Role of the Sphingosine-1-Phosphate Receptor EDG-1 in PDGF-Induced Cell Motility

See allHide authors and affiliations

Science  02 Mar 2001:
Vol. 291, Issue 5509, pp. 1800-1803
DOI: 10.1126/science.1057559


EDG-1 is a heterotrimeric guanine nucleotide binding protein–coupled receptor (GPCR) for sphingosine-1-phosphate (SPP). Cell migration toward platelet-derived growth factor (PDGF), which stimulates sphingosine kinase and increases intracellular SPP, was dependent on expression of EDG-1. Deletion of edg-1 or inhibition of sphingosine kinase suppressed chemotaxis toward PDGF and also activation of the small guanosine triphosphatase Rac, which is essential for protrusion of lamellipodia and forward movement. Moreover, PDGF activated EDG-1, as measured by translocation of β-arrestin and phosphorylation of EDG-1. Our results reveal a role for receptor cross-communication in which activation of a GPCR by a receptor tyrosine kinase is critical for cell motility.

Interest in SPP has accelerated recently with the discovery that it is the extracellular ligand for EDG-1, EDG-3, EDG-5, EDG-6, and EDG-8 (1). Although the biological functions of these GPCRs have not been completely elucidated, EDG-1 is implicated in cell migration, angiogenesis, and vascular maturation (2–4). Disruption of theedg-1 gene by homologous recombination in mice resulted in massive intra-embryonic hemorrhaging and intrauterine death caused by incomplete vascular maturation resulting from a failure of mural cells—vascular smooth muscle cells and pericytes—to migrate to arteries and capillaries and to reinforce them properly (4). Disruption of the PDGF-BBor PDGFR-β genes in mice resulted in a similar lethal phenotype (5, 6). Because in many different cell types, PDGF stimulates sphingosine kinase, leading to an accumulation of intracellular SPP (1, 7), we speculated that interplay between PDGF and SPP–EDG-1 signals might be required for cell migratory responses. In this study, we found that activation of EDG-1 by the PDGFR plays a crucial role in regulating cell motility. The results reveal a new paradigm for communication between tyrosine kinase receptors and GPCRs.

Human embryonic kidney (HEK) 293 cells, which only express EDG-3 and EDG-5, did not migrate toward SPP unless EDG-1 was expressed (2). EDG-1 overexpression also stimulated migration of HEK 293 cells toward PDGF-BB (Fig. 1A), whereas migratory responses to serum and fibronectin were unaffected (Fig. 1A). Conversely, migration of mouse embryonic fibroblasts (MEFs), which express transcripts for EDG-1, EDG-3, and EDG-5, but not EDG-6 or EDG-8 (4), toward PDGF-BB was reduced whenedg-1 was deleted (Fig. 1B). A smaller effect on migration toward serum was observed in these mutant fibroblasts, and migration toward fibronectin was unaffected (Fig. 1B), indicating thatedg-1 deletion does not disrupt all essential mechanisms of directed cell movement.

Figure 1

Requirement of EDG-1 for PDGF-induced migration. (A) Enhanced chemotaxis toward PDGF in cells overexpressing EDG-1. HEK 293 cells stably transfected with vector (white bars) or EDG-1 (filled bars) were allowed to migrate toward PDGF-BB (20 ng/ml), SPP (100 nM), fetal bovine serum (FBS, 20%), or fibronectin (FN, 10 μg/ml). Chemotaxis was measured in a modified Boyden chamber assay (2). The average number of cells in four random fields was determined and is presented as the average ± SD of three individual wells. Similar results were obtained in at least three independent experiments, and statistically different groups are indicated by asterisk (P < 0.05 by analysis of variance). (B) Chemotaxis toward PDGF is markedly reduced byedg-1 disruption. Wild-type (white bars) and EDG-1–/– (filled bars) MEFs were allowed to migrate toward PDGF-BB, SPP, serum, or FN as in (A). (C) Dependence of chemotaxis of ASMCs toward PDGF and SPP on EDG-1 expression. Human ASMCs were transfected without (white bars) or with 400 nM phosphothioate oligonucleotides [EDG-1 sense (filled bars), 5′-ATG GGG CCC ACC AGC GTC-3′; EDG-1 antisense (gray bars), 5′-GAC GCT GGT GGG CCC CAT-3′] using Lipofectamine (Life Technologies, Gaithersburg, Maryland) (3) and after 24 hours, were allowed to migrate toward SPP (100 nM), PDGF (20 ng/ml), or FBS (20%). Analysis by reverse transcriptase–polymerase chain reaction showed that EDG-1 mRNA expression was almost completely eliminated without affecting EDG-3 and EDG-5 expression.

