Insulin stabilizes microvascular endothelial barrier function via phosphatidylinositol 3-kinase/Akt-mediated Rac1 activation

ISSN: 1524-4636 Copyright © 2010 American Heart Association. All rights reserved. Print ISSN: 1079-5642. Online

7272 Greenville Avenue, Dallas, TX 72514Arteriosclerosis, Thrombosis, and Vascular Biology is published by the American Heart Association.

DOI: 10.1161/ATVBAHA.110.203901 25, 2010;

2010;30;1237-1245; originally published online MarArterioscler Thromb Vasc BiolThomas Noll and Muhammad Aslam

Erdogan, Mathias Grebe, Daniel Sedding, Hans Michael Piper, Harald Tillmanns, Dursun Gündüz, Johannes Thom, Imran Hussain, Diego Lopez, Frauke V. Härtel, Ali

Phosphatidylinositol 3-Kinase/Akt-Mediated Rac1 ActivationInsulin Stabilizes Microvascular Endothelial Barrier Function via

http://atvb.ahajournals.org/cgi/content/full/ATVBAHA.110.203901/DC1Data Supplement (unedited) at:

  http://atvb.ahajournals.org/cgi/content/full/30/6/1237

located on the World Wide Web at: The online version of this article, along with updated information and services, is

http://www.lww.com/reprintsReprints: Information about reprints can be found online at  

[email protected]. E-mail:

Fax:Kluwer Health, 351 West Camden Street, Baltimore, MD 21202-2436. Phone: 410-528-4050. Permissions: Permissions & Rights Desk, Lippincott Williams & Wilkins, a division of Wolters 

http://atvb.ahajournals.org/subscriptions/Biology is online at Subscriptions: Information about subscribing to Arteriosclerosis, Thrombosis, and Vascular

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

Insulin Stabilizes Microvascular Endothelial BarrierFunction via Phosphatidylinositol 3-Kinase/Akt-Mediated

Rac1 ActivationDursun Gunduz, Johannes Thom, Imran Hussain, Diego Lopez, Frauke V. Hartel, Ali Erdogan,

Mathias Grebe, Daniel Sedding, Hans Michael Piper, Harald Tillmanns,Thomas Noll, Muhammad Aslam

Objective—Insulin is a key regulator of metabolism, but it also confers protective effects on the cardiovascular system.Here, we analyze the mechanism by which insulin stabilizes endothelial barrier function.

Methods and Results—Insulin reduced basal and antagonized tumor necrosis factor-�–induced macromolecule perme-ability of rat coronary microvascular endothelial monolayers. It also abolished reperfusion-induced vascular leakage inisolated-perfused rat hearts. Insulin induced dephosphorylation of the regulatory myosin light chains, as well astranslocation of actin and vascular endothelial (VE)-cadherin to cell borders, indicating a reduction in contractileactivation and stabilization of cell adhesion structures. These protective effects were blocked by genistein orHydroxy-2-naphthalenylmethylphosphonic acid tris acetoxymethyl ester (HNMPA-[AM]3), a pan-tyrosine-kinase orspecific insulin-receptor-kinase inhibitor, respectively. Insulin stimulated the phosphatidylinositol 3-kinase (PI3K)/Aktpathway and NO production, and it activated Rac1. Inhibition of PI3K/Akt abrogated Rac1 activation andinsulin-induced barrier protection, whereas inhibition of the endothelial nitric oxide synthase/soluble guanylyl cyclasepathway partially inhibited them. Inhibition of Rac1 abrogated the assembly of actin at cell borders. Accordingly, itabolished the protective effect of insulin on barrier function of the cultured endothelial monolayer, as well as the intactcoronary system of ischemic-reperfused hearts.

Conclusion—Insulin stabilizes endothelial barrier via inactivation of the endothelial contractile machinery andenhancement of cell-cell adhesions. These effects are mediated via PI3K/Akt- and NO/cGMP-induced Rac1activation. (Arterioscler Thromb Vasc Biol. 2010;30:1237-1245.)

Key Words: capillary permeability � coronary circulation � endothelium � nitric oxide � vascular biology

Insulin is an essential hormone of metabolic homeostasis.Recent clinical findings show that intensive insulin therapy

has vasoprotective effects under inflammatory conditions1,2

and reduces major cardiovascular events in diabetics.3 Theseeffects of insulin seem to be independent of its metaboliceffects on endothelial cells.

Vascular endothelial cells forming the inner lining of allvessels play an important role in the regulation of vascularhomeostasis. They provide a semiselective barrier forwater, solutes, macromolecules, and blood-borne compo-nents and are also involved in regulating the trafficking ofblood cells across the vessel wall. This function can bealtered by a variety of diverse circulating vasoactiveinflammatory mediators and hormones such as insulin. Ithas recently been demonstrated that endothelial cells ofmacrovascular origin express insulin receptors.4 Binding

of insulin to the receptors induces nitric oxide productionvia phosphatidylinositol 3-kinase (PI3K)-mediated activa-tion of endothelial nitric oxide synthase (eNOS)5 andcauses relaxation of the smooth muscles. Recently, it hasbeen shown that insulin reduces mesenteric venular albu-min leakage on systemic insulin administration in rats.6

However, the underlying signaling mechanism of thiseffect is largely unknown.

The integrity of the endothelial barrier is highly dependenton the endothelial actomyosin-based contractile machinery7–9

and actin cytoskeleton-mediated adherens junctions consist-ing of VE-cadherin, which, together with other actin bindingproteins, seals the adjoining cells and thereby limits thepassage of macromolecules across the microvasculature.10,11

The activation of endothelial contractile machinery is pre-cisely regulated by phosphorylation state of the regulatory

Received on: June 29, 2009; final version accepted on: March 11, 2010.From Zentrum fur Innere Medizin, Abteilung fur Kardiologie/Angiologie, Universitatsklinikum Gießen und Marburg GmbH, Gießen, Germany (D.G.,

J.T., I.H., A.E., M.G., D.S., H.T.); Physiologisches Institut, Justus-Liebig-Universitat, Gießen, Germany (D.L., F.V.H., H.M.P., T.N., M.A.); Laboratoryof Experimental Cardiology, Heart Division, Hospital Universitari Vall d’Hebron and Research Institute, Barcelona, Spain (D.L.).

Correspondence to Dursun Gunduz, Zentrum fur Innere Medizin, Abteilung fur Kardiologie und Angiologie, Universitatsklinikum Gießen und MarburgGmbH, Klinikstraße 36, D-35392 Gießen, Germany. E-mail [email protected]

© 2010 American Heart Association, Inc.

