Synergism among Lysophosphatidic Acid, β1A Integrins, and Epidermal Growth Factor or Platelet-derived Growth Factor in Mediation of Cell Migration*

GD25 cells lacking the β1integrin subunit or expressing β1A with certain cytoplasmic mutations have poor directed cell migration to platelet-derived growth factor (PDGF) or epidermal growth factor (EGF), ligands of receptor tyrosine kinases, or to lysophosphatidic acid (LPA), a ligand of G-protein-coupled receptors (Sakai, T., Zhang, Q., Fässler, R., and Mosher, D. F. (1998) J. Cell Biol. 141, 527–538 and Sakai, T., Peyruchaud, O., Fässler, R., and Mosher, D. F. (1998) J. Biol. Chem. 273, 19378–19382). We demonstrate here that LPA synergizes with signals induced by β1A integrins and ligated EGF or PDGF receptors to modulate migration. When LPA was mixed with EGF or PDGF, migration was greater than with EGF or PDGF alone. The enhancement was greater for β1A-expressing cells than for β1-null cells. Cells expressing β1A with mutations of prolines or tyrosines in conserved cytoplasmic NPXY motifs had blunted migratory responses to mixtures of LPA and EGF or PDGF. The major effects on β1A-expressing cells of LPA when combined with EGF or PDGF were to sensitize cells so that maximal responses were obtained with >10-fold lower concentrations of growth factor and increase the chemokinetic component of migration. Sensitization by LPA was lost when cells were preincubated with pertussis toxin or C3 exotransferase. There was no evidence for transactivation or sensitization of receptors for EGF or PDGF by LPA. EGF or PDGF and LPA caused activation of mitogen-activated protein kinase by pertussis toxin-insensitive and -sensitive pathways respectively, but activation was not additive. These findings indicate that signaling pathways initiated by the cytoplasmic domains of ligated β1A integrins and tyrosine kinase receptors interact with signaling pathways initiated by LPA to facilitate directed cell migration.

Directed cell migration is of obvious importance in diverse physiological and pathological processes, including development, immunity, wound healing, and cancer metastasis (1)(2)(3). Cell migration requires fine control of cellular association with and release from the extracellular matrix (4 -6). Migration in a concentration gradient (chemotaxis) is accomplished by dynamic coordination among receptors for chemotactic agents, cellular adhesion receptors, and the actin-containing cytoskeleton. At any one time, a given cell likely encounters a variety of chemotactic substances and adhesive substrates. The mechanisms by which these signals are integrated to form a coordinated migratory response are poorly understood.
Integrins are transmembrane heterodimeric cell surface adhesion receptors composed of noncovalently associated ␣ and ␤ subunits. The fact that ligation of integrins by adhesive ligands can induce intracellular signaling events ("outside-in" signaling) and intracellular signaling pathways can control binding avidity of integrins for extracellular ligands ("inside-out" signaling) (7)(8)(9)(10)(11), makes integrins good candidates to mediate the adhesion-deadhesion events required for migration. In addition, integrins are linked to the cytoskeleton in a dynamic fashion by the molecules recruited to focal contacts (7)(8)(9)(10)(11)(12). Finally, cell migration may be facilitated by the cycling of integrins between cytoplasmic compartments and the cell surface (13,14).
