Insulin receptor substrate-1/SHP-2 interaction, a phenotype-dependent switching machinery of insulin-like growth factor-I signaling in vascular smooth muscle cells.

Insulin-like growth factor-I (IGF-I) plays a role in mutually exclusive processes such as proliferation and differentiation in a variety of cell types. IGF-I is a potent mitogen and motogen for dedifferentiated vascular smooth muscle cells (VSMCs) in vivo and in vitro. However, in differentiated VSMCs, IGF-I is only required for maintaining the differentiated phenotype. Here we investigated the VSMC phenotype-dependent signaling and biological processes triggered by IGF-I. In differentiated VSMCs, IGF-I activated a protein-tyrosine phosphatase, SHP-2, recruited by insulin receptor substrate-1 (IRS-1). The activated SHP-2 then dephosphorylated IRS-1 Tyr(P)-895, resulting in blockade of the pathways from IRS-1/Grb2/Sos to the ERK and p38 MAPK. Conversely, such negative regulation was silent in dedifferentiated VSMCs, where IGF-I activated both MAPKs via IRS-1/Grb2/Sos interaction-linked Ras activation, leading to proliferation and migration. Thus, our present results demonstrate that the IRS-1/SHP-2 interaction acts as a switch controlling VSMC phenotype-dependent IGF-I-induced signaling pathways and biological processes, and this mechanism is likely to be applicable to other cells.

Recent studies (8) have demonstrated the involvement of IGF-induced signals in the mutually exclusive processes of cell proliferation and differentiation. For instance, pharmacological study showed that although the MAPK pathway induced by IGF-I is involved in the mitogenic response of myoblasts, the PI3K/p70 S6K pathway is critical for myogenic differentiation. In hematopoietic cells, IGF-I signaling via IRS-1 plays a role in proliferation and via Shc in differentiation (9). In 3T3-L1 preadipocytes, the IGF-I-induced tyrosine phosphorylation of Shc, Shc-Grb2 complex formation, and ERK activation is observed in proliferating, but not differentiating, cells (10). In osteoblasts, IGF-IR-mediated signals via the PI3K, ERK, and p38 MAPK pathways are all involved in differentiation (11). Thus, the downstream pathways of IGF-I signaling seem to depend on the cellular context. IGF-I is also a potent mitogen and motogen for vascular smooth muscle cells (VSMCs) in vivo and in vitro (12,13); disruption of the IGF-I/IGF-IR interaction by a neutralizing anti-IGF-IR antibody (14) and suppression of IGF-IR expression by antisense oligonucleotides (15) and transcripts (16) inhibit IGF-I-induced VSMC proliferation and migration in culture. Targeted overexpression of IGF-I in arteries in vivo (17) and infusion of IGF-I into the circulatory system (18) have been shown to accelerate neointimal formation after vascular injury. On the other hand, we recently reported a primary culture system of VSMCs that exhibit a differentiated phenotype, in which isolated VSMCs are cultured on laminin under IGF-I-stimulated conditions (19,20). By using this culture system, we demonstrated that IGF-I is critically involved in maintaining the differentiated phenotype via the PI3K/PKB(Akt) pathway and that the coordinated activation of ERK and p38 MAPK triggered by other growth factors induces VSMC dedifferentiation. Taking these findings together, we concluded that changes in the balance between the strengths of the PI3K/PKB(Akt) and the ERK and p38 MAPK pathways determines the VSMC phenotype. However, none of these studies, including those using VSMCs, revealed the cellular phenotype-dependent switching machinery of the IGF-induced signaling pathways.
Here we studied the VSMC phenotype-dependent actions of IGF-I by characterizing the IGF-I-induced signaling pathways. In differentiated VSMCs, SHP-2 preferentially associated with IRS-1 and inhibited the interaction between IRS-1 and Grb2/ Sos, resulting in a blockade of the pathways from Ras to ERK and p38 MAPK. Therefore, in differentiated VSMCs, IGF-1 solely activates the PI3K/PKB(Akt) pathway, which plays a role in maintaining the differentiated phenotype. Conversely, the proliferation and migration of dedifferentiated VSMCs in response to IGF-I were because of an IGF-I-induced interaction between IRS-1 and Grb2/Sos, followed by the activation of Ras and its downstream kinases, ERK and p38 MAPK. This is the first report demonstrating that the cellular phenotype-dependent IRS-1/SHP-2 interaction functions as a switch for IGF-I signalings and distinct biological processes.
