The Lipoxin A4 Receptor Is Coupled to SHP-2 Activation

Mesangial cell proliferation is pivotal to the pathology of glomerular injury in inflammation. We have previously reported that lipoxins, endogenously produced eicosanoids with anti-inflammatory and pro-resolution bioactions, can inhibit mesangial cell proliferation in response to several agents. This process is associated with elaborate receptor cross-talk involving modification receptor tyrosine kinase phosphorylation (McMahon, B., Mitchell, D., Shattock, R., Martin, F., Brady, H. R., and Godson, C. (2002) FASEB J. 16, 1817–1819). Here we demonstrate that the lipoxin A4 (LXA4) receptor is coupled to activation and recruitment of the SHP-2 (SH2 domain-containing tyrosine phosphatase-2) within a lipid raft microdomain. Using site-directed mutagenesis of the cytosolic domain of the platelet-derived growth factor receptor β (PDGFRβ), we report that mutation of the sites for phosphatidylinositol 3-kinase (Tyr740 and Tyr751) and SHP-2 (Tyr763 and Tyr1009) recruitment specifically inhibit the effect of LXA4 on the PDGFRβ signaling; furthermore inhibition of SHP-2 expression with short interfering RNA constructs blocked the effect of LXA4 on PDGFRβ phosphorylation. We demonstrate that association of the PDGFRβ with lipid raft microdomains renders it susceptible to LXA4-mediated dephosphorylation by possible reactivation of oxidatively inactivated SHP-2. These data further elaborate on the potential mechanisms underlying the anti-inflammatory, proresolution, and anti-fibrotic bioactions of lipoxins.

There is a growing appreciation that novel, endogenously produced lipid mediators can regulate inflammatory processes and their resolution (1). Lipoxins (LXs) 3 are endogenously gen-erated eicosanoids with potent anti-inflammatory properties (2). There is increasing evidence that LXs promote the resolution of inflammation and can elicit distinct anti-fibrotic responses within an inflammatory milieu (3)(4)(5)(6). LXA 4 exerts its actions via interaction with its cognate G protein-coupled receptor (GPCR), the LXA 4 receptor (ALXR) (recently reviewed in Ref. 7). Aside from LXA 4 , a number of pleiotropic lipid and peptide agonists have been shown to activate the ALXR (8). The intracellular signaling pathways activated upon ALXR engagement in multiple cell types remain to be fully elucidated (2).
We have previously demonstrated that complex cross-talk exists between the ALXR and other GPCRs and receptor tyrosine kinases (RTKs) in primary human renal mesangial cells (2,4,9). In this context it is noteworthy that LX can inhibit MC proliferation in response to both growth factors and proinflammatory mediators such as the cysteinyl leukotriene D 4 (9). MC proliferation is a prototypic inflammatory response coupled to matrix accumulation and further production of cytokines that may culminate in the onset of many subacute and chronic forms of glomerulonephritis and eventual fibrosis (10).
Activation of the PDGFR␤ and subsequent signaling events play a pivotal role in renal inflammation and fibrosis (11). In this regard modulation of PDGFR␤ activity is considered an appropriate therapeutic target (10,12). PDGFR␤ activation leads to increased tyrosine phosphorylation and recruitment of Src homology 2 (SH2) domain-containing molecules to specific phosphotyrosine residues within the cytosolic domain of the receptor, initiating processes that promote cell proliferation and gene expression (13). We have previously reported that LXA 4 treatment of MC modulates PDGF-BB-stimulated tyrosine phosphorylation of the PDGFR␤ and subsequent mitogenic events (9). Here we report that the ALXR may be coupled to reactivation of a protein-tyrosine phosphatase (PTP), SHP-2, that specifically dephosphorylates the recruitment sites of the p85 subunit of PI 3-kinase on the PDGFR␤.

EXPERIMENTAL PROCEDURES
Reagents, antibodies, plasmids/primers information, and cell culture techniques are described in the supplemental text.