Lysophosphatidic acid (LPA) 1 is a product of activated platelets and cells and has been shown to mediate multiple cellular responses (15,16). LPA is the serum enhancement factor of fibronectin matrix assembly; enhancement of assembly closely correlates with LPA-induced actin stress fiber formation and cell contraction (17)(18)(19). LPA is a mitogen for a number of cells (16,20,21) and induces in vitro invasion across host cell monolayers by several types of tumor cells (22,23) and also stimulates random nondirectional migration (chemokinesis) of Rat-1 fibroblasts (24) or both chemokinesis and directional migration (chemotaxis) of mouse GD25 fibroblastic cells (25). Recently, a specific G-protein-coupled receptor (GPCR) for LPA, ventricular zone gene-1 (vzg-1) (26) or endothelial differentiation gene-2 (edg-2) (27), was identified. Edg-1, a homologous protein, is a receptor for sphingosine-1-phosphate (S1P), another lysophospholipid generated during platelet activation, and has a low affinity to LPA (28,29). Three other Edg GPCRs have also been identified, Edg-3 and Edg-5 as receptors for S1P and Edg-4 as a receptor for LPA, (16,30,31). LPA and S1P stimulate several transduction cascades through GPCRs, including activation of the botulinum C3 exotransferase-sensitive small GTP-binding protein p21 Rho (Rho) through G ␣ 12/13 and the Ras/mitogenactivated protein kinase (MAPK) pathway through pertussis toxin-sensitive G i (15,16). Expression of a dominant-negative Ras mutant inhibited migration of NIH(M17) cells in response to LPA as well as to other chemoattractants such as plateletderived growth factor (PDGF) (32). The latter results suggest that Ras plays a key role in regulation of both LPA-and growth factor-induced cell migration.
Growth factors such as PDGF and epidermal growth factor (EGF), acting via tyrosine kinase receptors (RTKs), stimulate cell migration in concert with integrin-extracellular matrix interactions (33). In NR6 fibroblasts expressing EGF receptor, EGF alters migration speed and directional persistence in a matrix-dependent manner (34). Mouse B82L-B3 fibroblasts expressing EGF receptor do not migrate in response to EGF alone but when co-presented with laminin or fibronectin, EGF mediates EGF receptor-mediated migration, suggesting an interaction between EGF receptor and integrins ligated by laminin or fibronectin (35). PDGF enhances migration of NIH 3T3 cells when plated on vitronectin, a ligand for ␣ v ␤ 3 ; other ␣ v integrins but not ␤ 1 integrins associate with activated PDGF receptor (36 -38).
We recently expressed a series of ␤ 1 A integrin subunits in GD25 mouse fibroblasts derived from ␤ 1 -null stem cells and demonstrated that restoration of ␤ 1 integrin function is required for efficient migration of GD25 cells across matrixcoated filters in response to LPA, PDGF, or EGF (25,39). Because LPA, PDGF or EGF, and integrin ligands stimulate cells via three distinct but overlapping signaling systems as described above, we questioned whether engagement of two of the three systems is sufficient for maximum cell migration. In the present study, we demonstrate that LPA greatly increases the migratory activity of EGF or PDGF and that a functional ␤ 1 A integrin is required for this enhancement.

EXPERIMENTAL PROCEDURES
Materials-The GD25 fibroblast line, which was established after differentiation of ␤ 1 -null stem cells and immortalization with SV40 large T antigen, and its derivatives transfected with ␤ 1 A or ␤ 1 As with mutations of the cytoplasmic domain were as described previously (39 -41). PDGF from porcine platelets (R&D Systems, Minneapolis, MN), recombinant human EGF (Upstate Biotechnology, Lake Placid, NY), 1-oleoyl-LPA and S1P (Avanti Polar Lipids, Birmingham, AL, and LC Laboratories, Woburn, MA, respectively), and pertussis toxin (List Biological Laboratories, Campbell, CA) were purchased. Also purchased were rabbit polyclonal antibodies against active MAPK (Promega, Madison, WI), rabbit polyclonal antibodies that recognized mouse EGF or PDGF receptors (Santa Cruz Biotechnology, Santa Cruz, CA), and mouse monoclonal antibody (mAb 4G10) against phosphotyrosine (Upstate Biotechnology). Recombinant C3 exotransferase was a generous gift from Dr. Tracee Panetti, University of Wisconsin (Madison, WI).
