| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(q/11)-coupled P2Y2 nucleotide receptor inhibits human keratinocyte spreading and migration
* CISMET, FRE CNRS 2737, Faculté de Pharmacie, Université d’Aix-Marseille, France;
Astellas Pharma Inc., Drug Discovery Research and Molecular Medicine Research Labs, Ibaraki, Japan; and
DIPTA, Aix en Provence, France.
1Correspondence: CISMET, FRE CNRS 2737, Faculté de Pharmacie, 27 Bd Jean Moulin, 13005 Marseille, France. E-mail: maxime.lehmann{at}pharmacie.univ-mrs.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
3 integrin expression at the cell periphery, and the dissolution of focal contacts as indicated by the alteration of v integrins and focal contact protein distribution. In addition, activation of P2Y2R prevents growth factor-induced phosphorylation of Erk1,2 and Akt/PkB. The use of a specific pharmacological inhibitor (YM-254890), the depletion of G(q/11) by siRNA, or the expression of a constitutively active G(q/11) mutant (Q209L) show that activation of G(q/11) is responsible for these ATP/UTP-induced effects. Finally, we report that ATP delays growth factor-induced wound healing of keratinocyte monolayers. Collectively, these findings provide evidence for a unique and important role for extracellular nucleotides as efficient autocrine/paracrine regulators of keratinocyte shape and migration during wound healing.—Taboubi, S., Milanini, J., Delamarre, E., Parat, F., Garrouste, F., Pommier, G., Takasaki, J., Hubaud, J-C., Kovacic, H., Lehmann, M. G(q/11)-coupled P2Y2 Nucleotide Receptor Inhibits Human Keratinocyte Spreading and Migration.
Key Words: GPCR • wound healing • PI3K/Akt • MAPK • UTP
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2)
. Within hours after wounding, keratinocytes undergo morphological changes and migrate from the wound edges to resurface the denuded lesion. Thus, keratinocyte shape changes and migration appear to be critical events for reepithelialization. They are orchestrated by matrix proteins, growth factors, cytokines, and proteases released after injury (3
, 4)
. In addition, some evidence indicates that extracellular nucleotides are secreted at high concentrations in the wound by platelets and damaged cells (5
6
7
8)
. ATP and other extracellular nucleotides are now considered to be important signaling molecules (9)
, they may thus have an essential function in wound healing.
Extracellular nucleotides bind to and activate P2 purinergic receptors. Purinergic receptors are divided into two categories distinguishable by their molecular structure, pharmacological properties, and signaling pathways: the ionotropic P2X receptors (P2X1–7) that are ATP-gated ion channels and the metabotropic P2Y receptors (P2Y1,2,4,6,11–14) that belong to the G protein-coupled receptor (GPCR) family (9)
.
Purinergic receptor expression is modulated in regenerating epidermis, which suggests that these receptors may have important functions during wound healing (6
, 10)
. Moreover, they are expressed in spatially distinct zones of the epidermis where they may regulate different cellular functions such as cell proliferation or differentiation (11)
. In vitro, both P2X receptors (P2X5–7) and P2Y receptors (P2Y1,2,4,6,11) are expressed in primary normal human keratinocytes and in the HaCat human keratinocyte cell line (12
, 13)
. Among them, the P2Y2 receptor (P2Y2R) was found to be the predominant functional receptor that mediates several biological responses induced by ATP and UTP, such as Ca2+ efflux, cell proliferation, and IL-6 production (5
, 11
12
13
14
15)
. Despite this, the physiological and/or pathophysiological role of P2Y2R in the epidermis remains poorly understood.
However, P2Y2R function and signal transduction have been extensively investigated in other cell types (16)
. Thus, P2Y2R has been reported to promote actin polymerization (17
, 18)
and cell motility (19
20
21
22
23
24
25
26)
. This receptor is mainly found coupled to the G(q/11) protein family, which activates phospholipase C-β and results in intracellular free Ca2+ increase and diacylglycerol-dependent activation of protein kinase C (16)
. Nevertheless, P2Y2R can also trigger G(q/11)-independent intracellular signals. For example, P2Y2R can interact with vβ3 integrin to stimulate G0/i(20
, 25
, 27)
. P2Y2R also contains a src-homology 3 binding site that can bind src and in turn trans-activate growth factor receptors (28
29
30)
. Thus, crosstalk between P2Y2R, integrins and growth factor receptors has been shown to stimulate cell migration by activating various signaling molecules, including Rho GTPases, phosphatidylinositol-3-kinase (PI3K), protein kinase B/Akt (Akt), mitogen-activated protein kinases (MAPK), and focal adhesion kinase (FAK) (17
, 21
, 25
, 27)
.
