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:
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:
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