vegf increases permeability of the endothelial cell monolayer by activation of pkb/akt, endothelial...

1Pwti(

Microvascular Research 62, 252–262 (2001)doi:10.1006/mvre.2001.2338, available online at http://www.idealibrary.com on

VEGF Increases Permeability of the Endothelial CellMonolayer by Activation of PKB/akt, Endothelial Nitric-Oxide Synthase, and MAP Kinase Pathways

Brajesh K. Lal, Shubha Varma, Peter J. Pappas, Robert W. Hobson, II,and Walter N. DuranProgram in Vascular Biology, Department of Pharmacology & Physiology, and Department of Surgery,UMDNJ–New Jersey Medical School; Newark, New Jersey 07103-2714

Received January 10, 2001; published online August 6, 2001

1

VEGF is a key regulator of vascular permeability. How-ever, its signaling pathways are incompletely under-stood. We tested the hypothesis that VEGF regulates en-dothelial cell (EC) permeability by activating PKB/akt,NOS, and MAP kinase dependent pathways using humanumbilical vein EC (HUVEC). Permeability was measuredfrom FITC–dextran 70-kDa flux across the EC monolayerat baseline and after VEGF at 0.034, 0.068, 1, 10, and 100nM. VEGF increased HUVEC permeability to FITC–dextran in a dose-dependent manner. VEGF (1 nM) in-creased permeability from 3.9 3 1026 6 0.7 3 1026 to

4.0 3 1026 6 1.7 3 1026 cm/s (mean 6 SEM; P < 0.001).ermeability changes were also assessed after treatmentith 1, 10, and 100 nM wortmannin (PI 3-kinase inhibi-

or); 0.01, 0.1, and 1.0 nM LY294002 (PI 3-kinase inhib-tor); 200 mM L-NMMA (NOS inhibitor); 2.7 mM AG126p42/44MAPK inhibitor); and 0.006, 0.06, and 0.6 mM

SB203580 (p38MAPK inhibitor). All inhibitors blockedVEGF-induced permeability changes. Our data demon-strate that (1) VEGF increases permeability of EC monolay-ers in a dose-dependent fashion, and (2) VEGF-induced per-meability is mediated through PI-3 kinase-PKB, NOS, and
MAP-kinase signaling cascades. These observations suggestthat microvascular hyperpermeability associated with in-flammation and vascular disease is mediated by activationof these EC signaling pathways. © 2001 Academic Press
RHct

252

Key Words: vascular endothelial growth factor; VEGF;microvascular permeability; biochemical signaling path-ways; protein kinase B; akt; endothelial NOS; nitric ox-ide; mitogen-activated protein kinases; MAP kinases.

INTRODUCTION

Signaling interactions between the vascular walland blood cells provide a unique way of communicat-ing, coordinating, and integrating an appropriatephysiologic response to the changing tissue environ-ment in health and disease. Studies in vivo have shownthat the signaling pathways for several inflammatoryagents involve the activation of protein kinase C(Kobayashi et al., 1994; Murray et al., 1991; Ramırez etal., 1996; Wu et al., 1999) and nitric oxide synthase(NOS; Hood et al., 1998; Mayhan, 1992, 1993, 1994,999; Ramırez et al., 1995; Yuan et al., 1993). These

studies suggest a central role for endothelial cells(EC1) in the regulation of microvascular permeability,

1 Abbreviations used: VEGF, vascular endothelial growth factor;
1 (flt-1), VEGF receptor 1; R2 (flk-1/KDR), VEGF receptor 2;UVEC, human umbilical vein endothelial cells; EC, endothelial
ells; PI 3-K, phosphatidylinositol 3-kinase; PKB or PKB/akt, pro-ein kinase B; MAPK, mitogen-activated protein kinase; FITC-D,

0026-2862/01 $35.00Copyright © 2001 by Academic Press

All rights of reproduction in any form reserved.

but do not exclude contributions from leukocytes,mast cells, and/or pericytes. Because of the potentialcomplexity of the in vivo signaling interactions, inves-tigations performed in the controlled environment oftissue culture are necessary. For this reason, we testedthe ability of EC to modulate permeability to macro-molecules across a monolayer using human umbilicalendothelial cells (HUVEC) as a model. We investi-gated signaling pathways associated with vascular en-dothelial growth factor (VEGF) activity in EC.

