FP Prostanoid Receptor Activation of a T-cell Factor/β-Catenin Signaling Pathway

FP prostanoid receptors are G-protein-coupled receptors (GPCR) that consist of two known isoforms, FPA and FPB. These isoforms, which are generated by alternative mRNA splicing, are identical except for their carboxyl-terminal domains. Previously we have shown that stimulation of both isoforms with prostaglandin F2α(PGF2α) activates the small G-protein Rho, leading to morphological changes consisting of cell rounding and the formation of cell aggregates. Following the removal of PGF2α, however, FPA-expressing cells show rapid reversal of cell rounding, whereas FPB-expressing cells do not. We now show that acute treatment of FPB-expressing cells with PGF2αleads to a subcellular reorganization of β-catenin, a decrease in the phosphorylation of cytoplasmic β-catenin, and persistent stimulation of Tcf/Lef-mediated transcriptional activation. This does not occur in FPA-expressing cells and may underlie the differences between these isoforms with respect to the reversal of cell rounding. The Tcf/β-catenin signaling pathway is known to mediate the actions of Wnt acting through the heptahelical receptor, Frizzled, and has not been associated previously with GPCR activation. Our findings expand the signaling possibilities for GPCRs and suggest novel roles for FP receptors in normal tissue development and malignant transformation.

The amino acid sequences of the ovine FP A and FP B prostanoid receptor isoforms are the same throughout their amino termini and seven-membrane-spanning domains, but the FP B isoform is truncated and lacks the last 46 carboxyl-terminal amino acids present in the FP A isoform (1). This is very similar to the EP 3 (2) and thromboxane A2 (3) prostanoid receptors in which alternative mRNA splicing gives rise to a variety of isoforms in humans and in other species (4). The physiological significance of these receptor isoforms is not clear, although differences have been shown to exist with respect to second messenger coupling and receptor desensitization. We have found that the FP A and FP B receptor isoforms have similar pharmacological properties and that prostaglandin F 2␣ (PGF 2␣ ) 1 stimulates phosphoinositide turnover to a similar extent in cells expressing these isoforms (1). In addition, stimulation of FP A -or FP B -expressing cells with PGF 2␣ activates Rho leading to the formation of actin stress fibers, phosphorylation of p125 focal adhesion kinase, and cell rounding (5). Cell rounding involves the retraction of filopodia and a change from an isolated dendritic appearance to one in which the cells are rounded and form small cobblestone-like aggregates (see Fig.  1A). Following the removal of PGF 2␣ , however, FP A -expressing cells return to their original dendritic morphology, but the FP B -expressing cells do not and remain rounded (6). We hypothesized that FP B -expressing cells might remain rounded because of prolonged signaling following the removal of agonist. However, a specific mechanism of this prolonged signaling was not established. Here we show that Tcf/␤-catenin-mediated transcriptional activation is elevated 16 h after an initial 1-h treatment of FP B -expressing cells with PGF 2␣ . This transcriptional activation is not observed in FP A -expressing cells and suggests that FP B -expressing cells remain rounded because of activation of a Tcf/␤-catenin signaling pathway.

