|Home | About | Journals | Submit | Contact Us | Français|
To understand the contribution of EGFR transactivation in GPCR agonist-induced signaling events, we have studied the capacity of thrombin in the activation of Gab1-SHP2 in vascular smooth muscle cells (VSMCs). Thrombin activated both Gab1 and SHP2 in a time- and EGFR-dependent manner. Similarly, thrombin induced both Rac1 and Cdc42 activation and these responses were suppressed when either Gab1 or SHP2 stimulation is blocked. Thrombin also induced PAK1 activation in a time- and EGFR-Gab1-SHP2-Rac1/Cdc42-dependent manner. Inhibition of activation of EGFR, Gab1, SHP2, Rac1, Cdc42 or PAK1 by pharmacological or genetic approaches significantly suppressed thrombin-induced VSMC stress fiber formation and motility. Thrombin activated RhoA in a time-dependent manner in VSMCs. LARG, a RhoA-specific GEF, was found to be associated with Gab1 and siRNA-mediated depletion of its levels suppressed RhoA, Rac1 and PAK1 activation. Dominant negative mutant-mediated interference of RhoA activation inhibited thrombin-induced Rac1 and PAK1 stimulation in VSMCs and their stress fiber formation and migration. Balloon injury induced PAK1 activity and interference with its activation led to attenuation of SMC migration from media to intima, resulting in reduced neointima formation and increased lumen size. Inhibition of thrombin signaling by recombinant Hirudin also blocked balloon injury-induced EGFR tyrosine phosphorylation and PAK1 activity. These results show that thrombin-mediated PAK1 activation plays a crucial role in vascular wall remodeling and it could be a potential target for drug development against these vascular lesions.
Besides its role in coagulation, thrombin elicits both mitogenic and motogenic actions in a variety of cell types, including vascular smooth muscle cells (VSMCs) (1–5). Thrombin mediates its effects via G protein-coupled protease-activated receptors (PARs), specifically the high affinity receptor, PAR-1 (3, 4, 6). In addition, thrombin-induced mitogenic and motogenic effects exhibit a requirement for transactivation of receptor tyrosine kinases (RTKs) such as epidermal growth factor receptor (EGFR), fibroblast growth factor receptor-1 (FGFR-1) and insulin-like growth factor-I receptor (IGF-IR) in various cell types, including VSMC (7–11). Furthermore, it was reported that transactivation of RTKs, particularly EGFR by GPCR agonists, is sufficient in the propagation of signaling events downstream to receptor activation (12, 13). In this aspect studies from others as well as our laboratory showed that transactivation of EGFR by thrombin leads to stimulation of extracellular signal-regulated kinases (ERKs) and phosphatidylinositol-3 kinase (PI3K) (2, 13). But it is less clear in regard to what extent RTK transactivation, e.g., EGFR, by GPCR agonists such as thrombin leads to activation of signaling events that are otherwise stimulated in response to a true ligand-induced RTK activation. One of the signaling events that are activated upon EGFR tyrosine phosphorylation is the recruitment of Grb2-associated binder 1 (Gab1) and its associated phosphatase, SHP2 onto the receptor (14, 15). Using tissue-specific knockout or knock-in mouse models, many studies have shown that Gab1-SHP2 plays a crucial role in a variety of cellular functions including cell proliferation and migration (15, 16).
Although a large number of studies have reported activation of EGFR by thrombin, it is less clear whether this GPCR agonist possess the capacity to activate Gab1/SHP2. Furthermore, the role of Gab1/SHP2 in the activation small GTPases such as Rac1/Cdc42 and their downstream target PAK1 is also not known. To test this, we have studied thrombin effects on Gab1-SHP2 activation and their involvement in Rac1/Cdc42-mediated PAK1 stimulation in VSMCs. We found that thrombin activates PAK1 via a signaling involving EGFR, Gab1/SHP2, LARG, RhoA and Rac1 as well as Cdc42 in VSMCs. Thrombin-induced PAK1 activation was also found to be crucial in mediating vascular wall remodeling in response to injury. In addition, phosphorylation of Thr423 but not Ser144 of PAK1 was observed to be correlated with its kinase activity in VSMCs in response to thrombin and in the artery in response to injury.
Isolation of rat VSMCs, Western blot analysis, rat carotid artery balloon injury, in vitro and in vivo VSMC migration and immunohistochemisrty were performed as described previously (17). All the animal protocols were performed in accordance with the relevant guidelines and regulations approved by the Internal Animal Care & Use Committee of the University of Tennessee Health Science Center, Memphis, TN.