VSMC Cultures-To obtain a large scale preparation of VSMCs, we modified our original method (20) as follows. Briefly, the aortae were quickly removed from ether-anesthetized 6-week-old Sprague-Dawley rats and were incubated for 30 min in the buffer containing 0.5 mg/ml collagenase (Yakult) and 0.5 mg/ml bovine serum albumin (BSA). Then the thin layers of adventitia and endothelium were carefully stripped off. The medial VSMC layers were minced well, and digested in Dulbecco's modified Eagle's medium containing 0.5 mg/ml collagenase, 0.5 mg/ml elastase (Sigma), and 5 mg/ml BSA for 60 min at 37°C. Isolated VSMCs were filtered through a nylon membrane. The cells thus obtained were plated on laminin-coated plastic dishes and cultured in Dulbecco's modified Eagle's medium supplemented with 2 mg/ml BSA and other additions, as indicated. Passaged VSMCs were prepared from the medial VSMC layers of rat aortae as described previously (22). VSMCs passaged 2-5 times were used as dedifferentiated VSMCs.
Expression of SMC Marker Genes-The expression of h-and l-CaDs, CN, and MHC SM2 was quantified by reverse transcriptase-PCR normalized to the expression of glyceraldehyde-3-phosphate dehydrogenase mRNA as described previously (20,21).
Kinase Assays-The activities of endogenous kinases or exogenously expressed FLAG-ERK and FLAG-p38 MAPK were determined after immunoprecipitation with anti-ERK, anti-p38 MAPK, anti-PKB(Akt), or anti-FLAG antibodies. The kinase activities in the immunoprecipitates were assayed as described elsewhere (19,23,24). Assay conditions for FLAG-ERK and FLAG-p38 MAPK activities were as follows. VSMCs were cotransfected with the expression plasmid for FLAG-ERK2 or -p38 MAPK and control plasmid or the expression plasmids for ⌬mSos, RasN17, RasV12, myc-SHP-2wt, myc-SHP-2 DN , IRS-1-HA wt , IRS-1-HA F1172/F1222 , or IRS-1-HA F895 . After 24 h, they were stimulated with IGF-I (100 ng/ml) for 10 min. To confirm the expression of exogenous recombinant proteins, the cell extracts were analyzed by immunoblot. The rat IRS-1 cDNA was kindly provided by Dr. T. Kadowaki (University of Tokyo). Mutant IRS-1 cDNAs were constructed using PCRassociated mutagenesis, and their sequences were confirmed. The expression plasmids for RasN17 and RasV12 were kindly provided by Dr. T. Hirano (Osaka University Graduate School of Medicine).
Interactions between Signaling Molecules-VSMCs of both phenotypes under quiescent culture conditions were stimulated with the indicated amounts of IGF-I and lysed in buffer containing 20 mM Tris-HCl, pH 7.5, 137 mM NaCl, 1% Nonidet P-40, 10% glycerol, 2 mM EDTA, 50 mM NaF, 1 mM vanadate, 10 mM ␤-glycerophosphate, 50 g/ml PMSF, 10 g/ml leupeptin, and 5 g/ml pepstatin, for 30 min at 0°C. The interaction of signaling molecules because of exogenously expressed wild-type or mutant SHP-2 or IRS-1 was analyzed as follows. VSMCs were cotransfected with the expression plasmid for FLAG-Grb2 and control plasmid or the expression plasmid for myc-SHP-2 wt or myc-SHP-2 DN (Fig. 5A), or they were transfected with the expression plasmid for IRS-1-HA wt , IRS-1-HA F1172/F1222 , or IRS-1-HA F895 (Fig. 5B). They were then stimulated with IGF-I (100 ng/ml) for 5 min. Equal amounts (400 g of protein) of the cell extracts were incubated with the indicated antibodies for 6 h at 4°C. The immune complexes were collected by incubation with protein A-or protein G-Sepharose beads for 2 h at 4°C. Proteins in the immunoprecipitates were detected by immunoblot by using the indicated antibodies. Target proteins were detected with a SuperSignal chemiluminescent detection kit (Pierce) and quantified using the NIH Image program.
Ras Activation Assay-The Ras activity was monitored using a Ras assay kit (Upstate Biotechnology, Inc.) according to the manufacturer's recommended procedures. In brief, the VSMC extracts were incubated with 15 g of the Ras binding domain (RBD) of Raf-1-agarose beads for 30 min. Ras bound to Raf-1 RBD was detected by immunoblot using an anti-Ras antibody and then quantified.
