Extracellular Nucleotides Induce Arterial Smooth Muscle Cell Migration Via Osteopontin

Several studies suggest that migration and proliferation of arterial smooth muscle cells (SMCs) play a prominent role in the development of atherosclerotic plaques and restenosis lesions.^1,2 Although SMC proliferation is an important feature in experimental arterial injury models, however, only few proliferating SMCs have been detected in human primary or secondary atherosclerotic plaques,^3,4 thus underlining the prominent role of the migration process in these pathological conditions.

SMC proliferation and migration are the result of multifactorial stimulation. Many growth-regulatory molecules and cytokines have been described in atherosclerotic plaques.⁵ Extracellular nucleotides have also recently been shown to be involved in SMC growth. Extracellular purine and pyrimidine nucleotides induce cell cycle progression and elicit growth of cultured vascular SMCs.^6–8 This mitogenic response involves the nucleotide binding to G protein–coupled P2Y receptor subtypes,^9–11 including the P2Y₂ receptor, which is activated by UTP and ATP.¹² Furthermore, the P2Y₂ receptor is upregulated in cytokine-stimulated SMCs¹³ and in rat aorta after balloon injury,¹⁴ suggesting that extracellular ATP and UTP could play a critical role in intimal hyperplasia or vascular remodeling. Vascular P2Y receptors are activated in an autocrine or paracrine manner by nucleotides that are released in the vascular wall from perivascular nerves, activated platelets, and mechanically stretched cells.^15–17

Because many mitogenic compounds for SMCs also demonstrate a chemoattractant activity,⁵ extracellular nucleotides could also exert a chemotactic effect on arterial SMCs. A potent role for extracellular nucleotides in SMC migration has also been suggested because it was previously shown that expression of the chemotactic protein osteopontin (OPN) is induced by ATP and UTP in cultured SMCs.^10,18 OPN is an RGD-containing extracellular matrix (ECM) protein involved in cell attachment and migration. Its activity necessitates its binding to integrin receptors or CD44.^19,20 It is detected in association with SMCs and macrophages of atherosclerotic plaques, and its expression is upregulated in neointimal thickening induced by balloon angioplasty of rat vessel.^21–23

The downstream signaling events after SMC exposure to extracellular ATP and UTP have not been fully established. The P2Y receptors are coupled to phospholipase C, IP₃ formation, and cytosolic Ca²⁺ concentration increase. The UTP-induced [Ca²⁺]_i increase is involved in mitogen-activated protein kinase (MAPK) (extracellular signal–regulated kinase [ERK]1/ERK2) phosphorylation in SMCs and other cell types.^11,24–26 A recent work demonstrates that SMC P2Y receptors are coupled to activation of the small GTPase RhoA.²⁷ Moreover, rho proteins (RhoA, Rac, and Cdc42) have recently been demonstrated in SMC migration and in ERK pathway activation.^28,29

The aim of the present study was to investigate the role of nucleotides in SMC migration and to determine the cascade of molecular events leading from receptor activation to SMC migration.

Materials and Methods

Cell Culture

Rat aortic SMCs were prepared from thoracic aortas of Wistar rats as previously described.⁷ Cells were cultured in DMEM containing 5% FCS, 150 mmol/L HEPES, 100 U/mL penicillin, and 100 μg/mL streptomycin (Life Technologies). SMCs from passages 6 to 12 were used.

Migration Assay

Cell migration was performed with the Transwell (Costar) system, which allows cells to migrate through 8-μm pore size polycarbonate membrane. Briefly, cells were trypsinized, washed, and resuspended in serum-free DMEM (5×10⁵ cells/mL). This suspension (100 μL) was added to the upper chamber of Transwells. The lower chamber was filled with 600 μL serum-free DMEM containing nucleotides (ATP, ADP, UTP, UDP, Sigma-Aldrich) in appropriate concentrations or not containing nucleotides. For checkerboard analysis, appropriate concentrations of nucleotides were added in either the lower, upper, or both chambers. After a 6-hour stimulation by nucleotides, filters were removed, and cells remaining on the upper surface of the membrane (ie, that had not migrated through the filter) were removed with a cotton swab. Then, membranes were washed with PBS, and cells present beneath the membrane were fixed with cold methanol for 15 minutes and stained with Hemalun. Cells were counted in 10 high-power microscope fields. Analysis was performed on 3 wells for each condition, and each experiment was repeated 3 times. For experiments using peptidic GRGDS and GRGES (Neosystem), nonpeptidic Ro64 (generous gift from Roche, Basel, Switzerland) integrin antagonists, the rho kinase (RhoK) 1/2 inhibitor Y27632 (Calbiochem), thapsigargin (Sigma-Aldrich), or the MEK inhibitor U0126 (Calbiochem), cells were placed in the Transwells 2 hours before antagonists were added to both chambers to avoid any interference of these molecules with the adhesion process. Monoclonal antibodies against OPN and vitronectin, MPIIIB10 (DHSB) and M4 (Biosource), respectively, and control IgG1 (Sigma-Aldrich) were added to the lower chamber at the same time as the nucleotide.

