Induction of Adherent Activity in Mastocytoma P-815 Cells by the Cooperation of Two Prostaglandin E2 Receptor Subtypes, EP3 and EP4*

In this study, we investigated the role of PGE2 in mouse mastocytoma P-815 cell adhesion to extracellular matrix proteins (ECMs) in vitro. We report that PGE2 accelerated ProNectin FTM (a proteolytic fragment of fibronectin)-mediated adhesion, which was abolished by addition of the GRGDS peptide, an inhibitor of the RDG binding site of ProNectin FTM. We show that the cAMP level and cAMP-regulated protein kinase (PKA) activity are critical mediators of this PGE2 effect, because the cell-permeable cAMP analogue 8-Br-cAMP accelerated P-815 cell adhesion to ProNectin FTM and the pharmacological inhibitor of PKA, H-89, blocked PGE2-mediated adhesion. Consistent with mRNA expression of the Gs-coupled EP4- and Gi-coupled EP3-PGE receptor subtypes, P-815 cell adhesion was accelerated by treatment with a selective EP4 agonist, ONO-AE1–329, but not a selective EP1/EP3 agonist, sulprostone. However, simultaneous treatment with ONO-AE1–329 and sulprostone resulted in augmentation of both the cAMP level and cell adhesion. The augmentation of EP3-mediated cAMP synthesis was dose-dependent, without affecting the half-maximal concentration for EP4-mediated Gs-activity, which was inhibited by a Gi inhibitor, pertussis toxin. In conclusion, these findings suggest that PGE2 accelerates RGD-dependent adhesion via cooperative activation between EP3 and EP4 and contributes to the recruitment of mast cells to the ECM during inflammation.

Differentiated mast cells (MCs), 1 which originate from bone marrow stem cells, traffic throughout the circulation and adhere to the extracellular matrix (ECM) in various tissues. MCs are widely distributed in tissues throughout the body, especially in connective tissues, serosal cavities, and on mucosal surfaces under normal physiological conditions. This characteristic distinguishes MCs from other bone marrow-derived hematopoietic cells, such as basophils, neutrophils, and eosin-ophils. MCs congregate around nerves, blood vessels, and lymphatic vessels. MCs therefore interact with not only the ECM but with other cells as well. As well known for rodent connective tissue-typed MCs and mucosal-typed MCs, the biological activity of MCs vary with their interactions with the ECM and other cells.
MCs are widely distributed along basement membranes, indicating that MCs might adhere to laminin. Supporting this fact, mouse bone marrow-derived mast cells (BMMC) have been reported to adhere to laminin, when the cells were activated by phorbol 12-myristate 13-acetate (PMA) (1)(2)(3) or antigen-stimulated aggregation of Fc⑀RI (4). In addition to laminin, MCs can adhere to other matrix components such as fibronectin (5) and vitronectin (6,7). As with laminin, the adherence of BMMC to fibronectin has been reported to occur through activation with PMA or after aggregation of Fc⑀RI. These adherence activities required calcium (3). In contrast to BMMC, the mouse PT18 cell line spontaneously adhered to laminin (1), and human skin mast cells also spontaneously adhered to laminin and fibronectin (8). These previous findings indicate that the interactions between MCs and matrix components may depend on the cells involved and the kinds of stimuli. PGE 2 , which is involved in inflammation section (9), affects both differentiation and growth of MCs in vitro. PGE 2 enhances mast cell differentiation from cord blood mononuclear cells (10) and in the fibroblast co-culture system (11). Very recently, Dormond et al. (12) reported that a COX-2 inhibitor suppressed ␣ V ␤ 3 -dependent HUVEC spreading, migration, and angiogenesis through Rac activation (12), and PGE 2 accelerated ␣V␤3mediated HUVEC responses in a cAMP-dependent manner (13). However, no reports have published the effects of PGE 2 on MC adhesion to the ECM.
The PGE 2 receptors (EP) are comprised of four subtypes, EP1, EP2, EP3, and EP4, which are coupled to different G proteins and signal pathways (14,15). Among these subtypes, EP2 and EP4 couple to G s , resulting in increases of intracellular cAMP concentrations (16,17), while EP3 couples to G i , causing a decrease in cAMP levels (18 -21). Very recently, we and other investigators have reported that activation of the G i -coupled EP3 receptor was able to augment adenylyl cyclase activity via stimulation of a G s -coupled receptor (22,23). We previously reported that the mouse EP3␤ receptor was able to augment EP2-induced adenylyl cyclase activity in both EP2and EP3-transfected COS-7 cells (22). Southall and Vasko (23) showed that the simultaneous depletion of rat EP3C and EP4 was essential to abolish PGE 2 -stimulated cAMP production and neuropeptide release in rat sensory neurons (23). These permissive actions between the two subtypes may be involved in the events of a number of physiological actions of PGE 2 (15).
