Glycogen Synthase Kinase-3 Phosphorylation, T-Cell Factor Signaling Activation, and Cell Morphology Change following Stimulation of Thromboxane Receptor α

Thromboxane receptor (TP) is a member of the G protein-coupled receptor (GPCR) superfamily. Two isoforms of human TP, TPα and TPβ, have been recognized. They differ in their carboxyl-terminal tails and their mechanisms of desensitization and internalization (Kinsella, 2001). TP is known to couple to multiple G proteins, including G_q/11, G_12/13, G_i, G_s, and G_h, and mediates many cellular responses (Halushka et al., 1995). One of the notable responses is the stimulation of mitogenic and hypertrophic growth of vascular smooth muscle. For example, U-46619, a TP agonist, stimulates the proliferation of rat coronary or aortic smooth muscle cells (Smith, 1992) and enhances the mitogenic effects of classic growth factors such as platelet-derived growth factor and basic fibroblast growth factor (Ali et al., 1993; Zhao and Foegh 1997). I-BOP, another TP agonist, triggers a protein kinase C (PKC)-, pertussis toxin-, Src-, and epidermal growth factor receptor (EGFR)-dependent activation of extracellular signal-regulated kinase (ERK) in a human endothelial cell line, ECV304 (Gao et al., 2001). U-46619 was also reported to induce c-Jun NH₂-terminal kinase activation and protein kinase A (PKA)-dependent cyclooxygenase-2 (COX-2) expression in porcine aortic smooth muscle cells (McGinty et al., 2000). Recently, Gallet et al. (2003) demonstrated that TP agonist-dependent transactivation of EGFR was mediated partially by matrix metalloproteinases and subsequent activation of ERK in human aortic smooth muscle cells. Stimulation of cell proliferation by TP agonists was observed beyond vascular smooth cells. Lin et al. (2005) reported that U-46619 stimulated the proliferation of oligodendrocyte progenitor cells and promoted the survival of mature oligodendrocyte in central nervous system. It seems that TP-mediated cell proliferation and survival are mainly dependent on two interactive pathways: one pathway is through transactivation of EGFR leading to phosphorylation of ERK; another pathway is activation of PKA-dependent pathway.

Some GPCRs are known to functionally link to T-cell factor (Tcf)/β-catenin signaling pathway. Activation of mouse F9 clones expressing the chimera receptor constructed from the β2 adrenergic receptor and the Frizzled-1 receptor with the β-adrenergic agonist isoproterenol stimulated stabilization of β-catenin, activation of a β-catenin-sensitive promoter, and formation of primitive endoderm (Liu et al., 2001). Thereafter, it was shown that activation of FP_B, an isoform of prostaglandin F_2α receptor, led to a decrease in the phosphorylation of cytoplasmic β-catenin and a stimulation of Tcf/Lef-mediated transcriptional activation (Fujino and Regan, 2001). The same group also demonstrated the activation of the Tcf/Lef signaling pathway through EP receptors, EP₂ and EP₄, using stably transfected receptor in HEK293 cells (Fujino et al., 2002). Recently, GPCR and Tcf/Lef cross-talk was expanded to colon cancer cells. Yang et al. (2005) showed that activation G protein-coupled lysophosphatidic acid receptors stimulated proliferation of colon cancer cells through the β-catenin pathway and induced cell proliferation.

An important enzyme in the β-catenin/Tcf signaling pathway is glycogen synthase kinase-3 (GSK-3) (Doble and Woodgett, 2003). This enzyme is responsible for the phosphorylation and degradation of β-catenin through ubiquitin proteosome pathway. Two isoforms of GSK-3, GSK-3α and GSK-3β, encoded by two distinct genes have been identified (Woodgett, 1990; Hoeflich et al., 2000). GSK-3 is well known to be phosphorylated by Wnt signal mediated by frizzled receptor (Nelson and Nusse, 2004). GSK-3 also can be phosphorylated by Akt/PKB, PKC, and PKA (Fang et al., 2002). Direct inhibition of the GSK-3 either by its inhibitors or by phosphorylation prevents the phosphorylation and degradation of β-catenin. Stabilized β-catenin accumulates in the cytosolic compartment of the cell, translocates to the nucleus, and binds to members of Tcf/Lef transcriptional factors and induces expression of several downstream genes that have important roles in the development and progression of carcinoma, including cyclin D1, c-myc, vascular endothelial growth factor, COX-2, matrix metalloproteinase 7, urokinase-type plasminogen activator, and CD44 (Wong and Pignatelli, 2002).

