Báo cáo khoa học: Plasminogen activator inhibitor type-1 inhibits insulin signaling by competing with avb3 integrin for vitronectin binding pptx

PRIORITY PAPER Plasminogen activator inhibitor type-1 inhibits insulin signaling by competing with avb3 integrin for vitronectin binding Roser Lo ´ pez-Alemany 1 , Juan M. Redondo 2 , Yoshikuni Nagamine 3 * and Pura Mun ˜ oz-Ca ´ noves 1,4 * 1 Institut de Recerca Oncolo ` gica (IRO), Centre d’Oncologia Molecular, L’Hospitalet de Llobregat, Barcelona, Spain; 2 C.B.M. Severo Ochoa, C.S.I.C., U.A.M., Facultat de Ciencias, Madrid, Spain; 3 Friedrich Miescher Institute for Biomedical Research, Novartis Research Foundation, Basel, Switzerland; 4 Centre de Regulacio ´ Geno ` mica (CRG), Programa de Diferenciacio ´ i Cancer, Barcelona, Spain Functional cooperation between integrins and growth factor receptors has been reported for several systems, one of which is the modulation of insulin signaling by avb3 integrin. Plasminogen activator inhibitor type-1 (PAI-1), competes with avb3 integrin for vitronectin (VN) binding. Here we report that PAI-1, in a VN-dependent manner, prevents the cooperation of avb3 integrin with insulin signaling in NIH3T3 fibroblasts, resulting in a decrease in insulin- induced protein kinase B (PKB) phosphorylation, vascular endothelial growth factor (VEGF) expression and cell migration. Insulin-induced HUVEC migration and angio- tube formation was also enhanced in the presence of VN and this enhancement is inhibited by PAI-1. By using specific PAI-1 mutants with either VN binding or plasminogen activator (PA) inhibiting activities ablated, we have shown that the PAI-1-mediated interference with insulin signaling occurs through its direct interaction with VN, and not through its PA neutralizing activity. Moreover, using cells deficient for uPA receptor (uPAR) we have demonstrated that the inhibition of PAI-1 on insulin signaling is inde- pendent of uPAR-VN binding. These results constitute the first demonstration of the interaction of PAI-1 with the insulin response. Keywords: plasminogen activator inhibitor type-1; vitro- nectin; insulin; angiogenesis; HUVEC. Insulin plays a central role in regulating metabolic pathways associated with energy storage and utilization. Its action is initiated by receptor-mediated tyrosine phosphorylation of ShcA and insulin receptor substrates (IRSs), and recruite- ment of these molecules to the intracellular receptor domain [1]. ShcA is linked to the Ras/Erk signaling pathway while IRS provides docking sites for several other signaling molecules including phosphatidylinositol 3-kinase (PI3-K), phospholipase Cc and Grb2, conveying insulin signals to various cellular events. IRS proteins couple the insulin receptor to both the PI3-K and MAPK pathways, and ShcA (through Gab1) also couples the insulin receptor to both pathways (reviewed in [2]). Perturbations of insulin- induced metabolic responses are associated with severe health complications, such as type 2 diabetes and obesity [3]. Under these pathological conditions, sensitivity of the insulin receptor to insulin is significantly reduced. Insulin stimulates endothelial cell survival [4] and promotes angio- genesis in vivo [5,6]. It is therefore noteworthy that type 2 diabetes is very often accompanied by cardiovascular complications, of which endothelial dysfunction is an early event [7]. Several lines of evidence indicate that integrin-mediated signaling processes synergize with growth factor responses [8,9]. In particular, integrin avb3, the main receptor of vitronectin (VN), potentiates platelet-derived growth factor (PDGF) and insulin/insulin-like growth factor (IGF) recep- tor signaling responses [10,11]. Smooth muscle cell migration in response to IGF1 is dependent on VN occupancy of avb3 intregrin [12,13]. IRS-1 binds to avb3 after insulin receptor activation in Rat1 fibroblasts, providing a potential mech- anism for the synergistic action between growth factors and extracellular matrix receptors [11]. Furthermore, Persad et al. have recently shown that the integrin-linked kinase (ILK) can directly phosphorylate PKB, suggesting a role as an upstream regulator for PKB [14]. VN occurs in the circulation as a complex with plasmi- nogen activator inhibitor-1 (PAI-1), the primary physiolo- gical inhibitor of fibrinolysis [15]. PAI-1 is synthesized in an active form but is rapidly converted into an inactive form [16]. Association with VN stabilizes PAI-1 and targets it to Correspondence to P. Mun ˜ oz-Ca ´ noves, Centre de Regulacio ´ Geno ` mica (CRG), Programa de Diferenciacio ´ iCa ´ ncer, Passeig Maritim 37-49, E-08003 Barcelona, Spain. Fax: + 34 93 224 08 99, Tel.: + 34 93 224 09 