PTPase Assay-VSMCs cultured under quiescent conditions were stimulated with IGF-I (100 ng/ml) or platelet-derived growth factor-BB (PDGF-BB) (20 ng/ml) for 5 min. The SHP-2 protein in the cell extracts was immunoprecipitated using an anti-SHP-2 antibody. The PTPase activities in the immunoprecipitates were assayed using tyrosine-phosphorylated Raytide as a substrate (25).
Cell Proliferation and Migration Assays-VSMCs of both phenotypes were cultured under quiescent conditions for 24 h and then stimulated with the indicated amounts of IGF-I. VSMC proliferation was determined by the incorporation of 5-bromo-2-deoxyuridine (BrdUrd) (20 M) for 20 h. Incorporated BrdUrd was visualized by staining with an anti-BrdUrd antibody, followed by labeling with secondary antibodies conjugated with Alexa 546 (Molecular Probes) and Hoechst. The migration of differentiated VSMCs (four 1.5-mm diameter fields in each experiment) was monitored for 48 h by a cooled CCD camera (Roper Scientific, Tucson, AZ) mounted on an Olympus IX-70 microscope. The migration of dedifferentiated VSMCs was assayed by wound healing. Confluent cultures of dedifferentiated VSMCs were wounded with a single-edged razor blade. After stimulation with the indicated amounts of IGF-I for 48 h, the numbers of migrated cells were counted. The proliferation of dedifferentiated VSMCs transfected with wild-type or mutant HA-tagged IRS-1-expressing plasmids was assayed as follows. Dedifferentiated VSMCs (40 -50% confluence) were transfected with the indicated expression plasmids and were cultured for 24 h. After serum starvation, they were cultured in the basal medium containing IGF-I (100 ng/ml) and BrdUrd for 20 h. The expression of HA-tagged IRS-I derivatives was visualized with anti-HA antibodies and Alexa 488 (Molecular Probes)-conjugated secondary antibodies. The migration of dedifferentiated VSMCs transfected with the indicated expression plasmids was monitored as described above. After stimulation with IGF-I (100 ng/ml) for 48 h, the expression of HA-tagged IRS-I derivatives was visualized as described above. The numbers of migrated cells and cells expressing each of HA-tagged IRS-1 derivatives were counted.

RESULTS
VSMC Phenotype-dependent IGF-I Signaling Pathways-As demonstrated previously (19,20), primary cultured VSMCs plated on laminin under IGF-I-stimulated conditions showed the differentiated phenotype, as indicated by a spindle-like cell shape, carbachol (CCH)-induced contractility (Fig. 1A), and high expression of SMC markers, such as h-CaD, CN, and MHC SM2, at the mRNA and protein levels (Fig. 1B). By contrast, passaged VSMCs, even when cultured under quiescent conditions, displayed a fibroblast-like shape change (Fig.  1A), loss of CCH-induced contractility (Fig. 1A), and downregulation of SMC marker expression in addition to isoform conversion of CaD from the h-to l-form (Fig. 1B), indicating the dedifferentiated phenotype. By using both phenotypes of VSMCs, we first examined the dose-dependent response to IGF-I ( Fig. 1, C-E). Consistent with our previous study (20), IGF-I (2 ng/ml) markedly activated PKB(Akt) in differentiated and dedifferentiated VSMCs, and its activation reached a maximum (more than 10-fold activation) at 20 ng/ml IGF-I (Fig.  1C). R 3 IGF-I, an IGF-I analog lacking affinity for IGF-binding proteins, showed a potency equivalent to IGF-I in activating PKB(Akt) in both phenotypes of VSMCs, at concentrations from 2 to 100 ng/ml (data not shown), suggesting that the IGF-I-induced signaling is IGF-binding protein-independent. The PKB(Akt) activation by IGF-I in both VSMC phenotypes was suppressed by LY294002 (10 -30 M) or wortmannin (10 -30 nM), but not by PD98059 (20 -30 M) or SB203580 (10 -20 M), indicating that IGF-I induces the PKB(Akt) activation via PI3K. Notably, IGF-I, even at high concentrations (100 ng/ml), did not activate ERK or p38 MAPK in differentiated VSMCs, whereas it dose-dependently activated them in dedifferentiated VSMCs; their activations reached a maximum (more than 15-and 5-fold activations of ERK and p38 MAPK, respectively) at 20 ng/ml IGF-I (Fig. 1, D and E). In particular, the activation rates of PKB(Akt) and ERK by IGF-I were nearly identical, but p38 MAPK showed a slightly different IGF-I dose responsibility. The IGF-I-induced ERK activation was suppressed by PD98059 (20 - IGF-I-induced Interactions between Signaling Molecules-Signaling molecules, including IGF-IR, IRS-1, PI3K, Gab1, PK-B(Akt), Sos, Grb2, Ras, ERK, p38 MAPK, SHP-2, SHPS-1, and Shc, have been shown to be involved in IGF-I signaling in a variety of cells, including VSMCs (1,3,14,16,17). We compared the expression of these signaling molecules (Fig. 2). The IGF-IR␤, IRS-1, Shc, Gab1, SHPS-1, Sos, and Ras proteins were markedly up-regulated in dedifferentiated VSMCs compared with differentiated VSMCs, whereas ERK, p38 MAPK, p85 PI3K, PKB(Akt), SHP-2, and Grb2 proteins were equally expressed in both VSMC phenotypes. Thus, some signaling molecules show different expression profiles in different VSMC phenotypes.