Since TP can activate Akt/PKB, PKC, and PKA (Miggin and Kinsella, 2002), we propose that there is a possible cross-talk between the TP and the Tcf/Lef signaling pathway. Using HEK293 cells stably transfected with TPα gene, we report here that the stimulation of TPα by its agonist I-BOP led to GSK-3 phosphorylation, Tcf/Lef signaling activation, and permanent cell morphology change. These effects seem to be related to cell transformation.

Materials and Methods

Materials. Culture medium, LipofectAMINE 2000, heat-inactivated fetal bovine serum (FBS), and restriction enzymes were from Invitrogen (Carlsbad, CA). I-BOP, U-46619, and SQ-29548 (SQ) were from Cayman Chemical (Ann Arbor, MI). H89, wortmannin, and GF109203X (GF) were obtained from Calbiochem (San Diego, CA). Other biochemicals, chemicals, and rabbit polyclonal anti-β-catenin antibody were obtained from Sigma-Aldrich (St. Louis, MO). Mouse monoclonal antibody specific to pERK, and β-catenin siRNA and control siRNA were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Rabbit polyclonal antibody specific to pGSK-3α/β was from Cell Signaling Technology Inc. (Beverly, MA). FuGENE6 was from Roche Diagnostics (Alameda, CA). pcDNA3 encoding TPα and rabbit polyclonal antibody specific to N-terminal sequence of the TP were generated as described previously (Zhou et al., 1999). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibodies were prepared as described previously (Wu et al., 1997). Rabbit monoclonal antibody specific to cyclin D1 was provided by Lab Vision (Fremont, CA). Horseradish peroxidase-linked goat anti-mouse and rabbit IgG were supplied by BD Transduction Laboratories (Lexington, KY). The HEK293 cell lines were supplied by the American Type Culture Collection (Manassas, VA). ECL Western blotting detection was supplied by the Amersham Pharmacia Biotech (Cardiff, UK). The Tcf/Lef reporter plasmids kit including TOP flash, containing two sets (with the second set in the reverse orientation) of the three copies of the Tcf site upstream of the luciferase open reading frame, and FOP flash, containing mutated Tcf elements, served as a background and was purchased from Upstate Biotechnology (Lake Placid, NY).

Cell Culture and Stable Transfectants. HEK293 cells were cultured as monolayer in DMEM supplemented with 10% heat-inactivated FBS, gentamicin, and antibiotic-antimycotic at 37°C in a humidified atmosphere of 95% air and 5% CO₂. The cells were subcultured in six-well plates or 12-well plates to achieve 80% confluence on the next day, and then the cells were transfected with pcDNA3-TPα using LipofectAMINE reagent. For stable transfection, the cells were allowed to grow in 10% FBS medium for 48 h after transfection and then the cells were diluted 10-fold and treated with 1 mg/ml Geneticin (G418) until the colonies formed as described previously (Zhou et al., 1999). The Western blot detection of the TPα and the [³H]SQ-29548 binding assay of the receptor were used to monitor the expression level among the different colonies. After the colony with a high level of expression of TPα was obtained, it was maintained in DMEM media supplemented with 10% heat inactivated FBS and 200 μg/ml G418.

Whole-Cell Radioligand Binding Assay. Cells were cultured in 10-cm plates and were harvested when 90% confluence was achieved. The cells were washed in ice-cold phosphate-buffered saline buffer three times. Then, 1 × 10⁶ cells were resuspended in 95 μl of phosphate-buffered saline buffer using 1.5-ml plastic tubes, and the binding assay was conducted in the same tubes. The total volume of the reaction was 100 μl, which contained [³H]SQ-29548 with the concentration from 2.6 × 10^–12 to 2.6 × 10^–7 M. The nonspecific binding was determined under the same conditions except 10 μM SQ-29548 was added to the reaction mixture. The binding reaction was performed at room temperature for 1 h and was terminated by placing the tubes on ice for 10 min. The cell suspension was then filtered under vacuum through a GF/C glass filter followed by washing three times. The radioactivity retained on the filter was counted in 10 ml of scintillation cocktail. The K_d and B_max were calculated using nonlinear regression method by GraphPad Prism 4.01 (Graph-Pad Software Inc., San Diego, CA).