33, E-mail: pura.munoz@crg.es Y. Nagamine, Friedrich Miescher Institute for Biomedical Research, Maubeleerstrasse 66, CH-4058 Basel, Switzerland. Fax: + 41 61 697 3976, Tel.: + 41 61 697 6669, E-mail: nagamine@fmi.ch Abbreviations: CsA, cyclosporin A; HUVEC, human umbilical vein endothelial cells; IGF, insulin-like growth factor; IRS, insulin receptor substrate; PAI-1, plasminogen activator inhibitor type-1; PI3-K, phosphatidylinositol 3-kinase; PKB, protein kinase B; tPA, tissue-type plasminogen activator; uPA, urokinase-type plasminogen activator; uPAR, urokinase-type plasminogen activator receptor; VEGF, vascular endothelial growth factor; VN, vitronectin. *Both authors contributed equally to this work. (Received 24 September 2002, revised 9 December 2002, accepted 7 January 2003) Eur. J. Biochem. 270, 814–821 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03453.x specific sites in the extracellular matrix. PAI-1 binds to the N-terminal domain of VN, adjacent to the RGD sequence (i.e. the integrin binding site), preventing VN from binding to avb3 [17]. VN interacts not only with PAI-1 but also with the urokinase receptor (uPAR) [18]. PAI-1 and uPAR share binding regions on the VN molecule and thus bind competitively [19]. In this work, we have tested whether PAI-1 might be modifying the cellular response to insulin. We demonstrate for the first time that PAI-1 modulates insulin signaling by preventing the binding of VN to avb3. This results in a decrease in insulin-induced PKB phosphorylation, VEGF expression, endothelial cell migration and angiotube formation. Experimental procedures Cells NIH3T3 cells were obtained from the American Type Culture Collection (ATCC, Rockville, MD, USA). HUVEC were isolated from umbilical veins and cultured as described previously [20]. All experiments were per- formed using HUVEC between passages 3 and 6. Mouse embryo fibroblasts (MEFs) from mice lacking urokinase- type plasminogen receptor gene (uPAR-/–) were gently provided by F. Blasi and M. Resnati (DIBIT, Milan, Italy). Proteins Murine vitronectin and recombinant murine PAI-1 in an active conformation were from Molecular Innovations (Royal Oak, MI). Mutants of PAI-1 were gently provided by D. Lawrence (American Red Cross, Rockville, MD, USA): PAI-1 containing a mutation of Gln123 to Lys (PAI- 1K) has a specific defect in VN binding; PAI-1 containing a mutation of Arg340 to Ala (PAI-1A) binds VN with wild- type affinity but does not inhibit plasminogen activation [21]. Insulin from bovine pancreas was from Sigma. Western-blot analysis Cells were grown to subconfluency and then kept overnight in serum-free medium. After treatments, cells were lysed in 20 m M TrisHClpH 8,150 m M NaCl, 2 m M EDTA, 100 l M Na 3 VO 4 ,10m M NaF, 25 l M b-glycerophosphate, 1% Triton X-100 and 100 m M phenylmethanesulfonyl fluoride. Phosphorylated PKB was detected with an anti-(phospho- PKB Ser473) polyclonal Ig (New England Biolabs). Cell lysates and conditioned media were collected and tested, for intracellular VEGF and secreted VEGF, respectively, with a polyclonal anti-VEGF Ig (sc-152, Santa Cruz Biotech- nology, Santa Cruz, CA, USA). Migration assay Cell migration assays were performed on Transwells (8-lm pore size), coated with Matrigel (Beckton and Dickinson, Bedford, MA, USA; 1 : 20 dilution in serum free medium). After blocking with 1% (w/v) BSA, Transwells were incubated with 5 lgÆmL )1 murine VN (Molecular Innova- tions, Royal Oak, MI, USA) or 5 lgÆmL )1 collagen type IV (Sigma). Cells (4 · 10 4 ) were added to the upper chamber of the Transwell. Insulin (170 n M ), 100 n M PAI-1, echistatin (100 n M ) or decorsin (100 n M ) were added to the media in the upper and lower chamber of the Transwell. After incubation at 37 °C, cells on the underside of the Transwell were fixed in methanol/acetic acid (75 : 25, v/v) and stained with 5% Trypan blue. The number of cells per high-power field that had migrated across the Matrigel were counted, after 16 h for NIH3T3 cells and uPAR–/– MEFs, and after 4 h for HUVEC. Data are presented as the mean values from at least five high-power fields. Proliferation assay Cells (1.0 · 10 4 cells per well) were seeded on 96-well plates previously coated with VN (5 lgÆmL )1 ) or collagen (5 lgÆmL )1 ). After 24 h of culture, complete medium was replaced by medium containing 1% (v/v) fetal bovine serum, 0.5% (w/v) BSA, 170 n M insulin and 0.2 lCi [ 3 H]thymidine per well (Amersham), and cells were cultured for a further 24 h for NIH3T3 cells or for 72 h for HUVEC. Cells were precipitated by the addition of 5% cold trichloroacetic acid and precipitates solubilized in 1 M NaOH. Radioactivity (c.p.m.) in the lysate was detected