Cell Lysis, Immunoprecipitation, and Immunoblotting-For immunoprecipitation of EGF or PDGF receptors, cells grown to 80 -90% confluency were starved in medium without serum for 16 h, released from the substrate, reseeded onto plates coated with 100 g/ml gelatin, incubated for 3 h in medium without serum, and then stimulated with agonists for 5 min or left unstimulated. For Western blotting of phosphorylated Erk1/2, cells were similarly seeded onto gelatin-coated plates, starved in serum-free medium for 3 h, and then stimulated with agonist for various time perioids. Cells were lysed on ice in buffer containing 1% (v/v) Triton X-100, 150 mM NaCl, 5 mM EDTA, 100 mM sodium fluoride,1 mM sodium orthovanadate, 0.5 mM sodium molybdate, 2 mM PMSF, 5 g/ml leupeptin, 0.1 g/ml pepstatin A, 0.4 mM pefabloc SC, and 20 mM Tris-HCl, pH 7.4. The same amounts of protein from different experimental samples were used for analyses, as determined using a BCA protein assay (Pierce). The proteins were run on SDS-polyacrylamide gel electrophoresis under reducing conditions. Immunoprecipitation analysis was performed as described previously elsewhere (42), with a slight modification. Briefly, the supernatants were precleaned with protein A-Sepharose 4 Fast-Flow (Pharmacia LKB Biotech, Sweden) and subsequently incubated with antibody. The complexes were precipitated with protein A-Sepharose 4 Fast-Flow, and the proteins were eluted from the resins by incubation with SDS-sample buffer. Samples were then subjected to SDS-polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membranes. For immunoblotting, the blots were probed with the primary antibody, then with a horseradish peroxidase-conjugated secondary antibody (Organon Teknika Corp., Westchester, PA; Promega, Madison, WI). Immunoreactive bands were developed using the enhanced chemiluminescence (ECL) substrate system (NEN Life Science Products).
Immunoprecipitation of EGF or PDGF receptors followed by immunoblotting for phosphotyrosine or ␤ 1 integrin was also carried out on cells lysed 60 min after plating on gelatin (100 g/ml), vitronectin, fibronectin, or laminin-1 (10 g/ml) coated substratum, or left in suspension.

LPA Sensitizes Cells to Growth Factors and Increases ␤ 1 Integrin-mediated Cell Migration
The effect of LPA in combination with EGF or PDGF on directed cell migration through gelatin-coated filters was analyzed with GD25 cells expressing wild type ␤ 1 A (Fig. 1). The cells did not migrate in response to EGF alone. PDGF alone stimulated modest migration at a concentration Ͼ1 ng/ml. 500 nM LPA alone stimulated robust migration. When EGF or PDGF were mixed together with LPA, cell migration was enhanced further by EGF or PDGF concentrations as low as 0.03 ng/ml (Fig. 1). The concentration of PDGF that supported maximal migration in the presence of both PDGF and LPA was 3 ng/ml as compared with Ͼ30 ng/ml in the absence of LPA.
The requirement for ␤ 1 A for migration of GD25 cells through

FIG. 1. Cell migration through gelatin-coated filters of GD25 cells expressing wild type ␤ 1 A in response to
LPA and/or EGF or PDGF. Migration was quantified for EGF alone (Ⅺ), PDGF alone (‚), LPA and EGF (f), and LPA and PDGF (OE). LPA (500 nM) and/or EGF or PDGF in concentrations from 0.003 to 30 ng/ml was added in the lower chamber, and the dose dependence of the growth factor was analyzed. Each symbol represents the mean of cell number/0.16-mm 2 field. Brackets indicate mean Ϯ S.D. of quadruplicate determinations.