Because the migration of smooth muscle cells appears to be aberrant in EDG-1 knockout mice (4), we also examined the role of EDG-1 in PDGF-directed migration of human aortic smooth muscle cells (ASMCs). Reduction of EDG-1 expression in ASMCs (which endogenously express EDG-1, EDG-3, and EDG-5) by EDG-1 antisense phosphothioate oligonucleotide (3), not only eliminated migration toward SPP but also reduced migration toward PDGF, but not serum (Fig. 1C). These results suggest that the loss of EDG-1 results in motility defects toward PDGF in diverse cell types. Dysfunctional migration of EDG-1–/– cells toward PDGF links this phenotype (4) to the PDGF-BB andPDGFR-β knockout phenotypes (5,6) at the final steps of vasculogenesis (4), underscoring the importance of cross-communication between PDGFR and EDG-1 in vascular maturation.

As it does in many other cell types (1,7), PDGF-BB stimulated sphingosine kinase activity in wild-type MEFs and had an even greater stimulatory effect in fibroblasts in which edg-1 was deleted (Fig. 2A). To investigate whether SPP generated in response to PDGF might be involved in PDGF-mediated chemotaxis, we used N,N-dimethylsphingosine (DMS), a competitive inhibitor of sphingosine kinase (8). DMS inhibited PDGF-directed chemotaxis of wild-type MEFs (Fig. 2B) but did not reduce PDGF-stimulated receptor tyrosine phosphorylation (Fig. 2C). In agreement with its inability to interfere with binding of SPP to EDG-1 and its activation (9), DMS did not significantly affect chemotaxis of cells toward a gradient of SPP. Similarly, DMS also blocked formation of SPP and inhibited PDGF-directed chemotaxis of HEK 293 cells overexpressing EDG-1 (Fig. 2D). As these results indicate potential cross-talk between PDGF and EDG-1 signaling, and EDG-1 is mainly coupled to Gi(9), cells were pretreated with pertussis toxin to inactivate Gi, which suppressed PDGF-induced chemotaxis of both wild-type MEFs (Fig. 2B) and HEK 293 cells overexpressing EDG-1 (Fig. 2D).

Figure 2

Requirement of sphingosine kinase for PDGF-induced motility. (A) Stimulation of sphingosine kinase by PDGF in wild-type and EDG-1–/– MEFs. Wild-type and EDG-1–/– MEFs were treated with PDGF-BB (20 ng/ml) for the indicated times, and sphingosine kinase activity was measured. Basal activity was 114 ± 5 and 133 ± 2 pmol/min per milligram for wild-type and EDG-1–/– fibroblasts, respectively. (B and D) Inhibition of chemotaxis toward PDGF by the sphingosine kinase inhibitor DMS or pertussis toxin. Wild-type MEFs (B) or HEK 293 cells transfected with vector or EDG-1 (D) were treated without (control) or with pertussis toxin (PTX, 200 ng/ml, 2 hours), or DMS (10 μM, 20 min), then allowed to migrate toward the indicated concentrations of PDGF-BB, and chemotaxis was measured. (C) Lack of effect of DMS on tyrosine phosphorylation induced by PDGF. Wild-type MEFs were serum-starved for 24 hours, and then treated without or with DMS (10 μM) for 20 min, before stimulation with PDGF-BB (60 ng/ml) for 5 min. Equal amounts of cell-lysate proteins were analyzed by Western blotting using antibody against phosphotyrosine.