by liquid scintillation counting. Each experimental point was determined five times. In vitro angiogenesis assay HUVEC (2.0 · 10 4 cells per well) were resuspended in OPTI-MEM (Life Technologies) supplemented with 1% (v/v) fetal bovine serum. Cells were treated with 100 n M PAI-1, 30 n M VN or 200 ngÆmL )1 cyclosporin A (CsA, from Sandoz) for 2 h at 37 °C. Matrigel was meanwhile diluted 1 : 2 in cold serum-free RPMI 1640 without growth factors and plated into 96-well plates and allowed to gel for 1–2 h at 37 °C before seeding. The cell suspension was then stimulated with 170 n M insulin or 50 ngÆmL )1 recombinant human VEGF 165 (PeproTech), and plated onto the surface of the Matrigel. After 12 h at 37 °C, cells were photo- graphed with a ZEISS inverted phase-contrast photomicro- scope. Capillary tubes were defined as cellular extensions linking cell masses or branch points, and tube formation was quantified from photographs of standardized fields from triplicate wells. Statistical analysis ANOVA test was used to determine whether there were significant (P < 0.05) differences in cell migration, cell proliferation and angiotube formation under different treatments. Results Vitronectin induction of insulin signaling is inhibited by PAI-1 Integrin-mediated signaling processes synergize with growth factor responses [8,9]. In this work, we first examined whe- ther VN was able to potentiate insulin-mediated responses in NIH3T3 cells. As shown in Fig. 1A, insulin-induced PKB Ó FEBS 2003 PAI-1 inhibition of insulin/vitronectin signaling (Eur. J. Biochem. 270) 815 phosphorylation was augmented when cells were preincu- batedwithVN.Echistatin,aspecificavb3 antagonist, prevented VN enhancement of insulin-induced PKB phos- phorylation. In contrast, decorsin, a structurally distinct disintegrin with high affinity for the IIb/IIIa platelet integrin but low affinity for avb3, had no effect (Fig. 1B), suggesting that the VN enhancement of insulin signaling is dependent on avb3 integrin. To assess the potential role of PAI-1 in the VN- mediated induction of insulin signaling, PAI-1 was included in preincubations with VN. There was a significant decrease in insulin/VN-induced phosphoryl- ation of PKB in the presence of PAI-1 (Fig. 1C), suggesting that PAI-1 might be competing with avb3 for VN binding. Moreover, we observed that urokinase- type plasminogen activator (uPA), which prevents PAI-1/ VN complex formation [22], was able to reverse the PAI-1-mediated inhibition of PKB phosphorylation in response to insulin/VN (Fig. 1C). To determine whether the PAI-1 effect on the insulin response was dependent on its binding to VN or on its plasminogen activator inhibiting activity, two different mutants of PAI-1 were used. A mutant of PAI-1 (PAI- 1A) that binds to VN normally, but does not inhibit plasminogen activators, inhibited phosphorylation of PKB in response to insulin/VN, in a manner identical to that of wild-type PAI-1. In contrast, a second PAI-1 mutant (PAI- 1K) that inhibits plasminogen activators normally, but has a significantly reduced affinity for VN, did not inhibit insulin/ VN phosphorylation of PKB (Fig. 1D). These results indicate that PAI-1 binding to VN is sufficient to block VN binding to avb3, and that the inhibition of insulin signaling by PAI-1 is independent of the antiproteolytic activity of PAI-1. VN interacts not only with PAI-1 but also with the urokinase receptor (uPAR) [18]. PAI-1 and uPAR share binding regions on the VN molecule and thus bind competitively [19]. We hypothesized that the interference of PAI-1 with avb3 integrins leading to the inhibition of the insulin response might occur via uPAR, in a uPA dependent or independent manner. To test this hypothesis we used mouse embryonic fibroblasts (MEFs) derived from uPAR- deficient mice. As seen in Fig. 1E, PAI-1 has the same inhibitory effect on PKB phosphorylation induced by insulin/VN in uPAR-devoid MEFs than in NIH3T3 fibroblast cells, indicating that uPAR does not play a role in the inhibition of insulin/VN signaling by PAI-1. To further assess the functional relevance of PAI-1/VN interaction on insulin signaling, we analyzed the expres- sion of an insulin-inducible gene, the VEGF gene. Insulin induces VEGF expression and secretion in a variety of cell types [23,24], including NIH3T3 cells, via the PI3-K/ PKB signaling cascade [25]. Western analysis of cell lysates demonstrated that VN enhanced insulin-stimulated VEGF expression in NIH3T3 cells, and this induction was inhibited by pretreatment with PAI-1 (Fig. 2A). Additionally, echistatin was also able to inhibit this enhancement, while decorsin had no effect. Moreover, VEGF secretion to the media, which was only