gelatin-coated filters is more stringent when the chemoattractant is LPA than when the chemoattractant is PDGF (25). The effects of ␤ 1 A integrin on cell migration induced by LPA in combination with EGF or PDGF were therefore investigated. When EGF and LPA were mixed together, cell migration was 2-3-fold greater for cells expressing ␤ 1 A than for cells lacking ␤ 1 A ( Fig. 2A). When PDGF and LPA were mixed together, the number of migrating cells was 1.2-1.5-fold greater for cells expressing wild type ␤ 1 A than for cells lacking ␤ 1 A (Fig. 2B). S1P, a phospholipid that, like LPA, reacts with Edg receptors (28,29), is unlike LPA in that it inhibits cell migration (43). S1P did not induce migration through gelatin-coated filters by either cell type and inhibited LPA-induced migration in cells expressing ␤ 1 A (Fig. 2C). These findings indicate that ␤ 1 A-dependent migration in response to LPA is enhanced by EGF or PDGF and attenuated by S1P. ␤ 1 A-dependent migration in response to LPA (25) or PDGF or EGF (39) is lacking in GD25 cells expressing ␤ 1 As with certain mutations in the cytoplasmic domain. These mutants fell into two groups, active (Y783F, Y795F, and Y783/795F) and inactive (T788P, P781A, P793A, and P781/793A), based upon the reactivity with anti-␤ 1 (9EG7) antibody, ␣ 6 ␤ 1 -dependent adhesion to laminin-1, and ability to support fibronectin assembly (25). LPA and EGF stimulated the same minimal migration of GD25 cells lacking ␤ 1 A and of GD25 cells expressing ␤ 1 A with the T788P, P781/793A, Y783/795F, Y783F, or Y795F mutations (Fig. 3). LPA and PDGF were less effective in causing migration of T788P, P781/793A, Y83/795F, Y783F, or Y795F cells than of GD25 cells lacking ␤ 1 A (Fig. 3). When LPA and EGF or PDGF were tested together on cells expressing ␤ 1 A with the activating D759A mutation, migration was similar to cells expressing wild type ␤ 1 A (Fig. 3).
We previously demonstrated that the magnitudes of the chemotactic response to EGF or PDGF of GD25 cells expressing wild type ␤ 1 A were 2-3-fold greater than the chemokinetic response (39), as opposed to the 1.2-1.5-fold difference seen with LPA (25). To learn if the enhancing effect of LPA on EGFor PDGF-induced migration is due to increased chemotaxis or chemokinesis in cells expressing wild type ␤ 1 A, we studied the effect of adding LPA with cells in the upper chamber or in both the upper and lower chambers (Fig. 4). When LPA was present at equal concentrations in both chambers, the synergistic effect of LPA on EGF-or PDGF-induced migration was 90 -95% of that observed when LPA and EGF or PDGF were present in just the lower chamber. When LPA was in the upper chamber and EGF or PDGF was in the lower chamber, more cells migrated toward the lower chamber than in the complete absence of LPA but less than when LPA alone was in the lower chamber. The same results were observed with filters coated with gelatin, vitronectin, or fibronectin (Fig. 4). These results indicate that LPA causes increased migration when mixed with EGF or PDGF mainly by a chemokinetic effect with some contribution of chemotaxis.

Signaling Pathways Induced by LPA and/or EGF or PDGF
Lack of Effect of LPA or ␤ 1 A-mediated Adhesion on Phosphorylation of EGF or PDGF Receptors-Treatment of Rat-1 fibroblasts, HaCaT keratinocytes, or COS-7 cells with high concentration of LPA (10 -25 M) causes phosphorylation of EGF receptors (44,45). GD25 cells lacking ␤ 1 A or expressing ␤ 1 A expressed comparable amounts of EGF and PDGF receptors when analyzed by SDS-polyacrylamide gel electrophoresis and immunoblotting, and both receptors in both cells were activated appropriately by concentrations of 3 ng/ml EGF or PDGF but not by 500 nM LPA when analyzed by anti-phosphotyrosine immunoblotting of immunoprecipitated receptors (25). We tested a wider range of growth factor concentrations to learn whether LPA sensitizes EGF or PDGF receptors to activation by low concentrations of EGF or PDGF. The amount of EGF or PDGF receptor phosphorylation in cells expressing wild type ␤ 1 A as assessed by anti-phosphotyrosine immunoblotting of immunoprecipitated receptors was increased in response to EGF or PDGF in a concentration-dependent manner (Fig. 5). When EGF or PDGF was mixed with 500 nM LPA together, there was no enhancement of EGF or PDGF receptor phosphorylation when compared with the stimulation by EGF or PDGF alone (Fig. 5). When 20 M LPA alone was added, no tyrosine phosphorylation of EGF or PDGF receptors was demonstrated (not shown). These experiments, therefore, indicate that GD25 cells do not respond to LPA with transactivation of EGF receptors like Rat-1, HaCaT, or COS-7 cells. We also compared the phosphorylation states of EGF and PDGF receptors in ␤ 1 GD25 cells adherent for 60 min to vitronectin, fibronectin, or laminin-1-coated surfaces. No phosphorylation was noted over and above the baseline phosphorylation present in suspended cells. In addition, no ␤ 1 A was co-immunoprecipitated with EGF or PDGF receptors (not shown).