To elucidate the molecular mechanism involved in VSMC phenotype-dependent IGF-I signalings, we examined the ty- rosine phosphorylation and interactions of the signaling molecules. After IGF-I (2 and 100 ng/ml) stimulation, an anti-IGF-IR␤ antibody specifically coimmunoprecipitated two tyrosine-phosphorylated proteins with molecular masses of 97 and 185 kDa in both VSMC phenotypes. We identified these proteins as IGF-IR␤ (97 kDa) and IRS-1 (185 kDa) by using their respective antibodies (Fig. 3, A and D). An anti-IRS-1 antibody also coimmunoprecipitated IGF-IR␤ in an IGF-I dose-dependent manner (data not shown). Tyrosinephosphorylated IGF-IR␤ and/or IRS-1 formed a complex with p85 PI3K but did not associate with Shc in either VSMC phenotype (Fig. 3, A and B, and D--F). We detected a stable complex of Grb2 and Sos in the anti-Grb2 (Fig. 3C) and anti-Sos (data not shown) immunoprecipitates of both VSMC phenotypes with or without IGF-I stimulation. Immunoblotting of the anti-Grb2 immunoprecipitates with anti-phosphotyrosine, anti-IRS-1, and anti-Sos antibodies showed that a complex of tyrosine-phosphorylated IRS-1 and Grb2/Sos formed in the dedifferentiated, but not in the differentiated, VSMCs (Fig. 3C), suggesting that the tripartite complex composed of IRS-1, Grb2, and Sos depends on the IGF-I-induced tyrosine phosphorylation of IRS-1. Immunoblotting of the anti-Sos immunoprecipitates with anti-IRS-1 and anti-Grb2 antibodies showed the same interactions in the IGF-I-stimulated dedifferentiated VSMCs (data not shown). The same tripartite complex was also detected in the anti-IRS-1 immunoprecipitates (Fig. 3B). Neither the IGF-I-induced tyrosinephosphorylated Shc nor an Shc and IGF-IR␤ or Grb2/Sos interaction was detected in either VSMC phenotype. As a control, PDGF-BB markedly induced tyrosine phosphorylation of Shc and an interaction between Shc and Grb2/Sos (Fig. 3F). The protein recognized by anti-phosphotyrosine antibody in the Shc immunoprecipitates from both phenotypes of VSMC was an IgG heavy chain, because anti-phosphotyrosine antibody cross-reacted with IgG heavy chain in rabbit antibodies (Fig. 3F). Because PDGF-BB or 18:1 lysophosphatidic acid potently induced the tyrosine phosphorylation of Shc in both VSMC phenotypes ( Fig. 3F and data not shown), IGF-I signaling in VSMCs is Shc-independent. Thus, our results suggest that tyrosine-phosphorylated IRS-1 preferentially recruits the Grb2-Sos complex in dedifferentiated but not in differentiated VSMCs.