detectable when cells were treated with insulin in the presence of VN, was inhibited by PAI-1 or echistatin, but not by decorsin (Fig. 2B). Furthermore, when the effect of mutants of PAI-1 were tested, only PAI-1A was able to inhibit VEGF secretion, as wild-type PAI-1, while mutant PAI-1K had no inhibitory effect (Fig. 2C). Taken together, these results indicate that PAI-1 can inhibit insulin signaling through a mechanism involving VN, and this inhibition is independent on its ability to inhibit plasminogen activation. PAI-1 inhibits cell migration in response to insulin/VN It has been reported that insulin stimulates migration of different cell types including NIH3T3 and human umbi- lical vein endothelial cells (HUVEC) [4,13,26]. Based on the above results, we tested the effect of PAI-1 on insulin- induced cell migration. In VN-containing Matrigel, NIH3T3 cell migration in response to insulin was increased twofold, and this was blocked when PAI-1 was added to the media (Fig. 3A). In contrast, PAI-1 had no effect on insulin-stimulated NIH3T3 cell migration on Fig. 1. Inhibition of insulin/VN induced PKB phosphorylation by PAI-1. NIH3T3 cells were incubated overnight with 100 n M echistatin (Ec) or decorsin (Dc), or for 4 h with 30 n M VN, 100 n M PAI-1 or 200 n M uPA, priorto5 mintreatmentwith170 n M insulin (I) as indicated. Celllysates (40 lg) were subjected to Western blotting with an anti-(phospho-PKB) or anti-PKB Ig (loading control). (A) Insulin-induced PKB phos- phorylation is increased by VN. (B) avb3 specificity in insulin/VN signaling. (C) PAI-1 inhibits insulin/VN-induced PKB phosphoryla- tion. (D) Effect of PAI-1 mutants on insulin/VN-induced PKB phos- phorylation. (E) Effect of insulin/VN/PAI-1 system in uPAR–/– MEFs. Results are representative of at least four independent experiments. 816 R. Lo ´ pez-Alemany et al. (Eur. J. Biochem. 270) Ó FEBS 2003 collagen-containing Matrigel. As expected, echistatin exerted a similar inhibitory effect on insulin-induced cell migration, while decorsin had no effect. When the mutants of PAI-1 were tested in a migration assay, only PAI-1A was able to inhibit insulin-induced migration on VN, while PAI-1K had no effect (Fig. 3B). Furthermore, the migration of uPAR-deficient MEFs on VN-containing Matrigel was also induced by insulin and inhibited by PAI-1 (Fig. 3C) indicating that the interference of PAI-1 with insulin-induced migration occurred through its interaction with VN/avb3, independently of uPAR. Insulin stimulated also HUVEC migration whether cells were plated on VN- or collagen-coated Transwells. However, PAI-1 inhibited insulin-stimulated cell migra- tion only when cells were plated on VN-coated plates (Fig. 3D). These results indicate that PAI-1 was able to block insulin-induced cell migration on VN-containing matrices, and this inhibition was dependent on its ability to bind to VN. Knowing that insulin also stimulates cell proliferation, we wondered whether PAI-1 had any effect on insulin/ VN-induced NIH3T3 and HUVEC proliferation. As shown in Fig. 4, insulin induced 4- and 10-fold [ 3 H]thymidine incorporation in NIH3T3 cells and HUVEC, respectively. In both cell types, this induction was not affected by treatment of cells with PA1-1, either on VN or collagen matrices, indicating that PAI-1 has no effect on insulin-induced proliferation of NIH3T3 and HUVEC proliferation. PAI-1 inhibition of endothelial cell angiogenesis induced by insulin/VN Recently, a role for PAI-1 in angiogenesis in vivo has been proposed [21,27], but the mechanism responsible for this action has yet to be defined. On the other hand, other authors have reported a role for insulin in angiogenesis by stimulating endothelial cell growth as well as tube formation through autocrine VEGF expression [24,28]. We therefore investigated the effect of PAI-1 on insulin-induced angio- tube formation of HUVEC on Matrigel, an in vitro model of angiogenesis. VEGF was used as a positive angiogenesis- inducing factor, and cyclosporin A (CsA) as an inhibitor of VEGF-induced angiogenesis. Consistently with results pre- viously published by our group [29], VEGF induced HUVEC to form capillary-like structures (increase of 39%), while CsA inhibited this induction (Fig. 5). Under identical experimental conditions, insulin induced a 59% increase in angiotube formation, which was markedly potentiated by VN (increase of 93%). PAI-1, in contrast, was able to completely inhibit the VN-mediated augmen- tation of insulin-induced angiotube formation. PAI-1A inhibited angiotube formation at the same level as wild-type PAI-1, while PAI-1K had no effect. Because PAI-1 is highly expressed in endothelial cells in vitro [30,31], it is