Arterioscler Thromb Vasc Biol is available at http://atvb.ahajournals.org DOI: 10.1161/ATVBAHA.110.203901

1237 at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

myosin light chains (MLC).7 MLC phosphorylation increasesactin-myosin interaction, cell contraction and barrier failure,whereas reduction in MLC phosphorylation causes stabiliza-tion of endothelial barrier.9,12

The members of the Rho family of GTPases are keyregulators of endothelial barrier function, controlling theendothelial contractile machinery, assembly of actin cytoskel-eton, and cell adhesion structures.11,13,14 It is well establishedthat the Rho GTPase Rac1 is required for the stability ofendothelial adherens junctions and barrier function.13,15 Sim-ilarly, activation of Rac1 strengthens VE-cadherin-based celladhesions and reduces the macromolecule permeability ofendothelial monolayers.16 Rac1 activity is regulated by gua-nine nucleotide exchange factors, which control the transitionof the GDP-bound inactive form to the GTP-bound activeform. The activation of Rac1 can be inhibited by the specificpharmacological inhibitor NSC23766, which specifically in-hibits Rac1 guanine nucleotide exchange factors Tiam1 andTrioN.17

The aim of the present study was to analyze whetherinsulin can enhance the barrier function of coronary micro-vascular endothelial cells. Special emphasis was laid on themolecular mechanism by which insulin may accomplish thisbarrier protection.

Materials and MethodsThe sources of reagents are listed in the supplemental Materials andMethods section, available online at http://atvb.ahajournals.org.

Isolation and culture of rat coronary microvascular endothelialcells was performed as described previously.18

Immunoprecipitation was performed by labeling protein-G–coated magnetic beads with anti-insulin-receptor �-subunit antibodyfollowed by Western blotting. Densitometric analyses of Westernblots were performed using the Quantity One image analyzersoftware (Bio-Rad). The procedure is described in detail in the onlinedata supplement.

Endothelial monolayer permeability was measured by the flux oftrypan blue–labeled albumin across endothelial monolayers grownon polycarbonate filters and myocardial water content of saline-perfused rat heart as an index of coronary leakage, as describedpreviously.19

Immunofluorescence MicroscopyThe endothelial monolayers cultured on glass coverslips wereexamined as described previously19 using a Zeiss LSM-510Minverted microscope.

Data are given as means�SD of 3 to 5 experiments usingindependent cell preparations. The comparison of means betweengroups was performed by 1-way analysis of variance (ANOVA)followed by a Student-Newman-Keuls post hoc test. Changes inparameters within the same group were assessed by multipleANOVA analyses. Probability (P) values of less than 0.05 wereconsidered significant.

ResultsEffect of Insulin on Insulin Receptors andEndothelial Monolayer PermeabilityThe presence of insulin receptors in rat coronary micro-vascular endothelial cells was confirmed by immunopre-cipitation with an antibody targeting the �-subunit of theinsulin receptor (Figure 1A). Insulin induced an increase intyrosine phosphorylation of the �-subunit, as shown by

reprobe of the Western blot membrane with an anti-phosphotyrosine antibody.

The unstimulated endothelial monolayers showed aconstant albumin permeability of 4.9�0.2�10�6 cm/s.The addition of insulin caused a prompt reduction inpermeability. It dropped within 10 minutes and remainedat that lower level for the rest of the period of observations.This reduction was concentration dependent (Figure 1Band 1C); it was significant as early as 0.001 IU/mL(equivalent to �5 nmol/L), half-maximal at 0.01 IU/mL,and at its maximum at 1 IU/mL. Therefore, this concen-tration was used for all further experiments.

To analyze whether insulin can also protect endothelialcells against agonist-induced hyperpermeability, cells werechallenged by tumor necrosis factor-� (TNF�). Exposure ofendothelial cells to TNF� (100 ng/mL) for 30 minutes led toa marked increase in macromolecule permeability, which wasreversed by 1 IU/mL insulin (Figure 1D).

In the next step, the involvement of insulin receptors in thebarrier protective effects of insulin was analyzed. Insulinreceptors are receptor-tyrosine-kinases; therefore, genistein, apan-tyrosine-kinase inhibitor, or HNMPA-(AM)3, a highlyspecific insulin-receptor-kinase inhibitor,20 was applied toblock insulin receptors. Preincubation of the cells for 30minutes with genistein (10 �mol/L) significantly attenuatedthe insulin effect on macromolecule permeability (Figure2A), whereas HNMPA-(AM)3 (10 �mol/L) completely abol-ished it (Figure 2B).

Effect of Insulin on the PI3K/Akt Pathway andMacromolecule PermeabilityInsulin mediates most of its effects in endothelial cells viaactivation of the PI3K/Akt pathway.5,21 In line with previ-ous reports, insulin induced a prompt increase in Aktphosphorylation in coronary microvascular endothelialcells in a concentration-dependent manner (Figure 3A).Akt phosphorylation reached its maximum within 1 minute andwas sustained over the whole period of observation (Figure 3B).Maneuvers that inhibit insulin-induced receptor-kinase activa-tion, such as genistein and HNMPA-(AM)3, reduced or abol-ished Akt phosphorylation, respectively (Figure 3C). Likewise,wortmannin (1 �mol/L), a PI3K inhibitor, completely abolishedinsulin-induced Akt phosphorylation. In line with this, wortman-nin also completely abrogated the insulin-mediated reduction inmacromolecule permeability (Figure 3D). Similar results wereobtained with LY294002 (data not shown).

Role of Insulin-Induced NO Production inInsulin-Mediated Barrier StabilizationIn line with previous reports,5 insulin induced NO produc-tion in a concentration-dependent manner in coronarymicrovascular endothelial cells (Figure 4A); its effect wasat its maximum at 0.1 IU/mL and was completely abol-ished by preincubation with the competitive eNOS inhibitorNG-nitro-L-arginine methyl ester (L-NAME) (100 �mol/L) for30 minutes. Preincubation of endothelial monolayers withL-NAME, or with NG-nitro-L-Arginine (L-NNA) (100 �mol/L)for 30 minutes attenuated but did not abolish the insulin effect onpermeability (Figure 4B). Because NO mediates many of its

1238 Arterioscler Thromb Vasc Biol June 2010

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

actions via activation of soluble guanylate cyclase (sGC) andgeneration of cGMP, endothelial monolayers were preincubatedwith 1H-(1,2,4)oxadiazolo(4,3-a)quinoxalin-1-one (ODQ)(100 �mol/L) or 8-bromo-4H-2,5-dioxa-3,9b-diaza-cyclopenta[a]naphthalen-1-one (NS2028) (100 �mol/L), 2 spe-cific sGC inhibitors, for 30 minutes. ODQ and NS2028 attenu-ated the insulin effect on macromolecule permeability to asimilar extent as the eNOS inhibitors (Figure 4C).

Effect of Insulin on Contractile Machinery andVE-Cadherin-Mediated Adherens JunctionsThe endothelial barrier is regulated by 2 principal mecha-nisms: the actomyosin-based endothelial contractile ma-chinery and VE-cadherin-dependent adherens junctions.Here, the effect of insulin on both elements was tested.Exposure of endothelial cells to insulin caused a reductionin MLC phosphorylation, which was maximum as early as0.1 IU/mL (Figure 5A). The effect on MLC dephosphor-ylation was rather delayed and was significant after 10minutes (Figure 5B).