Cell Transfection and Preparation of Lysates-Human embryonic kidney 293 (HEK293) cells and HEK293 cells stably overexpressing the HA epitope-tagged ALXR (designated ALXR cells) were transiently transfected with various plasmids using Genejuice transfection reagent as per the manufacturer's protocol. Briefly, the cells were plated to 70% confluency, allowed to adhere for 24 h, and subsequently transfected with 1 g of plasmid. 24 h post-transfection, the cells were serumrestricted in 0.01% FCS medium for 24 h prior to stimulation. For short interfering RNA (siRNA)-mediated knockdown of SHP-2 levels, ALXR cells were co-transfected with targeted SHP-2 PTP siRNA (sequence 5Ј-CCTCTGAAAGGTGGTTT-CATGGACATCTCTCTGGGAAAGAAGCAGAGAAAT-3Ј, residues 703-753 of the SHP-2 sequence) (Qiagen) using RNAfect transfection reagent as per the manufacturer's protocol. As a control, a nonspecific siRNA duplex with a random nucleotide sequence) (Qiagen) was used. Two additional siRNA sequences complementary to the SHP-2 gene were designed using custom SMARTPools from Dharmacon (Lafayette, CO) (Dharmacon catalogue numbers J-003947-09 and J-003947-10) and were transfected using Lipofectamine (Invitrogen) as per the manufacturer's protocol.
RTK Activity Assay-MCs were serum-restricted and exposed to various agents for 0 -60 min. The cells were lysed, and PDGFR␤ was immunoprecipitated as previously described. Thereafter, immunocomplexes were washed thrice in fresh RIPA buffer and once in kinase assay buffer (50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 1 mM dithiothreitol, 1 mM benzamidine, and 0.5 mM phenylmethylsulfonyl fluoride) before resuspending in 25 l of kinase assay buffer. 10 Ci of [␥-32 P]ATP was subsequently added, and the reactions were incubated for 30 min at 37°C. Immunocomplexes were then denatured in sam-ple buffer and were resolved by SDS/PAGE. After electrophoresis, the gels were dried and subsequently subjected to autoradiography. Autoradiograms were scanned using Chemidoc Quantity One software (Bio-Rad). Normalization of data were achieved by an anti-PDGFR␤ immunoblot using a small amount of each sample collected after the immunoprecipitation stage.
Analysis of SHP-2 Activity-SHP-2 activity was measured in SHP-2 immunoprecipitates using a Malachite Green PTP activity assay (Upstate Biotechnology Inc.) according to the manufacturer's instructions. Briefly, SHP-2 was immunoprecipitated from 200 -300 g of MC lysate using a specific monoclonal SHP-2 antibody (BD Transduction Laboratories) for 3 h at 4°C. Protein G-agarose beads were subsequently added to each immunoprecipitate and incubated for further 2 h at 4°C with constant rocking. Immunocomplexes were washed three times in fresh lysis buffer and then twice in the phosphatase assay buffer (Tris-HCl, 50 mM, pH 7.0, 0.1% bovine serum albumin) before finally resuspending in 100 l of assay buffer. To each sample, phosphopeptide substrate was added, and the samples were incubated for 30 min at 37°C. The reactions were terminated by the addition of 100 l of malachite green solution, allowing 15 min for color development followed by spectrophotometric analysis at 650 nm. The results were expressed as pmol of phosphate released per microgram of protein.
Sucrose Flotation Gradients-ALXR cells seeded in 150-mm plates were washed twice with ice-cold phosphate-buffered saline and scraped into 1 ml of phosphate-buffered saline. Thereafter, the cells were harvested and lysed with 1 ml of TNE buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 1%Triton X-100, 1ϫ Complete TM protease inhibitor mixture) for 30 min at 4°C. The lysates were subsequently homogenized with 15 strokes of a Dounce homogenizer, transferred to a microcentrifuge tube, and centrifuged at 4°C for 10 min at 830 ϫ g. 1 ml of the resultant cell supernatant was mixed with an equal volume of 80% (w/v) sucrose in TNE buffer, transferred to a 5-ml ultracentrifuge tube, and then carefully overlaid with 2 ml of 30% (w/v) sucrose in TNE buffer and finally with 1 ml of 5% (w/v) sucrose in TNE buffer. The tube was centrifuged at 4°C for 17 h at 134,400 ϫ g. Eight fractions (600 l) were collected from the top to the bottom of the gradient. Protein from each fraction was precipitated using a standard methanolchloroform protocol. The resulting precipitates were resuspended in reducing sample buffer, heated to 90°C for 5 min, and resolved on an SDS/PAGE gel or stored at Ϫ80°C until required.