Data analysis for statistical significance of variance was performed by Student’s t-test.
For a detailed Materials and Methods section, please see the online supplement at http://circres.ahajournals.org.
To understand the mechanisms by which thrombin induces VSMC migration, we have tested the role of Gab1 and SHP2. Thrombin at 0.5 U/ml induced tyrosine phosphorylation of Gab1 as measured by immunobloting of anti-Gab1 immunoprecipitates of control and various time periods of thrombin-treated VSMCs with anti-PY20 antibodies. Increases in tyrosine phosphorylation of Gab1 were observed at 10 min and peaked at 30 to 60 min (Figure 1A). Sequential probing of this membrane with anti-SHP2 antibodies showed a band with molecular moss of 72 kDa whose intensities were found to be higher in thrombin-treated VSMCs as compared to control suggesting association of SHP2 with tyrosine phosphorylated Gab1 in response to thrombin. In a converse experiment thrombin induced tyrosine phosphorylation of SHP2 as measured by immunobloting of anti-SHP2 immunoprecipitates of control and various time periods of thrombin-treated VSMCs with anti-PY20 antibodies (Figure 1B). Sequential probing of this membrane with anti-Gab1 antibodies showed a band with a molecular mass of 110 kDa whose intensities were found to be higher in thrombin-treated VSMCs compared to control indicating association of Gab1 with SHP2 in tyrosine phosphorylation-dependent manner. To identify the upstream mechanisms of Gab1 and SHP2 tyrosine phosphorylation, we examined the role of EGFR. Thrombin induced tyrosine phosphorylation of EGFR as measured by immunobloting of anti-EGFR immunoprecipitates of control and various time periods of thrombin-treated VSMC with anti-PY20 antibodies (Figure 1C). Maximum increases in tyrosine phosphorylation of EGFR occurred at 10 min and these increases were sustained at least for 2 hrs. Sequential probing of this membrane with anti-Gab1 and anti-SHP2 antibodies revealed their association with EGFR in tyrosine phosphorylation-dependent manner in response to thrombin. To test whether EGFR tyrosine kinase activity is required for thrombin-induced Gab1 and SHP2 tyrosine phosphorylation and/or their association, quiescent VSMCs were treated with and without thrombin (0.5 U/ml) in the presence and absence of 500 nM AG1478, a potent inhibitor of EGFR (8) and cell extracts were prepared. An equal amount protein from control and each treatment was analyzed for Gab1 tyrosine phosphorylation by immunoprecipitation with anti-Gab1 antibodies and immunoblotting with anti-PY20 antibodies. AG1478 significantly blocked thrombin-induced Gab1 tyrosine phosphorylation (Figure 1D). Reprobing of this membrane with anti-SHP2 antibodies revealed association of SHP2 with Gab1 in EGFR tyrosine kinase activity-dependent manner. To find the functional significance of Gab1 and SHP2 activation, we further tested their role in thrombin-induced VSMC F-actin stress fiber formation and migration. Thrombin treatment caused extensive F-actin stress fiber formation and these responses were completely blocked by adenovirus-mediated expression of dominant negative mutants of either Gab1 (dnGab1) or SHP2 (dnSHP2) (Figure 2A). Similarly, pretreatment with 500 nM AG1478 inhibited thrombin-induced F-actin stress fiber formation (Figure 2B). Adenovirus-mediated expression of dnGab1 or dnSHP2 or pretreatment with AG1478 also reduced thrombin-induced VSMC migration as measured by modified Boyden chamber method (Figure 2C & D).