Under control conditions, VE-cadherin was located at theborders of adjacent cells (Figure 5C and 5D). However,VE-cadherin staining is already enhanced at that site 5minutes after the addition of insulin (1 IU/mL), indicatingthat the strengthening effect of insulin on cell-cell adhesionstructures precedes inactivation of the contractile machinery.In line with the data on macromolecule permeability, trans-location of VE-cadherin to cell-cell junctions was abolished

by preincubation with insulin-receptor-kinase inhibitorHNMPA-(AM)3 (10 �mol/L) and the PI3K inhibitor wort-mannin (1 �mol/L) for 30 minutes. Likewise, insulin-inducedtranslocation of VE-cadherin is only partially inhibited by theNOS inhibitor L-NAME.

Effect of Insulin on Rac1 Activation andReorganization of Actin Cytoskeleton andBarrier StabilizationInsulin (1 IU/mL) induced a 3-fold increase in Rac1activation after 10 minutes (Figure 6A). This insulin effectwas abolished by insulin-receptor-kinase inhibitorHNMPA-(AM)3 (10 �mol/L) and the PI3K inhibitor wort-mannin (1 �mol/L), whereas the inhibitors of NOS andsGC, L-NAME (100 �mol/L) and ODQ (100 �mol/L),respectively, led only to partial inhibition of insulin-induced Rac1 activation. This corresponds to a partialinhibitory effect of L-NAME and ODQ on insulin-mediated barrier stabilization.

It is well established that Rac1 strengthens adherensjunctions via enhancement of the peripheral actin cytoskele-ton, accompanied by reduction of F-actin stress fibers. There-fore, the effect of insulin on actin rearrangement was ana-lyzed in early-confluent coronary microvascular endothelialmonolayers. Under these conditions, the cells characteristi-cally do not develop a prominent band of peripheral actin butstill exhibit F-actin stress fibers running across the cell. Thus,the maneuver inducing actin rearrangement can be easily

Figure 1. Effect of insulin on insulinreceptor activation and macromoleculepermeability of coronary microvascularendothelial monolayers. A, Effect of insu-lin on tyrosine phosphorylation of theinsulin receptor �-subunit (IR). Shown isa representative Western blot. Cells wereexposed to insulin (1 IU/mL) or vehicle(C) for 10 minutes, and insulin receptor�-subunit was immunoprecipitated (IP)and immunoblotted using an antibodytargeting the �-subunit of the insulinreceptor or anti-phosphotyrosine anti-body, respectively. B, Effect of insulin onmacromolecule permeability. Endothelialmonolayers were exposed to differentconcentrations of insulin or vehicle (con-trol) as indicated. C, Concentration-dependent response of permeability after10 minutes of exposure to insulin orvehicle (control) as in B. D, Effect ofinsulin on TNF�-induced hyperperme-ability on macromolecule permeability.Endothelial monolayers were exposed toinsulin (1.0 IU/mL), TNF� (100 ng/mL), orinsulin plus TNF� or vehicle (C). Data aremean�SD of 3 experiments with inde-pendent cell preparations. *P�0.05 ver-sus control; #P�0.05 versus TNF�.

Gunduz et al Insulin and Endothelial Permeability 1239

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

detected. As shown in Figure 6B, insulin induced formationof peripheral actin band and reduction in stress fibers after 5minutes of exposure. Preincubation of the monolayers withthe insulin-receptor-kinase inhibitor HNMPA-(AM)3

(10 �mol/L), the Rac1 inhibitor NSC23766 (100 �mol/L),17

and the PI3K inhibitor wortmannin (1 �mol/L) for 30minutes abolished peripheral actin band formation, whereasthe NOS inhibitor L-NAME only weakly antagonized insulin-induced actin reorganization. The role of Rac1 activation in

Figure 3. Effects of insulin on Akt phosphorylation and its role ininsulin-mediated reduction in permeability. A, RepresentativeWestern blots of Akt phosphorylation (P�Akt). Cells wereexposed to different concentrations of insulin for 10 minutes. B,Representative Western blots of P�Akt. Cells were exposed toinsulin (1 IU/mL) for different time intervals as indicated. C, Rep-resentative Western blots of P�Akt. Cells were exposed to insu-lin (1 IU/mL) for 10 minutes or to genistein (Gen; 10 �mol/L),HNMPA-(AM)3 (10 �mol/L; HNMP), or wortmannin (1 �mol/L;Wort) for 30 minutes with or without insulin for 10 minutes. D,Effect of insulin (1 IU/mL), wortmannin (1 �mol/L), and wortman-nin plus insulin or vehicle (control) on macromolecule permeabil-ity. Data are mean�SD of 5 experiments with independent cellpreparations. *P�0.05 versus control; #P�0.05 versus insulinalone.

Figure 2. Effect of insulin receptor inhibition on insulin-mediatedreduction in macromolecule permeability. A, Endothelial mono-layers were exposed to insulin (1 IU/mL), genistein (10 �mol/L),a pan-tyrosine-kinase inhibitor, insulin plus genistein, or vehicle(control) as indicated. B, Endothelial monolayers were exposedto insulin (1 IU/mL), HNMPA-(AM)3 (HNMP; 10 �mol/L), a spe-cific insulin-receptor-kinase inhibitor, insulin plus HNMP, orvehicle (control) as indicated. Data are mean�SD of 5 experi-ments with independent cell preparations. *P�0.05 versus con-trol; #P�0.05 versus insulin alone.

1240 Arterioscler Thromb Vasc Biol June 2010

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

insulin-mediated barrier stabilization was proven by using theRac1 inhibitor NSC23766. Preincubation of endothelial cellswith NSC23766 abolished the insulin effect on endothelialbarrier (Figure 6C).

A final series of experiments was performed in a well-established isolated perfused rat heart model19 to verify thebarrier protective effects of insulin in an intact coronarysystem. Myocardial water content was determined as an indexfor capillary leakage. Under control conditions, the myocar-dial water content of the normoxic perfused rat hearts was455 mL/100 g dry weight after 90 minutes (Figure 6D).Exposure of the hearts to ischemia for 60 minutes followedby 30 minutes of reperfusion caused an increase in myocar-dial water content to 554 mL/100 g dry weight. In contrast,addition of 0.1 IU/mL insulin, a concentration only one-tenththat applied in the cell culture model, during the first 10minutes of reperfusion abolished the reperfusion-inducedincrease in myocardial water content. In one set of experi-ments, 50 �mol/L NSC23766 was added 20 minutes beforethe onset and during the first 10 minutes of reperfusion. Thismaneuver abrogated the protective effect of insulin onreperfusion-induced increase in water content.

DiscussionIt is well established that intensive insulin therapy reducesvascular complications of the coronary system and othervascular provinces because of its antiatherosclerotic andantiinflammatory effects.1–3 Hyperpermeability of macro-and microvasculature is the hallmark of these disease states.In the present study, we show for the first time that insulinstabilizes the coronary microvascular endothelial barrierfunction in an in vitro model of coronary microvascularendothelial cells as well as in the intact coronary vascularsystem of isolated-perfused rat heart. Insulin not only reducedthe basal permeability of endothelial monolayers but alsoabolished hyperpermeability induced by the inflammatorymediator TNF�.