tempting to hypothesize that PAI-1 inhibition of insulin-mediated endothelial cell migration and angiotube formation can take place in vivo. Discussion We have demonstrated in this study that PAI-1 inhibits insulin signaling, affecting the phosphorylation of PKB, the expression of VEGF, one of key proteins in angiogenesis, and various cell activities including endothelial cell migra- tion and capillary formation. This inhibition is exerted through the interference of PAI-1 with the VN/avb3system. The results shown here provide the first demonstration of the interaction between PAI-1 and the insulin signaling, through the ability of PAI-1 to interact with VN/avb3 system. Using different mutants of PAI-1, we have demon- strated that the inhibition of insulin-induced responses by PAI-1 is due to its ability to form a complex with VN and not to its antifibrinolytic properties. Mutant PAI-1A, which binds normally to VN but lacks plasminogen activator inhibitory activity, was able to inhibit insulin/ VN-induced PKB phosphorylation, VEGF secretion, cell migration and angiotube formation, in a manner identical to that of wild-type PAI-1. In contrast, mutant PAI-1K, which retains the ability to inhibit plasminogen activation but does not bind to VN, had no inhibitory effect in the insulin/VN induced signaling. It is well known that uPAR binds VN and competes with PAI-1 for VN binding [18,19]; therefore PAI-1 could be interfering with insulin/VN through a uPAR/uPA-mediated effect. Because PKB phosphorytion and cell migration in response to insulin/VN was equally inhibited by PAI-1 in uPAR-deficient fibroblasts, we concluded that uPAR Fig. 2. Inhibition of insulin/VN-induced VEGF expression by PAI-1. After the indicated pretreatments, NIH3T3 cells were incubated or not for 4 h with 170 n M insulin. Cell lysates (A) or conditioned media (B and C) were subjected to Western blotting with an anti-VEGF Ig. An anti-tubulin Ig was used as loading control. Results are represen- tative of at least four independent experiments performed. Ó FEBS 2003 PAI-1 inhibition of insulin/vitronectin signaling (Eur. J. Biochem. 270) 817 does not play a role in the inhibition of the insulin/VN response by PAI-1. No effect of PAI-1 was observed on insulin/VN-induced proliferation of NIH3T3 fibroblasts and HUVEC, suggesting that the above described effects of PAI-1 on insulin signaling and insulin-induced proli- feration are uncoupled. As activated insulin receptor recruits several signaling molecules, including IRS and Shc, it is likely that insulin can activate different siganling pathways [32]. The point of divergence of the insulin- induced intracellular responses will be deciphered in future studies. Upon ligand binding, the insulin receptor recruits several signaling molecules, such as IRS. A number of reports suggest that full activation of IRS-1 requires avb3 integrin, although the precise mechanism underlying this cooperation remains to be elucidated [10,11]. In type 2 diabetes, reduced insulin sensitivity is not due to the decrease in insulin receptor number or activity, but rather to reduced activation (tyrosine phosphorylation) of IRS-1 [1]. Under these conditions, IRS-1-mediated insulin sign- aling is impaired while IRS-1-independent signaling still functions. A prominent feature of type 2 diabetic patients is elevated blood levels of PAI-1 [3]. It is therefore reasonable to propose that PAI-1 may contribute to the development of type 2 diabetes by interfering with IRS-1- dependent insulin signaling. In circulation, there is a great excess of VN over PAI-1 (estimated to be 1000–10 000-fold). VN exists both in free form and membrane-bound form; it is synthesized in the liver but it disseminates abundantly into the vessel wall from the plasma [33]. Several studies have demon- strated the increase of VN and PAI-1 during the progression of the atherosclerotic plaques [34,35]. Thus, it is tempting to speculate that PAI-1 concentration could be locally raised to a level enough to inhibit the VN/avb3 system-mediated insulin effects in the atherosclerotic plaque environment. It has been reported that the concentration of active PAI-1 in the atherosclerotic vessel wall is between 25 and 50 n M [36,37], which is in the same range of concentration used in this work, and that has been able to inhibit endothelial cell migration and angiotube formation (Figs 3 and 5). Thus, the inhibition of the insulin response by PAI-1 may well be taking place in vivo, in certain pathological conditions. Fig. 3. PAI-1 inhibits insulin-induced cell migration on VN. (A) Effect of PAI-1 on insulin-induced NIH3T3 cell migration. Transwells were coated with Matrigel containing collagen