The Rho family of small GTPases plays an important role in the regulation of F-actin stress fiber formation, which is essential for cell migration and proliferation (18–21). To identify the downstream effector molecules of EGFR-Gab1/SHP2 signaling, we next studied their role in the activation of GTPases, Rac1 and Cdc42 by thrombin. Quiescent VSMCs were treated with and without 0.5 U/ml thrombin for various time periods and cell extracts were prepared. An equal amount of protein from control and each treatment was subjected to pull-down assay using GST-PAK Sepharose-CL4B beads followed by Western blotting for either Rac1 or Cdc42 using their specific antibodies. As shown in Figure 3, thrombin induced activation of both Rac1 and Cdc42 in a time-dependent manner with maximum effect at 30 min to 60 min. Adenovirus-mediated expression of either dnGab1 or dnSHP2 or pretreatment with AG1478 inhibited thrombin-induced Rac1 and Cdc42 activation by 80% (Figure 3B & C). Interference with activation of Rac1 or Cdc42 via adenovirus-mediated expression of their dominant negative mutants (dnRac1 and dnCdc42, respectively) attenuated thrombin-induced VSMC F-actin stress fiber formation and migration (Figure 3D & E). Many studies have demonstrated that Rac1/Cdc42 target PAK1 in the mediation of cell migration (22–26). Therefore, to find if this was the case for the role of Rac1/Cdc42, we next studied the time course effect of thrombin on activation of PAK1. Thrombin while having no noticeable effect on Ser144 phosphorylation induced Thr423 phosphorylation of PAK1 in a time-dependent manner with maximum effect at 30 min to 60 min (Figure 4A). To confirm the activation of PAK1 by thrombin, we also measured its activity by immunocomplex kinase assay. Consistent with its effect on Thr423 phosphorylation, thrombin induced PAK1 activity in a time-dependent manner with a near maximum increase between 30 min and 60 min (Figure 4B). To test the role of PAK1 in thrombin-induced VSMC migration, cells were transduced with dominant negative PAK1 (dnPAK1) adenovirus at 40 moi, quiesced, treated with and without thrombin (0.5 U/ml) for appropriate time periods and tested for PAK1 activity, F-actin stress fiber formation and migration. Adenovirus-mediated expression of dnPAK1 suppressed thrombin-induced PAK1 activity, F-actin stress fiber formation and VSMC migration (Figure 4C, D & E). To unfold the mechanisms by which thrombin activates PAK1, we next tested the role of EGFR-Gab1/SHP2-Rac1/Cdc42 signaling. Blockade of Gab1, SHP2, Rac1 or Cdc42 activation by adenovirus-mediated expression of their respective dominant negative mutants or suppression of EGFR activity by AG1478 inhibited both thrombin-induced PAK1 Thr423 phosphorylation and its activity (Figure 5A–F).
Guanine nucleotide exchange factors (GEFs) play an important role in agonist-induced activation of GTPases (27). To understand the mechanisms by which Gab1 mediates GTPase stimulation, we tested the role of GEFs. Co-immunoprecipitation assays revealed that LARG, a RhoA-specific GEF (28), forms a complex with Gab1 in a time-dependent manner in response to thrombin (Figure 6A). In addition, siRNA-mediated depletion of LARG inhibited thrombin-induced Rac1 and PAK1 activation (Figure 6B & C). Furthermore, thrombin stimulated RhoA in a time-dependent manner (Figure 6D). Inhibition of EGFR by AG1478 or siRNA-mediated depletion of either Gab1 or LARG levels substantially reduced thrombin-induced RhoA activation (Figure 6E, F & G). Adenovirus-mediated expression of dominant negative mutant of RhoA attenuated thrombin-induced Rac1 and PAK1 activation and stress fiber formation, resulting in reduced VSMC migration (Figure 6H–K).
To understand the role of PAK1 in vascular wall remodeling in vivo, we have examined its involvement in injury-induced SMC migration and neointima formation. First, mechanical injury of rat carotid artery induced both Ser144 and Thr423 phosphorylation of PAK1 as early as 6 h after injury and peaked at 12 h after injury (Figure 7A). However, the steady-state levels of PAK1 were decreased by 30% to 40% at these time periods after injury as compared to its levels in uninjured arteries. Consistent with its phosphorylation, PAK1 activity was also increased in the arteries in response to injury (Figure 7B). To find whether activation of PAK1 occurs in SMC, double immunofluorescence staining was performed in the cryosections of injured and uninjured arteries for SMα-actin and Thr423 phosphorylated PAK1. As shown in Figure 7C, double immunofluorescence staining for SMα-actin and PAK1 Thr423 phosphorylation revealed PAK1 activation in SMC in response to injury. Adenovirus-mediated expression of dnPAK1 in the arteries suppressed only Thr423 phosphorylation of PAK1 and its activity (Figure 8A & B). Dominant negative mutant-mediated inhibition of PAK1 activation also reduced injury-induced SMC migration from medial to luminal surface and thereby attenuated neointima formation by 60% resulting in increased luminal size (Figure 8C & D). To find the link between thrombin, EGFR and PAK1 in balloon injury-induced vascular wall remodeling, we tested the effect of rHirudin. As compared to uninjured arteries, balloon injured arteries 16 hrs after injury showed increased EGFR tyrosine phosphorylation. Intravenous administration of a bolus dose of rHirudin (75 units) just before and 8 hrs after balloon angioplasty inhibited both injury-induced EGFR tyrosine phosphorylation and PAK1 activation (Figure 8E).