Immunoprecipitation experiments revealed the presenceof insulin receptors and phosphorylation of the �-subunitof the receptor on exposure to insulin. Inhibition of theinsulin receptor-tyrosine-kinase abrogated the barrier-enhancing effect of insulin, demonstrating a receptor-mediated effect. In line with the findings of the presentstudy, Sasaki et al6 have shown that insulin could reducemesentery venular leakage in rats. However, the molecularmechanism of this insulin effect on endothelial barrier hasnot been clear until now.

In line with previous reports,5 we show activation of thePI3K/Akt pathway in coronary microvascular endothelialcells, as demonstrated by Akt phosphorylation. The Aktphosphorylation is rapid and is sustained over longer periods

Figure 4. Role of NO pathway in insulin-mediated reduction ofmacromolecule permeability. A, Effect of insulin on NO produc-tion in endothelial cells. Cells were loaded with DAF2-FM andtreated with different concentrations of insulin or L-NAME(100 �mol/L) plus insulin (1.0 IU/mL) for 10 minutes, and thenDAF-2 triazole production, which is directly related to NO levels,was measured with a fluorescence plate reader. B, Effect ofNOS inhibitors L-NAME or L-NNA on insulin-mediated reductionin macromolecule permeability. Endothelial monolayers wereexposed to insulin (1 IU/mL), L-NAME (100 �mol/L), L-NNA(100 �mol/L), insulin plus L-NAME, insulin plus L-NNA, or vehicle

Figure 4 (Continued). (control). C, Effect of sGC inhibitors ODQor NS2028 on insulin-mediated reduction in endothelial perme-ability. Endothelial monolayers were exposed to insulin (1IU/mL), ODQ (100 �mol/L), NS2028 (100 �mol/L), insulin plusODQ, insulin plus NS2028, or vehicle (control). Data aremean�SD of 5 experiments with independent cell preparations.*P�0.05 versus control; §#P�0.05 versus insulin alone.

Gunduz et al Insulin and Endothelial Permeability 1241

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

of time, similar to its effect on barrier function. Moreover, wedemonstrate that this Akt phosphorylation is via receptor-mediated activation of PI3K. Accordingly, inhibition of PI3Kabolished insulin-mediated reduction in permeability, furthersupporting the notion that insulin strengthens the endothelialbarrier via receptor-mediated activation of PI3K/Aktpathway.

Furthermore, we show that insulin induces NO productionin endothelial cells under study and inhibition of eNOS withspecific inhibitors not only blocked insulin-induced NOproduction but also attenuated the insulin-mediated barrierstabilization. Accordingly, inhibition of cGMP/protein kinaseG (PKG) pathway by specific sGC inhibitors attenuatedinsulin-mediated barrier protection to a similar extent aseNOS inhibitors, confirming that this is a cGMP/PKG effect.The role of NO in the control of the endothelial barrier iscontroversial.22 The NO effect may differ depending on thestimulus, experimental conditions, vascular bed, time win-dow, and local concentration of NO. A number of studies inisolated perfused microvessels,23 as well as in eNOS knock-out mice,24,25 show that hyperpermeability induced by growth

factors or inflammatory mediators could be attenuated witheNOS inhibitors, supporting the concept that NO triggers thisincrease in permeability. We have previously demonstratedthat vascular endothelial growth factor has a biphasic effecton endothelial permeability.26 The initial transient barrierprotective effect is NO dependent, and inhibition of eNOSabolishes this initial reduction in permeability, suggestingthat the NO effect is context dependent and can beinfluenced by another repertoire of signaling, activatedsimultaneously by growth factors or inflammatory media-tors. Similarly, several other studies have elegantly dem-onstrated that NO donors and cGMP analogs can antago-nize agonist-induced hyperpermeability.27,28 Likewise, ithas been shown that even basal NO is required for themaintenance of vascular integrity.29 In line with this,Predescu et al, using eNOS knockout mice, have demon-strated that basal activity of eNOS is indeed required forendothelial adherens junction maintenance independent ofvascular bed.30 Accordingly, the present study demon-strates that insulin, in part, mediates its barrier protectiveeffect via the NO/cGMP pathway.

Figure 5. Effect of insulin on MLC phosphorylation (P�MLC) and localization of VE-cadherin in endothelial cells. A, Time-dependenteffect of insulin on P�MLC. Top: Representative Western blots of MLC phosphorylation. Cells were exposed to insulin (1.0 IU/mL) forindicated time points. Bottom: Densitometric analysis of Western blots of MLC phosphorylation. B, Concentration-dependent effect ofinsulin on P�MLCP. Top: Representative Western blots of MLC phosphorylation. Endothelial monolayers were exposed to differentconcentration of insulin for 10 minutes. Bottom: Densitometric analysis of Western blots of MLC phosphorylation. C, Effect of insulin onVE-cadherin localization. Representative immunofluorescence images of endothelial monolayers exposed to vehicle (control [C]), insulin(Ins; 1 IU/mL), insulin plus HNMPA-(AM)3 (HNMP; 10 �mol/L), insulin plus wortmannin (Wort; 1 �mol/L), or insulin plus L-NAME(100 �mol/L) for 5 minutes. Arrows denote VE-cadherin localized at cell borders (scale bar�20 �m; shown is a representative immuno-staining of 3 experiments with independent cell preparations). D, Quantification of the immunofluorescence staining of VE-cadherindetermined by image analysis. Data are mean�SD; n� 10 cells per endothelial monolayer of 3 independent experiments. *P�0.05 ver-sus C; #P�0.05 versus insulin alone.

1242 Arterioscler Thromb Vasc Biol June 2010

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

Insulin reduced the state of contractile activation andstrengthened adherens junctions in coronary endothelial cells,which explains the mechanistic aspects of this barrier stabi-lization. Insulin stimulates NO production and reduces MLCphosphorylation, the latter of which is rather delayed and wasobserved after 10 minutes. The maximum effect on bothparameters was obtained at 0.1 IU/mL, a concentrationone-tenth that observed for the maximal effect on permeabil-ity. Presently, the detailed mechanism of insulin-inducedcontractile inactivation is still elusive. It has previously beenshown that in isolated vessels, a maneuver activating cGMP/PKG signaling led to dephosphorylation of MLC via activa-tion of MLC phosphatase.31 Therefore, our data suggest that

insulin reduces MLC phosphorlyation via NO/cGMP-mediated pathway.

Insulin-induced stabilization of adherens junctions is me-diated via translocation of VE-cadherin, a major componentof endothelial adherens junctions, to cell-cell junctions. Thistranslocation was abrogated by inhibition of insulin receptor-tyrosine-kinase, PI3K and weakly attenuated by inhibition ofNOS, suggesting that insulin mediates its effect on endothe-lial adherens junctions via both the PI3K/Akt and the NOpathway.