Here we extended the analyses of the effects of PGE 2 on adhesion of mouse mastocytoma P-815 cells to the fibronectin component in a PKA-dependent manner, and we examined the correlation between cAMP levels and adhesion in P-815 cells to see whether EP3 could augment EP4-induced cAMP synthesis.

EXPERIMENTAL PROCEDURES
Materials-Sulprostone was a generous gift from Dr. M. P. L. Caton of Rhone-Poulenc Ltd. ONO-AE1-259 and ONO-AE1-329 were generous gifts from ONO Pharmaceuticals (Osaka, Japan). ProNectin F TM (Pronectin-F) and ProNectin L TM (Pronectin-L) were generous gifts from Sanyo Chemical Industries, Ltd. (Kyoto, Japan). The GRGDS peptide was obtained from the PEPTIDE Institute (Osaka, Japan). Collagen I-coated plates were from BD PharMingen Labware, the 125 Ilabeled cAMP assay system from Amersham Biosciences. Pertussis toxin (PT) from Seikagaku Kogyo (Tokyo, Japan), [5,6,8,11,12,14,  Reverse Transcription-Polymerase Chain Reaction-Mouse mastocytoma P-815 cells (P-815 cells) were maintained in Fischer's medium with 10% fetal bovine serum under humidified air containing 5% CO 2 at 37°C. Total RNA was isolated from P-815 cells using Isogen (Nippon Gene) according to the manufacturer's instructions. The reverse transcription (RT) reaction was performed using Molony murine leukemia virus reverse transcriptase (Invitrogen) in the presence of random hexamers. The polymerase chain reaction (PCR) was performed with Taq DNA polymerase (TOYOBO) using the first strand template. The primers were designed to be selective for the prostanoid receptors, and the sequences used were as follows: EP1 (base pairs (545 bp)), 5Ј-GCC ACT GAT CCA CGC GGC GCG CGT ATC TGT GGC C-3Ј (forward) and (forward) and 5Ј-CAG GTA GAT GAG CAG CTG GTG CTC TGT GGC-3Ј (reverse). The thermal cycle programs used were as follows: EP1, 94°C for 1 min and then 94°C for 1 min, 72°C for 3.5 min for the first 10 cycles, 94°C for 1 min, 69°C for 2 min, 72°C for 1.5 min for the next 10 cycles, 94°C for 1 min, 66°C for 2 min, 72°C for 1.5 min for the last 15 cycles, followed by 72°C for 7 min; EP2, EP3, EP4 and DP, 94°C for 1 min and then 30 cycles of 94°C for 0.8 min, 60°C for 0.7 min, 72°C for 1.5 min, followed by 72°C for 7 min; IP, 94°C for 1 min and then 30 cycles of 94°C for 0.8 min, 65°C for 0.7 min, 72°C for 1.5 min, followed by 72°C for 7 min; FP and TP, 94°C for 1 min and then 94°C for 1 min, 72°C for 3.5 min for the first 10 cycles, 94°C for 1 min, 67°C for 2 min, 72°C for 1.5 min for the next 10 cycles, 94°C for 1 min, 63°C for 2 min, 72°C for 1.5 min for the last 15 cycles, followed by 72°C for 7 min. Samples were visualized on a 1.5% agarose gel.
Measurement of cAMP Formation-Cyclic AMP levels in P-815 cells were determined as reported previously (25). P-815 cells (2 ϫ 10 6 cells/assay) were suspended with 10 M indomethacin in HEPES-buff-ered saline containing 140 mM NaCl, 4.7 mM KCl, 2.2 mM CaCl 2 , 1.2 mM MgCl 2 , 1.2 mM KH 2 PO 4 , 11 mM glucose, and 15 mM HEPES (pH 7.4), and preincubated for 10 min at 37°C. Reactions were started by addition of test agents along with 500 M 3-isobutyl-1-methyl-xanthine (Sigma). After incubation for 10 min at 37°C, reactions were terminated by the addition of 10% trichloroacetic acid. The level of cAMP was measured by radioimmunoassay with an Amersham Biosciences cAMP assay system.