The important findings of the present study are as follows: 1. Thrombin stimulated tyrosine phosphorylation of Gab1 and SHP2 in a time-dependent manner in VSMCs. 2. Thrombin also induced tyrosine phosphorylation of EGFR with a time course similar to those of Gab1 and SHP2 in VSMCs. 3. Both Gab1 and SHP2 were found to be associated with tyrosine-phosphorylated EGFR and inhibition of EGFR tyrosine kinase activity by AG1478 suppressed the phosphorylation of Gab1 and its association with SHP2. 4. Adenovirus-mediated expression of dominant negative Gab1 or SHP2 or pretreatment with AG1478 suppressed thrombin-induced VSMC F-actin stress fiber formation and migration. 5. Thrombin activated both Rac1 and Cdc42 in a manner that is dependent on activation of EGFR and Gab1/SHP2 and blockade of Rac1 and Cdc42 inhibited thrombin-induced VSMC F-actin stress fiber formation and migration. 6. Thrombin stimulated Thr423 phosphorylation of PAK1 and its activity in a time-dependent manner and EGFR tyrosine kinase inhibitor AG1478 and adenoviral-mediated expression of dominant negative mutants of Gab1, SHP2, Rac1 or Cdc42 reduced both PAK1 phosphorylation and activity. 7. Adenovirus-mediated expression of dominant negative mutant PAK1 also blocked thrombin-induced VSMC F-actin stress fiber formation and migration. 8. LARG, a RhoA-specific GEF, was found to be associated with Gab1 in response to thrombin and siRNA-mediated depletion of its levels blocked thrombin-induced Rac1 and PAK1 activation. 9. Thrombin induced RhoA activation in EGFR-Gab1-LARG-dependent manner and interference with its stimulation via adenovirus-mediated expression of its dominant negative mutant substantially attenuated thrombin-triggered Rac1 and PAK1 activation in VSMCs and their F-actin stress fiber formation and migration. 10. Balloon injury of rat carotid artery induced phosphorylation and activity of PAK1 in SMC six hours after post injury and adenovirus-mediated expression of dnPAK1 suppressed BI-induced SMC migration from media to intima, resulting in reduced neointima formation. 12. BI also induced EGFR tyrosine phosphorylation. 13. Inactivation of thrombin via intravenous administration of rHirudin suppressed BI-induced EGFR tyrosine phosphorylation and PAK1 activation. Together, these findings demonstrate that thrombin signaling involving EGFR, Gab1/SHP2, LARG, RhoA, Rac1/Cdc42 and PAK1, most likely in this axis plays a crucial role in VSMC migration influencing vascular wall remodeling.
It is well established that activation of EGFR leads to recruitment and tyrosine phosphorylation of Gab1/SHP2 (14, 15). It is also demonstrated that GPCR agonists, LPA and thrombin influence tyrosine phosphorylation of EGFR and interference with activation of this receptor reduces the actions of these agents on cell proliferation and migration (7–9, 12). The present results reveal that stimulation of EGFR by thrombin is sufficient to activate the signaling events downstream to the receptor. In addition to SHP2, Gab1 has been shown to recruit PI3K and via different interacting effector molecules it targets the development of various organs. Specifically, it was demonstrated that the recruitment of PI3K by Gab1 is essential for EGFR-mediated embryonic eyelid closure and keratinocyte differentiation whereas Gab1 association with SHP2 is required for Met receptor function in placental development and muscle progenitor cell migration to the limbs (15, 16). Since thrombin activation of EGFR also caused the recruitment of Gab1/SHP2 onto the receptor, it is possible that this signaling complex participates in the regulation of VSMC motility. Evidence in support of this possibility comes by the finding that activation of Gab1/SHP2 is required for thrombin stimulation of Rac1/Cdc42, whose functions have been shown to be essential for F-actin stress fiber formation (18–21). The present findings also provide the first mechanistic evidence for the role of Gab1 in Rac1 activation. Specifically, Gab1 recruits LARG, a RhoA-specific GEF (28), which, in turn, via its influence on RhoA activation leads to stimulation of Rac1. Although an antagonism was observed between RhoA and Rac1 in some cell types in response to many agonists (29), a potential role for RhoA in the activation of Rac1 has also been demonstrated in Swiss 3T3 fibroblasts (30). Based on these findings, it can be further speculated that Gab1 or SHP2 via recruiting and influencing either GEFs, GTPase activating proteins (GAPs) or guanine nucleotide dissociation inhibitors (GDIs) may be facilitating Cdc42 activation by thrombin. Indeed, some reports showed that SHP2 via dephosphorylating p190-B RhoGAP mediates RhoA activation during myogenesis (31). A large body of data suggests that Rho GTPases via influencing the regulation of F-actin stress fiber formation play an important role in the mediation of cell motility (26–29). In fact, our finding that disruption of EGFR-dependent Gab1/SHP2-mediated RhoA, Rac1 or Cdc42 activation signaling aborts thrombin-induced VSMC F-actin stress fiber formation and migration suggests a role for this signaling axis in the regulation of GPCR agonist-induced cell motility.