Rac1, a member of Rho family of GTPases, is known toregulate assembly of peripheral actin and stimulates theformation of adherens junctions.14,15,32 The results of the

Figure 6. Effect of insulin on the GTPase Rac1, actin cytoskeleton, and endothelial barrier function of cultured endothelial mono-layer and intact coronary system of the isolated-perfused heart. A, Effects of insulin on Rac1 activation. Top: RepresentativeWestern blots of Rac1-GTP and Rac1 total. The cells were exposed to insulin (Ins; 1 IU/mL), Ins plus HNMPA-(AM)3 (HNMP;10 �mol/L), Ins plus L-NAME (L-NA; 100 �mol/L), Ins plus wortmannin (Wort; 1 �mol/L), Ins plus ODQ (100 �mol/L), or vehicle (C;Control) for 10 minutes. Bottom: Densitometric analysis of the Western blots. Data are mean�SD of 3 experiments with indepen-dent cell preparations. *P�0.05 versus control; #P�0.05 versus insulin alone. B, Effect of insulin on F-actin cytoskeleton. Repre-sentative immunofluorescence images of endothelial monolayers exposed to insulin (1 IU/mL), insulin plus HNMPA-(AM)3 (HNMP;10 �mol/L), insulin plus wortmannin (Wort; 1 �mol/L), insulin plus L-NAME (100 �mol/L), insulin plus NSC23766 (NSC; 100 �mol/L, Rac1 in-hibitor), or vehicle (control). Arrows denote the cell periphery and arrowheads F-actin stress fibers. Scale bar�20 �m; shown are repre-sentative images of 3 experiments with independent cell preparations. C, Effect of the Rac1 inhibitor NSC23766 on insulin-mediatedreduction of macromolecule permeability. Endothelial cells were exposed to insulin (1 IU/mL), NSC23766 (NSC; 100 �mol/L), insulinplus NSC, or vehicle (control) as indicated. Data are mean�SD of 3 experiments with independent cell preparations. *P�0.05 versuscontrol; #P�0.05 versus insulin alone). D, Effect of insulin on myocardial water content after ischemia-reperfusion of saline-perfused rathearts. Hearts were exposed for 60 minutes to ischemia followed by 30 minutes of reperfusion (Rep) or 90 minutes of normoxia (Nor).Insulin (Ins; 0.1 IU/mL) was added at the onset of reperfusion during the first 10 minutes. In one set of experiments, NSC23766 (NSC;50 �mol/L) was added 20 minutes before the onset and in the first 10 minutes of reperfusion. Data are mean�SD of 5 separate experi-ments with independent organ preparations. *P�0.05 versus Nor; #P�0.05 versus Rep. E, Insulin stabilizes microvascular endothelialbarrier function: proposed signaling pathways.

Gunduz et al Insulin and Endothelial Permeability 1243

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

present study clearly show that insulin activates Rac1 viaPI3K/Akt and also in part via the NO/cGMP pathway. Thiscould be mediated via activation of the Rac1-specificguanine nucleotide exchange factors Tiam1 and TrioN,because a specific inhibitor of these guanine nucleotideexchange factors, NSC23766,17 abolished the insulin ef-fect. Inhibition of PI3K abolished Rac1 activation andreorganization of the peripheral actin cytoskeleton,whereas inhibition of either NOS or sGC led only to partialinhibition of Rac1 and a weak antagonistic effect onperipheral actin. In line with our findings, 2 reports onfibroblasts show that activation of either PI3K/Akt33 orNO/cGMP34 can induce Rac1 activation.

Inhibition of Rac1 with a specific inhibitor, NSC23766,17

abolished the insulin-mediated barrier stabilization effect.Remarkably, the effect of inhibition of Rac1 was much moreeffective than inhibition of NOS or sGC, indicating that Rac1is a central signaling element of the insulin-mediated barrierstabilization beyond NO/cGMP pathway. Furthermore, theinability of eNOS and sGC inhibitors to block the initialeffect of insulin on permeability and the delayed effect ofinsulin on MLC phosphorylation suggest that NO signalingdoes not contribute to the acute but rather to the sustainedeffect of insulin.

The barrier protective effect of insulin is further supportedby our data obtained in an isolated, saline-perfused rat heartmodel.19 In this model, insulin caused a strong reduction ofthe reperfusion-induced increase in myocardial water contentif applied for only a short period of time during the earlyphase of reperfusion. These data indicate that insulin is also apotent stabilizer of the vascular permeability barrier in theintact coronary system and that its application during reper-fusion can protect the heart against an imminent edema. Thisantiedematous effect of insulin was lost if the heart wasperfused with insulin in the presence of the Rac1 inhibitorNSC23766, clearly demonstrating that insulin-mediated bar-rier protection is via Rac1.

Hyperglycemia is a major contributor to endothelial barrierfailure35,36 and an early feature of diabetic microangiopathy.36

Substitution of insulin has been shown to be vasoprotective indiabetic rats.37 Several mechanisms for hyperglycemia-induced vascular leakage have been proposed, includingdamage to endothelial glycocalyx.38 Although an acute effectof insulin on glycocalyx has not been analyzed until today,recently it has been shown that long-term insulin administra-tion could protect P-glycoproteins in experimental diabeticrats.37

In summary, the data of the present study support theconcept that insulin stabilizes barrier under basal conditionsand protects against imminent failure induced by inflamma-tory mediators or ischemia-reperfusion injury via inhibitionof the contractile machinery and strengthening of cell-celladhesion structures. This barrier stabilizing effect is depen-dent on PI3K/Akt, NO/cGMP, and Rac1, which play adecisive role in insulin-mediated actin cytoskeleton rear-rangement and stabilization of cell-cell adhesions. The pro-posed signal transduction pathways suggested by the data ofthe present study are illustrated in Figure 6E.

AcknowledgmentsTechnical support by Sabine Schaefer, Daniela Reitz, Henrike Thomas,Herman Holztrager, and Anna Reis is gratefully acknowledged.

Sources of FundingThe study was supported by an Anschubsfinanzierung and Rontgen-Behring-Stiftung grant to Dr Gunduz, an Excellence Cluster Cardio-pulmonary System postdoctoral grant to Dr Aslam, and ExcellenceCluster Cardiopulmonary System and Deutsche Forschungsgemein-schaft grants SFB 547 and GRK 534 to Dr Noll.

DisclosuresNone.

References1. Langouche L, Vanhorebeek I, Vlasselaers D, Vander PS, Wouters PJ,

Skogstrand K, Hansen TK, Van den Berghe G. Intensive insulin therapyprotects the endothelium of critically ill patients. J Clin Invest. 2005;115:2277–2286.

2. Dandona P, Chaudhuri A, Ghanim H, Mohanty P. Insulin as an anti-inflammatory and antiatherogenic modulator. J Am Coll Cardiol. 2009;53:S14–S20.

3. Nathan DM, Cleary PA, Backlund JY, Genuth SM, Lachin JM, OrchardTJ, Raskin P, Zinman B. Intensive diabetes treatment and cardiovasculardisease in patients with type 1 diabetes. N Engl J Med. 2005;353:2643–2653.