Detection of Reactive Oxygen Species-Reactive oxygen species (ROS) were detected in MCs stimulated with various agents using an Image-iT TM LIVE Green ROS detection kit. Briefly, subconfluent cultures of MC were plated on 8-well chamber slides and serum-restricted in 0.2% FCS for 48 h. Thereafter, MCs were treated with vehicle (30 min), PDGF-BB (10 ng/ml, 30 min), LXA 4 (10 nM, 30 min), or with LXA 4 (10 nM, 30 min) followed by PDGF-BB (10 ng/ml, 30 min). The cells were washed once with warm Dulbecco's phosphate buffered saline (DPBS) solution. Carboxy-H 2 DCFDA (10 mol/liter) was added to each well, and the cells were further incubated for 30 min at 37°C. In addition, the nuclei were countered stained with a Hoechst 33342 nuclei stain (1 M) for 5 min at 37°C. Subsequently, the slides were gently immersed and washed three times in a warm DPBS solution, after which one drop of Slowfade Light Antifade solution was added before placing a coverslip over the wells. The fluorescence of the oxidized form of carboxy-H 2 DCFDA was visualized using a M.1 Zeiss microscope, and the images were captured using an Axiocam system and Axiovision Rel 4.4 software (Carl Zeiss).
Data Analysis-Statistical analysis was carried out using the unpaired Student's t test. All of the values are expressed as either the means Ϯ S.D. or the means Ϯ S.E. p values of less than or equal to 0.05 were considered to indicate a statistically significant difference.

RESULTS
Treatment of MCs with LXA 4 resulted in a concentrationdependent (100 pM to 100 nM) decrease in activated PDGFR␤ upon stimulation with PDGF-BB. (The EC 50 was ϳ10 nM LXA 4 ; data not shown). Similarly, the amount of immunodetected PDGFR␤ in phosphotyrosine immunoprecipitates was decreased by 52 Ϯ 8% following preincubation with LXA 4 (10 nM) prior to stimulation with PDGF ( Fig. 1A, inset). To investigate whether altered receptor phosphorylation reflected decreased autokinase activity or possible phosphatase activation, the effect of LXA 4 on RTK activation was investigated. As shown in Fig. 1B, LXA 4 did not impact receptor tyrosine kinase activity.
The ALXR binds both lipid and peptide ligands (8). Recent data have shown that the annexin 1-derived peptide Ac2-26 can bind the ALXR (14), suggesting that some of the anti-inflammatory pro-resolution effects of dexamethasone (15) may be mediated via the ALXR (16). Interestingly, we have found that Ac2-26 mimics the effect of LXA 4 on PDGFR activation (Fig.  1C). In addition, we investigated whether LXA 4 could impact EGFR activation, another important pathologic event in renal dysfunction (17). As shown in Fig. 1D, preincubation of MC with LXA 4 (10 nM) caused a decrease in the EGF-activated tyrosine phosphorylation of the EGFR. Similarly, specific PDGFR and EGFR protein-tyrosine kinase inhibitors AG1296 and AG1478 inhibited PDGF-and EGF-induced activation of the PDGFR␤ and EGFR, respectively ( Fig. 1, C and D).
To establish the molecular basis of ALXR-PDGFR␤ crosstalk, we generated a HEK293 cell line stably overexpressing the ALXR (ALXR cells) in a PDGFR␤ null background. Thereafter, we examined the effect of LXA 4 on PDGFR␤ phosphorylation in ALXR and control cells that were transiently transfected with PDGFR␤. Consistent with data generated in MCs, LXA 4 signif-icantly reduced PDGF-BB stimulated PDGFR␤ phosphorylation ( Fig. 2A). In contrast, whereas stimulation of HEK293 cells transiently transfected with PDGFR␤ resulted in PDGFR␤ phosphorylation, preincubation of those cells with LXA 4 had no significant effect on receptor phosphorylation (Fig. 2B). These data indicate that the observed effects of LXA 4 are mediated via the ALXR.
To define the locus/loci at which LXA 4 modulates growth factor receptor tyrosine phosphorylation and subsequent mitogenic events, a series of PDGFR␤ mutants were generated whereby specific tyrosines were substituted for phenylalanine at positions 579, 771, 740/751, and 763/1009. Although stimulation of ALXR cells transiently expressing Y579F and Y771F mutants displayed a clear increase in receptor phosphorylation when stimulated with PDGF-BB, preincubation with LXA 4 reduced the extent of phosphorylation similar to that of vehicle ( Fig. 2, C and D). These residues are specific for recruitment of Src and Ras GTPase-activating protein, respectively (13), and therefore we concluded that LXA 4 was not acting at these loci. In contrast, mutation of the PDGFR␤ binding sites for p85 recruitment (Y740F/Y751F) blocked the effect of LXA 4 on receptor phosphorylation (Fig. 2E). Interestingly, the effect of LXA 4 in cells expressing the Y740F alone was maintained, whereas in cells expressing Y751F, the effect of LXA 4 was partially reduced (supplemental data). These data suggest that the high affinity binding site for the recruitment of p85 (Tyr 751 ) is the site most likely to be modulated by LXA 4 , although both sites must be mutated before a complete block in the effect of LXA 4 is observed. These data are particularly noteworthy given that p85 is a component of the PI 3-kinase complex, and our previous data have demonstrated that LXA 4 inhibits MC PI 3-kinase activity in response to leukotriene D 4 and PDGF, respectively (4,18). Consistent with our observations in ALXR cells expressing the Y740F/Y751F mutant, PDGF-induced phosphorylation of Tyr 751 was significantly inhibited by preincubation with LXA 4 in primary MC as assessed by immunoblot analysis using an antibody specific against PDGFR␤-Y751 phosphorylation (supplemental data).