Many reports showed that RhoA, Rac1 and Cdc42 play a role in the activation of PAK1 (32, 25). However, it is not known whether Gab1/SHP2 targets PAK1 in regulating either cell proliferation or migration in response to RTK or GPCR agonists. In this regard, the present study reveals that thrombin-induced Gab1/SHP2 leads to activation of PAK1. In addition, since blockade of EGFR or Gab1/SHP2 activation suppresses Rac1 and Cdc42 activation and inhibition of these GTPases attenuates PAK1 Thr423 phosphorylation and activity, it is quite likely that thrombin induced PAK1 activation involves EGFR, Gab1/SHP2-Rac1/Cdc42 signaling axis and facilitates F-actin rearrangement and stress fiber formation in VSMCs leading to their migration. Although, how EGFR-Gab1/SHP2 leads to activation of Cdc42 is yet to be explored, it appears that LARG, a RhoA-specific GEF, connects EGFR-Gab1 signaling to PAK1 activation via RhoA-dependent Rac1 stimulation. Furthermore, since no noticeable changes are observed in the Ser144 phosphorylation of PAK1 by thrombin, it is likely that EGFR-Gab1/SHP2-LARG-RhoA-Rac1/Cdc42 signaling does not affect the phosphorylation of this residue. It also appears that Ser144 phosphorylation is not required for PAK1 activity as there was no correlation between these two events in response to thrombin in VSMCs. It is noteworthy that adenovirus-mediated expression of kinase-dead PAK1 while enhancing Ser144 phosphorylation reduced Thr423 phosphorylation and activity of endogenous PAK1, a finding that suggests a correlation between Thr423 phosphorylation and kinase activity. This result indicates that overexpression of kinase-dead PAK1 somehow sequesters endogenous PAK1 from being phosphorylated at Thr423 residue. It also suggests that Thr423 is present in the catalytic domain (25). In regard to PAK1 activation in the arteries in response to injury in vivo, its levels were decreased by about 30% but its phosphorylation both at Ser144 and Thr423 and activity were increased very robustly. This finding suggests that although PAK1 levels were reduced by injury, its activity was increased by enhanced phosphorylation and this appears to be sufficient to activate its downstream signaling events necessary for SMC migration. Since downregulation of PAK1 activation significantly blocked injury-induced neointima formation, it is possible that PAK1 may also be involved in SMC multiplication in response to injury. In fact, a large body of data suggests that PAK1 plays a role in the regulation of cell growth (33). Given the role of PAK1 in the regulation of both cell proliferation and migration, and downregulaion of its activity inhibited injury-induced neointima formation, it is quite likely that PAK1 plays a crucial role in vascular wall remodeling. In addition, since downregulation of thrombin activity via rHirudin inhibited injury-induced EGFR tyrosine phosphorylation and PAK1 activity, it is conceivable that endogenously produced thrombin activates both EGFR and PAK1 in vascular wall as well, contributing to neointima formation following angioplasty. The role of thrombin in injury-induced neointima formation has also been reported previously, but the underlying mechanisms were not explored (34, 35). In this aspect, the present data provides a mechanistic evidence for the role of thrombin in vascular wall remodeling.
PAK1 via its downstream effector substrate, LIM kinase, phosphorylates and inhibits the activity of cofilin, an actin depolymerizing protein, which in turn promotes F-actin filament stabilization and stress fiber formation (36). Since blockade of PAK1 inhibits VSMC migration both in vitro in response to thrombin and in vivo in response to injury, it would be interesting to find which of its downstream substrates are critical in mediating vascular wall remodeling.
This work was supported by NIH grant HL64165 to GNR.