4. Nitert MD, Chisalita SI, Olsson K, Bornfeldt KE, Arnqvist HJ. IGF-I/insulin hybrid receptors in human endothelial cells. Mol Cell Endocrinol.2005;229:31–37.

5. Zeng G, Nystrom FH, Ravichandran LV, Cong LN, Kirby M, MostowskiH, Quon MJ. Roles for insulin receptor, PI3-kinase, and Akt in insulin-signaling pathways related to production of nitric oxide in humanvascular endothelial cells. Circulation. 2000;101:1539–1545.

6. Sasaki R, Whitt SP, Huxley VH. Permeability response of the rat mes-enteric microvasculature to insulin. Clin Hemorheol Microcirc. 2006;34:259–263.

7. Wysolmerski RB, Lagunoff D. Regulation of permeabilized endothelialcell retraction by myosin phosphorylation. Am J Physiol. 1991;261:C32–C40.

8. Goeckeler ZM, Wysolmerski RB. Myosin light chain kinase-regulatedendothelial cell contraction: the relationship between isometric tension,actin polymerization, and myosin phosphorylation. J Cell Biol. 1995;130:613–627.

9. Garcia JG, Davis HW, Patterson CE. Regulation of endothelial cell gapformation and barrier dysfunction: role of myosin light chain phosphor-ylation. J Cell Physiol. 1995;163:510–522.

10. Lampugnani MG, Corada M, Caveda L, Breviario F, Ayalon O, Geiger B,Dejana E. The molecular organization of endothelial cell to cell junctions:differential association of plakoglobin, beta-catenin, and alpha-cateninwith vascular endothelial cadherin (VE-cadherin). J Cell Biol. 1995;129:203–217.

11. Dejana E, Orsenigo F, Lampugnani MG. The role of adherens junctionsand VE-cadherin in the control of vascular permeability. J Cell Sci.2008;121:2115–2122.

12. Gunduz D, Hirche F, Hartel FV, Rodewald CW, Schafer M, Pfitzer G,Piper HM, Noll T. ATP antagonism of thrombin-induced endothelialbarrier permeability. Cardiovasc Res. 2003;59:470–478.

13. Wojciak-Stothard B, Potempa S, Eichholtz T, Ridley AJ. Rho and Rac butnot Cdc42 regulate endothelial cell permeability. J Cell Sci. 2001;114:1343–1355.

14. Wojciak-Stothard B, Ridley AJ. Rho GTPases and the regulation ofendothelial permeability. Vascul Pharmacol. 2002;39:187–199.

15. Waschke J, Baumgartner W, Adamson RH, Zeng M, Aktories K, Barth H,Wilde C, Curry FE, Drenckhahn D. Requirement of Rac activity formaintenance of capillary endothelial barrier properties. Am J PhysiolHeart Circ Physiol. 2004;286:H394–H401.

16. Waschke J, Burger S, Curry FR, Drenckhahn D, Adamson RH. Activationof Rac-1 and Cdc42 stabilizes the microvascular endothelial barrier.Histochem Cell Biol. 2006;125:397–406.

17. Gao Y, Dickerson JB, Guo F, Zheng J, Zheng Y. Rational design andcharacterization of a Rac GTPase-specific small molecule inhibitor. ProcNatl Acad Sci U S A. 2004;101:7618–7623.

1244 Arterioscler Thromb Vasc Biol June 2010

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

18. Piper HM, Spahr R, Mertens S, Krutzfeldt A, Watanabe H. Microvascularendothelial cells from heart. In: Piper HM, ed. Cell Culture Technique inHeart and Vessel Research. Heidelberg: Springer; 1990:158–177.

19. Noll T, Wozniak G, McCarson K, Hajimohammad A, Metzner HJ, InserteJ, Kummer W, Hehrlein FW, Piper HM. Effect of factor XIII on endo-thelial barrier function. J Exp Med. 1999;189:1373–1382.

20. Saperstein R, Vicario PP, Strout HV, Brady E, Slater EE, Greenlee WJ,Ondeyka DL, Patchett AA, Hangauer DG. Design of a selective insulinreceptor tyrosine kinase inhibitor and its effect on glucose uptake andmetabolism in intact cells. Biochemistry. 1989;28:5694–5701.

21. Hermann C, Assmus B, Urbich C, Zeiher AM, Dimmeler S. Insulin-mediated stimulation of protein kinase Akt: a potent survival signalingcascade for endothelial cells. Arterioscler Thromb Vasc Biol. 2000;20:402–409.

22. Yuan SY. New insights into eNOS signaling in microvascular perme-ability. Am J Physiol Heart Circ Physiol. 2006;291:H1029–H1031.

23. Wu HM, Yuan Y, Zawieja DC, Tinsley J, Granger HJ. Role of phospho-lipase C, protein kinase C, and calcium in VEGF-induced venular hyper-permeability. Am J Physiol. 1999;276:H535–H542.

24. Fukumura D, Gohongi T, Kadambi A, Izumi Y, Ang J, Yun CO, BuerkDG, Huang PL, Jain RK. Predominant role of endothelial nitric oxidesynthase in vascular endothelial growth factor-induced angiogenesisand vascular permeability. Proc Natl Acad Sci U S A. 2001;98:2604 –2609.

25. Bucci M, Roviezzo F, Posadas I, Yu J, Parente L, Sessa WC, Ignarro LJ,Cirino G. Endothelial nitric oxide synthase activation is critical forvascular leakage during acute inflammation in vivo. Proc Natl Acad SciU S A. 2005;102:904–908.

26. Marumo T, Noll T, Schini-Kerth VB, Harley EA, Duhault J, PiperHM, Busse R. Significance of nitric oxide and peroxynitrite in per-meability changes of the retinal microvascular endothelial cellmonolayer induced by vascular endothelial growth factor. J Vasc Res.1999;36:510 –515.

27. Kurose I, Wolf R, Grisham MB, Granger DN. Modulation of ischemia/reperfusion-induced microvascular dysfunction by nitric oxide. Circ Res.1994;74:376–382.

28. Johnston B, Gaboury JP, Suematsu M, Kubes P. Nitric oxide inhibitsmicrovascular protein leakage induced by leukocyte adhesion-

independent and adhesion-dependent inflammatory mediators.Microcirculation. 1999;6:153–162.

29. Kurose I, Kubes P, Wolf R, Anderson DC, Paulson J, Miyasaka M,Granger DN. Inhibition of nitric oxide production. Mechanisms ofvascular albumin leakage. Circ Res. 1993;73:164–171.

30. Predescu D, Predescu S, Shimizu J, Miyawaki-Shimizu K, Malik AB.Constitutive eNOS-derived nitric oxide is a determinant of endothelialjunctional integrity. Am J Physiol Lung Cell Mol Physiol. 2005;289:L371–L381.

31. Nakamura K, Koga Y, Sakai H, Homma K, Ikebe M. cGMP-dependentrelaxation of smooth muscle is coupled with the change in the phosphor-ylation of myosin phosphatase. Circ Res. 2007;101:712–722.