Western Blot Analysis. Cells were cultured in six-well plates to achieve approximately 80% confluence. For ERK and GSK-3 activation, the cells were starved in DMEM media without FBS for 16 h, and then cells were challenged by 50 nM I-BOP for 15 min to achieve significant phosphorylation of ERK, and for 0 to 60 min to monitor the time course of the activation of GSK-3α/β. For inhibition study, various inhibitors were preincubated with cells for 20 min following treatment with I-BOP for 15 min. Cells were harvested and lysed in lysis buffer (1% Nonidet P-40 in 150 mM NaCl, 50 mM HEPES, pH 7.4, 5 mM NaF, 5 mM pyrophosphate, 1 mM sodium orthovanadate, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride) for 1 h on ice. Lysate was cleared by centrifugation at maximum speed on a benchtop centrifuge and then subjected to 12% SDS-PAGE. Proteins were then electrophoretically transferred onto PVDF membrane. The membrane was blocked with 5% nonfat milk in 30 mM Tris-HCl, pH 7.4, containing 120 mM NaCl (TBS) at room temperature for 1 h. It was then incubated for 2 h at room temperature with a primary antibody in TBS with 5% nonfat milk, and then the membrane was washed three times with TBS buffer containing 0.05% Tween 20, following incubation with horseradish peroxidase-linked goat anti-mouse or rabbit IgG for 1 h at room temperature. Finally, the membrane was washed with TBS buffer containing 0.05% Tween 20 three times. The immunoreactive bands were detected using ECL Western blotting detection system.

Tcf/Lef Reporter Gene Assay. HEK293-TPα cells were grown in six-well plates and transiently transfected with 1 μg/well of either TOP flash or FOP flash plasmids using FuGENE6 procedure as described in the manufacturer's protocol. Cells were preincubated with inhibitors at indicated concentrations or vehicle (0.1% dimethyl sulfoxide) for 20 min at 37°C followed by challenging with 50 nM I-BOP or vehicle for 1 h at 37°C. Cells were then washed three times each with 1 ml/well of FBS-free DMEM media and then incubated with FBS-free DMEM media for 16 h at 37°C. Cell lysate was prepared for luciferase assay. Luciferase activity was measured using a microplate luminometer with 5 μg of protein per sample. The specific reporter gene activity was corrected for background activity by subtraction of FOP flash values from corresponding TOP flash values.

siRNA Treatment. Cells were transiently transfected with β-catenin siRNA at the concentration of 20 pmol/well in 12-well plates using LipofectAMINE reagent. The cells were treated with the siRNA for 48 h to achieve the maximal blockade of the expression of the β-catenin.

Phase-Contrast Microscope. For phase-contrast microscopic studies, HEK293 or HEK293-TPα cells were grown in 12-well plates to achieve approximately 40 to 50% confluence using DMEM media containing 10% FBS and the same antibiotics as mentioned above. Cells were visualized under phase-contrast microscope using a Zeiss Axiovert S100 inverted microscope. Images were captured using a Kodak digital camera.

Results

Stable Transfection of TP in HEK293 Cells and Characterization of the Functional Expression of the Receptor. HEK293 cell line was chosen to construct TPα stably transfected cells since it has minimal basal expression of TP. As shown in Fig. 1A, Western blotting using antibody specific to N-terminal peptide of TP demonstrated that HEK293-TPα expressed significantly TPα compared with HEK293 cells transfected with vector alone. Similar to previous reports, HEK293-TPα had a high affinity to TP antagonist SQ-29548, with a K_d value of 14.6 ± 1.4 nM and B_max value of 5.3 ± 0.31 pmol/mg protein characterized by [³H]SQ-29548 binding assay (Fig. 1B). The functional assay of the TPα was characterized by agonist-stimulated ERK activation. As shown in Fig. 1C, the TP agonist I-BOP stimulated significant phosphorylation of ERK in HEK293-TPα cells, whereas minimal response was observed in HEK293 cells transfected with vector alone. TP agonist-induced activation of ERK has been well documented in several cell lines (Gao et al., 2001; Miggin and Kinsella, 2001). Our results are consistent with previous reports and provide evidence that HEK293-TPα cells expressed functionally active TPα.