VSMC Phenotype-dependent Interaction of Tyrosine-phosphorylated IRS-1 and SHP-2-IRS-1 is phosphorylated on multiple tyrosine residues in response to IGFs. Among them, phosphorylated Tyr-895 of IRS-1 has been identified as a binding site for Grb2 (26). We analyzed the phosphorylation of IRS-1 Tyr-895 by using an anti-IRS-1 Tyr(P)-895 antibody. The IRS-1 Tyr-895 was markedly phosphorylated by IGF-I stimulation only in dedifferentiated VSMCs (Fig. 4A). This is consistent with the IGF-I-induced tripartite complex formation between tyrosine-phosphorylated IRS-1 and Grb2/Sos in dedifferentiated, but not in differentiated, VSMCs (Fig. 3). It is well documented that tyrosine-phosphorylated IRS-1 interacts with SHP-2, the ubiquitous PTPase containing the SH2 domain, leading to the positive or negative regulation of receptor tyrosine kinase-linked events (27)(28)(29)(30)(31). IRS-1 has two tyrosine residues for SHP-2 binding, Tyr-1172 and Tyr-1222 (32). We characterized the tyrosine phosphorylation of these SHP-2-binding sites using anti-IRS-1 Tyr(P)-1172 and Tyr(P)-1222 antibodies, respectively. In a preliminary experiment, the former antibody specifically recognized IRS-1 Tyr(P)-1172, but the latter did not recognize Tyr(P)-1222 (data not shown). We therefore analyzed the phosphorylated Tyr-1172 of IRS-1 in IGF-I-stimulated VSMCs. In response to IGF-I stimulation, IRS-1 Tyr-1172 was significantly phosphorylated in both VSMC phenotypes (Fig.  4A). We then examined the possible involvement of SHP-2 in the VSMC phenotype-dependent interaction between IRS-1 and Grb2/Sos. In both VSMC phenotypes under quiescent and IGF-I-stimulated conditions, almost all of the SHP-2 protein was extractable by nonionic detergents. Notably, IGF-I dosedependently enhanced the PTPase activity of SHP-2 in differentiated VSMCs, whereas its effect was less significant in dedifferentiated ones. As a control, PDGF-BB enhanced the PTPase activity of SHP-2 in both the VSMC phenotypes (Fig.  4B). The PTPase activities might be mainly originated from the SHP-2 PTPase activity itself, because exogenously expressed Myc-tagged PTPase-dead (dominant-negative) SHP-2 (myc-SHP-2 DN ) (34) in IGF-I-stimulated differentiated VSMCs showed the low PTPase activity; 11.2 Ϯ 1.0% of the wild-type SHP-2 (myc-SHP-2 wt ) PTPase activity (Fig. 4B). Thus, the IGF-I-induced PTPase activation is VSMC phenotype-specific. Consistent with this, an interaction between SHP-2 and tyrosinephosphorylated IRS-1 was detected only in IGF-I-stimulated, differentiated VSMCs. The interaction reached a maximum 5 min after IGF-I stimulation, and decreased thereafter (Fig.  4C). This interaction process coincided well with the progressive changes in phosphorylation of Tyr-1172 (Fig. 4D). Thus, the extent of this interaction in differentiated VSMCs may depend on the phosphorylation of Tyr-1172 and possibly Tyr-1222 in IRS-1. No IRS-1/SHP-2 interaction was detected in dedifferentiated VSMCs, even when IRS-1 Tyr-1172 was phosphorylated by IGF-I stimulation (Fig. 4D), suggesting that an additional inhibitory mechanism of this interaction may exist in dedifferentiated VSMCs. No interactions between SHP-2 and other tyrosine-phosphorylated adaptor proteins, such as SHPS-1, Gab1, and Shc, were detected in either VSMC phenotype, even under the reported assay conditions (33, data not shown). Together, our results suggest that SHP-2 recruited to IRS-1 might rapidly dephosphorylate the IRS-1 Tyr-895 in differentiated VSMCs only.
Regulation of the IRS-1 and Grb2/Sos Interaction by SHP-2-We examined the effects of wild-type and dominant-negative SHP-2 on the IGF-I-induced interaction between IRS-1 and Grb2/Sos. The extracts from IGF-I-stimulated, differentiated, and dedifferentiated VSMCs cotransfected with expression plasmids for FLAG-tagged Grb2 (FLAG-Grb2) and myc-SHP-2 wt or myc-SHP-2 DN were immunoprecipitated with an anti-FLAG antibody, followed by immunoblotting with an anti-IRS-1 antibody (Fig. 5A). Overexpressed myc-SHP-2 DN in differentiated VSMCs caused a new interaction between IRS-1 and Grb2, but myc-SHP-2 wt did not. These results suggest that overexpressed SHP-2 DN competes with endogenously activated SHP-2, leading to induction of the IRS-1 and Grb2/Sos interaction. In dedifferentiated VSMCs, overexpressed myc-SHP-2 wt markedly suppressed the formation of this complex, but myc-SHP-2 DN did not. This negative regulation in dedifferentiated VSMCs may be due to dephosphorylation of the IRS-1 Tyr(P)-895 by the overexpressed cytosolic SHP-2 wt .
