transfected cells (Fig. 1D). Although the majority of Wnt-1 in the intestine is believed to come from nonepithelial mesenchymal cells or, more specifically, intestinal subepithelial myofibroblasts (41), we also tested whether LTD 4 stimulation could affect the level of Wnt-1 in the human intestinal epithelial cell line Int 407. We did not find any change in the cellular level of Wnt-1 after LTD 4 stimulation (Fig. 1E).
Increased TCF/LEF Transcription upon LTD 4 Stimulation-It is known that free ␤-catenin, made available by the deregulation of Wnt-1 signaling or disruption of the APC complex, translocates to the nucleus where it acts as a regulator of the TCF/LEF family of transcription factors (19,42). We therefore examined whether the cellular localization of ␤-catenin and its transcriptional activity is effected by LTD 4 . First, we investigated whether the increase in the level of ␤-catenin upon LTD 4 stimulation leads to an increase of ␤-catenin in the nuclear frac-tion. Nuclear fractions were isolated from cells that had been stimulated with LTD 4 for different periods of time ( Fig. 2A). We observed by immunoblotting a 2-fold, statistically significant, increase in the amount of ␤-catenin in the nuclear fraction after 60 min of LTD 4 stimulation ( Fig. 2A). This LTD 4 -induced elevation was still evident after 24 h ( Fig.  2A). We also found a clear reduction of the LTD 4 -induced ␤-catenin nuclear localization in cells transfected with GSK-3␤ 9A compared with cells transfected with an empty vector (Fig. 2B). The nuclear localization of ␤-catenin suggests that LTD 4 stimulation can affect gene transcription. To determine whether this was the case we analyzed TCF-dependent transcriptional activity in cells transfected with ␤-catenin/TCF transcriptional reporter TOPFlash, or inactive mutant FOPFlash, luciferase reporter constructs. We found that LTD 4 increased TCF/LEF activity after 30 min and that the statistically significant effect still remained after 24 h of stimulation (Fig. 2C). In accordance with the effect of LTD 4 on GSK-3␤ phosphorylation status and on the level of ␤-catenin (Fig. 1B), we also found that the LTD 4 -induced TCF/LEF activity was abolished by CysLT 1 receptor antagonists, pertussis toxin, and PI 3-kinase inhibitors (Fig. 2D).
The LTD 4 -induced redistribution of ␤-catenin to the nuclei was confirmed by confocal laser scanning microscopy imaging of intact intestinal cells (Fig. 3). Cells grown on coverslips were stimulated with LTD 4 and stained with the ␤-catenin antibody, as well as an antibody directed against Lamin B, a well known marker of the inner nuclear membrane previously used in this context in this cell line (6). In unstimulated control cells, ␤-catenin was localized predominantly to the plasma membrane, but with extending duration of LTD 4 treatment we observed increased ␤-catenin accumulation in the nucleus (Fig. 3). Although there is a detectable increase in ␤-catenin accumulation in the nuclei early on, the robust increase is first seen after 24 h. This difference in kinetics between the results obtained by Western blot analysis ( Fig. 2A) and immunofluorescence (Fig. 3) could be because the epitope for the  antibody is subject to an altering environment in the latter case. A novel and interesting observation made in Fig. 3 was that LTD 4 treatment of these intestinal cells also caused translocation of ␤-catenin to a punctate extranuclear structure that was thought to be mitochondria. 4 Stimulation-The above observation that ␤-catenin was potentially localized to the mitochondria after 1 h of LTD 4 stimulation is interesting in relation to a recent publication in which ␤-catenin was suggested to be involved in regulation of cytochrome c release from mitochondria (29). Consequently, we further investigated a possible mitochondrial presence of ␤-catenin in Int 407 cells by an alternative approach. Isolation of a mitochondrial fraction from these cells clearly revealed that the amount of ␤-catenin in this fraction was increased significantly after 1 h of stimulation with LTD 4 (Fig. 4). Additionally, the mitochondrial localization of ␤-catenin upon LTD 4 stimulation was confirmed by confocal microscopy using cells stained with both Mitotracker Red and ␤-catenin antibodies. In control cells, ␤-catenin was localized mainly to the plasma membrane (Fig. 5A, upper panels), but after 1 h of LTD 4 stimulation it had translocated to the mitochondria as indicated by its colocalization with Mitotracker Red (Fig. 5A, lower right panel). In light of our previous finding that LTD 4 stimulation also increases the amount of Bcl-2 in Int 407 cells (16) and the fact that mitochondria are the primary site of Bcl-2 function, we investigated and found Bcl-2 in the mitochondria of both nonstimulated (Fig. 5B, upper panels) and LTD 4 -stimulated cells (Fig. 5B, lower panels). Next we investigated the possibility of an interaction between ␤-catenin and Bcl-2 at this location. The cells were stimulated or not with LTD 4 , after which ␤-catenin immunoprecipitates were immuno-blotted with an anti-Bcl-2 antibody (Fig. 6A). We observed a