EXPERIMENTAL PROCEDURES
Immunofluorescence Microscopy-HEK-293 cells stably expressing the ovine FP A and FP B prostanoid receptor isoforms (5) were split and grown in six-well plates containing 22-mm round glass coverslips for 3-4 days. Cells were treated with either vehicle (sodium carbonate, 0.002% final) or 1 M PGF 2␣ and were rapidly washed, fixed, and incubated with a 1:1000 dilution of a mouse monoclonal antibody to ␤-catenin (Transduction Laboratories). They were then washed and incubated with a 1:4000 dilution of an fluorescein isothiocyanate-conjugated goat anti-mouse secondary antibody (Sigma). Nuclei were stained with 4Ј,6-diamidino-2-phenylindole (DAPI, Sigma). Cells were visualized by phase contrast and epifluorescence microscopy as described previously (6).
Immunoprecipitation and Blotting-Cells were scraped and sonicated in a lysis buffer consisting of 20 mM Tris-HCl (pH 7.5), 10 mM EDTA, 2 mM EGTA, 2 mM phenylmethylsulfonylfluoride, 0.1 mg/ml leupeptin, and 2 mM sodium vanadate. Samples were centrifuged (16,000 ϫ g) for 15 min at 4°C, the supernatant (cytosolic fraction) was removed, and the pellet (particulate fraction) was solubilized with lysis buffer containing 0.2% Triton X-100 and then centrifuged again to remove insoluble debris. For immunoprecipitation, samples were rotated for 2 h at 4°C with antibodies to ␤-catenin followed by the addition of protein G-Sepharose (Amersham Pharmacia Biotech) and rotation for another hour. The Sepharose was washed with lysis buffer and then resuspended with SDS-polyacrylamide gel electrophoresis sample buffer and boiled. Samples were electrophoresed on 7.5% SDSpolyacrylamide gels, transferred to nitrocellulose membranes, and incubated with either antibodies to ␤-catenin or a mixture of mouse monoclonal antibodies to phosphoserine (Sigma) and phosphothreonine (Sigma). The membranes were washed, incubated with horseradish peroxidase-conjugated goat anti-mouse secondary antibodies, and visualized by enhanced chemiluminescence (SuperSignal, Pierce). The resulting films were scanned at high resolution (300 dpi) as positive transparencies (Microtek ScanMaker4) and saved as TIFF files. Quantitation was performed using the Gelplot2 macro in Scion Image for Windows (beta version 4.02). Nuclear extracts were prepared according to the method of Dignam as modified by Westin et al. (7).
RT-PCR-RT was done using the Superscript Preamplification System (Life Technologies, Inc.) and 1 g of RNA/sample that had been * This work was supported in part by grants from the National Institutes of Health (EY11291) and Allergan Inc. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed. Tel.: 520-626-2181; Fax: 520-626-2466; E-mail: regan@pharmacy.arizona.edu. pretreated with DNase I. This was followed by PCR using an initial incubation at 94°C for 5 min, followed by 20 cycles of 94°C, 60°C, and 68°C each for 2 min, and a final incubation at 68°C for 10 min. The human ␤-catenin and GAPDH primer pairs were exactly according to Rezvani and Liew (8). Product sizes were 521 base pairs for ␤-catenin and 737 base pairs for GAPDH and were resolved by electrophoresis on 1.5% agarose gels. Preliminary experiments were done to find the optimal conditions for quantitative amplification of ␤-catenin and GAPDH mRNA.
Tcf/Lef Reporter Gene Assay-Cells were split into 10-cm dishes and the next day were transiently transfected using FuGENE-6 (Roche Molecular Biochemicals) and either 10 g/dish of the wildtype Tcf/Lef reporter plasmid TOPflash or the mutant plasmid FOPflash. FOPflash differs from TOPflash by the mutation of its Tcf binding sites and serves to differentiate Tcf/␤-catenin-mediated signaling from background (Upstate Biotechnology, Inc.). Cells were incubated overnight and were treated for 1 h at 37°C with either vehicle or 1 M PGF 2␣ . They were then rapidly washed three times each with 2 ml of Opti-MEM (Life Technologies, Inc.) as described previously (6) and incubated for 16 h at 37°C in 10 ml of Opti-MEM containing 250 g/ml geneticin, 200 g/ml hygromycin B, and 100 g/ml gentamicin. Cells were placed on ice and rinsed twice with ice-cold phosphate-buffered saline, and extracts were prepared using the Luciferase Assay System (Promega). Luciferase activity in the extracts (ϳ500 ng protein/sample) was measured using a Turner TD-20/20 luminometer and was corrected for background by subtraction of FOP-FLASH values from corresponding TOP-FLASH values. Fig. 1A shows phase contrast microscopy of HEK cells stably expressing either the ovine FP A prostanoid receptor (panels a and b) or the ovine FP B prostanoid receptor (panels c and d) following a 1-h treatment with either vehicle (panels a and c) or 1 M PGF 2␣ (panels b and d). It can be appreciated that in both FP A -and FP B -expressing cells treatment with PGF 2␣ resulted in morphological changes consisting of a loss of filopodia and formation of cell aggregates. We have previously shown that these morphological changes involve the activation of Rho and phosphorylation of p125 focal adhesion kinase (5). However, following the removal of PGF 2␣ the FP A -expressing cells show a rapid (within 1 h) reversal of these morphological changes, whereas the FP B -expressing cells remain rounded even after 48 h (6). To investigate the possible role of other adhesion proteins in this process, we used immunofluorescence microscopy to examine the localization of E-cadherin and ␤-catenin in HEK cells stably expressing either the FP A or FP B isoforms following treatment with 1 M PGF 2␣ . Although the effects on E-cadherin localization were not apparent (data not shown), Fig. 1B shows that PGF 2␣ treatment resulted in a marked accumulation of ␤-catenin in regions of cell-to-cell contact in FP B -expressing cells (panels c and d) but not in FP A -expressing cells (panels a and b). Both cell lines, however, showed agonistdependent cell rounding following treatment with PGF 2␣ (Fig.  1A), indicating that the process of cell rounding itself was not responsible for the increased contiguous accumulation of ␤-catenin in the FP B -expressing cells.