Our work aimed to examine whether P2YR, and especially P2Y2R, regulates keratinocyte motility. We report that, in contrast to other cell types, extracellular ATP and UTP induce via P2Y2R, a potent inhibition of keratinocyte cell spreading and lamellipodium dynamics, and disorganize the actin cytoskeleton and focal contacts. Our results indicate that P2Y2R mediates these effects by activating G(q/11) protein, which in turn inhibits growth factor-induced phosphorylation of MAPK/extracellular signal-regulated kinase-1,2 (Erk1,2) and Akt. Moreover, we show that extracellular ATP inhibits keratinocyte migration in wound healing assays, whereas it stimulates endothelial cell migration. Thus, our results reveal a previously unknown role for P2Y2R and G(q/11) in the regulation of keratinocyte migration during wound repair.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3)
and (phospho)-Erk1,2 (Thr-202/Tyr-204) were purchased from Cell Signaling (Beverly, MA, USA). Rabbit polyclonal antibodies raised against G(q/11), Erk1,2 (K-23), and (phospho)-Tyr (PY-99) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Mouse monoclonal antibodies C3VLA3 (anti-3 integrin) and AMF7 (anti-v integrin) and rat monoclonal antibody GoH3 (anti-6 integrin) were from Immunotech (Marseille, France). Fibronectin and collagen I were from Sigma-Aldrich. Horseradish peroxidase (HRP)-conjugated anti-rabbit antibody and enhanced chemiluminescence reagents were from Amersham Biosciences (Buckinghamshire, UK). Fluorescent secondary antibodies conjugated with AlexaFluor®-488 were from Molecular Probe, Invitrogen (Cergy Pontoise, France). Monoclonal antibody against paxillin was from Chemicon, monoclonal antibody raised against FAK was from BD Transduction Laboratory. Rhodamine-conjugated phalloïdin and antivinculin antibody (clone hVIN-1) were obtained from Sigma-Aldrich. The selective G(q/11) inhibitor, YM-254890, was isolated from the culture broth of Chromobacterium sp. strain QS3666 (31
, 32)
. Cells were pretreated with YM-254890 (10 µM) 5 min before the experiment. HA-tagged forms of wild-type and Q209L mutant of Gq were generated as described previously (31)
. Plasmid encoding the pleckstrin homology domain (PH) of Akt fused with green fluorescent protein (GFP) was a gift of Dr. J. Downward, London, UK, and was generated as described previously (33)
.
Cell culture and transfections
Immortalized human keratinocyte HaCaT cells (34)
were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal calf serum (FCS) (Invitrogen) in a humidified atmosphere of 5% CO2 at 37°C. Normal human keratinocytes (NHK) were obtained by dispase and trypsin dissociation of human foreskins, followed by culture in keratinocyte serum-free medium (KSFM, Invitrogen) supplemented with epidermal growth factor (EGF) (50 ng/ml) and pituitary extracts. Cells were used between the second and fourth subculture. Human squamous carcinoma cells SCC-15 (obtained from Dr. L. Larue, Institut Curie, Paris, France) were cultured at 37°C and 5% CO2 in DMEM supplemented with 10% FCS. Mouse keratinocytes (MK) were given by Dr. S. Rodius and Dr. E. Georges-Labouesse (Illkirch, France). They were isolated and immortalized as described previously (35)
and were cultured in EMEM (BioWhittaker) supplemented with 4% Ca2+ chelated-FCS, EGF (Gibco BRL), cholera toxin (ICN Biomedical), 3,3',5-triiodo-L-thyronine (Sigma), and INF-
(Gibco BRL). Human umbilical vein endothelial cells (HUVECs) were obtained and cultured as described previously (36)
. For transfection experiments, single cell suspensions were prepared by treatment of cell monolayers with trypsin-EDTA. Cells in suspension were then nucleofected with Amaxa Nucleofector (Amaxa GmbH, Cologne, Germany) accordingly to the manufacturer’s optimized protocol.
Gq/11 siRNA experiments
The target sequence selected for G(q/11) silencing was 5'-GATGTTCGTGGACCTGAAC-3', corresponding to positions 932 to 950 relative to the start codon of human Gq and G11 (37
, 38)
. G(q/11) siRNA and "AllStars Negative Control siRNA" were from Qiagen. HaCat cells were nucleofected using Amaxa Nucleofector and the transfection efficiency (80–85%) was monitored using 3'-AlexaFluor 555 probed siRNAs. Kinetics and dose-effect experiments were performed to determine the optimal conditions for G(q/11) silencing: siRNA was applied at 3 µg and cells were tested 48 h after nucleofection.