Western Blot Analysis

After incubation, the culture medium was removed, and SMCs were lysed directly in the dish at 4°C for 15 minutes with RIPA buffer (50 mmol/L Tris-HCl [pH 7.5], 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS) containing phosphatase and protease inhibitors (1 mmol/L Na₃VO₄, Sigma-Aldrich; 1 mmol/L AEBSF, Interchim). Cell debris was eliminated by a 2-minute centrifugation at 10 000g. The protein concentration of the cell lysate was determined by the microBCA method (Pierce). Proteins (10 μg) were separated by SDS-PAGE on a 10% acrylamide gel under reducing conditions and blotted onto polyvinylidine difluoride membrane (Millipore). Membranes were probed with primary antibody, then with an anti-mouse IgG-peroxidase conjugate (Amersham Pharmacia). Signals were visualized by chemoluminescence with an enzyme-linked chemiluminescence kit (Amersham Pharmacia). The homogeneity of sample loading was checked by probing with anti–α-tubulin monoclonal antibody (clone B-5-1-2, Sigma-Aldrich). The monoclonal antibody MPIIIB10 (1/1000) was used for OPN detection.

ERK Assay

MAPK phosphorylation was assayed by Western blot analysis. The effect of nucleotides on MAPK phosphorylation was checked 30 minutes after the replacement of the serum-containing medium by serum-free medium, thereby eliminating MAPK activation due to serum. For experiments using the MEK inhibitor U0126, cells were preincubated for 2 hours in culture medium containing the inhibitor before stimulation by UTP. For experiments using the B toxin, cells were pretreated for 24 hours in culture medium with the toxin before incubation with UTP. Western blot was essentially done as described above. Proteins (25 μg) were loaded on the acrylamide gel, and nitrocellulose membranes (Hybond C Extra, Amersham Pharmacia) were used. Membranes were probed with an antibody against phospho-ERK1/ERK2 (rabbit anti-active MAPKK, Promega). The homogeneity of sample loading was checked by Coomassie blue gel staining.

Northern Blot Analysis

Cells were incubated for 6 hours with various nucleotide concentrations or in serum-free DMEM. For experiments using the MEK inhibitor U0126 or thapsigargin, a pretreatment of cells with the inhibitors for 2 hours or 3 minutes, respectively, was performed before incubation with UTP. Total RNAs were isolated from SMCs by the guanidinium isothiocyanate/phenol/chloroform extraction method.³⁰ RNA (30 μg) was submitted to a 1% agarose gel electrophoresis and blotted onto a nylon membrane (Hybond N⁺, Amersham Pharmacia). Blots were first hybridized with an OPN-specific cDNA probe as previously described,¹⁸ then stripped and hybridized again with an oligonucleotide probe against ribosomal 18S RNA.

Statistics

ANOVA and unpaired Student’s t test were performed for statistical analysis. Probability values of P<0.05 were considered statistically significant. Data are expressed as mean±SD.