RESULTS
In addition to its role in cell adhesion, ␤-catenin is well recognized as a signaling molecule that undergoes stimulus-dependent translocation from the cytosol to the nucleus where it is involved in the regulation of Tcf/Lef-mediated gene transcription (9 -11). We, therefore, used immunoblotting to examine both particulate and cytosolic fractions for changes in ␤-catenin expression following treatment of FP A -and FP Bexpressing cells with PGF 2␣ . Fig. 2A shows that the expression of ␤-catenin is ϳ3-fold higher in the particulate fraction and ϳ2-fold higher in the cytosolic fraction from FP B -expressing cells compared with FP A -expressing cells. Furthermore, treatment with PGF 2␣ caused a slight increase the levels of cytosolic ␤-catenin in both the FP A -and FP B -expressing cells but had little effect on the levels of ␤-catenin in the particulate fraction.

FP Receptor Activation of ␤-Catenin Signaling 12490
Reverse transcription (RT) followed by polymerase chain reaction (PCR) was used to determine whether there were any differences in ␤-catenin mRNA levels under these same experimental conditions. Fig. 2B shows that ␤-catenin and GAPDH mRNA levels were the same for both cell lines and were not affected by PGF 2␣ , indicating that the observed differences in ␤-catenin expression appear to be the result of changes in translation and/or protein turnover.
Serine/threonine phosphorylation of ␤-catenin by glycogen synthase kinase-3␤ (GSK-3␤) marks ␤-catenin for degradation and is a critical factor in the regulation of its signaling activity (12,13). Thus, under most conditions cytosolic ␤-catenin is phosphorylated, leading to an association with the tumor suppressor protein, adenomatous polyposis coli (APC), and the scaffolding protein, axin, which is then followed by ubiquitination and proteasomal degradation (14). Using immunoprecipitation and immunoblotting, we examined serine/threonine phosphorylation of ␤-catenin following treatment of either FP Aor FP B -expressing cells with PGF 2␣ . Fig. 3 d, upper panel). The ratio of phosphorylated to total ␤-catenin in the cytoplasm, therefore, shows dramatic differences following activation of these two FP receptor isoforms. Thus, in FP A -expressing cells this phosphorylation ratio increases from 0.5 to 1.6 with agonist treatment, whereas in FP B -expressing cells it falls from 2.3 to 0.1. It would therefore be expected that degradation of cytosolic ␤-catenin would be favored at the expense of nuclear translocation in FP Aexpressing cells, whereas the opposite would be true in FP Bexpressing cells. This appears to be confirmed in Fig. 3 where immunoblotting of nuclear extracts shows a 3-fold higher level of ␤-catenin in FP B -expressing cells following treatment with PGF 2␣ (lane d,

bottom panel) as compared with FP A -expressing cells (lane b, bottom panel).
Following nuclear translocation, ␤-catenin is known to interact with members of the Tcf/Lef family of transcription factors (15). Because of this signaling potential, we were interested in the possibility that the failure of FP B -expressing cells to return to their original dendritic morphology following removal of PGF 2␣ might represent a ␤-catenin-mediated switch in gene expression. To examine an effect on gene epression, we transiently transfected either FP A -or FP B -expressing cells with a Tcf/Lef-responsive reporter plasmid (16) and measured luciferase reporter gene activity following treatment with 1 M PGF 2␣ . Initially we found that basal levels of luciferase activity were elevated (ϳ3-fold) in FP B -expressing cells as compared with FP A -expressing cells and that luciferase activity was not stimulated immediately following a 1-h treatment with PGF 2␣ in either cell line (data not shown). However, as shown in