Thus, EDG-1 appears to be necessary for PDGF-mediated chemotaxis. Therefore, we determined whether PDGF signaling might activate EDG-1 by regulating sphingosine kinase activity and accumulation of SPP, which in turn activates EDG-1. β-Arrestins are cytosolic proteins that bind with high affinity to agonist-activated, phosphorylated GPCRs to terminate receptor to G protein coupling. They also mediate receptor endocytosis (10, 11) and initiation of a second wave of signaling (12, 13). SPP promoted rapid redistribution of β-arrestin2 tagged with green fluorescent protein (GFP) from the cytoplasm to the plasma membrane only in EDG-1–expressing HEK 293 cells (Fig. 3A, part b). Treatment of cells overexpressing both EDG-1 and sphingosine kinase type 1 (SPHK1) with PDGF and sphingosine increased intracellular SPP and induced translocation of β-arrestin2 to the plasma membrane (Fig. 3B, part c), as did exposure of cells to exogenous SPP (Fig. 3B, part b). Availability of sphingosine is a limiting factor that influences levels of cellularly generated SPP (14). Thus, sphingosine-induced translocation of β-arrestin was dependent on overexpression of SPHK1 (note the lack of translocation in Fig. 3A, part c). Treatment with PDGF or sphingosine alone, or transfection with SPHK1 stimulated production of SPP by 2-, 6-, and 10-fold, respectively (Fig. 3C), but these concentrations of SPP did not result in significant translocation of β-arrestin. Treatment of SPHK1-expressing HEK 293 cells with a high concentration of sphingosine (Fig. 3B, part d) or sphingosine together with PDGF (Fig. 3B, part c), which increased SPP levels by 60- and 30-fold, respectively, resulted in translocation of β-arrestin to activated EDG-1. To determine the effect of PDGF alone, we used a more sensitive assay of GPCR activation because of the observation that activated GPCRs become phosphorylated before β-arrestin binding (10,12). To enhance sensitivity of detection, HEK 293 cells were cotransfected with expression plasmids encoding Flag epitope–tagged EDG-1 and PDGFR-β, labeled in situ with intracellular phosphate-32, and EDG-1 was immunoprecipitated with antibody against Flag. As did SPP (15), PDGF increased phosphorylation of EDG-1 in these cells (Fig. 3D), whereas no phosphorylation could be detected in vector-transfected cells (16).

Figure 3

SPP produced intracellularly can act in an autocrine or paracrine fashion to activate EDG-1. (A) β-Arrestin translocation in HEK 293 cells cotransfected with expression plasmids encoding β-arrestin2–GFP and EDG-1. β-Arrestin2–GFP fluorescence was visualized (10) after treatment with vehicle (a), 100 nM SPP (b), or 5 μM sphingosine (c). (B) Translocation of β-arrestin in HEK 293 cells cotransfected with β-arrestin2–GFP, EDG-1, and SPHK1. β-Arrestin2–GFP fluorescence was visualized after treatment with vehicle (a), 100 nM SPP (b), 2.5 μM sphingosine and 20 ng/ml PDGF-BB (c), or 5 μM sphingosine (d). (C) Cellular levels of SPP. Levels of SPP were measured in vector-transfected (white bar) and in SPHK1-transfected (filled bars) HEK 293 cells after treatment with 2.5 μM sphingosine, 20 ng/ml PDGF-BB, or both for 10 min (20). (D) PDGF-induced EDG-1 phosphorylation. Vector or Flag–EDG-1–expressing HEK 293 cells were transfected with PDGFR-β and cultured in 10% charcoal-stripped FBS for 24 hours, metabolically labeled in phosphate-free Dulbecco's modified Eagle's medium with [32P]orthophosphate (70 μCi/ml) for 2.5 hours at 37°C, then stimulated with SPP (100 nM) or PDGF (20 ng/ml). Cell lysates were prepared and immunoprecipitated with antibody against Flag M2 (Sigma) as described (15). Immunoprecipitates were either separated on 10% SDS-PAGE, transblotted to nitrocellulose, and autoradiographed (upper panel) or immunoblotted with antibody against Flag (lower panel). (E) β-Arrestin translocation in HEK 293 cells cotransfected with β-arrestin2–GFP, EDG-1, and empty vector for SPHK1 (green cells). Conditioned medium from β-arrestin2–RFP, EDG-1, and SPHK1 transfectants (red cells) that had been treated with 5 μM sphingosine to generate intracellular SPP (a) or the cells themselves (b and c) were added to “green” cells. β-Arrestin2–GFP and β-arrestin2–RFP fluorescence was visualized by using dual excitation (488 and 568 nm) and emission (515 to 540 nm, GFP; 590 to 610 nm, RFP) filter sets (10).