Results

Extracellular Nucleotides Induce SMC Migration

The chemotactic effect of ATP, ADP, UTP, and UDP on cultured SMCs was checked with a modified Boyden chamber assay. Cells that had migrated below the membrane were counted. The majority of cells that had migrated remained hanging below the membrane. Only rare cells (maximum to 1 to 2 cells) were found in the bottom of the lower chamber and did not modify results significantly. The 4 nucleotides at 1 μmol/L significantly enhanced SMC migration (P<0.05) compared with the control in serum-free medium (Figure 1A). UMP, AMP, uridine, and adenosine did not induce any migratory response (data not shown). The strongest migratory effect was elicited by UTP. The order of potency for SMC migration was UTP>UDP>ATP>ADP. UTP induced a concentration-dependent increase in rat SMC migration from 10 nmol/L (P<0.05 versus control), with a maximal response for 100 μmol/L (Figure 1B). At this concentration, SMC migration was ≈4 times higher than that observed in the control. A significant increase in cell migration was observed 4 hours after UTP stimulation, and the maximal ratio (UTP-stimulated/control cells) was reached at hour 6 (not shown). This effect was not the consequence of a modulation of SMC adhesion by nucleotides, because the same number of cells had adhered after 2 or 4 hours of seeding in the presence or absence of UTP. Moreover, the number of migrated cells was identical when a 2-hour adhesion period was performed before addition of UTP (Figure 4A and 4B). A checkerboard analysis was performed to assess whether the response to the extracellular nucleotide was the result of a directed migration rather than a random movement. The presence of UTP in both the upper and lower chambers resulted in a migratory response 2-fold lower than that observed when UTP was present only in the lower chamber (Figure 1C). Only a few cells migrated through the membrane when UTP was present only in the upper chamber. Therefore, UTP acts as a chemotactic agent but also induces chemokinetic effects.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

UTP-Directed Migration of SMCs Is OPN-Dependent

A previous work demonstrated that UTP 100 μmol/L induced OPN expression in cultured SMCs.¹⁸ Moreover, OPN is known to be a chemotactic protein for various cell types, including SMCs.²⁰ Therefore, the ability of UTP to induce OPN gene expression in the range of concentrations that induced a migratory response was studied. SMCs were incubated for 6 hours either in serum-free medium or in serum-free medium containing UTP concentrations from 0.01 to 10 μmol/L. Northern blot analysis (Figure 2A) revealed a concentration-dependent OPN mRNA accumulation in UTP-stimulated SMCs even at concentrations as low as 0.01 μmol/L. The OPN mRNA steady-state level reached its maximum between hours 4 and 8, as previously shown.^10,18 In parallel, OPN protein was strongly increased in SMCs from hour 4 after addition of UTP (Figure 2B).

Migration assays with UTP were performed in the presence of a blocking monoclonal antibody against OPN (MPIIIB10) to determine the involvement of OPN in UTP-stimulated SMC migration. Addition of 10 μg/mL MPIIIB10 in the lower chamber resulted in a total inhibition of UTP-induced SMC migration, whereas an identical concentration of a control mouse IgG1 did not significantly decrease it (Figure 3). In these conditions, SMC adhesion and spreading were not modified by addition of MPIIIB10 (data not shown). Moreover, addition of a vitronectin-blocking antibody (M4) had no inhibitory effect on UTP-induced migration.

The involvement of OPN in UTP-induced SMC migration was also evaluated by use of α_vβ₃ and α_vβ₅ integrin blockers. The peptidic integrin antagonist GRGDS added to the culture medium simultaneously with UTP 2 hours after SMC seeding inhibited SMC migration by ≈50% when used at 40 μmol/L. In identical conditions, the inactive control peptide GRGES did not demonstrate any inhibitory effect (Figure 4A). The major role of β₃/β₅ integrins in UTP-induced SMC migration was confirmed by the total inhibitory action of the nonpeptidic antagonist Ro64 (Figure 4B). These experiments demonstrated that α_vβ₃ and α_vβ₅, the main OPN receptors involved in cell migration, are responsible for the full activity of OPN in UTP-induced SMC migration.

MAP Kinase Phosphorylation, Ca²⁺ Release From Internal Stores, and Rho Protein Activation Are Required for UTP-Induced SMC Migration

ERK1/ERK2 phosphorylation, Ca²⁺ release from internal stores, and RhoA protein activation have been shown to be induced by UTP in SMCs. Therefore, we tested the involvement of these 3 signaling events in UTP-induced SMC migration.

The MEK inhibitor U0126 was used to prevent ERK1/ERK2 phosphorylation induced by UTP in SMCs and consequently to determine whether this activation was also required for UTP-stimulated SMC migration. To prevent disorders during adhesion, SMCs were seeded in Transwells 2 hours before the 2-hour preincubation with U0126. Then UTP 100 μmol/L was added to the lower chamber to stimulate migration. Inhibition of the MAPK pathway by 0.5 and 1 μmol/L U0126 resulted in a 27% and 88% inhibition of SMC migration, respectively (Figure 5). U0126 at 5 μmol/L drastically decreased the UTP-induced MAPK phosphorylation (data not shown) and induced an almost total inhibition of UTP-directed migration.