rylation. Consistent with this hypothesis, the observed reduction in phosphorylation of the PDGFR␤ in MC in response to preincubation with LXA 4 was abolished by the broad spectrum, nonspecific PTP inhibitors vanadate and phenylarsine oxide (Fig. 3A). Interestingly, incubation of MC with the inhibitor mixture alone resulted in a 4-fold maximal increase in basal activation of the PDGFR␤ (data not shown), suggesting a tonic inhibition of RTK activation by PTPs in quiescent MC. To test whether SHP-2 contributes to LXA 4 modulation of PDGF-induced PDGFR␤ phosphorylation, ALXR cells were transfected with a double mutant for the recruitment sites for SHP-2 PTP (Y763F/Y1009F). Consistent with the nonspecific PTP inhibitor data, LXA 4 (10 nM, 30 min) had no effect on PDGF-induced receptor phosphorylation (Fig. 3B), suggesting that LXA 4 may require the recruitment of SHP-2 PTP to the PDGFR␤ to exert its effect.
To further investigate the role of SHP-2 on LXA 4 modulation of PDGF-induced PDGFR␤ phosphorylation, specific SHP-2 activity was measured in SHP-2 immunoprecipitates from differentially treated MC. As seen in Fig. 3C, LXA 4 induced significant increases in SHP-2 activity. Consistent with previous reports (22,23), PDGF significantly increased SHP-2 activity; however, the kinetics at which it did so were considerably different to LXA 4 . Maximal activation in response to PDGF was observed at 5 min, after which no significant activation was detected. In contrast, LXA 4 stimulated a prolonged activation of SHP-2 (Fig. 3C). To evaluate the role of the ALXR in the activation of SHP-2 in response to LXA 4 , MCs were treated with an ALXR antagonist, Boc-2 (24) prior to treatment with LXA 4 . Pretreatment with Boc-2 inhibited the SHP-2 activation in response to LXA 4 (Fig. 3D).
Collectively, our data suggest that LXA 4 regulation of PDGFR␤ activation is mediated via ALXR activation of SHP-2. To explore this hypothesis directly, we investigated whether inhibition of SHP-2 expression could block the effect of LXA 4 on PDGFR␤ activation in ALXR cells. To do so, we used SHP-2-specific siRNA to silence the expression of SHP-2 by RNA interference. Treatment of ALXR cells with siRNA for 0 -96 h resulted in strong reduction of SHP-2 levels relative to nontreated cells or cell treated with scrambled siRNA (Fig. 4A), with maximal inhibition at 50 nM occurring after 48 h. Thereafter, we compared the ability of LXA 4 to modulate PDGFR␤ activation in cells transfected with SHP-2 siRNA and in cells transfected with scrambled siRNA (Fig. 4C). Importantly these effects on inhibition of SHP-2 expression were also seen with other SHP-2 siRNA constructs (supplemental Fig. S3). Consistent with Y740F/Y751F mutant data, LXA 4 had no effect on PDGF-induced receptor phosphorylation in SHP-2 knock-out ALXR cells, whereas the effect of LXA 4 was maintained in ALXR cells transfected with scrambled siRNA (Fig. 4C and supplemental data).