Expression of the Adenylyl Cyclase Subtypes-Expression of adenylyl cyclase II, IV, and VII were determined by immunoblot analysis of whole cell detergent lysates of P-815 cells using rabbit polyclonal anti-A cyclase II (C-20) antibody, rabbit polyclonal anti-A-cyclase IV (C-20) antibody, and goat polyclonal anti-A cyclase (M-20) antibody. P-815 cells (2 ϫ 10 8 cells) were washed twice in PBS. The cell pellet was suspended in 1 ml of radioimmune precipitation assay buffer containing 30 mM HEPES-NaOH (pH 7.3), 150 mM NaCl, 1% Triton X-100, 1% deoxycholate, and 0.1% SDS, and incubated for 2 h at 4°C. For protection against proteolytic degradation, a mixture of protease inhibitors (0.2 mM phenylmethylsulfonyl fluoride, 100 M benzamidine, 10 g/ml leupeptin, 10 g/ml aprotinin, 10 g/ml E-64, and 1 g/ml pepstatin A) was added. The mixture was then centrifuged at 10,000 ϫ g for 15 min at 4°C. The resulting supernatant was dissolved in Laemmli buffer and heated for 30 min at 60°C. Aliquots (500 g of protein) were then subjected to SDS-PAGE (7.5%) as described by Laemmli (26), and the separated proteins were transferred electrophoretically onto a polyvinylidene difluoride membrane in 25 mM Tris base containing 40 mM 6-aminohexanoic acid, 0.02% SDS, and 20% methanol at room temperature for 30 min at 15 V. The membrane fraction was rinsed in Trisbuffered saline (TBS) containing 20 mM Tris-HCl (pH 7.5) and 150 mM NaCl, and then preincubated overnight in TBS containing 5% nonfat milk at 4°C. The membrane fraction was then incubated with anti-A cyclase antibodies (A cyclase II, IV, 1:200; A cyclase VII, 1:100) in TBS containing 5% nonfat milk for 1 h at 37°C. After washing three times with TTBS (TBS containing 0.05% Tween 20) at room temperature, the membrane fraction was incubated with peroxidase-conjugated antirabbit or -goat IgG in TTBS for 1 h at room temperature, and then detected with the ECL Western blot detection reagent.
Cell Adhesion Assay-P-815 cells were suspended in Fischer's medium with test agents along with 10% fetal bovine serum and 10 M indomethacin at a density of 5 ϫ 10 5 cells per ml, and 1 ml per well was dispensed into 12-well tissue culture plates. The 12-well tissue culture plates were precoated with 10 g/ml of Pronectin-F or Pronectin-L for 5 min at room temperature. Nonspecific binding was determined by using wells precoated with bovine serum albumin (3%) for 1 h at room temperature. The P-815 cells were incubated for 8 h at 37°C in a CO 2humidified atmosphere and non-adherent P-815 cells were suspended in 2 ml of PBS (non-adherent cells). The adherent P-815 cells were treated with PBS containing 0.02% EDTA and 0.25% trypsin at 37°C for 30 min, and the collected cells were suspended in 2 ml of PBS containing 0.02% EDTA and 2% fetal bovine serum (adherent cells). The numbers of non-adherent and adherent cells were counted using the COULTER Z1 cell counter (Beckman Coulter). The percentage of cell adhesion was calculated with the following formula shown in Equation 1.

P-815 Cell Adhesion to Pronectin-F-coated Plates-When
P-815 cells were cultured in Fischer's medium containing 10% fetal bovine serum together with 10 M indomethacin, almost no attachment to non-coated plates was observed. However, after incubation with 1 M PGE 2 for 8 h, ϳ11% of P-815 cells were retained on the non-coated plates as adhered cells, after being washed three times with PBS containing 0.02% EDTA (Fig. 1A). We next examined the effect of fibronectin on cell adhesion using Pronectin-F, which consists of multiple copies of the RGD cell attachment ligand of fibronectin (27). The number of adhered P-815 cells did not significantly change between the Pronectin-F-coated plates and the non-coated plates when the cells were incubated without PGE 2 , but markedly increased upon incubation with 1 M PGE 2 (Fig. 1B). PGE 2 -induced adherent activity to Pronectin-F-coated plates almost completely disappeared with addition of the GRGDS peptide (Fig. 1C), which is a soluble inhibitor for the RDG site of fibronectin as well as Pronectin-F (28). Furthermore, this PGE 2 -treated cell adhesion was not affected by further incubation with 0.02% EDTA for 30 min, but the cells could be detached by 0.25% trypsin treatment.