• Download figure
• Download PowerPoint

Moreover, in the same conditions as used previously, c-Jun N-terminal kinase (JNK) inhibitor SB203580 (1 μmol/L) and p38 MAPK inhibitor (curcumin 10 μmol/L) did not demonstrate any effect on UTP-induced SMC migration (not shown).

To evaluate the part played by α_vβ₃ integrin–mediated signal on MAPK phosphorylation, we studied the UTP-induced ERK1/ERK2 phosphorylation in the presence of Ro64 αvβ3 inhibitor. Figure 6A demonstrates that ERK1/ERK2 phosphorylation is identical in both conditions.

• Download figure
• Download PowerPoint

B toxin inhibits rho protein activation (RhoA, RhoB, RhoC, Rac, and Cdc42 proteins) by glycosylation of these proteins.³¹ To assess the involvement of rho proteins in UTP-induced SMC migration, SMCs were pretreated with B toxin 1, 5, or 10 ng/mL for 24 hours before migration assay. The same concentrations were also added in both chambers during a UTP-induced migration test. B toxin inhibited UTP-induced rat SMC migration in a concentration-dependent manner (Figure 5). A significant inhibition was observed from B toxin 1 ng/mL. B toxin 5 ng/mL induced an 84% inhibition, and 10 ng/mL totally blocked the UTP-stimulated SMC migration.

The UTP-induced [Ca²⁺]_i increase in SMCs is essentially mediated by calcium release from reticulum. Depletion of the Ca²⁺ store by addition of thapsigargin 1 μmol/L in the culture medium for 30 minutes after SMC adhesion and spreading in the Transwells inhibited the increase of [Ca²⁺]_i and fully abolished UTP-directed SMC migration (Figure 5).

These results indicated that Ca²⁺ mobilization, rho protein activation, and MAPK phosphorylation induced by UTP were involved in UTP-directed rat SMC migration.

Rho Protein Activation Is Involved in UTP-Mediated ERK1/ERK2 Phosphorylation

Because rho protein activation was shown to be involved in ERK pathway activation in mechanically stretched rat SMCs,²⁸ we examined whether this signaling event might be involved in UTP-mediated MAPK activation in SMCs. Therefore, cells were treated with B toxin as described in the migration assay, and MAPK phosphorylation was studied by Western blot analysis 30 minutes after UTP stimulation. Pretreatment with B toxin decreased UTP-induced ERK phosphorylation in a concentration-dependent manner (Figure 6B), demonstrating that rho protein activation is involved in UTP-induced MEK pathway activation. However, even at 10 ng/mL, B toxin did not totally inhibit UTP-induced ERK phosphorylation.

RhoK and MAPK Pathways Are Required for UTP-Induced OPN Expression

Because Ca²⁺ mobilization, rho protein activation, and ERK phosphorylation were involved in UTP-directed SMC migration and OPN plays a key role in this process, we verified whether these events were involved in the signaling pathway leading to UTP-induced OPN expression. SMCs were stimulated for 6 hours with UTP, and ERK1/ERK2 phosphorylation, RhoK cascade, or Ca²⁺ mobilization was inhibited by the MAPK inhibitor U0126, the RhoK 1/2 inhibitor Y27632, or thapsigargin, respectively. Northern blot analysis revealed that U0126 exerted a concentration-dependent inhibition of UTP-induced OPN mRNA accumulation in SMCs. A total inhibition was observed with 5 μmol/L U0126 (Figure 7A). RhoK inhibitor (10 μmol/L) also strongly inhibited UTP-induced OPN expression, whereas Ca²⁺ release inhibitor demonstrated no effect.

• Download figure
• Download PowerPoint

Thus, taken together, these results showed that rho protein activation and MAPK phosphorylation are intermediary events leading from P2 receptor stimulation to OPN expression, then to SMC migration, whereas Ca²⁺ signaling is involved in UTP-induced SMC migration by an OPN-independent pathway.

Discussion

The present work demonstrates for the first time that the extracellular nucleotides ATP, ADP, UTP, and UDP serve as directional cues for SMC migration. The recently reported ability of nucleotides to induce chemotaxis of human neutrophils³² and rat mast cells³³ suggests that this property could be a general feature of these compounds.