32. Wojciak-Stothard B, Tsang LY, Paleolog E, Hall SM, Haworth SG. Rac1and RhoA as regulators of endothelial phenotype and barrier function inhypoxia-induced neonatal pulmonary hypertension. Am J Physiol LungCell Mol Physiol. 2006;290:L1173–L1182.

33. Qian Y, Zhong X, Flynn DC, Zheng JZ, Qiao M, Wu C, Dedhar S, ShiX, Jiang BH. ILK mediates actin filament rearrangements and cellmigration and invasion through PI3K/Akt/Rac1 signaling. Oncogene.2005;24:3154 –3165.

34. Hou Y, Ye RD, Browning DD. Activation of the small GTPase Rac1by cGMP-dependent protein kinase. Cell Signal. 2004;16:1061–1069.

35. Hempel A, Maasch C, Heintze U, Lindschau C, Dietz R, Luft FC,Haller H. High glucose concentrations increase endothelial cell per-meability via activation of protein kinase C alpha. Circ Res. 1997;81:363–371.

36. Yuan SY, Breslin JW, Perrin R, Gaudreault N, Guo M, Kargozaran H,Wu MH. Microvascular permeability in diabetes and insulin resistance.Microcirculation. 2007;14:363–373.

37. Liu H, Liu X, Jia L, Liu Y, Yang H, Wang G, Xie L. Insulin therapyrestores impaired function and expression of P-glycoprotein inblood-brain barrier of experimental diabetes. Biochem Pharmacol. 2008;75:1649–1658.

38. Zuurbier CJ, Demirci C, Koeman A, Vink H, Ince C. Short-term hyper-glycemia increases endothelial glycocalyx permeability and acutelydecreases lineal density of capillaries with flowing red blood cells. J ApplPhysiol. 2005;99:1471–1476.

Gunduz et al Insulin and Endothelial Permeability 1245

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

1 Supplement Material

Materials and Methods

Materials: HRP-conjugated anti-mouse IgG and rabbit IgG antibodies were from

Amersham Biosciences (Heidelberg, Germany); anti-phospho-tyrosine antibody

(mouse monoclonal IgG) and 96-well ELISA plates were from Becton Dickinson

GmbH (Heidelberg, Germany); anti-insulin-receptor antibody (clone 29B4, mouse

monoclonal IgG), benzonase, genistein, HNMP(AM)3 [Hydroxy-2-

naphthalenylmethylphosphonic acid tris acetoxymethyl ester], an inhibitor of insulin

receptor (IR) tyrosine kinase activity, 1 L-NAME, L-NNA, LY294002, and wortmannin

were from Calbiochem (Darmstadt, Germany); anti-phospho-Akt (Ser473) and anti-

phospho-MLC (rabbit polyclonal) antibodies were from Cell Signaling (Danvers,

USA); Rac1 pulldown assay kit and anti-Rac1 antibody (rabbit polyclonal) were from

Cytoskeleton (Denver, USA); DAF-FM diacetate [4-amino-5-methylamino- 2 , 7 -

difluorofluorescein diacetate], Protein G magnetic beads and Alexa Flour 488

conjugated secondary (anti rabbit IgG) antibody were from Invitrogen (Karlsruhe,

Germany); M199 medium was from Dianova (Hamburg, Germany); fetal calf serum

[FCS] and neonatal calf serum [NCS] were from PAA (Pasching, Austria); ECL

solution was from Pierce (Rockford, USA); Complete® [protease inhibitor cocktail]

was from Roche (Mannheim, Germany); anti-actin antibody (clone C4, mouse IgG),

insulin solution (recombinant from yeast, human) [290 IU/ml in 25 mM HEPES], anti-

pan-cadherin antibody (rabbit polyclonal), phalloidin-TRITC, and anti-vinculin

antibody (clone hVIN-1, mouse IgG) were from Sigma (Steinheim, Germany);

NSC23706 [N6-[2-[[4-(Diethylamino)-1-methylbutyl]amino]-6-methyl-4-pyrimidinyl]-2-

methyl-4,6-quinoline diamine trihydrochloride], NS-2028 [8-bromo-4H-2,5-dioxa-

3,9b-diaza-cyclopenta[a]naphthalen-1-one], and ODQ [1H-(1,2,4)oxadiazolo(4,3-

a)quinoxalin-1-one], were from Tocris bioscience (Bristol, UK); Costar Transwell®

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

2 Supplement Material

polycarbonate membrane filters (24-mm round) were from Vitaris (Baar, Germany).

All other chemicals were of the best available quality, usually analytical grade.

Cell culture: The investigation conforms to the Guide for the Care and Use of

Laboratory Animals published by the US National Institutes of Health (NIH

Publication No. 85–23, revised 1996). Microvascular coronary endothelial cells were

isolated from 200–250 g male Wistar rats and cultured as previously described. 2-4

Experimental protocols: The basal medium used in incubations was modified

Tyrode's solution (composition in mM: 150 NaCl, 2.7 KCl, 1.2 KH2PO4, 1.2 MgSO4,

1.0 CaCl2, and 30.0 N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid [HEPES];

pH 7.4, 37° C). Agents were added as indicated. Stock solution of insulin was in 25

mM HEPES and stock solutions of L-NAME and L-NNA were prepared immediately

before use with basal medium. Stock solutions of genistein, HNMP(AM)3, and

wortmannin were prepared with dimethyl sulfoxide (DMSO). Appropriate volumes of

these solutions were added to the cells yielding final solvent concentrations < 0.1%

(vol/vol). The same final concentrations of HEPES, DMSO, or basal medium were

included in all respective control experiments.

In a set of pilot experiments concentration-response relationships were

determined to find the optimal effective concentration. Unless otherwise stated the

following agents were applied in their optimal effective concentrations: genistein (10

µM), insulin (1 IU/ml), HNMP(AM)3 (10 µM), L-NAME (100 µM), L-NNA (100 µM) and

wortmannin (1 µM).

Rac1 Pulldown Assay: Activity of Rac1 was measured with a pulldown Rac1

activation assay biochem kit according to the manufacturer’s protocol.

3 Supplement Material

Macromolecule permeability measurement: Rat coronary microvascular endothelial

monolayer macromolecule permeability was measured as described previously. 3, 5, 6

Myocardial water content: Hearts from 250-g male Wistar rats were mounted

immediately after isolation on a Langendorff perfusion system in a temperature-

controlled chamber (37 °C), as previously described 3 with some modifications.