VEGF increases microvascular permeability andplays a key role in normal neonatal vasculogenesis.VEGF also plays a role in pathologic disease statessuch as tumor angiogenesis, tissue ischemia, diabeticretinopathy, granulation tissue formation, and isch-emia–reperfusion injury (Dvorak et al., 1999). Mostinvestigations have primarily focused on the mito-genic properties of VEGF. Therefore, little is knownregarding the cellular signaling pathways involved inthe VEGF-mediated increase in EC permeability.

VEGF binds to transmembrane tyrosine kinase re-ceptors of which R1 (flt-1) and R2 (flk-1/KDR) are wellcharacterized (Quinn et al., 1993; Millauer et al., 1993).Additionally, neuropilin 1 and 2 bind VEGF but theircharacteristics are still being defined (Gluzman-Pol-torak et al., 2000). Once VEGF receptor binding occurs,several intracellular kinases are stimulated (Risau,1997). Phosphatidylinositol 3-kinase (PI 3-K) and pro-tein kinase B (PKB/akt) have been reported to beVEGF sensitive (Guo et al., 1995). PI 3-K stimulationresults in the phosphorylation of inositol phosphates(Toker and Cantley, 1997), which bind to the serine–threonine kinase PKB/akt (Franke et al., 1997). Wehypothesize that VEGF-mediated hyperpermeabilityis regulated via signaling cascades involving PI 3-Kand PKB. Furthermore, we propose that PI 3-K- andPKB-mediated signals converge at eNOS resulting indownstream stimulation of MAP (mitogen activatedprotein) kinases.

VEGF-induced stimulation of the MAP kinases p42/44MAPK and p38MAPK increases EC proliferation and dif-

VEGF Signaling in Vascular Hyperpermeability

fluorescein isothiocyanate–dextran; eNOS, endothelial nitric oxidesynthase; EBM, endothelial basal medium; LC, luminal chamber;ALC, abluminal chamber.

ferentiation (Rousseau et al., 1997; Marshall, 1995).Similarly, one report suggests that p42/44MAPK maymediate VEGF-induced EC permeability (Kevil et al.,1998). Therefore, in order to determine whetherVEGF-induced hyperpermeability is regulated via PI3-K/PKB stimulation of eNOS and downstream acti-vation of MAP kinases, we measured permeabilitychanges of HUVEC monolayers using a vertical diffu-sion chamber system and inhibition of key signalingmolecules with selective antagonists.

METHODS

Cell culture. Passage 1 HUVEC were obtainedfrom Clonetics (San Diego, CA) and grown in T-75tissue culture flasks at 37° under 5% CO2 and 95% air.Cells were grown in 15 ml of endothelial growth me-dium (Clonetics) containing fetal bovine serum, bo-vine brain extract, human epidermal growth factor,hydrocortisone, gentamycin, and amphotericin B.Confluent cells were trypsinized with 2 ml of trypsin–EDTA (Gibco-BRL, Grand Island, NY) and seededonto fibronectin-coated 12-mm, 0.4-mm polycarbonatemembranes (Snapwell, Costar, Cambridge, MA) at 2 3105 cells/ml. Seeded EC were allowed to grow toconfluence on the membranes over 5–7 days. HUVECbetween passages 2 and 4 were used for the experi-ments.

Measurement of permeability. Endothelializedmembranes were rinsed with serum-free endothelialbasal medium (EBM, Clonetics) and mounted onto aprewarmed system of vertical diffusion chambers(Navicyte, San Diego, CA). The diffusion system con-tained six units, with each unit having two chambers.Each chamber was filled with 5 ml of warm serum-freeEBM. A circulating water bath (VWR Scientific, SanFrancisco, CA) was used to maintain the chambers at37°. A mixture of air and 5% CO2 was bubbledthrough the chambers to stir each side of the mem-

253

brane. The system was allowed to equilibrate for 15min. For operational purposes, the chamber whoselumen faced the HUVEC was defined as the luminalchamber (LC). Consequently, the chamber facing the

Copyright © 2001 by Academic PressAll rights of reproduction in any form reserved.

atn

swmosddPP

ct

HUVEC-devoid side of the membrane was defined asthe abluminal chamber (ALC).