Cell spreading and cell adhesion assay
Laminin-5-enriched extracellular matrix (LM-5 ECM) was prepared as described previously (39)
. Briefly, HaCat cells were seeded at confluency and cultured for 5 days. Cells were removed by exposure to 40 mM ammonium hydroxyde for 30 min, and the resulting coated dishes were extensively washed with sterile water. As a control, it was verified that HaCat cell adhesion to LM-5 ECM was efficiently inhibited by function blocking antibodies against 3β1 and 6β4 integrins, the two keratinocyte LM-5 receptors (not shown).
For the cell spreading assays shown in Fig. 1
, single cell suspensions were obtained by treatment of growing cells with 0.53 mM EDTA in PBS. After centrifugation, cells were washed and resuspended in DMEM containing 0.2% BSA. When indicated, cells were incubated for 10 min at 37°C in the presence of PPADS (a P2XR receptor antagonist), MRS-2179 (a P2Y1R antagonist), or CGS-15943 (an adenosine receptor antagonist) (100 µM each). Extracellular nucleotides were spotted onto the wells just prior cell seeding. Cells were plated on LM-5-coated wells (15000 cells/cm2) and incubated for 20 min, then fixed with 3% formaldehyde. Alternatively, (Figs. 2
and 6
), HaCat cells were cultured at 3000 cells/cm2 for 48 h in serum-containing medium, then treated as indicated on the figures. Three random fields per well from duplicate wells were pictured under 10x objective, and the cell surface of 
100 cells per experiment was measured using the ImajeJ software (NIH, http://rsb.info.nih.gov/ij).
|
|
|
|
|
|
Cell adhesion assays were performed in serum-free medium as described previously (40)
. The amount of adherent cells (% of total input per well) was evaluated, and the data were expressed as the mean +/ – SD from triplicate wells. The plot presented is representative of three independent experiments.
Video-microscopy and kymography
HaCat cells were cultured in serum-containing medium on glass coverslips for 48 h. Coverslips were mounted in microscope chambers maintained at 37 ± 1°C and observed using an inverted fluorescence microscope (Leica DM-IRBE) with a 40x plan-Fluotar objective lens. Phase-contrast images were acquired at 2-minute intervals using a Coolsnap FX digital camera (Princeton Instruments) driven by Metamorph® software (Universal Imaging Corporation). For each experiment, 10 images were acquired before and 120 images were acquired after extracellular nucleotide addition. Quantification of lamellipodium size was performed every 3 images as described previously (41)
. Three independent experiments were performed, and 15 cells per experiment were examined. Data are expressed as the mean +/ – SD.
To quantify lamellipodium dynamics, time-lapse phase contrast movies were made. Recordings were 10 min long with a frequency of 1 frame/3 s. For each microscopic field, a first movie was made 10 min before stimulation, a second one 40 min, and a third one 60 min after nucleotide addition. Images of the three time-lapses were then stacked and two 1-pixel-thick lines were drawn over the lamellipodium. This ensured that exactly the same region of the cell was analyzed over the 3 different periods of time. Kymographs were then built using Metamorph® software. A total of 24 kymographs for ATP stimulation and 30 kymographs for UTP stimulation were made from three independent experiments. For quantitative analysis, a straight line was drawn on each kymograph from the beginning to the end of both the protrusion and retraction edges. The slope of the line was used to calculate the length, the persistence and the velocity of lamellipodium protrusions and lamellipodiun retractions as described previously (42)
. The event frequency was also measured. Data are expressed as the mean +/ – SEM and statistically analyzed using the paired Student’s t test.
Wound-healing assay
Keratinocytes were cultured to confluency in 24 well-plates. Then, the cell monolayer was wounded using a sterile tip and extensively rinsed to remove cell debris. Cells were then allowed to migrate for 24 h in the indicated medium. Four different fields per well made in triplicate were pictured with a 10x objective by using a CoolSnapFx CCD camera and Metamorph® software (Universal Imaging Corporation). The same fields were pictured just after wounding and 24 h later. Cell migration was evaluated by measuring the wounded surface at the two time points with ImageJ software. Data were expressed as the mean of the recovered surface in µm2 +/ – SEM. Experiments were performed in presence of mitomycin-C (5 µg/ml) to ensure that the observed effects were not due to cell proliferation. Data presented are from an experiment representative of three to four independent experiments.
Immunofluorescence
For actin labeling, cells were treated as described elsewhere (43)
. For immunofluorescence, cells were labeled overnight at 4°C with primary antibodies at 5 µg/ml and then with the appropriate fluorescent secondary antibody. Cells were then observed under immersion oil 40x or 100x Plan-Fluotar objectives on an inverted Leica DM-IRBE microscope and pictured with a CoolSnapFx CCD using Metamorph® software.