Negative Regulation of ERK and p38 MAPK by SHP-2-To confirm the negative regulation of ERK and p38 MAPK by SHP-2, we cotransfected both phenotypes of VSMCs with expression plasmids for myc-SHP-2 (wt or DN) or IRS-1-HA (wt or indicated mutants) and FLAG-ERK or FLAG-p38 MAPK, and we assayed the ERK and p38 MAPK activities in the anti-FLAG immunoprecipitates (Fig. 6). IGF-I activated ERK and p38 MAPK in differentiated VSMCs expressing myc-SHP-2 DN or IRS-1-HA F1172/F1222 but not in those expressing myc-SHP-2 wt or IRS-1-HA wt . By contrast, IGF-I failed to activate ERK and p38 MAPK in dedifferentiated VSMCs expressing myc-SHP-2 wt or IRS-1-HA F89 , but barely affected either MAPK in dedifferentiated VSMCs expressing myc-SHP-2 DN , IRS-1-HA wt , or IRS-1-HA F1172/F1222 . Thus, the VSMC phenotype-dependent interaction between IRS-1 and Grb2/Sos, which is negatively regulated by SHP-2, is critical for the IGF-I-induced activation of both MAPKs.

Ras as an Upstream Effector for the IGF-I-induced
Activation of ERK and p38 MAPK-Although the Ras/Raf/MEK/ERK cascade is well known in various cell types (35), the pathway from Ras to p38 MAPK, however, remains unclear. Our present results suggest that Ras may be involved in the IGF-I-induced activation of both ERK and p38 MAPK in dedifferentiated VSMCs. We monitored the IGF-I-induced Ras activation with a

IRS-1/SHP-2 Interaction Regulates IGF-I Signaling
Raf-1 binding assay. Consistent with the kinase activities (Fig.  1C) and interactions between signaling molecules (Fig. 3), IGF-I dose-dependently activated Ras in dedifferentiated, but not differentiated, VSMCs (Fig 7A). This activation was transient; Ras in dedifferentiated VSMCs was maximally activated at 5-10 min after IGF-I stimulation and the activation decreased thereafter. By contrast, IGF-I (2-100 ng/ml) maximally activated ERK and p38 MAPK for 10 min, and their activities remained high 4 h after stimulation (Fig. 7B). Thus, the transient activation of Ras induced by IGF-I led to rapidly rising and falling activities of both MAPKs, but the activities remained above the basal levels even 4 h after stimulation. To address the direct involvement of Ras in the IGF-I-induced MAPK activation in dedifferentiated VSMCs, we cotransfected dedifferentiated VSMCs with ⌬mSos (36), RasN17, or RasV12 (37), and FLAG-ERK or FLAG-p38 MAPK, and we assayed for the ERK and p38 MAPK activities in the anti-FLAG immunoprecipitates. ⌬mSos or RasN17 markedly suppressed the IGF-I-induced activation of both MAPKs; RasN17 inhibited both MAPK activities to less than basal levels, whereas ⌬mSos did them to nearly the basal levels (Fig. 7C). In contrast, their activities were greatly enhanced by RasV12 without IGF-I stimulation (Fig. 7C). Thus, the Grb2/Sos/Ras cascade induced by IGF-I activates both ERK and p38 MAPK in dedifferentiated VSMCs.