Expression of the EP3 and EP4 Subtypes in P-815 Cells-The RT-PCR experiment revealed that P-815 cells express mRNAs for EP3, EP4, and IP ( Fig. 2A). We confirmed the expression of the EP3 and EP4 receptor proteins by the replacement of the binding of 4 nM [ 3 H]PGE 2 to membrane fractions with 10 M PGE 2 , sulprostone (EP1 and EP3 agonist), or ONO-AE1-329 (EP4 agonist). The order of replacement potency was found to be PGE 2 Ͼ sulprostone Ͼ ONO-AE1-329 (Fig. 2B). These re-sults indicate that PGE 2 can bind to EP3 and EP4 present in the membrane fractions of P-815 cells.

PGE 2 -induced cAMP and Protein Kinase A-dependent P-815
Cell Adhesion-Along with PGE 2 , 8-Br-cAMP, and IBMX accelerated the adhesion of P-815 cells to Pronectin-F-coated plates (Fig. 3A). PGE 2 -mediated cell adhesion was greatly reduced by pretreatment with 10 M H-89, a PKA inhibitor (Fig.  3B). On the other hand, phorbol 12-myristate 13-acetate (PMA) had no effect on cell adhesion in the absence of PGE 2 , and in addition, PGE 2 did not change the intracellular Ca 2ϩ level (data not shown). These results suggest that the PGE 2 -induced adhesion of P-815 cells to Pronectin-F is via a cAMP-dependent and cAMP-protein kinase A-dependent pathway.
Sulprostone-induced Augmentation of ONO-AE1-329-induced cAMP Formation in P-815 Cells-Since P-815 cells express the G i -coupled EP3 as well as the G s -coupled EP4, it is possible that PGE 2 may bind to both subtype receptors at the same time, and hence the cAMP level may reflect the difference between G s -and G i -activated adenylyl cyclase activity. We therefore examined the cAMP levels in P-815 cells when treated with the EP1/EP3 agonist sulprostone and/or the EP4 agonist ONO AE1-329 (Fig. 4A). The cAMP level was increased by treatment with ONO-AE1-329 and reached a plateau level at 0.1 M. However, cAMP accumulation induced by various concentrations of the EP4 agonist were all augmented by the simultaneous addition of sulprostone (1 M), without affecting the half-maximal concentration, although sulprostone alone did not stimulate cAMP accumulation (Fig. 4A). The half-maximal concentration of sulprostone for EP3-mediated augmentation in EP4-induced cAMP synthesis was calculated to be ϳ5 ϫ 10 Ϫ8 M (data not shown). Furthermore, sulprostoneinduced augmentation of ONO-AE1-329-activated cAMP formation was absolutely inhibited by treatment with PT, and the cAMP level was almost equal to that observed in the cells stimulated by ONO-AE1-329 alone (Fig. 4B). These findings suggest that EP3-mediated augmentation of G s activity is achieved through G␤␥ subunits resulting from G i/o protein activation (29), although EP4-mediated G s activity is not.
Sulprostone-induced Augmentation of ONO-AE1-329-stimulated P-815 Cell Adhesion to Pronectin-F-We next examined whether EP3 stimulation could augment EP4-mediated adhe- sion of P-815 cells to Pronectin-F. 1 M sulprostone further augmented ONO-AE1-329 induced P-815 cell adhesion to Pronectin-F (Fig. 5), and the level of EP3-mediated augmentation of this adhesion was almost equal to the level of cAMP augmentation by the EP3 subtype (Fig. 4B). These findings suggest that PGE 2 -stimulated P-815 cell adhesion to Pronectin-F is mediated through the cooperative action of the EP3 and EP4 subtypes. DISCUSSION PGE 2 was previously reported to affect both differentiation and growth of mast cells in vitro. For example, PGE 2 is essential for the differentiation of IL-3-dependent mouse mast cells from spleen cells (30), PGE 2 enhanced human mast cell differentiation from cord blood mononuclear cells by inhibiting the production of macrophage-derived GM-CSF (10), and PGE 2 enhanced the growth of differentiated mast cells in a fibroblast co-culture system (11). It is therefore possible that PGE 2 plays a role in the interaction of mast cells with fibroblastic cells and extracellular matrix components. The interaction between mast cells and extracellular matrix components have profound influences on the targeting of mast cell progenitors to specific locations, the distribution of mast cell subsets, and the biological responsiveness of mast cells in tissues. The ability of mast cells to adhere to fibronectin, which involves the RGD sequence located within the cell-attachment domain of the fibronectin molecule, may play a role in the migration of mast cells in various tissues (5). In the present report, we found that PGE 2 is able to stimulate the adhesion of P-815 cells to Pronectin-F through the RGD cell attachment domain of fibronectin, and that this adherent activity is mediated via a cAMP-dependent pathway induced by the activation of the EP4 and/or EP3 receptors. It is possible that the cAMP-protein kinase A pathway may be involved in the induction of these cell attachment molecules, as PGE 2 -induced cell attachment was inhibited by treatment with H-89 (Fig. 3B) and cycloheximide (data not shown). However, neither PGE 2 nor 8Br-cAMP augmented the expression of VLA-5, one of the fibronectin receptors (data not shown). Therefore, cycloheximide may affect the signaling pathway of the fibronectin receptors. Further experiments are required to clarify these points.