Preparation of whole cell lysates and immunobloting
Cells were incubated in lysis buffer (25 mM Tris-HCl, pH 7.6; 150 mM NaCl; 1% Triton X-100; 0.1% sodium deoxycholate; 4 mM EDTA; 50 mM NaF; 1 mM sodium orthovanadate; 10 mM sodium pyrophosphate; 1 mM PMSF; 1 µg/ml leupeptin; 1 µg/ml aprotinin) for 15 min at 4°C. Cell lysates were centrifuged for 10 min at 10,000 g. Equal amounts of protein were resolved by SDS-PAGE and transferred to Hybound-C nitrocellulose membranes (Amersham Pharmacia Biotech, France). Membranes were then probed with the appropriate primary antibody (2 µg/ml), and bound antibodies were detected according to the enhanced chemiluminescence protocol from Amersham.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. When plated on LM-5, HaCat cells spread in 20 min (Fig. 1A
). Remarkably, addition of either ATP or UTP at 100 µM, i.e., the optimal concentration for calcium release in keratinocytes (12)
, strongly inhibited HaCat cell spreading (Fig. 1A
). To identify the receptor mediating these effects, we performed pharmacological analysis (Fig. 1B
). ATP is a broad agonist of P2Y and P2X receptors, whereas UTP is selective for P2Y2R and P2Y4R. ATP-s can activate P2Y2R but is a more potent agonist of P2Y11R; ADP activates P2Y1R; 2-MeSATP is selective for P2Y1R and P2X3R; UDP activates P2Y6R; and BzATP is a P2Y11R and P2X7R agonist (9)
. As shown in Fig. 1B
, the potency rank order for HaCat spreading inhibition was ATP=UTP>ATP-s>>UDP, ADP, 2-MeSATP, and BzATP. It should be noted that the IC50 concentration for both ATP and UTP was 0.8 µM (Fig. 1B
), i.e., a physiological concentration that is reached in the wound bed (7)
. In agreement with the inefficiency of Bz-ATP, PPADS (100 µM), a broad P2XR antagonist, did not protect HaCat cells from ATP (10 µM)- or UTP (10 µM)-induced effects (Fig. 1C
). Moreover, the effects of ATP (10 µM) were not antagonized by MRS2179 (100 µM), excluding the involvement of P2Y1R. Finally, CGS-15943 (100 µM), an adenosine receptor antagonist, did not prevent ATP inhibition of cell spreading (10 µM) (Fig. 1C
), which suggests that adenosine is not involved in ATP-induced effects. In addition, both ATP and UTP inhibited HaCat cell spreading on other ECM proteins such as collagen I and fibronectin (Fig. 1D
). Finally, cell adhesion assays on LM-5 ECM (Fig. 1E
) or collagen (not shown) indicated that inhibition of cell spreading was not a consequence of a loss of HaCat adhesion to the substratum. Together, these results strongly suggest that ATP and UTP stimulate P2Y2R, and possibly P2Y4R. Moreover, activation of these receptors leads to the inhibition of HaCat cell spreading on various ECM proteins without affecting cell adhesion.
ATP and UTP induce transient lamellipodium withdrawal
As lamellipodium formation is an essential process for cell spreading, we next sought to determine whether it might be affected by ATP. For this purpose, we used time-lapse video-microscopy to observe HaCat cells that were cultured on glass-coverslips in the presence of serum growth factors. Figure 2A
, top panels, depicts phase contrast frames of HaCat cells before and after stimulation with ATP (100µM) (see also Supplemental Movie 1). Before nucleotide application, HaCat cells displayed a well-organized and dynamic lamellipodium with ruffle-rich membranes. ATP addition induced a dramatic collapse of HaCat cell lamellipodium and the disappearance of membrane protrusions and ruffles. Lamellipodium destruction was never associated with cell contraction and was observed regardless of cell density. Quantification of lamellipodium length indicated that ATP-induced effects were maximal 6–12 min after addition and remained persistent for up to 20 min (Fig. 2A
, bottom). Later on, lamellipodium regrew and remained present during all of the 70-minute recording time. Similar results were obtained when UTP (100 µM) was used instead of ATP (not shown). The fact that lamellipodium destabilization was transient may be explained by ATP/UTP-induced desensitization shown in Fig. 2B
. Indeed, HaCat cell spreading inhibition by ATP and UTP was prevented by a 1 h preincubation of cells with either ATP or UTP (100 µM) but not with ,β-meATP (100 µM). Finally, we measured the impact of UTP on the presence of membrane ruffles at the periphery of HaCat cells, mouse keratinocytes (MK), and human squamous carcinoma cells (SSC-15) (Fig. 2C
). Ten minutes after UTP treatment, ruffles disappeared in nearly 90% of HaCat cells, 60% of MK cells, and 80% of SCC-15 cells.