Activation of Erk1/2 Stimulated by LPA and/or EGF or PDGF-LPA signaling is mediated through Ras and known to result in activation of MAPK (15,16). We therefore studied whether cell migration in response to LPA and EGF or PDGF is associated with the hyperactivation of MAPK in GD25 cells wild type ␤ 1 A. To replicate conditions of the cell migration assay, cells were incubated for 3 h without serum on a gelatincoated substratum and then stimulated with agonists for various time periods. Doubly phosphorylated Erk1 and Erk2, which represent the activated forms of these proteins, were analyzed using anti-active MAPK antibodies. Basal activation of Erk1/2 was minimal (Fig. 6). In response to the stimulation of LPA, EGF, or PDGF alone, activation of Erk1/2 was increased in a concentration-dependent manner (Fig. 6). When LPA and EGF or PDGF were added together, only modest additive effects of Erk1/2 activation were demonstrated when compared with LPA, EGF, or PDGF alone (Fig. 6). This finding was true after both 5 and 30 min of stimulation.
Correlation Between Cell Migration and Ras-or Rho-mediated Signaling Pathway-Treatment of GD25 cells expressing ␤ 1 A with pertussis toxin, which modifies G i and blocks stimulation of Ras induced by LPA (20), caused nearly complete loss of migration through gelatin-coated filters in response to LPA alone at a toxin dose of 3 ng/ml (Fig. 7A). The 50% inhibitory dose of pertussis toxin, determined in a separate experiment, was 0.01-0.03 ng/ml (not shown). When LPA and EGF or PDGF were added together to cells treated with pertussis toxin, enhancement by LPA of growth factor-induced migration was completely lost at 3 ng/ml toxin, such that the number of migrated cells were the same as in response to PDGF or EGF alone (Fig. 7A). Treatment of pertussis toxin caused downregulation of Erk1/2 activation in response to LPA but did not significantly affect Erk1/2 activation in response to EGF or PDGF alone or to a combination of LPA and EGF or PDGF (Fig. 8).