Relationships between IGF-I Signaling Pathways and Biological Actions of IGF-I on Distinct Phenotypes of VSMCs-To reveal the relationship between VSMC phenotype-dependent IGF-I signaling and its biological actions, we analyzed the IGF-I-induced cell proliferation and migration. Even under IGF-Istimulated conditions, differentiated VSMCs never underwent cell proliferation, as monitored by BrdUrd incorporation (Fig.  8A) or migration (Fig. 8B), suggesting that IGF-I acts solely to maintain the differentiated phenotype but not as a mitogen or motogen. Conversely, IGF-I dose-dependently enhanced the proliferation and migration of dedifferentiated VSMCs (Fig. 8,  A and B). Treatment with either PD98059 (30 M) or SB203580 (20 M) partially inhibited the BrdUrd incorporation and migration, but simultaneous treatment with both inhibitors completely suppressed them (Fig. 8, A and B), suggesting that these biological actions depend on the IGF-I-induced activation of ERK and p38 MAPK. We further analyzed the effects of the IGF-I-dependent interaction between IRS-I and Grb2/Sos on the BrdUrd incorporation and migration of dedifferentiated VSMCs by overexpressing IRS-1-HA wt , IRS-1-HA F895 , or IRS-1-HA F1172/F1222 . Dedifferentiated VSMCs expressing IRS-1-HA F895 showed markedly less BrdUrd incorporation and migration compared with those expressing either IRS-1-HA wt or IRS-1-HA F1172/F1222 (Fig. 9). These results indicate that the negative regulation of ERK and p38 MAPK by blockade of the IRS-1 and Grb2/Sos interaction results in inhibition of the proliferation and migration of dedifferentiated VSMCs. DISCUSSION Several studies have suggested that IGF-I is a potent mitogen and motogen for VSMCs, because IGF-I-induced activation of PI3K and/or MAPKs is involved in the proliferation and migration of VSMCs (12,13). Conversely, we demonstrated by using our VSMC culture system that IGF-I plays a critical role in maintaining the differentiated phenotype of VSMCs via the PI3K/PKB(Akt), but not the ERK or p38 MAPK, pathway (19,20). Here we demonstrate the distinct upstream signaling pathways, from IGF-IR to PI3K/PKB(Akt) and MAPKs, in different VSMC phenotypes and that the inconsistency regarding the distinct biological actions of IGF-I on VSMCs is because of the different VSMC phenotypes used. IGF-I potently activated ERK and p38 MAPK, via IRS-I-associated Grb2/Sos and Ras only in dedifferentiated VSMCs (Figs. 1 and 5-7), resulting in induction of the VSMC proliferation and migration (Figs. 8 and 9). Previous in vivo studies demonstrated that targeted overexpression of IGF-I in VSMCs (17) or IGF-I infusion into the circulatory system (18) enhances neointimal formation after vascular injury. Based on our present results, these in vivo observations might be explained by the idea that once VSMCs are primed by dedifferentiation stimuli such as vascular injury, IGF-I might further accelerate the proliferation and migration of dedifferentiated VSMCs via the signaling pathway shown in Fig. 10.
In differentiated VSMCs, IRS-I-associated SHP-2 negatively regulates the activation of both ERK and p38 MAPK (Fig. 6). It has been reported that a variety of PTPases are involved in the positive and negative regulation of signal transduction via receptor tyrosine kinases (38). For instance, CD45 positively regulates the T-cell antigen receptor signaling in lymphocytes (39). In hematopoietic cell lines, SHP-1 negatively regulates mitogenic signals through the colony-stimulating factor 1 and stem-cell factor receptors (40). SHP-2 can modulate receptor tyrosine kinase-mediated signaling in a positive or negative manner. Positive regulation by SHP-2 has been reported in mitogenic signaling through the insulin receptor (27,29). However, SHP-2 also negatively regulates insulin receptor-mediated mitogenic signaling (31) and gp130-dependent transcriptional activation (30). Our present study demonstrates that SHP-2 is critical for the VSMC phenotype-dependent regulation of IGF-I signaling; in differentiated VSMCs, an IRS-1/ SHP-2 interaction enhances the PTPase activity of SHP-2 ( Fig.  4) and suppresses the recruitment of Grb2/Sos to IRS-1 by the dephosphorylation of IRS-1 Tyr-895, resulting in inhibition of the activation of both MAPKs (Figs. 5 and 6). In contrast, SHP-2 remains inactive in IGF-I-stimulated dedifferentiated VSMCs (Fig. 4). The quiescence of SHP-2 PTPase activity in dedifferentiated VSMCs is a specific feature of the IGF-I signaling, given that PDGF-BB markedly enhanced the SHP-2 