Thus, ATP and UTP transiently inhibit the formation of lamellipodium and membrane ruffles in HaCat cells as well as in other keratinocyte cell types. The cross-desensitization between ATP and UTP indicates that both nucleotides mediate these effects through the same receptor, i.e., P2Y2R.
ATP and UTP activation inhibits lamellipodium dynamics
As described above, 30 to 45 min after addition of extracellular nucleotides, lamellipodium had regrown (Fig. 2A
). However, the neosynthetized lamellipodium was much less dynamic, and ruffles appeared both thinner and smaller (Supplemental Movie 1). To study lamellipodium dynamics, we performed kymography assays on control and nucleotide-treated cells. A kymograph is a time-line representation of lamellipodium that allows quantitation of four parameters: velocity, length, persistence, and frequency of the protrusions and retractions (42)
. Figure 3
A presents three typical kymographs obtained from a single microscopic field taken before (control), 40 min (40/50), and 60 min (60/70) after ATP addition. Kymographs of control cells appeared as a succession of "shark fin"-like structures as described previously by Bear and colleagues (44)
. Analogous structures were also observed in kymographs from ATP-treated cells, but they were of smaller size (Fig. 3A
), reflecting an ATP-induced inhibition of membrane protrusion dynamics. Quantitative analysis of kymographs revealed that ATP and UTP significantly reduced the length as well as the velocity of formation and withdrawal of the protusions. By contrast, extracellular nucleotides did not affect protrusion frequency or persistence (Fig. 3B
).
Thus, kymography analysis clearly shows that ATP and UTP induce a prolonged inhibition of lamellipodium dynamics.
ATP and UTP disrupt actin cytoskeleton organization and integrin adhesion sites
Lamellipodium formation is driven by actin polymerization coupled with integrin-mediated adhesion to the ECM. F-actin labeling of control HaCat cells revealed the presence of a well-developed cortical actin network supporting the lamellipodium that contained microspikes and actin-enriched membrane ruffles. Stress fibers were barely present (Fig. 4
, time 0). Ten minutes after addition, ATP and UTP induced an important disorganization of the actin cytoskeleton. Lamellipodium collapse was associated with a dismantling of the cortical actin network, with only a few actin filaments surrounding the cell border (Fig. 4
, time 10 min). Thirty minutes after nucleotide addition, lamellipodium was resynthesized and cortical actin reappeared while, after 90 min, cells regained an aspect similar to unstimulated cells (Fig. 4)
. Thus, in keratinocytes, activation of P2YR by ATP or UTP elicits signals that profoundly disorganize the actin cytoskeleton.
Because keratinocyte spreading on a complex ECM involves several integrins such as 6β4 and 3β1 (receptors for laminin-5) or v (receptors for fibronectin), we compared the localization of these integrins before and after HaCat stimulation with UTP. First, it should be noted that extracellular nucleotides did not modify the total amount of 6, 3, or v integrins expressed at the cell surface (not shown). As expected (45)
, 6β4 accumulated in hemidesmosome-like adhesion sites (Supplemental Fig. S1). These adhesive structures, which are linked to intermediate filaments, were not affected by UTP treatment. As described elsewhere (46)
, 3β1 integrin labeling decorated cell/cell contacts and lamellipodium edges. UTP-treatment removed 3β1 integrin from the cell borders without affecting its presence at intercellular junctions (Supplemental Fig. S1). Focal contacts are large molecular complexes that link scaffolding and signaling proteins to the actin cytoskeleton and mediate signals that are critical for lamellipodium stabilization and cell spreading (47)
. v Integrins were found in adhesion sites resembling focal contacts. Interestingly, v-bearing focal contacts were disrupted by ATP or UTP treatment (Supplemental Fig. S1). To confirm the effects of UTP on these structures, HaCat cells were labeled with antibodies against three major components of focal contacts, i.e., paxillin, vinculin, and FAK. In the presence of serum, all these proteins were detected in focal contacts (Fig. 5
, left panels). As observed above for v integrins, distribution of paxillin, vinculin, and FAK was strongly altered 10 min after UTP addition. Although these proteins were still present at the cell border, focal contacts were smaller than in control cells (Fig. 5
, right panels). A hallmark of focal contacts is the intense activity of protein tyrosine kinases. As expected, labeling of phosphotyrosine residues decorated both focal complexes and cell-cell junctions in control cells. However, in UTP-stimulated cells, phosphotyrosine labeling was strongly decreased at the cell edge and only small punctate adhesion sites remained detectable (Fig. 5
, bottom), while tyrosine phosphorylated proteins remained localized at intercellular adhesion sites. Note that similar results were obtained with ATP (not shown).