Several RTKs and GPCRs as well as their downstream effector proteins have been reported to be present in or recruit to lipid rafts following activation. Indeed, there is a growing body of evidence that caveolae-containing lipid rafts may provide platforms for specific signaling molecules (26). To investigate whether such microdomains are implicated in ALXR-PDGFR␤ cross-talk, we used a cholesterol-binding agent, methyl-␤-cy-clodextrin, to sequester cholesterol and cause disassembly of cholesterol-rich rafts and caveolae. Incubation of MC with 10 mM methyl-␤-cyclodextrin for 30 min did not alter cell morphology or integrity as measured by trypan blue exclusion (data not shown). Fig. 5A shows that LXA 4 -dependent dephosphorylation of the PDGFR␤ was significantly reduced and approached nontreated, PDGF-stimulated levels after depletion of plasma membrane cholesterol. In contrast, methyl-␤cyclodextrin did not significantly decrease PDGF-BB-stimulated PDGFR␤ phosphorylation (Fig. 5A), implying that it is the transinactivation process that requires cholesterol-rich microdomains. To examine whether the apparent importance of cholesterol for LXA 4 activity reflects the association of the ALXR with lipid microdomains, we used sucrose gradient centrifugation to separate detergent-resistant rafts, which include caveolae, from ALXR cell lysates. Immunoblotting of the eight different fractions obtained revealed that fraction 3 was highly enriched in both caveolin-1 and flotillin-1 (Fig.  5B), whereas proteins such as clathrin and ␤-actin, which are thought not to reside within caveolae in situ, localized to fractions 4 -8 (bulk plasma membrane). Importantly, fraction 3 contained the highest levels of the ALXR, whereas significant levels of transfected PDGFR␤ and SHP-2 are also shown to be present (Fig. 5B).
To investigate whether the LXA 4 -driven dephosphorylation of the PDGFR␤ is localized to the lipid raft fraction, we isolated the soluble (fraction 7) and lipid raft fractions (fraction 3) from the sucrose gradient and immunoblotted for the aforementioned proteins. The PDGFR␤ in both fractions was phosphorylated in response to PDGF-BB treatment, but only the PDGFR␤ in the lipid raft fraction was dephosphorylated in response to preincubation with LXA 4 (Fig. 5C). This reduction in PDGFR␤ phosphorylation in response to LXA 4 was accompanied by concomitant increases in levels of SHP-2 within the lipid raft fraction. Interestingly, SHP-2 levels in the soluble fraction were increased in response to PDGF-BB (Fig. 5C), suggesting that the different ligands activate different subpopulations of the PTP.
Previous reports have suggested that the interaction between PTPs and GPCRs/RTKs might be controlled by regulated distribution of PTPs to lipid rafts (27). To investigate possible recruitment of SHP-2 to the lipid raft-bound ALXR, we immunoprecipitated the HA-tagged ALXR from the lipid raft (Fraction 3) and soluble (Fraction 7) fractions of ALXR cells and immunoblotted for SHP-2. As Fig. 5D demonstrates, SHP-2 was constitutively associated with the ALXR in the lipid raft fraction (fraction 3) of quiescent ALXR cells, whereas no association was seen within the soluble (fraction 7) fraction or with the mock-transfected HEK293 cell lysate control. Moreover, the SHP-2 association seen in the lipid raft fraction was increased upon activation of the receptor by LXA 4 and returned to basal levels 30 min post-stimulation. In contrast, no SHP-1 recruitment to the ALXR was observed in either the lipid raft fraction (fraction 3) or in the soluble fraction (fraction 7) (data not shown).
To understand the molecular mechanism by which PDGF signaling is down-regulated by ALXR activation of SHP-2 within lipid microdomains, we analyzed endogenous H 2 O 2 production, which has been shown to mediate site-selective amplification of PDGFR phosphorylation (25). Recent studies have implicated the involvement of a cellular protease, 2-Cys peroxiredoxin type II (Prx II), in the down-regulation of PDGF signaling via a decrease in localized endogenous H 2 O 2 production (25). To assess the role of LXA 4 within this signaling system, we examined the induction of ROS in MC. Upon PDGF stimulation, LXA 4 -pretreated cells exhibited a significant reduction in the amount of ROS released compared with MC treated with PDGF alone, as measured by reduced 5-(and-6)-carboxyl 2Ј,7Јdichlorodihydrofluorescein diacetate oxidation (Fig. 6A). Whereas MC exhibited significant PDGF-induced Prx II protein expression, pretreatment of cells with LXA 4 prior to PDGF stimulation reduced Prx II expression to levels that were not significantly different from those generated by nontreated cells (Fig. 6B).

DISCUSSION
The mechanisms underlying the resolution of glomerular inflammation remain enigmatic. However, it has been proposed that regulation of MC proliferation by LX may play a key role (2,3,28).