TP Agonist-Stimulated Phosphorylation of GSK-3α/β in HEK293-TPα Cells. Previous reports showed that several prostaglandins such as PGE₂ and PGF₂α could induce phosphorylation of GSK-3α/β through their specific receptors (Fujino et al., 2002; Shao et al., 2005). We used HEK293-TPα cells to study GSK-3α/β phosphorylation upon TP agonist stimulation. As shown in Fig. 2A, both GSK-3α and GSK-3β were phosphorylated in 2 min following exposure to 50 nM I-BOP in HEK293-TPα cells. Phosphorylation of GSK-3α/β was achieved maximally at 5 min after stimulation by I-BOP and was gradually decreased during the following 50 min. The time-course studies showed that phosphorylation of GSK-3α/β still persisted at high level even after 60 min of incubation with I-BOP, indicating that the signal induced by a TP agonist existed for a relatively long time. However, no detectable phosphorylation of GSK-3α/β was found in HEK293 cells (data not shown). As shown in Fig. 2B, I-BOP induced phosphorylation of GSK-3α/β in a concentration-dependent manner. I-BOP at 1 nM clearly led to phosphorylation of GSK-3α/β and phosphorylation reached to a maximum at 50 nM. We then used the specific kinase inhibitors to determine which pathway(s) was involved in the GSK-3α/β phosphorylation. As shown in Fig. 2C, PKA inhibitor H89 not only totally blocked I-BOP-induced GSK-3α/β phosphorylation but also inhibited basal level of pGSK-3α/β, suggesting that PKA is required to maintain the pGSK-3α/β in unstimulated cells. TP antagonist SQ-29548 can block I-BOP-induced phosphorylation of GSK-3α/β, suggesting that the signal transduction is directly through TP. A concentration-dependent inhibition by H89 is shown in Fig. 2D. However, wortmannin, an inhibitor of PI-3 kinase, partially blocked GSK-3α/β phosphorylation. Surprisingly, the well known PKC activation following TP activation was not involved in this process, since its inhibitor GF109203X had no effect on I-BOP-induced GSK-3α/β phosphorylation.

β-Catenin Accumulation in Cytoplasmic Fraction and Tcf/Lef Reporter Gene Activation following Stimulation of TPα. Phosphorylation of GSK-3α/β is known to inactivate GSK-3α/β. Consequently, I-BOP-induced phosphorylation of GSK-3α/β may stimulate accumulation of β-catenin in the cytosolic compartment and enhance translocation of β-catenin from the cytosol to the nucleus, leading to the Tcf/Lef transcription factor activation. As shown in Fig. 3, I-BOP induced an increase in the amount of β-catenin in the cytosolic fraction of HEK293-TPα cells. I-BOP consistently stimulated an 8-fold increase in the Tcf/Lef reporter gene expression in HEK293-TPα cells as shown in Fig. 4. Since H89 and wortmannin were found to block I-BOP-induced GSK-3α/β phosphorylation, they were also used to examine the effect on the activation of the Tcf/Lef reporter gene expression. As shown in the same figure, I-BOP-induced activation of the Tcf/Lef reporter gene was totally blocked by H89 and partially blocked by wortmannin, but it was not affected by GF compound. Again, the TP antagonist SQ-29548 can block the TCF/Lef activation. This is consistent with the results of GSK-3α/β phosphorylation induced by I-BOP in HEK293-TPα cells. It is clear that I-BOP activates Tcf/Lef transcription factor through GSK-3α/β phosphorylation and that PKA and PI-3 kinase/Akt pathways were involved in this process. However, TPα coupled to G_q, leading to PKC activation had no direct effect on this process in our system.