FP Receptor Activation of ␤-Catenin
Signaling 12491 erase activity (column d) that is roughly 6.5-fold higher than either the vehicle control (column c) or PGF 2␣ -treated FP A cells (column b). The failure of FP B -expressing cells to show reversal of cell rounding is not because of changes in the kinetics of PGF 2␣ binding or in its removal during the washout procedure (6). DISCUSSION We show that FP B -expressing cells differ in several important regards from FP A -expressing cells in terms of their potential for activation of Tcf/␤-catenin-mediated signaling. First, FP B -expressing cells show PGF 2␣ -stimulated accumulation of ␤-catenin at their contiguous cell boundaries that is not evident in FP A -expressing cells. Second, although both FP A and FP Bexpressing cells show PGF 2␣ -stimulated increases in cytosolic ␤-catenin, in FP A -expressing cells this is accompanied by increased ␤-catenin phosphorylation and in FP B -expressing cells by decreased ␤-catenin phosphorylation. Third, FP B -expressing cells show a stimulation of Tcf/Lef reporter gene activity 16 h after agonist removal that is essentially absent in FP A -expressing cells. A key control point could be in the differential phosphorylation of ␤-catenin. Thus, it is possible that the agoniststimulated accumulation of ␤-catenin at the cell boundaries of FP B cells results in enhanced interactions with E-cadherin. In turn, this could initiate E-cadherin outside-in signaling leading to the sequential activation of phosphatidylinositol 3-kinase and Akt kinase (17). This is meaningful because phosphorylation of GSK-3␤ by Akt kinase is inhibitory (18) and could lead to the decreased phosphorylation of ␤-catenin found in agonisttreated FP B cells.
Recently Meigs et al. (19) reported that constitutively active mutants of G ␣12 and G ␣13 interact with E-cadherin resulting in a release of ␤-catenin and stimulation of Tcf/Lef reporter gene activity in a mutant cell line lacking APC. This link between heterotrimeric G-proteins and the Tcf/␤-catenin signaling pathway is novel, but its physiological relevance might be questioned because of the altered nature of the model. In light of the present findings, however, it appears likely that both GPCRs and heterotrimeric G-proteins will be involved with activation of this signaling pathway. We have shown that FP receptors activate Rho through the probable activation of G 12 and/or G 13 (5). Both receptor isoforms were equally effective in this regard, and therefore it appears unlikely that activation of G 12 and/or G 13 could be solely responsible for the present findings because activation of Tcf/␤-catenin signaling was observed only for cells expressing the FP B isoform.
We believe that activation of Tcf/␤-catenin signaling by PGF 2␣ in cells expressing the FP B receptor is involved with a phenotypic transformation that is morphologically similar to, but fundamentally different from, the cell rounding observed in agonist-treated FP A cells. Thus, maintenance of shape change in FP A -expressing cells depends on continuous stimulation by PGF 2␣ , and following its removal the cells revert back to their original morphology. In contrast, although shape change in FP B -expressing cells is initiated by PGF 2␣ , its maintenance is independent of further PGF 2␣ stimulation. In this manner the FP B prostanoid receptor functions as one would expect of a trigger in a developmental or malignant transformation pathway.
The present findings have significance for the signaling potential of FP prostanoid receptors and possibly for other GPCRs as well. For example, in sheep it is known that PGF 2␣ is the physiological signal for regression of the corpus luteum but only during a short window of the luteal cycle. Thus, if preg-nancy occurs the corpus luteum is maintained and loses sensitivity to the luteolytic actions of PGF 2␣ (20). Interestingly, the expression of FP A receptors does not change during this transition (21). Brief expression of a small population of FP B receptors during the sensitive phase of the luteal cycle could explain the luteolytic actions of PGF 2␣ .
Another condition that might involve the FP B isoform or a homologue is in colorectal cancer. It is known that aberrant activation of Tcf/␤-catenin signaling is associated with the development of this disease (22)(23)(24) and that inhibition of cyclooxygenase by NSAIDs can slow tumor progression (25). However, the specific mechanism of this beneficial effect is vague because of the large number of prostanoid metabolites that are affected. Our findings support a mechanism in which NSAID-mediated decreases in PGF 2␣ would decrease Tcf/␤catenin signaling by FP B prostanoid receptors. This conclusion is supported by animal models of skin carcinogenesis in which PGF 2␣ reversed the anti-tumor-promoting activity of indomethacin (26). Although a human homologue of the ovine FP B receptor has not yet been identified other mechanisms could give rise to functional FP B isoforms. Thus, much like the known mutations of APC, truncation of the human FP A receptor by allelic variation, somatic mutations or proteolytic cleavage could give rise to receptors capable of producing activation of Tcf/␤-catenin signaling. The possible role of FP B receptors in these and other physiological processes is intriguing and awaits future studies.