We have previously demonstrated that the anti-proliferative effects of LXA 4 are mediated through activation of the ALXR and result in modulation of growth factor (PDGF and EGF) receptor phosphorylation leading to down-regulation of the PI 3-kinase signaling pathway (4,9,18). Here, we report that the ability of LXA 4 to reduce the tyrosine phosphorylation level of the PDGFR␤ is independent of inhibition of RTK activity but dependent on SHP-2 PTP activation. Moreover, we demonstrate that the effect of LXA 4 is replicated by a low affinity agonist of the ALXR, Ac2-26, (K d for Ac2-26 is 0.9 M versus 1.7 nM for LXA 4 ), a peptide derivative of annexin 1, and that these effects are not restricted to the PDGFR signaling pathway but encompass other pleiotropic RTK such as the EGFR.
The PDGFR␤ is activated by binding of PDGF and undergoes autophosphorylation at multiple tyrosine residues. The tyrosine-phosphorylated receptor associates with numerous SH2 domain-containing signal transduction molecules that include   1-3) or were co-transfected with scrambled siRNA (all lanes 2) or siRNA for SHP-2 (all lanes 3) for 48 h prior to serum restriction in 0.01% FCS for 24 h. Thereafter, the cells were treated with vehicle (5 min), PDGF-BB (10 ng/ml, 5 min), or LXA 4 (10 nM, 30 min) followed by PDGF-BB (10 ng/ml, 5 min). The cells were lysed and immunoblotted for SHP-2 (upper panel). Alternatively, PDGFR␤ was immunoprecipitated and was subsequently immunoblotted for phosphotyrosine (middle panel). Total cellular PDGFR␤ was determined by immunoblotting for PDGFR␤ (lower panel). The blots are representative of three independent experiments. WB, Western blot; P-Y, phosphotyrosine; Veh, vehicle; IP, immunoprecipitation.
phospholipase C␥1, the GTPase-activating protein of Ras, the p85 subunit of PI 3-kinase, the phosphotyrosine phosphatase SHP-2, and several other proteins (29,30). Indeed, individual mutation of these specific tyrosine residues prevents association of its cognate effectors without altering the binding of other enzymes to the receptor. Thus, this phenomenon has afforded us a convenient method to assess the role of LXA 4 transinhibition on the receptor-associated effector-binding sites of the PDGFR␤ in HEK293 cells overexpressing both the ALXR and wild type/mutant PDGFR␤ receptors. Our results suggest that, among the sites for binding of effector molecules (i.e. PI 3-kinase, Ras GTPase-activating protein, Src kinase, and SHP-2), only mutation of the sites for PI-3-kinase and SHP-2 affected the dephosphorylation of the PDGFR␤ by LXA 4 pretreatment. Our data indicate that the observed LXA 4 -mediated transinhibition of PDGFR␤ does not reflect a diminution in receptor autokinase activity but a specific effect on particular phosphotyrosines residues.
SHP-2, a ubiquitously expressed PTP, is a critical intracellular regulator of cytokine and growth factor-induced cell survival, proliferation, and differentiation (31). Several PTPs have been shown to bind PDGFR␤, with SHP-2 being the most efficiently binding PTP (32). SHP-2 binds to Tyr(P) 763 and Tyr(P) 1009 on activated PDGFR␤ and in turn may dephosphorylate selective docking sites that partake in the down-regulation of receptor signaling, including Tyr(P) 751 and Tyr(P) 740 (p85 recruitment sites) (33). However, the effect of SHP-2 on the PDGFR␤ is contradictory. Depending on the cell type and the nature of the agonist, SHP-2 either augments or antagonizes the signaling pathway initiated by activation of the PDGFR␤. For example, SHP-2 can act as a positive regulator for PDGF signaling by acting as an adapter between the PDGF receptor and the Grb2-Sos complex, leading to activation of the Ras signaling pathway (34). In contrast, Zhao and Zhao (35) demonstrated that catalytically active forms of SHP-2 decreased the tyrosine phosphorylation of the PDGFR␤ but enhanced PDGF-induced activation of the mitogen-activated protein kinase pathway (35). From our results, activation of SHP-2 PTP by LXA 4 via the ALXR leading to reduced PDGFR␤ phosphorylation is supported by multiple lines of evidence. First, the effect of LXA 4 on PDGFR␤ phosphorylation in MC was diminished in the presence of PTP inhibitors vanadate and phenylarsine oxide. Second, we detected sustained elevation of SHP-2 PTP activity in immunoprecipitates from LXA 4 -treated MC. This effect was mediated via the ALXR given that inclusion of the nonspecific ALXR antagonist, Boc-2, blocked LXA 4induced SHP-2 activation. Third, we demonstrate that mutation of the SHP-2 recruitment sites on the PDGFR␤ results in an inhibition of the LXA 4 effect and therefore suggests that SHP-2 is recruited to the PDGFR␤ in response to activation by LXA 4 . Furthermore, siRNA-directed modulation of SHP-2 expression resulted in inhibition of the LXA 4 effect on the