Although these results suggest that endogenously generated SPP can activate EDG-1 and consequent translocation of β-arrestin, no significant release of SPP into the extracellular medium could be detected by mass measurements (<0.4 nM), even after treatment of SPHK1-expressing HEK 293 cells with PDGF-BB and sphingosine to increase SPP. To examine the possibility that amounts of SPP below our mass detection limits may in fact be secreted into the vicinity of EDG-1 and may activate it, SPHK1-expressing cells were transfected with β-arrestin2 fused to red fluorescent protein (RFP) at the NH2-terminus (red) to differentiate them from β-arrestin2–GFP (green) transfectants. The β-arrestin2–RFP transfectants were treated with 5 μM sphingosine to increase intracellular SPP maximally. Conditioned medium from these SPP-producing cells did not induce translocation of β-arrestin2–GFP in EDG-1–transfected cells (Fig. 3E, part a). Nevertheless, coculturing of “red” SPP-producing cells with “green” EDG-1–transfected cells induced translocation of β-arrestin2–GFP to the plasma membrane on adjacent cells (Fig. 3E, part b) and on distant cells (Fig. 3E, part c). These results suggest that endogenously generated SPP can activate EDG-1 in an autocrine or paracrine manner.

Although deletion of edg-1 or uncoupling Gi inhibited PDGF-directed motility, there were no significant differences in PDGFR expression or PDGF-stimulated receptor tyrosine phosphorylation in wild-type compared with EDG-1 null fibroblasts (Fig. 4A). Rac, a member of the Rho family of small guanosine triphosphatases (Rac, Cdc42, and Rho), plays a critical role in cell motility by regulating formation of new lamellipodial protrusions at the leading edge (17). PDGF-BB rapidly activated Rac (Fig. 4B), but not Cdc42, in wild-type fibroblasts. Deletion of edg-1 or inhibition of sphingosine kinase in wild-type MEFs decreased Rac activation induced by PDGF. These results suggest that Rac may participate in integration of PDGFR and EDG-1 signaling to promote cell migration and that the SPP signaling pathway may be important to amplify activation of Rac.

Figure 4

Effect of edg-1 deletion on PDGF signaling. (A) Deletion of edg-1 has no effect on PDGF-induced tyrosine phosphorylation of PDGFR. Wild-type and EDG-1–/– MEFs were serum-starved for 24 hours, and then treated without or with PDGF-BB (20 ng/ml) for 5 min. Equal amounts of cell-lysate proteins were analyzed by Western blotting with antibody against phosphotyrosine. Blots were then stripped and reprobed with polyclonal antibody against PDGFR (Upstate Biotechnology, Lake Placid, New York). (B) Deletion ofedg-1 or inhibition of sphingosine kinase diminishes PDGF-mediated Rac activation. Wild-type and EDG-1–/– MEFs were treated with PDGF-BB (50 ng/ml) for the indicated times in the absence or presence of pretreatment with DMS (20 μM) for 20 min as indicated. Cell lysates were incubated with immobilized PAK-1 binding domain (Upstate Biotechnology) and associated GTP-Rac was determined by Western blotting using a specific Rac antibody or used without affinity immunoprecipitation to determine total Rac levels as shown below (4). (C) Activation of EDG-1 by PDGF. Scheme for intracellular communication between tyrosine kinase growth factor receptor (PDGFR) and GPCR (EDG-1) signaling pathways.

Various agonists for GPCRs can activate growth factor tyrosine kinase receptors in the absence of added growth factors (18). Although this type of cross-communication is important for regulation of cell growth (18), our results suggest that cell motility is regulated by a reciprocal mechanism of receptor cross-talk. Thus, a tantalizing notion is that spatially and temporally localized generation of SPP by activation of sphingosine kinase in response to PDGF results in restricted activation of the GPCR EDG-1 that in turn activates Rac (Fig. 4C). Rac may then amplify the initial receptor signals (19), thus creating a positive feedback loop at the leading edge of the cell.

  • * These authors contributed equally to this report.

  • To whom correspondence should be addressed. E-mail: spiegel{at}


View Abstract

Stay Connected to Science

Navigate This Article