or VN. The extent of NIH3T3 cell migration through Matrigel is shown with and without 100 n M PAI-1 or disintegrins, in the presence of 170 n M insulin. Cells that had crossed the Matrigel membrane were counted after 16 h migration. (B) Effect of mutants of PAI-1 on insulin-induced NIH3T3 cell migration on VN. (C) Effect of PAI-1 on insulin-induced uPAR–/– MEFs migration. (D) Effect of PAI-1 on insulin-induced HUVEC migration. Transwells were coated with 5 lgÆmL )1 collagen, 5 lgÆmL )1 VN or BSA. The extent of HUVEC migration is shown through Transwells induced by 50 ngÆmL )1 VEGF, 170 n M insulin (± 100 n M PAI-1) and PAI-1 alone. Cells that had crossed the Transwell membrane were counted after 4 h of migration. The data represent the average of four experiments. *P < 0.05 compared to the value vs. control. **P < 0.05 compared to the value corresponding to insulin. 818 R. Lo ´ pez-Alemany et al. (Eur. J. Biochem. 270) Ó FEBS 2003 The interference of PAI-1 with the VN/avb3systemand its consequences on insulin signaling, shown in this work, may constitute a general mechanism by which PAI-1, the main fibrinolytic regulator, could influence insulin-induced responses, providing a potential explanation for the fre- quently observed positive correlation between impairment of the fibrinolytic system and insulin resistance syndrome in diabetes and obesity. Indeed, PAI-1 is highly expressed in diabetic and obese patients and amelioration of insulin- resistance is achieved by diet leading to the reduction of fat tissues, the main source of PAI-1 [38]. PAI-1 has been recognized as a marker of endothelial dysfunction in diseases associated with impaired angio- genesis, including atherosclerosis and diabetic vasculo- pathy [39]. Although regulation of in vitro and in vivo angiogenesis by PAI-1 has been reported, it is not clear whether PAI-1 affects angiogenesis via its antiproteolytic activity or by interacting with VN [21,27]. Our observa- tion that PAI-1 can inhibit insulin-induced HUVEC migration and angiotube formation by interfering with the VN/avb3 system suggests that the latter mechanism is likely. Given that PAI-1 and VN are highly expressed in the atherosclerotic plaques, it is tempting to propose that Fig. 5. PAI-1 inhibits insulin/VN-induced angiotube formation. HUVEC were preincubated with and without CsA (200 ngÆmL )1 )orPAI-1, PAI-1A or PAI-1K (100 n M ),andthenseededin96-wellplatespre- coated with Matrigel containing or not 5 lgÆmL )1 VN. Next, cells were stimulated with 50 ngÆmL )1 VEGF or 170 n M insulin as indicated. Tube formation was quantified 6 h after plating on Matrigel by counting the number of tubular structures in four to six fields. (A) Fold-increase of tubes formed in comparison with the control. Each column represents the mean value ± SD of four different experiments. *P < 0.05 compared to the value vs. insulin; **P <0.05 compared to the value insulin + VN. (B) Representative photographs (original magnification · 100) of four different fields corresponding to the experiment showed in (A). Results are representative of at least four independent experiments performed. Fig. 4. PAI-1 has no effect on insulin/VN induced cell proliferation. The effect of PAI-1 on insulin-induced NIH3T3 cells (A) and HUVEC (B) proliferation. Cells were seeded on VN or collagen-coated plates, and incubated in the absence or presence of 170 n M insulin or 100 n M PAI-1. Cell proliferation was determined by [ 3 H]thymidine incorporation after 24 h of incubation for NIH3T3 cells and after 72 h of incubation for HUVEC. The data represent the average of three experiments. Ó FEBS 2003 PAI-1 inhibition of insulin/vitronectin signaling (Eur. J. Biochem. 270) 819 PAI-1 can contribute to the development of vascular complications associated with diabetes, such as diabetic retinopathy, impaired wound healing, and cardiovascular complications. Acknowledgements We want to thank Dr D. Lawrence for the generous gift of PAI- mutants, and Drs F. Blasi and M. Resnati for uPAR–/– cells. This work was supported by grants from the European Union (1999/C361/06), Fundacio ´ La Marato ´ -TV3, MCyT (SAF2001-0482) and FIS (01/1474), PM 990116 and FIS01/218. We thank A. Blanco for technical assistance. Y.N. is supported partly by the Roche Research Foundation. References 1. Withers, D.J. & White, M. (2000) Perspective: The insulin sig- naling system – a common link in the pathogenesis of type 2 diabetes. Endocrinology 141, 1917–1921. 2. Giovannone, B., Scaldaferri, M.L., Federici, M., Porzio, O., Lauro,D.,Fusco,A.,Sbraccia,P.,Borboni,P.,Lauro,R.&Sesti, G. (2000) Insulin receptor substrate (IRS) transduction system: distinct and overlapping signaling potential. Diabetes Metab. Res. Rev. 16, 434–441. 