Thus, extracellular nucleotides disorganize adhesion sites containing 3β1 and v integrins that are linked to the actin network and involved in lamellipodium formation and stabilization.
G(q/11) signaling is required for the P2Y2R-induced morphological effects
P2Y2R is mainly coupled to the G(q/11) heterotrimeric G protein family (16)
. Thus, we sought to determine whether G(q/11) activation by P2Y2R could contribute to the blockage of lamellipodium formation and actin remodeling in keratinocytes. To this end, we first used YM-254890, a cyclic depsipeptide that selectively inhibits G(q/11) (31)
. As shown on Fig. 6
A, YM-254890 alone (10 µM) had no effect on the actin network or lamellipodium formation in control cells. Remarkably, in cells stimulated for 10 min with ATP, actin depolymerization was completely prevented by inhibiting G(q/11) with YM-254890 (Fig. 6A
). YM-254890 also prevented ATP- or UTP-induced inhibition of cell spreading (Fig. 6B
) and lamellipodium withdrawal (Fig. 6C
). In an additional series of experiments, Gq and G11 were silenced using a siRNA sequence previously validated in human cells (37
, 38)
. Forty-eight hours after G(q/11) siRNA nucleofection, G(q/11) was depleted by 80% in HaCat cells. Moreover, G(q/11) silencing was found to prevent UTP (Fig. 6D
) and ATP (not shown) inhibition of cell spreading. Finally, HaCat cells were transfected to transiently express HA-tagged forms of wild-type or constitutively active Gq mutant (GqQ209L). Cells expressing wild-type Gq showed normal F-actin labeling with a cortical actin network and lamellipodium (Fig. 6E
, left panels). By contrast, expression of GqQ209L mutant led to a strong disruption of actin microfilaments (Fig. 6E
, right panels). These keratinocytes were devoid of cortical actin and did not develop any lamellipodium. Moreover, actin polymerized into thin filaments that did not resemble stress fibers.
Together, these data indicate that ATP/UTP-induced alteration of actin cytoskeleton organization, lamellipodia stability and cell spreading via P2Y2R is transduced by G(q/11)-controlled signaling pathways.
P2Y2R/G(q/11) signaling down-regulates PI3K/Akt and Erk1,2 pathways
Convergent signals between integrins and growth factor receptors activate Erk1,2 and PI3K/Akt, two signaling pathways that are critical for keratinocyte spreading (48
, 49)
. It was, therefore, important to determine whether ATP/UTP-activated P2Y2R may regulate these two pathways. ATP and UTP transiently inhibited serum-induced phosphorylation of Erk1,2 and Akt (a main substrate phosphorylated on PI3K activation) (Fig. 7
A). Dose-response experiments showed that 10 µM of ATP or UTP was sufficient to inhibit Erk1,2 and Akt phosphorylation (Supplemental Fig. S2A). The equipotency of ATP and UTP again suggests that P2Y2R was involved in this phenomenon (see also comments on Fig. 1
). ATP and UTP also transiently decreased the growth factor-induced phosphorylation of Erk1,2 in MK cells (Supplemental Fig. S2B). In agreement with our results on the ATP/UTP-induced morphological changes (see Fig. 6
), the G(q/11) inhibitor YM-254890 did not affect Erk1,2 or Akt phosphorylation induced by serum alone. By contrast, YM-254890 completely abolished ATP/UTP-mediated inhibition of Erk1,2 and Akt phosphorylation (Fig. 7B
). Involvement of G(q/11) was further demonstrated by the incapacity of UTP and ATP to block, after siRNA-mediated G(q/11) depletion, the serum-induced phosphorylation of Erk1,2 and Akt (Fig. 7C
).
|
Immunofluorescence experiments revealed that phosphorylated Erk1,2 was located at the edges of HaCat cells (Fig. 7D
, top panel), which is in agreement with a previous report (50)
. However, this labeling was no longer observed in 10-minute-UTP-treated cells (Fig. 7D
, top panels). PI3K activation leads to a rise in PI(3,4,5)P3 at the plasma membrane, a docking signal for several proteins containing a pleckstrin-homology domain (PH). The PH domain of Akt fused to green fluorescent protein (GFP-PH-Akt) was, therefore, used to probe for the presence of PI(3,4,5)P3 at the plasma membrane (Fig. 7D
, bottom). GFP-PH-Akt protein accumulated at the lamellipodium in control cells but not in UTP-stimulated cells. The effects of UTP on phospho-Erk1,2 and GFP-PH-Akt distribution were transient, and a peripheral labeling reappeared 30 min after nucleotide addition (not shown). Similar data were also obtained with ATP (unpublished data).