novel interaction between ␤-catenin and Bcl-2 after stimulation with LTD 4 . This finding was further confirmed by analyzing Bcl-2 immunoprecipitates from cells stimulated or not with LTD 4 by Western blotting with an anti-␤-catenin antibody (Fig. 6B). As seen in Fig. 6, A and B, the LTD 4induced association between ␤-catenin and Bcl-2 is barely present in unstimulated control cells. We performed the same type of experiments with a Bcl-2-glutathione S-transferase fusion protein and using this approach also pulled down a significant amount of ␤-catenin (data not shown). Further analysis of this protein-protein interaction was performed by overexpressing either wild type HA-␤-catenin or mutant HA-S33Y-␤-catenin in Int 407 cells. Following stimulation or not with LTD 4 we precipitated ␤-catenin with an anti-HA antibody and analyzed these precipitates by immunoblotting with an anti-Bcl-2 antibody (Fig.  6C). Already in unstimulated cells overexpressing HA-␤-catenin a significant amount of Bcl-2 was pulled down with an anti-HA antibody, but after LTD 4 stimulation the amount of associating Bcl-2 increased significantly (Fig. 6C). In accordance, in unstimulated cells, overexpression of a mutant, constitutively active form of ␤-catenin (HA-S33Y-␤-catenin, mimicking the action of LTD 4 ) increased the amount of Bcl-2 (185%) that was pulled down by immunoprecipitation with an anti-HA antibody (Fig. 6C). These data clearly show that ␤-catenin and Bcl-2 can interact with each other and that increased amounts of nonphosphorylated ␤-catenin, obtained after LTD 4

stimulation of intestinal cells boosts this protein-protein interaction.
Overexpression of Bcl-2 Affects TCF/LEF Transcription Activity-The interaction between the survival protein Bcl-2 and the multifunctional protein ␤-catenin could indicate that they functionally affect each other's activities. To address this possibility we first overexpressed the Bcl-2 protein in Int 407 cells and stimulated them or not with LTD 4 for different periods of time. Interestingly enough, overexpression of Bcl-2 increases both basal TCF/LEF transcriptional activity as well as LTD 4induced activity, the latter already after 30 min of stimulation (Fig. 7A). In agreement with the situation in nontransfected cells the LTD 4 -in-  duced TCF/LEF activity was significantly impaired by CysLT 1 receptor antagonists, pertussis toxin, and PI 3-kinase inhibitors (Fig. 7B).
Overexpression of ␤-Catenin Increases Survival of Intestinal Epithelial Cells-To investigate a possible influence of ␤-catenin on cell survival, we transfected cells with different ␤-catenin constructs and investigated their effect on intracellular caspase-3 activity. We tested the response to NS-398-induced apoptosis, a selective COX-2 inhibitor that has been used previously as an experimental tool to induce caspase-3dependent apoptosis in intestinal cells (16,35). Bcl-2 is a well known upstream inhibitor of caspase-3 (43), and therefore we analyzed the degree of apoptosis in this series of experiments by measuring the activity of this caspase. Overexpression of wild type ␤-catenin did not significantly affect the basal survival rate of Int 407 cells. However, overexpression of constitutively active S33Y-␤-catenin caused a statistically significant decrease in basal caspase-3 activity (Fig. 8A). NS-398 induced a 50% increase in caspase-3 activity compared with untreated control cells (Fig. 8B). This increase was abolished and the activity of caspase-3 reduced below basal level by transient transfections with either wild type ␤-catenin or even more so by mutant S33Y-␤-catenin (Fig. 8B). The reduction in caspase-3 activity caused by wild type ␤-catenin was enhanced even further by the presence of LTD 4 (Fig. 8B). LTD 4 on its own caused a 30% decrease even further To validate further the results obtained in Fig. 8 we also performed trypan blue exclusion experiments. We found an almost 40% reduction of viability in NS-398treated cells, an effect that was reversed by the addition of LTD 4 (Fig.  9A). The reducing effect of NS-398 on cell viability was not evident in cells transfected with S33Y-␤-catenin (Fig. 9A). We also stained the cells with Hoechst as another means of quantitating the degree of apoptosis. We found that NS-398 induced typically fragmented and condensed nuclear chromatin in cells transfected with empty vectors. These effects of NS-398 were much less prominent in cells transfected with ␤-catenin S33Y (Fig. 9B). Finally, LTD 4 had no effect on cell viability as determined by trypan blue and Hoechst staining in ␤-catenin S33Y-transfected cells (Fig. 9). To validate further that the effect of LTD 4 on intestinal cell apoptosis was at least in part mediated by inhibition of GSK-3␤ and accumulation of ␤-catenin, we transfected cells with the constitutively active GSK-3␤ 9A. We found that LTD 4 had no effect on caspase-3 activity (Fig. 10A), cell viability (Fig. 10B), or Hoechst staining (Fig. 10C) in cells transfected with the constitutively active GSK-3␤ 9A. These data suggest a role for GSK-3␤ and ␤-catenin in LTD 4 regulation of intestinal cell survival.