To initiate the experiment, the medium in each LCwas replaced with 5 ml of warm serum-free EBMcontaining 13.3 mg/ml FITC-labeled dextran (70 kDa)(Sigma Chemicals, St. Louis, MO). Baseline EC layerpermeability (P) was then measured for 60 min. Dur-ing this period, 5-ml samples from the LC at 0 and 60min, and 20-ml samples from each ALC at 0, 15, 30, 45,nd 60 min were drawn. To determine the impact ofopical VEGF on HUVEC permeability the test ago-ist, recombinant human VEGF165 (R&D Systems, Inc.,

Minneapolis, MN), was added to both chambers attime 60 min and additional measurements were per-formed for the next 60 min. During this experimentalperiod a 5-ml sample from the LC at 120 min, and 20-mlamples from each ALC at 65, 75, 90, and 120 min wereithdrawn. In all cases (baseline and stimulated per-eability measurements), we replaced equal volumes

f warm serum-free EBM in each chamber as theamples were drawn. All FITC–dextran samplesrawn from the chambers were diluted in 1 ml ofistilled water and their fluorescence was read in aerkin–Elmer fluorescence spectrophotometer (LS-3,erkin–Elmer, Norwalk, CT).Calculation of permeability. Permeability was cal-

ulated by measuring fluorescence intensity as a func-ion of time according to the standard formula: P 5

J/ADC, where P 5 permeability; J 5 solute flux; A 5area of membrane; and DC 5 concentration differenceacross the diffusion membrane. Flux of FITC–dextran( J) was estimated as the derivative of the fluorescenceintensity of samples from the ALC (I) with respect totime (dI/dt); A 5 1.13 cm2, and DC 5 FITC–dextranconcentration difference across the HUVEC layer andwas estimated as the mean fluorescence intensity inthe luminal chamber. This approximation for DC isadequate because the concentration of FITC–dextranin ALC is negligible compared to LC.

An initial series of experiments was conducted tostandardize our system. Subsequently, the dose (con-centration)–response characteristics for VEGF were

254

determined. Finally, the cellular signaling pathwayswere explored by blocking specific molecules.

Standardization protocols. For initial standard-ization baseline permeability across fibronectin-coated


polycarbonate membranes was assessed. The mea-surement of the flux of FITC–dextran in unseeded,naked membranes for 120 min served to demonstratethat the pore size of the fibronectin-coated polycar-bonate membrane offered no resistance to the trans-port of the test macromolecule. As a second procedureof standardization the baseline flux of FITC–dextranwas measured across an unstimulated HUVEC mono-layer over 240 min (twice the total period of the ex-periment). This experiment served to assess the influ-ence of time on the experimental HUVEC preparationand confirmed that the preparation did not degradeover the experimental period (data not shown).

Dose–response protocols. To determine the dose–response characteristics of VEGF, baseline permeabil-ity was measured across nonstimulated HUVECmonolayers for 60 min. Various doses of VEGF werethen added simultaneously to both LC and ALC at 60min and the permeability was measured again. Thefinal VEGF concentrations in the diffusion chamberswere 0.034, 0.068, 1, 10, and 100 nM. Because thevertical diffusion system has six units, the experimentswere planned so that some units received no treat-ment and served as naive controls. The experimentswere designed in a paired modality so that permeabil-ity measurements across the HUVEC monolayers dur-ing the experimental period (61–120 min) could becompared to their respective baseline from the first 60min.

Signaling pathway protocols. Our working hy-pothesis is that VEGF-mediated hyperpermeability isregulated via signaling cascades involving PI 3-K andPKB, with activation of eNOS and MAP kinases. Theproposed signaling pathways are depicted in Fig. 1.To test the elements of our hypothesis, specific signal-ing molecules were blocked with pharmacologicagents (Fig. 1) and their impact on the permeabilityresponse stimulated by VEGF was determined. Mem-branes seeded with EC were placed in the diffusionchambers, pharmacologic blocking agents were ap-plied to each chamber for 60 min, and permeabilitywas measured. At time 60 min, VEGF was added to

Lal et al.

each chamber to a final concentration of 1 nM andpermeability was assessed again. The efficacy of theblocking agents was also determined by exploringtheir dose–response characteristics.

pStSSicpwia

2-4-Morpholinyl-8-phenyl-4H-1-benzopyran-4-one(LY294002,Calbiochem, San Diego, CA) at concentra-tions of 0.01, 0.1, and 1.0 nM and wortmannin (Sigma)at 1, 10, and 100 nM were used to block PI 3-K, andindirectly PKB. l-N-Monomethyl arginine (l-NMMA,Calbiochem) at 200 mM was applied to inhibit NOS(Ramırez et al., 1995, 1996). 3-Hydroxy-4-nitrobenzyli-dene malononitrile (AG126, Biomol, Plymouth Meet-ing, PA) at 2.7 mM was used to block p42/44MAPK

(Kevil et al., 1998), and SB203580 (Calbiochem) at0.006, 0.06, and 0.6 mM was used to block p38MAPK.Serum-free EBM was added to a set of chambers ineach experiment as a control.