At identical concentration, the most powerful migratory response observed for nucleotides was elicited by UTP. This migration is the consequence of both chemotaxis and chemokinesis and may result either from the activation of one particular P2 nucleotide receptor or from activation of the several P2 receptors present at the SMC surface. Our previous work showed that the P2Y₂ receptor is the most abundant P2Y receptor expressed by cultured SMCs. The activity of UTP at submicromolar levels to stimulate SMC migration supports the hypothesis that this response could be physiological and is essentially mediated by P2Y₂ receptor activation without excluding participation of other P2Y receptor subtypes. The difference in capacity of UTP and ATP to elicit SMC migration could be due to the inhibition of nucleotide-induced cell migration by adenosine generated from ATP catabolism by cellular ectonucleotidases. Conversely, uridine did not demonstrate any inhibitory effect (data not shown).

SMCs need an ECM to migrate. In our migration assay, Transwell membranes were not coated with ECM proteins, and the cells had to synthesize and secrete these proteins before migrating. Therefore, the migratory capacity of extracellular nucleotides could be mediated by inducing ECM protein expression. The present study complements our previous work demonstrating that UTP 100 μmol/L could induce OPN mRNA accumulation, because it shows that UTP even at low concentrations is able to enhance OPN expression in SMCs. Moreover, we now demonstrate that the increase in OPN expression plays a key role in UTP-induced SMC migration, because a monoclonal antibody against OPN fully abolished it, whereas an antibody against vitronectin, another ECM protein also involved in SMC migration,³⁴ did not. This finding corroborates a previous result demonstrating the inhibition of angiotensin II–induced migration by the same antibody,³⁵ thereby underlining the specific role of OPN in the migration process. Furthermore, the inhibition of migration with α_vβ₃/β₅ integrin antagonists confirms the involvement of OPN as an autocrine or paracrine effector.³⁶

The ERK1/ERK2 pathway has been shown to be a critical event for cell movement on ECM by enhancement of myosin light chain kinase activity, leading to phosphorylation of myosin light chains³⁷ and gene transcriptional events.³⁸ Moreover, various studies have demonstrated that ERK1 and ERK2 are involved in the regulation of cell motility.³⁹ The present study demonstrates that UTP-induced SMC migration is dependent on ERK1/ERK2 activation, an intracellular pathway also involved in platelet-derived growth factor–directed SMC migration.⁴⁰ In addition, inhibition of the ERK1/ERK2 but not that of p38 or JNK pathways leads to the blockade of UTP-induced OPN expression and can explain the inhibition of SMC migration by MAPK inhibitors. Conversely, OPN expression is certainly not the only mechanism involved in MAPK-induced cell migration.

In a recent work, the RhoA was shown to be coupled with P2Y receptors.²⁷ In our study, the role of rho proteins in UTP-induced SMC migration is evidenced, because B toxin, a rho protein inhibitor, or Y27632, a RhoK 1/2 inhibitor, blocked the SMC migration triggered by UTP. Although rho proteins may directly induce migration by acting on actin cytoskeleton, these proteins are involved in OPN expression and may thus induce migration in this way. Furthermore, our work suggests that UTP-induced OPN expression is not dependent on the cytoplasmic [Ca²⁺]_i increase, because its blockade by thapsigargin did not lead to a decrease of UTP-induced OPN expression.

Because OPN mediates integrin activation during UTP-induced SMC migration, it is important to discriminate whether rho activation, Ca²⁺ release, and MAPK phosphorylation belong to the integrin or UTP signaling. Figure 2 demonstrated that OPN protein was not increased before hour 4 after UTP stimulation and thus cannot induce rapid signaling events, such as Ca²⁺ release or rho activation. Moreover, ERK1/ERK2 phosphorylation after UTP stimulation was not modified in the presence of α_vβ₃ integrin inhibitor. All together, these results suggested that rho activation, Ca²⁺ release, and MAPK phosphorylation are the result of UTP stimulation.

Taken together, these results showed that UTP via P2Y receptors induces [Ca²⁺]_i increase, rho protein activation, and the MAPK phosphorylation cascade. Thus, OPN transcription is activated by still unknown transcription factors, and the protein is produced. Finally, OPN activates α_vβ₃/β₅ integrin–mediated migration via a paracrine or autocrine pathway.

This ability of nucleotides to act as chemoattractant for arterial SMCs in a concentration range potentially found in pathological vessels⁴¹ and the findings of previous studies demonstrating the mitogenic ability of extracellular nucleotides for these cells suggest that nucleotides released from mechanically stretched vascular or damaged cells during the angioplasty process may participate in arterial wall remodeling.

Footnotes