Cyclin D1 Expression in HEK293-TPα Cells upon TP Agonist Stimulation. Previous reports demonstrated that β-catenin regulated cyclin D1 expression (Tetsu and McCormick, 1999). β-Catenin/TCF/Lef complex could bind to cyclin D1 promoter to modulate cell cycle progression and proliferation. In HEK293-TPα cells, I-BOP caused a rapid and persistent (up to 8 h) increase in cyclin D1 expression (Fig. 5A). Inhibition of β-catenin expression by siRNA approach as indicated in Fig. 5B attenuated I-BOP-induced cyclin D1 expression as shown in Fig. 5C. Control siRNA had no effect.

Cell Morphology Change in HEK293-TPα Cells following Stimulation by TP Agonist. Activation of β-catenin/TCF/Lef is often linked to cell phenotypic transformation leading to developmental or malignant transformation. Cell morphology change in HEK-TPα cells began to show after 30 min of incubation with I-BOP and was clearly observed following 1 h of incubation as shown in Fig. 6. Incubation of I-BOP for 1 h resulted in cell rounding and aggregation in HEK293-TPα cells (Fig. 6A, c and d), whereas I-BOP had no effect on HEK293 cells transfected with pcDNA3 vector alone under the same conditions (Fig. 6A, a and b). After 1-h incubation with I-BOP, the HEK293-TPα cells were washed with DMEM without FBS to remove I-BOP, but the cells could not revert back to the normal shape even after 24 h (Fig. 6B, a and b), suggesting that a short time incubation with TP agonist was sufficient to induce a permanent change in cell morphology. The permanent morphology change is likely the consequence of the agonist-induced gene expression mentioned above. It may be recalled that in the Tcf/Lef reporter gene assay, H89 and wortmannin had an inhibitory effect on the activation of Tcf/Lef. In accordance with this, preincubation of HEK293-TPα cells with H89 (Fig. 7, E and F) prevented the I-BOP-induced cell morphology change, whereas wortmannin (Fig. 7, G and H) and GF109203X (Fig. 7, I and J) had little preventive effect. Blockade of the TP signaling by antagonist SQ-29548 could prevent I-BOP-induced permanent cell morphology change, suggesting that the TP was directly involved in the signal transduction leading to cell transformation.

Discussion

We used HEK293 cells stably transfected with TPα to study the signaling mechanism and its relevance to cell transformation. We demonstrated that the activation of TPα led to GSK-3 phosphorylation, T-cell factor signaling activation, and cell morphology change. A detailed scheme depicting the signaling pathway of TP activation is summarized in Fig. 8. Previous report indicated that two isoforms of TP may be coupled differentially to G_s and G_i (Hirata et al., 1996). TPα is coupled to G_s, leading to an increase in intracellular level of cAMP and activation of PKA, whereas TPβ is coupled to G_i, leading to a decrease in intracellular level of cAMP. It seems that the activation of Tcf/Lef pathway by TPα is mainly mediated by a PKA-dependent mechanism. A reasonable prediction is that TPβ may not induce the activation of Tcf/Lef pathway or TPβ activation of Tcf/Lef pathway is mediated by other mechanism(s). It has been shown that the activation of both EP₂ and EP₄ may induce the stimulation of Tcf/Lef pathway. Activation of Tcf/Lef pathway by EP₂ was shown to be mediated by a PKA mechanism, whereas stimulation of Tcf/Lef pathway by EP₄ was demonstrated to be mediated by a PKB/Akt mechanism (Fujino et al., 2002). This is because activation of EP₂ induces a significantly higher level of intracellular cAMP compared with that of EP₄ (Fujino et al., 2002). Whether the activation of TPβ may lead to the stimulation of Tcf/Lef pathway or other pathway(s) involved in the activation of Tcf/Lef needs to be elucidated in TPβ stably transfected cells.