LPA signaling through G ␣ 12/13 causes direct activation of Rho (46). Migration of cells expressing wild type ␤ 1 A in response to LPA alone or PDGF alone was decreased by the treatment of cells with C3 exotransferase, which inactivates Rho in a concentration-dependent manner up to exotransferase doses of 100 g/ml (Fig. 7B). When LPA and EGF or PDGF were added together to cells treated with C3 exotransferase, the enhanced effect of EGF or PDGF on LPA-induced migration was lost at an exotransferase dose as low as 10 g/ml (Fig. 7B). The treatment of C3 exotransferase did not block LPA-stimulated Erk1/2 activation at any of the toxin concentrations tested (not shown). DISCUSSION ␤ 1 A with an intact cytoplasmic tail, including intact NPXY motifs, is important for chemotaxis of GD25 fibroblasts in response to EGF or PDGF (39). The requirement for ␤ 1 A is more stringent when the chemotactic agent is LPA than when the chemotactic agent is EGF or PDGF (25). In this study, we describe a tripartite interaction among ligated ␤ 1 A integrins, LPA GPCRs, and EGF or PDGF RTKs. Although signaling induced by two of the three receptor systems supports chemotaxis, engagement of all three receptor systems leads to greater migration. The major effects of LPA when combined with EGF or PDGF were to sensitize cells so that maximal responses were obtained with Ͼ10-fold lower concentrations of growth factor and to increase the chemokinetic component of migration. The synergism between LPA and EGF or PDGF was lost when cells were preincubated with pertussis toxin or C3 exotransferase. S1P mimics some of the effects of LPA (47)(48)(49)(50), and both lipids signal via Edg receptors. GD25 cells lacking ␤ 1 A and expressing wild type ␤ 1 A have been shown to express edg-2 and edg-1 receptors by reverse transcription-polymerase chain reaction (25). In contrast to the effect of LPA, neither cell type was induced to migrate by S1P. S1P inhibits invasion and motility of melanoma cells and PDGF-induced chemotaxis of human smooth muscle cells, whereas LPA induces invasion of rat hepatoma cells and migration of NIH 3T3 or human skin fibroblasts (22,32,43,51,52). Similarily S1P inhibited LPAinduced migration of ␤ 1 A-GD25 cells. Presumably, the two bioactive lipids activate a different balance of signaling pathways, thus leading to different effects on cellular motility.
The results indicate that the ␤ 1 A integrin must not only be active but have intact NPXY motifs in the cytoplasmic domain. The latter requirement differentiates the LPA effect on migration from its effect on fibronectin matrix assembly, which is up-regulated by LPA in GD25 cells expressing ␤ 1 A with con-FIG. 5. Dose-response analysis of tyrosine phosphorylation of EGF or PDGF receptor stimulated by EGF or PDGF and/or LPA in GD25 cells expressing wild type ␤ 1 A. Serum-starved cells were seeded onto plates coated with 100 g/ml gelatin, incubated for 3 h in medium without serum, and then stimulated with agonists for 5 min as indicated or left unstimulated (Ϫ). The EGF or PDGF receptor was immunoprecipitated from cell lysates and analyzed by immunoblotting with anti-phosphotyrosine antibody 4G10 (upper panels). The same samples were analyzed by immunoblotting with anti-EGF or anti-PDGF antibody (lower panels).
FIG. 6. Dose-response analysis of phosphorylation of Erk1/2 stimulated by EGF or PDGF and/or LPA in GD25 cells expressing wild type ␤ 1 A. Cells were seeded onto plates coated with 100 g/ml gelatin and starved in medium without serum for 3 h and then stimulated with agonists for various time periods or left unstimulated (Ϫ). Cell lysates were analyzed by immunoblotting with anti-active MAPK polyclonal antibody.
servative Tyr to Phe substitutions (39). An unexpected finding of the present experiments was the deleterious effect of mutations of the NPXY motifs on migration due to LPA and PDGF when compared with migration of cells lacking ␤ 1 A completely.
This result suggests that the mutant ␤ 1 A tails depress migration mediated by other cell surface receptors.