PDGFR␤. Importantly, the EGFR, like the PDGFR␤, also contains tyrosine residues for the recruitment of SHP-2 (37), suggesting that multiple phosphorylated growth factor receptors may act as substrates for SHP-2 and in turn may be modulated by LXA 4 . MCs were also preincubated with M␤C (10 mg/ml, 30 min) and/or LXA 4 (10 nM, 30 min) prior to PDGF-BB (10 ng/ml, 5 min). The lysates were harvested, and PDGFR␤ was immunoprecipitated from all conditions, followed by immunoblotting as indicated. B, ALXR cells were transiently transfected with pcDNA3.1-PDGFR␤ prior to serum restriction in 0.01% FCS for 24 h. The cells were subsequently treated with vehicle (30 min) and lysed with TNE buffer, and a sucrose flotation experiment was carried out as described under "Experimental Procedures. " Thereafter, immunoblotting for caveolin-1, flotillin-1, ␤-actin, clathrin, HA-tagged ALXR, SHP-2, and PDGFR␤ was carried out on all gradient fractions. The blots shown are representatives of three independent experiments. C, ALXR cells transiently transfected with pcDNA3.1-PDGFR␤ were serumrestricted in 0.01% FCS for 24 h prior to treatment with vehicle (5 min), PDGF-BB (10 ng/ml, 5 min), or LXA 4 (10 nM, 30 min) followed by PDGF-BB (10 ng/ml, 5 min). The cells were lysed with TNE buffer, and a sucrose flotation experiment was carried out as described under "Experimental Procedures. " Lipid raft (Fraction 3) and soluble fractions (Fraction 7) were analyzed by immunoblotting for HA-tagged ALXR, SHP-2, PDGFR␤, phosphotyrosine, caveolin-1, and ␤-actin. D, ALXR cells were serum-restricted in 0.01% FCS for 24 h prior to treatment with vehicle (5 min) or LXA 4 (10 nM, 5-30 min). Thereafter, the cells were lysed with TNE buffer, and a sucrose flotation experiment was carried out as previously described under "Experimental Procedures. " To analyze the association of the ALXR with SHP-2, lipid raft fraction (Fraction 3) lysates were subject to immunoprecipitation with anti-HA serum to immunoprecipitate the ALXR. Soluble fractions (Fraction 7), lysates from mock-transfected HEK-293 cells and mouse IgG served as negative controls. Immunoprecipitates were resolved by SDS-PAGE/immunoblotting and screened with anti-SHP-2 as described under "Experimental Procedures. " All of the blots are representatives of three independent experiments. WB, Western blot; P-Y, phosphotyrosine; Veh, vehicle; IP, immunoprecipitation; M␤CD, methyl-␤-cyclodextrin; Cav-1, caveolin-1. The temporal requirements of exposure to LXA 4 for significant inhibition of phosphorylation of RTK (30 -60 min) suggested that the effect may reflect RTK trafficking from the cell membrane (combined with PTP activation). Receptor downregulation appears to be an important element in the control of cellular proliferation, because mutations that impair the ability of some receptors to internalize and degrade cause cell transformation (41). We have investigated the involvement of cholesterol-rich lipid microdomains in the LXA 4 effect on the PDGFR␤. These microdomains participate in endocytosis and are thought to facilitate recruitment and organization of downstream effector molecules into specialized signaling complexes (26). We have employed a cholesterol-depleting agent to disrupt lipid rafts and demonstrate that intact lipid rafts are necessary for LXA 4 transinhibition of the PDGFR␤. From our data, we identify that caveolae-enriched fractions of HEK293 cells contain both the ALXR and PDGFR␤, and we conclude that the ALXR requires association with lipid rafts to initiate both the recruitment of SHP-2 and the dephosphorylation of the PDGFR␤. The targeting of GPCRs and RTKs to lipid rafts following activation can occur through different mechanisms. The PDGFR␤ is known to be recruited to lipid raft fractions upon ligand activation (42), whereas other receptors exist in lipid rafts apparently as a result of their random distribution in the plasma membrane (43). Both the ALXR and PDGFR␤ were found to reside in the lipid raft fraction in ALXR cells. However, we found no enrichment of either receptor in the lipid raft fraction in response to ligand activation. A previous study has identified distinct mechanisms by which SHP-2 is recruited to the cell membrane in a multiprotein complex that is distributed differently inside and outside lipid rafts (44). In our study we found increased levels of SHP-2 in the lipid raft fraction of ALXR cells in response to LXA 4 treatment, whereas PDGF stimulated an increase of SHP-2 levels in the soluble fraction, suggesting that both ligands target different cellular pools of the PTP. Interestingly, only the lipid raft-associated PDGFR␤ was found to be dephosphorylated in response to LXA 4 pretreatment, indicating a specific deactivation process involving the lipid raft population of SHP-2.