Together, these data indicate that ATP and UTP activate P2Y2R and G(q/11) to transiently inhibit serum-induced activation of Erk1,2 and PI3K/Akt pathways in keratinocytes.
ATP inhibits growth factor-induced migration of keratinocytes in wound-healing assays
We finally addressed the possibility that extracellular nucleotides may impact keratinocyte migration. To this end, we analyzed HaCat and NHK cell migration in wound healing assays (Fig. 8
). After 24 h, and in the absence of exogeneous growth factors, HaCat and NHK cells partially recovered the denuded surface. Stimulation of HaCat cells with serum (Fig. 8A
) or NHK cells with epidermal growth factor and insulin-like growth factor-I (Fig. 8B
) strongly stimulated wound healing. Remarkably, ATP (100 µM) inhibited 75–80% of the growth factor-induced wound closure in both HaCat and NHK cells. These experiments were performed in the presence of mitomycin-C (3 µM), such that ATP-induced inhibition of wound healing was not due to an alteration of keratinocyte proliferation. In contrast, when similar experiments were performed with growth factor-stimulated human umbilical vein endothelial cells (HUVEC), both ATP and UTP (100 µM) substantially increased the rate of cell migration (Fig. 8C
), which is in complete agreement with the report of Kaczmarek and colleagues (21)
. Together, these data indicate that extracellular ATP strongly and specifically inhibits keratinocyte migration, thus noticeably delaying wound closure.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. In the present article, we report that extracellular ATP strongly inhibits the motogenic activity of growth factors in in vitro keratinocyte wound-healing assays (Fig. 8)
. This inhibition is associated with an ATP/UTP-induced alteration of several hallmark features of cell migration, especially, cell spreading, lamellipodium formation and dynamics, actin cytoskeleton organization, focal contact assembly, and PI3K/Akt and Erk1,2 signaling pathways. Resting keratinocytes release nanomolar concentrations of ATP (51)
, i.e., concentrations too low to activate their own purinergic receptors. However, it has been estimated that the concentration of ATP released from injured cells may exceed 100 µM at the wound site (8)
. Such a concentration is largely superior to the IC50 (0.8 µM) required to inhibit keratinocyte cell spreading (Fig. 1)
or to the concentration (1 µM) required to inhibit PI3K/Akt and Erk1,2 signaling pathways (Supplemental Fig S2). Thus, it is likely that in vivo, ATP released after wounding may act as an autocrine/paracrine factor (26)
, that activates keratinocyte purinergic receptors and alters keratinocyte shape changes and migration.
Although the absolute identification of P2Y2R requires further experiments, its predominant involvement in these phenomena is supported by i) the equipotency of ATP and UTP in the assays performed during this study and the pharmacological profile of extracellular nucleotides in HaCat cell spreading inhibition (Fig. 1)
. The latter is similar to the one reported for the stimulation of Ca2+ efflux, cell proliferation and IL-6 production in human keratinocytes (5
, 11
12
13
14
15)
; ii) the cross-desensitization between ATP and UTP (Fig. 2)
, indicating that these nucleotides share a common receptor to transduce their effects; and iii) the identification of G(q/11) as the trimeric G protein that conveys the ATP and UTP inhibitory signals (Figs. 6
and 7)
. All these data suggest that signals transduced by P2Y2R, although early and transient, trigger and maintain long-term processes that impede cell migration. In such a context, we show that lamellipodium dynamics is durably inhibited following P2Y2R activation, which may in turn considerably reduce keratinocyte motility (Fig. 3)
.
To our knowledge, this is the first evidence that activation of P2Y2R in keratinocytes is antimotogenic. In contrast to our findings, previous reports have indicated that P2Y2R stimulates cell migration in other cell types (18
19
20
21
22
23
24
25)
. This discrepancy is not likely to be due to our experimental protocol since we confirm in wound healing assays that extracellular ATP and UTP enhance migration of human endothelial cells (Fig. 8C
). Additionally, the β2-adrenergic receptor, another GPCR, inhibits keratinocyte motility (50
, 52)
but conversely stimulates migration of dermal fibroblasts (53)
. Extracellular ATP has also been reported to be either chemoattractive or chemorepulsive, depending on the subpopulation of dendritic cells examined (54
, 55)
. Thus, we propose that during skin wound healing, extracellular nucleotides may have a dual function, i.e., inhibiting keratinocyte motility as shown here, and facilitating migration of other cell types.