activity in dedifferentiated VSMCs (Fig. 4), and SHP-2 never interacted with signaling molecules such as IRS-1 (Fig. 4), SHPS-1, Gab1, or Shc (data not shown) in the IGF-I-stimulated dedifferentiated VSMCs. Maile and Clemmons (41) have re-ported that SHP-2 in passaged porcine VSMCs associates with SHPS-1, IGF-IR␤, and IRS-1 in response to IGF-I stimulation, and these interactions depend on integrin-associated protein. It is reasonable to consider that the pathway via SHPS-1 might not function in differentiated VSMCs, given that no integrinassociated protein expression is detected in differentiated VSMCs in vivo (42). Our results using dedifferentiated VSMCs also showed this pathway to be silent. Presently, the reasons for this discrepancy remain unknown. The IRS-1 Tyr-1172 was significantly phosphorylated in IGF-I-stimulated dedifferentiated VSMCs. However, the IRS-1/SHP-2 interaction is not detected in dedifferentiated VSMCs despite such a tyrosine phosphorylation state of IRS-1 (Fig. 4). The IGF-I-induced interaction between IRS-1 and Grb2/Sos and the activation of ERK and p38 MAPK are markedly inhibited in dedifferentiated VSMCs overexpressing SHP-2 wt (Figs. 5 and 6). These results suggest that in dedifferentiated VSMCs, the function of SHP-2 would be blocked by interacting with other tyrosine-phosphorylated proteins. Therefore, excess SHP-2 wt might overcome such trapping of endogenous SHP-2 or dephosphorylate IRS-1 Tyr-895, resulting in blockade of the interaction between IRS-1 and Grb2/Sos.
Ras is a key transducer of mitogenic signaling in many cell types, including VSMCs, in vitro and in vivo. It is well documented that Ras activates ERK in a variety of cells, whereas stress-activated protein kinases, including p38 MAPK, are poorly activated by the Ras-coupled cascade (35). We now demonstrate that the activation of Ras functions as an upstream effector for both ERK and p38 MAPK in IGF-I-stimulated dedifferentiated VSMCs (Fig. 7). However, unlike the ERK pathway mediated through Ras/Raf/MEK, the p38 MAPK pathway might be indirectly affected by Ras, because IGF-I dose-responsive activation rates of both MAPKs were slightly different (Fig. 1, D and E). It has been reported that in some cell lines, including myoblast lines PKB(Akt), phosphorylate Raf, leading to suppression of the Ras/Raf/MEK/ERK axis (43). In our present study by using both phenotypes of VSMCs, treatment with PI3K inhibitor never increased the ERK activation by IGF-I ( Fig. 1), suggesting that the PKB(Akt)-mediated negative regulation of the ERK pathway may not be involved.
Our present study provides a novel insight into the signal transduction in VSMCs and into the pharmacological approaches that might be applicable for preventing vascular disorders such as atherosclerosis and restenosis. As described in the Introduction, IGF-I signal affects the mutually exclusive biological actions, such as proliferation and differentiation of hematopoietic cells, adipocytes, myoblasts, and osteoblasts. In these cells, the signaling pathways analogous to those of VSMCs (Fig. 10) may be involved in the distinct biological actions of IGF-I. In our preliminary experiments, we have identified the similar pathways regulating the IGF-I-induced proliferation and differentiation of myogenic cells. 2 Thus, the IGF-I-induced pathways demonstrated here appear to represent a novel mechanism regulating the discriminating signals involved in cell proliferation and differentiation, not just in VSMCs, but in other cells as well.

FIG. 10. IGF-I plays a bifunctional role in the different VSMC phenotypes that is mediated through different signaling pathways.
In differentiated (D) VSMCs, IGF-I induces the tyrosine phosphorylation of IGF-IR␤ and IRS-1, but Grb2/Sos never associates with tyrosinephosphorylated IRS-1, leading to inactivation of the ERK and p38 MAPK pathways. This negative regulation depends on SHP-2 associated with phosphorylated Tyr-1172 and Tyr-1222 of IRS-1, which dephosphorylates Tyr-895 of IRS-1. Tyrosine-phosphorylated IRS-1 also recruits p85 PI3K and then activates the PI3K/PKB(Akt) pathway, resulting in maintenance of the differentiated VSMC phenotype. In dedifferentiated (DD) VSMCs, the expression of some signaling molecules involved in the MAPK pathways is markedly up-regulated. In these cells, such IGF-I-induced negative regulation by SHP-2 is silent. Thus, phosphorylated Tyr-895 of IRS-1 recruits the Grb2/Sos and activates Ras followed by both MAPKs, leading to cell proliferation and migration.