DISCUSSION
We have observed previously that prolonged exposure of nontumor intestinal epithelial cells to the inflammatory mediator LTD 4 results in increased expression of two proteins associated with colon cancer,  ␤-catenin and COX-2 (16). A further indication of a role for LTD 4 in the development of colon cancer came from the finding that its receptor, CysLT 1 , is up-regulated in human colon cancer tissue (8). Here we outline a plausible mechanism whereby the previously observed CysLT 1 receptor-induced increase in ␤-catenin, and subsequent rise in COX-2 levels in intestinal cells, can be explained (44). Our work reveals that LTD 4 -induced activation of CysLT 1 receptor signaling increases ␤-catenin levels via a PI 3-kinase-dependent phosphorylation of GSK-3␤, the kinase targeting ␤-catenin for subsequent proteosomal degradation. This phosphorylation of the Ser-9 residue of GSK-3␤ enables further downstream signaling of ␤-catenin. Traditionally ␤-catenin is known as an effector protein in the Wnt signaling pathway (45). Increased stabilization and thus accumulation of ␤-catenin can result both from increased canonical Wnt signaling and from mutations in either ␤-catenin itself or in any of its regulatory gene products, such as APC. In these situations the canonical Wnt pathway can be considered constitutively active, even though the mechanism responsible varies. All of these situations are more or less associated with different cancer conditions, including colon cancer (19,25). Wnt receptors, known as Frizzled, belong to the large family of G protein-coupled receptors that also includes the LTD 4 receptor CysLT 1 . The present evidence that the CysLT 1 receptor can regulate ␤-catenin levels is supported by recent findings demonstrating that other receptors are capable of transducing intracellular signals leading to ␤-catenin stabilization (46). Activation of the G protein-coupled receptor for histamine has recently been found to initiate Wnt-like signaling in HeLa cells, and also the human androgen receptor, which belongs to the nuclear receptor superfamily, has been shown to regulate ␤-catenin levels in human embryonic kidney cells HEK 293, a cell line that exhibits epithelial morphology (18,47). Our data suggest that PI 3-kinase and GSK-3␤ are signaling molecules necessary for the LTD 4 -mediated up-   ). B, the cells were transiently transfected with HA-wt-␤-catenin, HA-S33Y-␤-catenin, or empty vectors and then treated or not with 100 M NS-398 and/or 80 nM LTD 4 for 4 h. Thereafter, the cells in both A and B were lysed and caspase-3 activities assayed. The fluorescent intensity data are calculated as a percentage of untreated control cells and given as the means Ϯ S.E. of three separate experiments (the mean value of fluorescent intensity (AU) in the control was 638 in A and 530 in B). Statistical significances (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.005) were calculated between cells transfected with empty vectors and cells transfected with either wild type ␤-catenin or S33Y-␤-catenin (A). In B these calculations were done between untreated controls and cells either exposed to LTD 4 (gray bar) or NS-398 (lined bar) and between nonstimulated (ns) and LTD 4 -stimulated cells transfected with empty vector, wild type ␤-catenin, or S33Y-␤-catenin. All white bars (***) were statistically different from NS-398treated cells (lined bar).
regulation of ␤-catenin in human intestinal epithelial cells. The role of PI 3-kinase in this context is most likely to cause activation of the Akt kinase in intestinal epithelial cells, a previously demonstrated property of LTD 4 (39). Such an activation of Akt could explain the PI 3-kinasedependent phosphorylation of the Ser-9 residue of GSK-3␤ and its subsequent inactivation (48). In support of our data, it was shown recently that the heregulin receptor, ErbB3, can increase ␤-catenin-dependent transcription of cyclin D1 via a PI 3-kinase-and GSK-3␤-dependent signaling pathway in the non-small lung cancer cell line H441, and the lung adenocarcinoma cell line H1373 (49).