Erk1,2 and PI3K/Akt signaling pathways are, among the signals transduced by growth factors, essential regulators of keratinocyte spreading and migration and of epidermal wound repair (48
49
50)
. Here, we report that activation of P2Y2R by extracellular nucleotides strongly inhibits growth factor-induce Akt and Erk1,2 phosphorylation (Fig. 7
and Supplemental Fig. S2). ATP/UTP induces lamellipodium collapse and inactivates Erk and Akt with a similar dose response sensitivity and kinetics. Thus, it is likely that inhibition of Erk1,2 and PI3K/Akt pathways by ATP/UTP-activated P2Y2R mediates the alteration of keratinocyte morphology and motility. This negative crosstalk between growth factor receptors and P2Y2R is quite remarkable. It is indeed widely reported that, like other GPCRs, P2Y2R, cooperatively with tyrosine kinase receptors and integrins, transduces intracellular signals leading to a convergent activation of PI3K/Akt and Erk1,2 pathways (16)
. For instance, activation of Erk1,2 and PI3K/Akt sustains P2Y2R-dependent migration of endothelial and astrocyte cells (21
, 25)
. However, the opposite has been also observed in astrocytes, where stimulation of P2Y receptor by ATP can inhibit the activation of cRaf-1 and MAPK/Erk-kinase via the fibroblast growth factor-2 (56)
. Similarly, ADP has been reported to elicit PI3K activation in serum-starved glioma cells presumably through the P2Y1 receptor but also to moderate PI3K signaling when cells are cultured in presence of serum (57)
. Therefore, activation of P2Y2R can trigger opposite intracellular signals that may crosstalk either positively or negatively with growth factor receptors, depending on the cell type, its state of differentiation, and the current network of autocrine/paracrine regulatory molecules present in the cell microenvironment. Such a dual signaling function may also explain the opposing activities of extracellular nucleotides on cell motility as discussed above.
Importantly, this study identifies G(q/11) as the signal transducer mediating the inhibitory activity of P2Y2R on keratinocyte shape changes (Fig. 6)
and serum-induced activation of PI3K/Akt and Erk1,2 pathways (Fig. 7)
. The role of G(q/11) was assessed by using a pharmacological inhibitor (YM-254890), specific siRNA sequences, and the expression of an active mutant (Q209L). Other GPCR have previously been reported to inhibit PI3K/Akt pathways in various cell types (58
59
60)
. It is important to note that, like P2Y2R, these GPCR (e.g., 1-adrenergic receptor, angiotensin II type I receptor and m1 muscarinic acetylcholine receptor) are all coupled to G(q/11). Increasing evidence also suggests that G(q/11) can negatively regulate the PI3K/Akt pathway (61
, 62)
. Thus, G(q/11)-dependent activation of PLC-β has been shown to decrease the availability of phosphoinositide 4,5-bisphosphate, the PI3K substrate (63)
. However, G(q/11) can also transduce PLC-β independent signals (64)
. Lin’s group has shown that activated G(q/11) can bind to p110, which is the catalytic subunit of PI3K, and inhibit its enzymatic activity (65
66
67)
. Finally, the present work is the first description of a link between G(q/11) and the inhibition of serum growth factor-induced Erk1,2 phosphorylation. However, the molecular mechanism of such an inhibition remains to be elucidated.
In conclusion, this work shows that ATP and UTP activate P2Y2R in keratinocytes, which down-regulates the PI3K/Akt and Erk1,2 signaling pathways and brings a mechanistic support for the inhibitory activity of P2Y2R on lamellipodium dynamics and formation of focal contacts. We further demonstrate that these unique and unexpected functions of P2Y2R are mediated by the G(q/11) protein family. Importantly, all these events are linked to the capacity of extracellular ATP to delay keratinocyte migration in scratch wound healing assays. It has been previously reported that extracellular ATP may regulate the balance between proliferation, differentiation, and apoptosis in keratinocytes (5
, 11)
. Our present findings demonstrate a novel function for extracellular ATP that may be important for the regulation of epidermal homeostasis and wound healing. Future work should help determine whether pharmacological modulators of the P2Y2R/G(q/11) pathway may constitute useful therapeutic tools in pathological processes involving cell migration, such as impaired wound healing or cancer cell invasiveness.
|
ACKNOWLEDGMENTS |
|---|
Received for publication November 16, 2006. Accepted for publication May 31, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