Recent evidence suggests at least three mechanisms of "cross-talk" among signaling pathways. First, there is evidence of cross-talk between GPCRs and RTKs. LPA and other GPCR agonists can induce activation of EGF or PDGF receptors in the absence of EGF or PDGF with subsequent activation of the Ras-MAPK cascade (44,45,(53)(54)(55). G ␣ 13-mediated Rho activation induced by LPA has also been suggested to stimulate EGF RTK activity in Swiss 3T3 cells (56). Second, there is evidence of cross-talk between RTKs and ␤ 1 integrins. Stimulation of ␤ 1 integrin on human AG1518 fibroblasts by interaction with extracellular matrix induces ligand-independent tyrosine phosphorylation of PDGF receptors, but not of EGF receptors (38). Similarly, adhesion of human skin fibroblasts, endothelial cells to matrix proteins, or antibody to ␤ 1 integrin stimulates ligand-independent tyrosine phosphorylation of EGF receptors in the absence of receptor ligands and association of ␤ 1 integrins with EGF receptors (57). Evidence of crosstalk between GPCRs and EGF or PDGF receptors appears to be cell type-specific and dependent, e.g. on the presence or absence of cell surface EGF or PDGF receptors (55). We found no LPA-induced phosphorylation of EGF and PDGF receptors in GD25 cells despite good expression of both receptors and thus have no evidence for transactivation of receptor tyrosine kinases through GPCRs in these cells. Further, we found no evidence for adhesion-dependent phosphorylation of EGF or PDGF receptors. However, we cannot exclude the possibility that LPA treatment or ␤ 1 -dependent adhesion caused activation of an undetectably small but still functionally significant subpopulation of receptors.
A third type of cross-talk is between GPCRs and adhesion receptors. Activation of Rho by LPA stimulation results in reorganization of actin stress fibers, recruitment or phosphorylation of different focal contact proteins including integrins, and formation of focal contacts (58). PDGF or EGF initially activate Rac, which stimulates membrane ruffling, and also leads to Rho-dependent responses (59). Stimulation of multiple coordinated pathways of actin remodeling and coupling of remodeled actin to cell surface complexes would explain the increased migration that we observed. Pretreatment of GD25 cells expressing wild type ␤ 1 A with pertussis toxin ablated both LPA-stimulated Erk1/2 activation and LPA-induced cell migration, whereas C3 exotransferase pretreatment partially inhibited LPA-induced cell migration but did not influence MAPK activation. These results indicate that MAPK activation by LPA in GD25 cells is a G i -Ras-Raf-mediated pathway, but that the G i -Ras-Raf pathway alone is not sufficient for cell migration induced by LPA. Importantly, LPA did not sensitize cells pretreated with pertussis toxin or C3 exotransferase to growth factors. This finding suggests the potential participation of Rho in pathways leading to Ras and Erk activation (60). Thus, both FIG. 7. Effect of pertussis toxin (A) or C3 exotransferase (B) on LPA-and/or EGF-or PDGF-induced cell migration through gelatin-coated filters of GD25 cells expressing wild type ␤ 1 A. Cells were pretreated overnight with pertussis toxin or C3 exotransferase at the concentrations of 0, 3, 10, 30, and 100 ng/ml pertussis toxin, or 0, 3, 10, 30, and 100 g/ml (C3 exotransferase). LPA (500 nM) and/or EGF or PDGF (3 ng/ml) were in the lower chamber. Migration was then quantified in response to LPA alone (E), EGF alone (Ⅺ), PDGF alone (‚), LPA and EGF (f), and LPA and PDGF (OE). Each symbol represents the mean of cell number/0.16-mm 2 field. Brackets indicate mean Ϯ S.D. of quadruplicate determinations.
FIG. 8. Effect of pertussis toxin on LPA-and/or EGF-or PDGF-induced phosphorylation of Erk1/2 in GD25 cells expressing wild type ␤ 1 A. Cells were pretreated overnight with pertussis toxin at concentrations of 0, 3, 10, 30, and 100 ng/ml, seeded onto plates coated with 100 g/ml gelatin, starved in medium without serum for 3 h, and then stimulated with agonists (LPA (500 nM) and/or EGF or PDGF (3 ng/ml)) for 5 min or left unstimulated (Ϫ). Cell lysates were analyzed by immunoblotting with anti-active MAPK polyclonal antibody. G i -Ras-Raf and G ␣ 12/13-Rho mediated pathways seem critical for the sensitization induced by LPA.
The phenomenon described herein may be of considerable importance for the migration of one group of cells in response to a growth factor diffusing from a second distant group of cells. The sensitivity and specificity of such communication would be enhanced considerably if the first group of cells was sensitized to the product of the second group.