Increased levels of free ␤-catenin result in an increased translocation of this protein to the nucleus, where its main task is to activate the TCF/LEF family of transcription factors (50). These factors regulate the transcriptional activities of several genes associated with carcinogenesis, such as cyclin D1, c-myc, and COX-2 (28,51,52). Furthermore, they also regulate proliferation in many different cell types (53). The present study reveals, through different methodological approaches, that LTD 4 not only increases the level of ␤-catenin, but also causes an increased nuclear accumulation of ␤-catenin and activation of the TCF/LEF transcription factor family. Taken together this means that LTD 4 stimulation triggers Wnt-like downstream ␤-catenin signaling in the nuclei of intestinal epithelial cells, which could relate to the simultaneous up-reg-ulation of COX-2 (16) and the effect of LTD 4 on proliferation of intestinal and other cell types (15, 54 -56).
During our analysis of the cellular distribution of ␤-catenin we made the intriguing observation that LTD 4 induces translocation of ␤-catenin not only to the nuclei but also to the mitochondria. This is interesting in the context of recent reports indicating that ␤-catenin signaling might affect functions separate from those associated with traditional canonical Wnt signaling. For example, Shin and co-workers (29) demonstrated that Wnt-induced ␤-catenin signaling can mediate protection from cytochrome c leakage. To validate a signaling role for ␤-catenin in mitochondria we searched for a possible regulatory/interacting partner. Earlier findings from our group have revealed that LTD 4 induces upregulation of Bcl-2 in parallel with ␤-catenin and COX-2 in intestinal epithelial cells (16). Bcl-2 is a well known cell survival signaling protein, and interestingly enough its predominant location is the mitochondria (57). With this in mind, our novel finding that LTD 4 signaling induces an association between ␤-catenin and Bcl-2 is all the more fascinating and opens up for a role of this interaction in protecting cells from apoptosis. In addition, in cells that overexpress the active (S33Y) form of ␤-catenin, apoptosis was inhibited, as indicated by increased caspase-3 activity, suggesting that active ␤-catenin can potentiate survival of these cells. We have shown previously that LTD 4 protects intestinal cells from caspase-3-driven apoptosis via a Bcl-2-dependent mechanism (16,58). A different but related finding is that overexpression of Bcl-2 causes an increase in TCF/LEF transcriptional activity, although the mechanism behind this effect is presently not understood.
In summary we show that the inflammatory mediator LTD 4 induces accumulation of free ␤-catenin and the activation of TCF/LEF transcription factors via PI 3-kinase-dependent phosphorylation of the FIGURE 9. Increased ␤-catenin levels enhance cell survival. A, cells were transfected with empty vector or S33Y-␤-catenin and then exposed or not to 100 M NS-398 and 80 nM LTD 4 for 3 days as indicated in the figure. They were then stained with trypan blue, and the percentage of viable cells was calculated. The values represent the means Ϯ S.E. of three separate experiments. Statistical significances were calculated between nonstimulated (black bar) and NS-398-treated cells (gray bar) or between NS-398-treated cells (gray bar) and LTD 4 -stimulated cells transfected with S33Y-␤-catenin. No significance between S33Y-␤-catenin-transfected cells stimulated or not with LTD 4 was obtained. B, representative fluorescence micrographs of cells transfected with empty vector (top two micrographs) or with S33Y-␤-catenin (lower two micrographs) and then exposed to NS-398 with or without LTD 4 (as indicated in the figure). Finally the cells were stained with Hoechst 34580 before examination in a fluorescent microscope. The images shown are representative of three separate experiments. inhibitory Ser-9 residue of GSK-3␤, as well as a physical interaction between ␤-catenin and Bcl-2. We hypothesis that this novel association appears to serve as a mechanism whereby ␤-catenin can augment Bcl-2-mediated survival signaling. Aberrant ␤-catenin-assisted TCF/LEF transcription is an integral part of carcinogenesis, as is the inability of cells to undergo apoptosis. Hence, our finding that the inflammatory mediator LTD 4 is capable of regulating both of these key events via an increased accumulation of ␤-catenin constitutes a significant contribution to the understanding of potential mechanisms in cancer occurrence under inflammatory conditions.