3. Juhan-Vague, I., Alessi, M.C. & Vague, P. (1996) Thrombogenic and fibrinolytic factors and cardiovascular risk in non-insulin- dependent diabetes mellitus. Ann. Med. 28, 371–380. 4. Hermann, C., Assmus, B., Urbich, C., Zeiher, A.M. & Dimmeler, S. (2000) Insulin-mediated stimulation of protein kinase Akt: a potent survival signaling cascade for endothelial cells. Arterioscler. Thromb. Vasc. Biol. 20, 402–409. 5. Smith, L.E., Shen, W., Perruzzi, C., Soker, S., Kinose, F., Xu, X., Robinson, G., Driver, S., Bischoff, J., Zhang, B., Schaeffer, J.M. & Senger, D.R. (1999) Regulation of vascular endothelial growth factor-dependent retinal neovascularization by insulin-like growth factor-1 receptor. Nature Med. 5, 1390–1395. 6. Treins, C., Giorgetti-Peraldi, S., Murdaca, J. & Van Obberghen, E. (2001) Regulation of vascular endothelial growth factor expression by advanced glycation end products. J. Biol. Chem. 276, 43836–43841. 7. McLaren, M., Elhadd, T.A., Greene, S.A. & Belch, J.J. (1999) Elevated plasma vascular endothelial cell growth factor and thrombomodulin in juvenile diabetic patients. Clin. Appl. Thromb. Hemost. 5, 21–24. 8. Ruoslahti, E. & Engvall, E. (1997) Integrins and vascular extra- cellular matrix assembly. J. Clin. Invest. 99, 1149–1152. 9. Miyamoto, S., Teramoto, H., Gutkind, J.S. & Yamada, K.M. (1996) Integrins can collaborate with growth factors for phos- phorylation of receptor tyrosine kinases and MAP kinase activa- tion: roles of integrin aggregation and occupancy of receptors. J. Cell Biol. 135, 8856–8860. 10. Bartfield, N.S., Pasquale, E.B., Geltosky, J.E. & Languino, L.R. (1993) The alpha V beta 3 integirn associates with a 190-kDa protein that is phosphorylated on tyrosine in response to platelet- derived growth factor. J. Biol. Chem. 268, 17270–17276. 11. Vuori, K. & Ruoslahti, E. (1994) Association of insulin receptor substrate-1 with integrins. Science 266, 1576–1578. 12. Jones, J.I., Prevette, T., Gockerman, A. & Clemmons, D.R. (1996) Ligand occupancy of the alphaVbeta3 integrin is necessary for smooth muscle cells to migrate in response to insulin-like growth factor I. Proc. Natl Acad. Sci. USA 93, 2482–2487. 13. Zheng, B. & Clemmons, D.R. (1998) Blocking ligand occupancy of the alphaVbeta3 integrin inhibits insulin-like growth factor I signaling in vascular smooth muscle cells. Proc. Natl Acad. Sci. USA 95, 11217–11222. 14. Persad,S.,Attwell,S.,Gray,V.,Mawji,N.,Deng,J.T.,Leung, D., Yan, J., Sanghera, J., Walsh, M.P. & Dedhar, S. (2001) Regulation of protein kinase B/Akt-serine 473 phosphorylation by integrin-linked kinase: critical roles for kinase activity and amino acids arginine 211 and serine 343. J. Biol. Chem. 276, 27462–27469. 15.Carmeliet,P.,Kieckens,L.,Schoonjans,L.,Ream,B.,van Nuffelen, A., Prendergast, G., Cole, M., Bronson, R., Collen, D. & Mulligan, R.C. (1993) Plasminogen activator inhibitor-1 gene- deficient mice. I. Generation by homologous recombination and characterization. J. Clin. Invest. 92, 2746–2755. 16. Loskutoff,D.J.,Sawdey,M.&Mimuro,J.(1989)Type1plas- minogen activator inhibitor. In Progress in Hemostasis and Thrombosis, Vol 9 (Coller, B.S., eds), pp. 87–115. Saunders, Philadelphia, USA. 17. Deng, G., Royle, G., Wang, S., Crain, K. & Loskutoff, D.J. (1996) Structural and functional analysis of the plasminogen activator inhibitro-1 binding motif in the somatomedin B domain of vitronectin. J. Biol. Chem. 271, 12716–12723. 18. Wei,Y.,Waltz,D.A.,Rao,N.,Drummond,R.J.,Rosenberg,S.& Chapman, H.A. (1994) Identification of the urokinase receptor as an adhesion receptor for vitronectin. J. Biol. Chem. 269, 32380–32388. 19. Deng, G., Curriden, S.A., Wang, S., Rosenberg, S. & Loskutoff, D.J. (1996) Is plasminogen activator inhibitor-1 the molecular switch that governs urokinase receptor-mediated cell adhesion and release? J. Cell Biol. 134, 1563–1571. 20. Munoz, C., Castellanos, M.C., Alfranca, A., Vara, A., Esteban, M.A., Redondo, J.M. & de Landazuri, M.O. (1996) Transcrip- tional up-regulation of intracellular adhesion molecule-1 in human endothelial cells by the antioxidant pyrrolidine dithiocarbamate involves the activation of activating protein-1. J. Immunol. 157, 3587–3597. 21. Stefansson, S., Petitclerc, E., Wong, M.K., McMahon, G.A., Brooks, P.C. & Lawrence, D.A. (2001) Inhibition of angiogenesis in vivo by plasminogen activator inhibitor-1. J. Biol. Chem. 276, 8135–8141. 22. Lawrence, D.A., Palaniappan, S., Stefansson, S., Olson, S.T., Francis-Chmura, A.M., Shore, J.D. & Ginsburg, D. (1997) Characterization of the binding of different conformational forms of plasminogen activator inhibitor-1 to vitronectin. J. Biol. Chem. 272, 7676–7680. 