Prostaglandin E 2 acts through binding to its specific receptors, which are comprised of four subtypes, EP1, EP2, EP3, and EP4 (14,15). Among them, EP3 and EP4 couple to G i and G s , and result in inhibition and stimulation of adenylyl cyclase activity, respectively (17)(18)(19)(20)(21). The mouse EP3 receptor is comprised of three isoforms, EP3␣, EP3␤, and EP3␥, which differ in their C terminus. Among them, EP3␤, which was used in this experiment is known to be coupled to the G i protein (19). Very recently, activation of a G i -coupled receptor has been reported to augment the adenylyl cyclase activity induced by the stimulation of a G s -coupled receptor in COS-7 cells. For example, activation of G i -coupled receptors such as ␣2 adrenoreceptor (29) and bradykinin B2 receptor (32) lead to the augmentation of G s -stimulated adenylyl cyclase in COS-7 cells. This synergistic effect has not been clearly shown in mammalian cells. Southall and Vasko (23) showed that the simultaneous deple- tion of both rat EP3C and EP4 was essential for abolishing PGE 2 -stimulated cAMP production and neuropeptide release in rat sensory neurons (23). In the current experiment we showed that the G i -coupled EP3 agonist sulprostone augmented cAMP formation induced by the EP4 agonist ONO-AE1-329 in P-815 cells. Similarly, we found that sulprostone was able to augment cAMP formation in P-815 cells activated by an IP agonist, carbacyclin, suggesting that the augmentative effects of EP3 can be observed irrespective of G s activation. The mechanism underlying these phenomena was thought to be via G␤␥-mediated activation of type IV adenylyl cyclase (29), because pretreatment with PT inhibited the augmentation by the activation of the G i -coupled EP3 receptor, and P-815 cells express type IV adenylyl cyclase (data not shown). Therefore, G␤␥-mediated activation of type IV adenylyl cyclase may be involved in EP3-mediated signaling to augment EP4-stimulated adenylyl cyclase activity in P-815 cells.
Although the involvement of phosphatidylinositol 3-kinase (PI3K) in cell adhesion to matrix proteins is shown in a variety of cell types (33), the role of PI3K in mast cell adhesion is still unknown. Kinashi et al. (31) reported that the adhesion of platelet-derived growth factor receptor-expressed bone marrow-derived mast cells was inhibited by the addition of the PI3K inhibitor wortmannin. To understand whether PI3K is activated by PGE 2 during P-815 cell adhesion, we examined the effect of the PI3K inhibitors, wortmannin (100 nM) and LY294002 (10 M) on EP3/EP4-agonist induced cell adhesion to fibronectin. As a result, these inhibitors showed an additive effect on the augmentation of PGE 2 -induced cell adhesion (wortmannin: 81.7 Ϯ 0.1% and LY294002: 82.4 Ϯ 0.2%, compared with PGE 2 stimulation alone: 24.8 Ϯ 0.6%). Furthermore, each inhibitor alone had an effect on cell adhesion (wortmannin: 20.4 Ϯ 0.1% and LY294002: 24.1 Ϯ 0.1%, compared with the absence of these inhibitors (3.3 Ϯ 0.1%)). Therefore, these results indicate that P-815 cell attachment to fibronectin is regulated by two independent signaling pathways involving PKA and PI3K. Further experiments are necessary to clarify the fundamental differences in the signaling involving PKA and PI3K in P-815 cell attachment to fibronectin. In summary, this study clearly demonstrates that two subtypes of the PGE 2 receptor, EP3 and EP4, are cooperatively involved in PGE 2evoked and cAMP-mediated functions of P-815 cell adhesion.