Statistical analysis. All data are reported as rela-tive permeability compared to control and shown as

FIG. 1. Schematic of signaling pathways. VEGF-induced hyper-ermeability is mediated by its KDR (VEGF R2) specific receptor.timulation of KDR activates PI-3K and PKB/akt sequentially. Inurn, PKB/akt phosphorylates eNOS and leads to NO production.oluble guanylate cyclase produces cGMP upon activation by NO.ubsequently, cGMP-dependent kinase (PKG) stimulates the activ-ty of MAP kinases. It is assumed that MAP kinases interact withytoskeletal and cell junction proteins to regulate microvascularermeability. Selective inhibitors of specific signaling moleculesere applied to block the activity of the target molecule. Inhibition

s indicated by the symbol —X— and the target is shown by anrrowhead. Inhibitors: Anti KDR 5 monoclonal antibody directed

against KDR (VEGF R2); LY294002 and wortmannin 5 inhibitors ofPI 3-K; l-NMMA 5 l-nitro-monomethyl arginine 5 inhibitor ofnitric oxide synthase; AG126 5 inhibitor of p42/44MAPK; SB203580 5

inhibitor of p38MAPK.


means 6 standard error of the mean. Statistical anal-ysis was performed using Graphpad Prizm 3.0 (SanDiego, CA). Samples, with n $ 6 for each experiment,were compared using a one-way analysis of variance

with Tukey post hoc testing. Significance was deter-mined for all treatments compared with the appropri-ate control and accepted at P , 0.05.

RESULTS

Dose Response

We verified that a linear relationship exists betweenfluorescence intensity and FITC–dextran 70 concen-tration in our experimental setup, thus validating theuse of fluorescence measurements in our approach.Our initial experiments also showed that the pore sizeof the polycarbonate membrane did not offer resis-tance to the transfer of FITC–dextran. These data con-firmed that confluent HUVEC are responsible for thebarrier function between the diffusion chambers.

Application of VEGF increased HUVEC permeabil-ity to FITC–dextran in a dose–response manner (Fig.2). The permeability enhancing effect of VEGF wasrapid in onset. A significant increase in permeabilitywas noted as early as 15 min after the addition ofVEGF. The effect was limited in duration and resolved3 to 4 h after the stimulus (data not shown). VEGF at0.034 nM did not enhance permeability, as the mea-sured values were not significantly different from con-trols. We obtained a significant response with VEGF at0.068 nM. A maximal response was achieved afterVEGF at 1 nM. The difference in permeability changesinduced by VEGF at 0.068 and at 1 nM was significant(P , 0.05). VEGF at 1 nM produced a significant3.6-fold increase in HUVEC permeability from a base-line mean of 3.9 3 1026 6 0.7 3 1026 cm/s to aVEGF-treated mean permeability of 14.0 3 1026 61.7 3 1026 cm/s during the second hour of the exper-iment (P , 0.001). The permeability response de-clined, relative to 1 nM, at VEGF concentrations of 10and 100 nM. However, these responses were signifi-

255

cantly higher than controls (P , 0.05). Based on theseresults, we tested the involvement of signaling mole-cules in the VEGF-stimulated pathways using 1 nMVEGF as the test agonist concentration.


d

apratpnL

layerse period

PI 3-K and PKB as Signaling Molecules in VEGF-Induced Permeability

We inhibited PI 3-K by application of wortmanninand LY294002 in separate experiments to test its rolein VEGF-induced permeability. It is generally ac-cepted that inhibition of PI 3-K serves also to studyindirectly the role of PKB in the signaling cascade.Wortmannin, applied to HUVEC during the base-line measurement period, efficaciously inhibited the1 nM VEGF-induced increase in permeability. Thiseffect was noted at all three wortmannin dosestested—1, 10, and 100 nM (Fig. 3). Interestingly, thestrongest inhibition was obtained with 1 nM wort-mannin (P , 0.001 relative to VEGF), the lowest

ose tested.Similarly, incubation of HUVEC with LY294002,

n inhibitor considered more specific for PKB, com-letely eliminated the VEGF-induced permeabilityesponse (Fig. 4). The strongest inhibition waschieved with 0.01 nM LY294002 (P , 0.001 relativeo VEGF alone), even though the magnitude of the

FIG. 2. VEGF: Permeability dose response. Endothelial cell monoxpressed as a percentage increase in permeability over the baseline

3, P , 0.05.