23. Zelzer, E., Levy, Y., Kahana, C., Shilo, B.Z., Rubinstein, M. & Cohen, B. (1998) Insulin induces transcription of target genes through the hypoxia-inducible factor HIF-1alpha/ARNT. EMBO J. 17, 5085–5094. 24. Yamagishi, S., Kawakami, T., Fujimori, H., Yonekura, H., Tanaka, N., Yamamoto, Y., Urayama, H., Watanabe, Y. & Yamamoto, H. (1999) Insulin stimulates the growth and tube formation of human microvascular endothelial cells through autocrine vascular endothelial growth factor. Microvasc. Res. 57, 329–339. 25. Miele, C., Rochford, J.J., Filippa, N., Giorgetti-Peraldi, S. & Van Obberghen, E. (2000) Insulin and insulin-like growth factor-I induce vascular endothelial growth factor mRNA expression via different signaling pathways. J. Biol. Chem. 275, 21695–21702. 26. Shigematsu, S., Yamauchi, K., Nakajima, K., Iijima, S., Aizawa, T. & Hashizume, K. (1999) D -Glucose and insulin stimulate migration and tubular formation of human endothelial cells in vitro. Am. J. Physiol. 277, E433–E438. 27. Bajou, K., Masson, V., Gerard, R.D., Schmitt, P.M., Albert, V., Praus, M., Lund, L.R., Frandsen, T.L., Brunner, N., Dano, K., Fusenig, N.E., Weidle, U., Carmeliet, G., Loskutoff,D., Collen, D., Carmeliet,P.,Foidart,J.M.&Noel,A.(2001)Theplasminogen activator inhibitor PAI-1 controls in vivo tumor vascularization by 820 R. Lo ´ pez-Alemany et al. (Eur. J. Biochem. 270) Ó FEBS 2003 interaction with proteases, not vitronectin. Implications for anti- angiogenic strategies. J. Cell. Biol. 152, 777–784. 28. Dunn, S.E. (2000) Insulin-like growth factor I stimulates angio- genesis and the production of vascular endothelial growth factor. Growth Horm. IGF Res. 10, S41–S42. 29. Hernandez, G.L., Volpert, O.V., Iniguez, M.A., Lorenzo, E., Martinez-Martinez, S., Grau, R., Fresno, M. & Redondo, J.M. (2001) Selective inhibition of vascular endothelial growth factor- mediated angiogenesis by cyclosporin A: roles of the nuclear factor of activated T cells and cyclooxygenase 2. J. Exp Med. 193, 607–620. 30. Olofson, B., Korpelainen, E., Pepper, M.S., Mandriota, S.J., Aase, K., Kumar, V., Gunji, Y., Jeltsch, M.M., Shibuya, M. & Alitalo, K. (1998) Vascular endothelial growth factor B (VEGF-B) binds to VEGF receptor-1 and regulates plasminogen activator activity in endothelial cells. Proc. Natl Acad. Sci. USA 95, 11709–11714. 31. Pepper, M.S., Belin, D., Montesano, R., Orci, L. & Vassalli, J.D. (1990) Transforming growth factor-beta 1 modulates basic fibro- blast growth factor-induced proteolytic and angiogenic properties of endothelial cells in vitro. J. Cell. Biol. 111, 743–755. 32. Saltiel, A.R. & Pessin, J.E. (2002) Insulin signaling pathways in time and space. Trends Cell Biol. 12, 65–71. 33. Dufourcq,P.,Louis,H.,Moreau,C.,Daret,D.,Boisseau,M.R., Lamaziere, J.M. & Bonnet, J. (1998) Vitronectin expression and interaction with receptors in smooth muscle cells from human atheromatous plaque. Arterioscler Thromb Vasc Biol. 18, 168–176. 34. Robbie, L.A., Booth, N.A., Brown, A.J. & Bennett, B. (1996) Inhibitors of fibrinolysis are elevated in atherosclerotic plaque. Arterioscler Thromb Vasc Biol. 16, 539–545. 35. Sato, R., Komine, Y., Imanaka, T. & Takano, T. (1990) Monoclonal antibody EMR1a/212D recognizing site of depos- ition of extracellular lipid in atherosclerosis. Isolation and char- acterization of a cDNA clone for the antigen. J. Biol Chem. 265, 21232–21236. 36. Padro, T., Emeis, J.J., Steins, M., Schmid, K.W. & Kienast, J. (1995) Quantification of plasminogen activators and their inhibitors in the aortic vessel wall in relation to the presence and severity of atherosclerotic disease. Arterioscler Thromb Vasc Biol. 15, 893–902. 37. Underwood, P.A., Bean, P.A. & Cubeddu, L. (2001) Human endothelial cells grow poorly on vitronectin: role of PAI-1. J. Cell Biochem. 82, 98–109. 38. Primrose, J.N., Davies, J.A., Prentice, C.R.M., Hughes, R. & Johnston, D. (1992) Reduction in factor VII, fibrinogen and plasminogen activator inhibitor-1 activity after surgical treatment of morbid obesity. Thrombosis Haemostasis 68, 396–399. 39.Brodsky,S.,Chen,J.,Lee,A.,Akassoglou,K.,Norman,J. & Goligorsky, M.S. (2001) Plasmin-dependent and -independent effects of plasminogen activators and inhibitor-1 on ex vivo angiogenesis. Am. J. Physiol. Heart Circ. Physiol. 281, H1784–H1792. Ó FEBS 2003 PAI-1 inhibition of insulin/vitronectin signaling (Eur. J. Biochem. 270) 821 . PRIORITY PAPER Plasminogen activator inhibitor type-1 inhibits insulin signaling by competing with avb3 integrin for vitronectin binding Roser Lo ´ pez-Alemany 1 ,. signaling by avb3 integrin. Plasminogen activator inhibitor type-1 (PAI-1), competes with avb3 integrin for vitronectin (VN) binding. Here we report that PAI-1,