Hearts were perfused with Krebs-Henseleit buffer (composition in mM: 140.0 NaCl,

24.0 NaHCO3, 2.7 KCl, 0.4 KH2PO4, 1.0 MgSO4, 1.8 CaCl2, 5.0 glucose, pH 7.4) for

30 minutes (10 ml/min) before each experiment and then exposed to one of the

following protocols: (1) Normoxic conditions for 90 minutes, (2) 60 minutes of hypoxia

followed by 30 minutes of reperfusion, (3) 60 minutes of hypoxia followed by 30 of

reperfusion with 0.1 IU/ml insulin, during first 10 minutes of reperfusion, (4) 60

minutes of hypoxia followed by 30 minutes of reperfusion in which Rac1 inhibitor,

NSC23766 was added to the perfusion medium during last 20 min of hypoxia and

first 10 minutes of reperfusion with insulin (during first 10 minutes of reperfusion

only). The normoxic perfusion (10 ml/min) was with Krebs-Henseleit buffer gassed

with 95% O2 [vol/vol]/5% CO2 [vol/vol]), the chamber was flushed with humidified air,

and hypoxic perfusion with Krebs-Henseleit buffer with humidified 95% N2

[vol/vol]/5% CO2 [vol/vol]). At the end of each experiment, wet weight and after 24 h,

dry weight of the perfused rat hearts were measured.

Immunoprecipitation: immunoprecipitation was carried out as described previously. 7

Briefly, confluent endothelial monolayers grown in 10-cm dishes were stimulated as

indicated in the text. Cells were lysed for 10 min on ice (composition of lysis buffer:

1% (vol./vol.) Triton X-100, 0.5% (vol./vol.) Nonidet P-40, 150 mM NaCl, 1 mM EDTA,

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

4 Supplement Material

1 mM EGTA, 1 mM Na-orthovanadate, 0.5 mM PMSF, protease inhibitor

Complete™, 50 mM Tris/HCl pH 7.4). Lysates were cleared by centrifugation (1000 ×

g, 5 min, 4 °C). The supernatant was incubated with insulin receptor (IR)-specific

antibody pre-immobilized on magnetic protein-G beads overnight at 4 °C. After that

beads were washed three times with PBS (pH 7.4) containing 0.1% (vol./vol.) Tween

20. The beads were collected and bound proteins were eluted in Laemmli sample

buffer 8, 9 for Western blot analysis using pY20 antibody.

Nitric oxide (NO) assay: Intracellular NO was measured using the NO-specific

fluorescence probe DAF-FM diacetate. The cell permeable DAF-FM diacetate

diffuses freely across the cell membrane and once within the cells it is deacetylated

by endogenous esterases, resulting in the formation of DAF-FM. DAF-FM reacts with

the NO oxidation product N2O3 and generates the highly fluorescent DAF-FM

triazole, which was detected with a ‘‘Infinite® 200’’ fluorescent plate reader (Tecan,

Austria) using excitation wavelength of 495 nm and emission wavelength of 515 nM

(λex = 495 nm, λem = 515 nm).

The confluent cultures of coronary microvascular endothelial cells in 96-well plates

were washed twice with PBS and incubated with 5 μM DAF-FM diacetate in DMEM

for 30 min, washed again and incubated with HBSS for 30 min to allow time for DAF-

FM diacetate de-esterification. After 30 min incubation, cells were exposed to insulin

or vehicle and DAF-FM triazole fluorescence was measured. The results were

corrected by subtracting the non-specific fluorescence detected in wells that had not

been treated with DAF-FM diacetate or which did not contain cells. Four

measurements per treatment group were performed on cells from an individual

culture and averaged to yield one value. Experiments were done on 3 individual

cultures.

5 Supplement Material

Immunofluorescence. Confluent endothelial monolayers were rinsed three times with

PBS (pH 7.4), fixed with 100% methanol for 20 min at -20°C or 4% paraformaldehyde

(PFA) for 20 min at room temperature, and washed again thrice with PBS and

permeabilized with 0.2% Triton X-100 for 20 min. The cells were incubated with an

anti-pan-cadherin antibody (1:100) overnight at 4°C. Afterward, the coverslips were

washed thrice with PBS and incubated with FITC-conjugated anti-rabbit IgG (1:200)

for 1h at 37°C. The coverslips were finally mounted on glass slides with a drop of

buffered glycerol (pH 8.5).

Confocal microscopy and image analysis. Confocal images were obtained by laser

scanning microscopy (LSM 510; Zeiss, Jena, Germany). Fluorophores were excited

using He-Ne (545 nm) and argon (492 nm) lasers. Image acquisition and analyses

were carried out using software provided with the confocal microscope. For

quantification of VE-cadherin distribution, we quantified fluorescence in a zone that

included only the cell junction using the line feature of the image analysis software to

trace the cell junction along its contours. Fluorescence data obtained reflects

analyses on 10 cells per experiment from three experiments.

Statistical analysis. Data are given as means + S.D. of 3-5 experiments using

independent cell preparations. The comparison of means between groups was

performed by one-way analysis of variance (ANOVA) followed by a Student-

Newman-Keuls post-hoc test. Changes in parameters within the same group were

assessed by multiple ANOVA analysis. Probability (P) values of less than 0.05 were

considered significant.

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from

6 Supplement Material

Reference List

(1) Saperstein R, Vicario PP, Strout HV, Brady E, Slater EE, Greenlee WJ, Ondeyka DL, Patchett AA, Hangauer DG. Design of a selective insulin receptor tyrosine kinase inhibitor and its effect on glucose uptake and metabolism in intact cells. Biochemistry. 1989;28:5694-701.

(2) Piper HM, Spahr R, Mertens S, Krützfeldt A, Watanabe H. Microvascular endothelial cells from heart. In: Piper HM, ed. Cell Culture Technique in Heart and Vessel Research.Heidelberg: Springer; 1990. p. 158-77.

(3) Noll T, Wozniak G, McCarson K, Hajimohammad A, Metzner HJ, Inserte J, Kummer W, Hehrlein FW, Piper HM. Effect of factor XIII on endothelial barrier function. J Exp Med. 1999;189:1373-82.

(4) Gündüz D, Kasseckert SA, Härtel FV, Aslam M, Abdallah Y, Schäfer M, Piper HM, Noll T, Schäfer C. Accumulation of extracellular ATP protects against acute reperfusion injury in rat heart endothelial cells. Cardiovasc Res. 2006;71:764-73.

(5) Noll T, Holschermann H, Koprek K, Gündüz D, Haberbosch W, Tillmanns H, Piper HM. ATP reduces macromolecule permeability of endothelial monolayers despite increasing [Ca2+]i. Am J Physiol. 1999;276:H1892-H1901.

(6) Gündüz D, Hirche F, Härtel FV, Rodewald CW, Schäfer M, Pfitzer G, Piper HM, Noll T. ATP antagonism of thrombin-induced endothelial barrier permeability. Cardiovasc Res. 2003;59:470-8.

(7) Härtel FV, Rodewald CW, Aslam M, Gündüz D, Hafer L, Neumann J, Piper HM, Noll T. Extracellular ATP induces assembly and activation of the myosin light chain phosphatase complex in endothelial cells. Cardiovasc Res. 2007;74:487-96.

(8) Bindewald K, Gündüz D, Härtel F, Peters SC, Rodewald C, Nau S, Schäfer M, Neumann J, Piper HM, Noll T. Opposite effect of cAMP signaling in endothelial barriers of different origin. Am J Physiol Cell Physiol. 2004;287:C1246-C1255.

(9) Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature. 1970;227:680-5.

at Universitaet Giessen on May 20, 2010 atvb.ahajournals.orgDownloaded from