256

ermeability change induced by VEGF was not sig-ificantly different from controls at 0.1 and 1 nMY294002.


Signaling through eNOS in VEGF-InducedPermeability

Having shown that PI 3-K and PKB are importantsignaling molecules in the cascade leading to VEGF-enhanced permeability, we explored the participationof eNOS signaling in this process. The approach isappropriate because it has been recently demonstratedthat PKB can directly phosphorylate eNOS and induceNO production (Fulton et al., 1999; Dimmeler et al.,1999) and also that VEGF can phosphorylate eNOS(Wu et al., 1999). Based on our in vivo experience(Ramırez et al., 1995, 1996), we preincubated HUVECwith 200 mM l-NMMA during the baseline period andsubsequently tested the ability of VEGF to increasepermeability in the presence of l-NMMA. Figure 5demonstrates that blockade of eNOS efficiently andcompletely inhibits VEGF-induced enhancement ofpermeability (P . 0.05).

Role of MAPK in VEGF-Induced Hyperpermeability

were treated with 0.034, 0.068, 1, 10, and 100 nM VEGF. Data areof no treatment. With respect to control, w, P , 0.001; }, P , 0.01;

Lal et al.

Normally, production of NO by eNOS stimulatessoluble guanylyl cyclase with an increase in cGMPand PKG activity. The activity of PKG in turn modu-lates MAP kinases. For this reason, we investigated

A

ySpa

cHpVSit0

hibits, VEGFpercen1; R, P

the possible involvement of MAPK in the signalingcascade associated with VEGF-stimulated hyperper-meability. To ascertain the role of p42/44MAPK orp38MAPK, we applied 2.7 mM AG126 and several dosesof SB203580 to HUVEC. Figure 6 displays the inhibi-tion of VEGF-induced hyperpermeability obtained byblocking p42/44MAPK with AG126. We applied a single

G126 concentration of 2.7 mM based on the results ofour preliminary studies and published data (Kevil etal., 1998). AG126 significantly inhibited the impact ofVEGF application on HUVEC permeability to FITC–dextran. The increase in FITC–dextran transportachieved with VEGF after application of AG126 wasnot significantly different from baseline.

Because AG126 is a specific blocker of tyrosine phos-phorylation of p42/44MAPK, we explored the possiblerole of p38MAPK using SB203580, a tyrosine phosphor-

FIG. 3. PI 3-kinase and PKB/akt blockade with wortmannin inpreincubated with wortmannin at 1, 10, and 100 nM. Subsequentlypermeability to FITC–D 70 was measured. Data are expressed as theto control: w, P , 0.001. With respect to 1 nM VEGF: 3, P , 0.00


lation blocker for this specific signaling molecule.B203580 was dissolved initially in DMSO. To test aossible direct action of DMSO, we investigated thection of DMSO in the dilution yielding the final con-

entration of SB203580. DMSO did not alter baselineUVEC permeability (results not shown). Due to theharmacodynamics of SB203580, we coapplied 1 nMEGF with the blocking agent. With this approach,B203580 inhibited VEGF-induced hyperpermeabilityn a dose-dependent manner (Fig. 7), with no inhibi-ion at 6 nM and strongest inhibition at 600 nM (P ,.01 with respect to VEGF).

DISCUSSION

We have demonstrated that the signaling pathwaysregulating VEGF-induced hyperpermeability of endo-thelial cells in vitro include PKB/akt, NOS, and

VEGF-induced hyperpermeability. Endothelial monolayers werewas added to achieve a concentration of 1 nM and the change in

tage increase in permeability over the baseline period. With respect, 0.01; 1, P , 0.05.

257

MAPK. The strength of the demonstration is based onthe significant reduction in VEGF-induced hyperper-meability caused by specific pharmacologic inhibitorsto PI 3-K, PKB, NOS, and MAPK.