We have shown that I-BOP leads to a rapid phosphorylation of both GSK-3α/β, activation of Tcf/Lef signaling, and cell morphology change in TPα stably transfected HEK293 cells. Similar results were obtained with U-46619, another stable agonist of TPα, in these cells. Furthermore, we also observed I-BOP-induced rapid phosphorylation of GSK-3α/β and activation of Tcf/Lef signaling in human lung cancer A549 cells (data not shown). Previously, it was also shown that the activation of EP₂ and EP₄ led to significant phosphorylation of GSK-3α but not that of GSK-3β (Fujino et al., 2002). However, in our studies both forms of GSK-3 were rapidly phosphorylated upon stimulation of TPα. The difference in response is not clear and needs to be further explored. In addition to cAMP/PKA pathway, we also found that PI-3 kinase inhibitor wortmannin partially inhibited phosphorylation of GSK-3α/β. The activation of Akt/PKB by TP agonists was reported previously by Miggin and Kinsella (2002). Akt/PKB is known to be one of upstream kinases of GSK-3 inducing phosphorylation and inhibition of GSK-3 (Cross et al., 1995). In fact, Akt/PKB-dependent activation of the β-catenin/Tcf pathway was reported in ECV304 carcinoma cells (Maupas-Schwalm et al., 2005). One possible explanation is that PKA is an upstream kinase of Akt/PKB pathway and may indirectly phosphorylate GSK-3α/β (Filippa et al., 1999). This may account for why H89 totally blocked the phosphorylation of GSK-3, whereas wortmannin partially attenuated it.

The accumulation of β-catenin in the cytosol is a direct result from the phosphorylation and inhibition of GSK-3α/β, as shown in this study. Intracellular β-catenin level, however, can be regulated by other pathways as well. It is known that Wnt signaling also regulates intracellular β-catenin level. A large multiprotein complex that includes proteins of the Adenomatous polyposis coli and Axin families normally facilitates the phosphorylation of β-catenin by GSK-3 (Nelson and Nusse, 2004). Phosphorylated β-catenin is then rapidly degraded by the ubiquitin-mediated proteosome pathway. When Wnt signal is absent, the signal transduction pathway is off because β-catenin is rapidly destroyed. When cells are exposed to Wnt signal, it binds to cell surface receptors of the Frizzled family. Receptor activation antagonizes the Adenomatous polyposis coli-Axin “destruction complex” by an unknown mechanism that seems to be mediated by the Dishevelled protein. This blocks the phosphorylation of β-catenin and its subsequent degradation. Unphosphorylated β-catenin forms a pool in the cytosol and then translocates to the nucleus where it activates the genes regulated by Tcf/Lef transcription factors. Our findings and reports of studies on other GPCRs such as EP₂ and FP_B expanded the Wnt pathway to other signaling pathways that could directly phosphorylate and inactivate the GSK-3α/β and regulate β-catenin level.

The linkage between β-catenin and cyclin D1 was established from accumulated reports especially in human cancers (Utsunomiya et al., 2001). The cyclin D1 is associated with cancer growth, invasion, and metastasis through influencing the cell cycle progression. We show here that I-BOP stimulates overexpression of cyclin D1, and blockade of β-catenin expression by siRNA approach attenuates this effect. Furthermore, our preliminary data showed that TPα agonist stimulated a permanent change in cell morphology, such as cell rounding and aggregation. A same phenomenon was found in the activation of FP receptors stably transfected in HEK293 cells (Fujino and Regan, 2001). In this report, FP_B isoform that activated the Tcf/Lef transcription factors induced a permanent change in cell morphology, whereas FP_A, which could not activate Tcf/Lef, only caused a temporary change in cell morphology. A previous report showed that overexpression of COX-2 and microsomal prostaglandin E synthase had the potential to affect cellular proliferation and morphology change (Murakami et al., 2000). Many studies demonstrated that expression of COX-2 is associated with cell proliferation, antiapoptosis, and differentiation (Mann et al., 2001). As mentioned in previous reports, a permanent change in cell morphology was an indication of developmental or malignant transformation. Thromboxane A₂ is one of the downstream metabolites of COX-2 and is involved the proliferation of several cell lines. It also has roles in the tumor growth and metastasis (Daniel et al., 1999; Nie et al., 2000; Pradono et al., 2002). Therefore, our findings have expanded COX-2/PGE₂ pathway to COX-2/thromboxane A₂ pathway. To verify the possible roles of TPα in cancer, lung cancer A549 cells stably overexpressing TPα receptor and lung cancer H157 cells stably transfected with TPα siRNA were prepared. Preliminary data indicated that I-BOP induced COX-2 expression in A549 cells overexpressing TPα, suggesting that a positive feedback might occur in COX-2- and TPα-expressing cells. Generation of tumors in the nude mice using these cell lines is currently in progress in our laboratory.