In further studies, we uncovered a constitutive association of SHP-2 with the lipid raft-bound ALXR in HEK293 cells stably expressing the ALXR. This association was increased in response to ligand activation and returned to basal levels in a time-dependent manner. In this context, it is noteworthy that SHP-2 itself has been shown to be directly recruited and subsequently activated by the immunoreceptor tyrosine-based inhibitor motif (21) sequence in the bradykinin B2 receptor in MC, resulting in global dephosphorylation of tyrosine residues and inhibition of proliferation (20). We have identified a consensus immunoreceptor tyrosine-based inhibitor motif (45) in the ALXR (supplemental data), located at the interface of the N-terminal part of the second transmembrane domain and the first intracellular domain. Previous studies have demonstrated that occupation of the SH2 domain of SHP-2 by a ligand may stimulate both subcellular relocalization of the enzyme to the vicinity of substrates or may also stimulate the phosphatase activity (46). Thus, we envision that SHP-2 is recruited to the ALXR in lipid microdomains via Tyr(P)-SH2 interaction, and subsequent ALXR dephosphorylation eventually leads to dephosphorylation of the SH2-binding domain-binding sites and disruption of the complex.
Numerous studies suggest that ROS such as O 2 . and H 2 O 2 function as mitogenic mediators of activated growth factor receptor signaling such as the PDGFR (38,39). PDGF-induced tyrosine phosphorylation and DNA synthesis was inhibited when the growth factor-stimulated rise in H 2 O 2 concentration was blocked. Moreover, Choi et al. (25) report that the mammalian Prx II allows oxidatively inactivated protein-tyrosine phosphatases to be reactivated by removing endogenous H 2 O 2 , resulting in decreased phosphorylation of key tyrosine residues in both PDGFR␤ and the downstream signaling molecule phospholipase C␥1 (25). We have observed robust PrxII up-regulation in response to PDGF after 30 min of stimulation, a response similar to that reported for serum-stimulated epithelial cells (47). In an effort to elucidate the molecular mechanism by which PDGF signaling is down-regulated by ALXR activation of SHP-2, we examined the influence of LXA 4 on ROS and Prx II levels produced in response to PDGF in MC. LXA 4 attenuated PDGF-stimulated ROS production, consistent with previous data from experiments carried out in human polymorphonuclear leukocytes (40). Interestingly, although we have shown that LXA 4 results in the prolonged activation of SHP-2, LXA 4 abolished PDGF-induced Prx II protein expression. This leads us to suggest that LXA 4 attenuation of ROS/H 2 O 2 production and hence activation of SHP-2 may be independent of Prx II. Conversely, Prx II down-regulation is a direct result of LXA 4 attenuation of ROS production (21,38). Therefore, we propose that LXA 4 possibly mimics Prx II in that it potentially reactivates oxidatively inactivated SHP-2 PTP by removing endogenous H 2 O 2 .
In summary, based on our general understanding of PDGFR signaling, we now propose a model for SHP-2 regulation of the PDGFR␤ in response to LXA 4 (Fig. 7). LXA 4 binding to the lipid raft-bound ALXR induces a recruitment to and reactivation of oxidatively inactivated SHP-2 in the lipid raft region. In a time-dependent manner, recruited SHP-2 is released from the ALXR whereby it is free to bind to and dephosphorylate the lipid raft population of the PDGFR␤ and hence diminish its subsequent mitogenic activities. Interestingly, SHP-2 has been shown to exert siteselective dephosphorylating activity, showing preference for the PI 3-kinase-binding sites (33). Our results support this hypothesis as mutation of the PDGFR␤-binding sites for p85 recruitment (Y740F/Y751F) blocked the effect of LXA 4 on receptor phosphorylation. Taken together, these data further elaborate on the potential mechanisms underlying the anti-inflammatory, pro-resolution, and anti-fibrotic bioactions of LX.