EGF-ib s addet increac , P ,

wm

Experimental Preparation

Investigations of endothelial cell permeability utiliz-ing in vitro cell systems and HUVEC are well-recog-

FIG. 4. Inhibition of PI 3-kinase–PKB/akt with LY294002 blocks Vated with LY294002 at 0.01, 0.1, and 1.0 nM. Subsequently, VEGF wao FITC–D 70 was measured. Data are expressed as the percentageontrol: w, P , 0.01. With respect to 1 nM VEGF: }, P , 0.001; 3

FIG. 5. VEGF-induced hyperpermeability is blocked by inhibitionof NOS with l-NMMA. Endothelial monolayers were preincubatedwith l-NMMA at 200 mM. Subsequently, VEGF was added to

258

achieve a concentration of 1 nM and the change in permeability toFITC–D 70 was measured. Data are expressed as the percentageincrease in permeability over the baseline period. With respect tocontrol: w, P , 0.01. With respect to 1 nM VEGF: }, P , 0.05.


nized (Shasby et al., 1982; Kevil et al., 1998). The ad-vantage of HUVEC is that they grow well in cultureand their phenotype is well characterized and remainsstable through several passages (at least up to passage

nduced hyperpermeability. Endothelial monolayers were preincu-d to achieve a concentration of 1 nM and the change in permeabilityse in permeability relative to the baseline period. With respect to0.01.

FIG. 6. Inhibition of p42/44MAP kinase tyrosine phosphorylationith AG126 blocks VEGF-induced hyperpermeability. Endothelialonolayers were preincubated with 2.7 mM AG126. Subsequently,

VEGF was added to achieve a concentration of 1 nM and the change

Lal et al.

in permeability to FITC–D 70 was measured. Data are expressed asthe percentage increase in permeability over the baseline period.With respect to control: w, P , 0.001. With respect to 1 nM VEGF:R, P , 0.05.

mpWtatditlt1

c

tcpurdHptapattt

r2E

noarecprpta

4, as used in our experiments). In addition, they ex-press the VEGF binding receptors R1 and R2. For thesereasons, we employed a HUVEC monolayer grown onSnapwell membranes and placed vertically in a diffu-sion chamber. The major advantages of this systemover that of the Transwell system are that they allowfor built-in control of temperature through a warmwater-jacket and for stirring the medium through aCO2 gas port. This preserves the monolayer environ-

ent better and eliminates errors associated with sam-ling from unstirred layers for fluorescence analysis.e performed a dose–response experiment in which

he relationship between VEGF and HUVEC perme-bility to FITC–dextran was determined at concentra-ions reported to induce proliferation of EC. We choseextrans as probes due to their similarity to proteins

n terms of effective hydrodynamic radius and the facthat fluorescence intensity of FITC–dextrans correlatesinearly with dextran concentration in the experimen-

FIG. 7. Blockade of tyrosine phosphorylation of p38MAP kinasewith SB203580 inhibits VEGF-induced permeability. Baseline per-meability of endothelial cell monolayers was measured during thefirst hour. Subsequently, 1 nM VEGF with different doses ofSB203580 (6, 60, 600 nM) was coapplied and the change in perme-ability to FITC–D 70 was measured. Data are expressed as thepercentage increase in permeability over the baseline period. Withrespect to control: w, P , 0.001. With respect to 1 nM VEGF: 1, P ,

0.01; 3, P , 0.001.


al range (Gawlowski and Duran, 1986; Kim et al.,993).VEGF at 1 nM elicited maximal permeability

hanges in our diffusion chamber system. The onset of

he response started within 15 min of the VEGF appli-ation. Our results differ markedly from those re-orted by Chang et al. (2000), who reported that stim-lation of HUVEC with VEGF (0.5–2.0 nM) did notesult in increased endothelial cell permeability. Theifferent permeability observations may be related toUVEC passage number. We used HUVEC up toassage 4, while Chang et al. employed HUVEC be-

ween passages 4 and 6. Whether R1 and R2 receptorsnd an intact signaling pathway for VEGF are ex-ressed in passage 4–6 HUVEC is unknown. We alsodded VEGF to both luminal and abluminal chamberso enhance access to the EC VEGF R1 and R2 recep-ors, because the receptors may exhibit asymmetry inheir location.

The concentration of VEGF in normal human serumanges from 2 to 7 pM (Kraft et al., 1999; Kranz et al.,000). This concentration of VEGF does not enhanceC proliferation above baseline in vitro (Bernatchez et

al., 1999). VEGF concentrations ranging from 0.023 to0.185 nM have been reported in the serum of patientswith various malignancies (Kraft et al., 1999). The ED50

for VEGF to induce proliferation of EC in vitro is 0.04M (Myoken et al., 1991), with most studies reportingptimal doses for proliferation from 0.25 (Myoken etl., 1991) to 1.2 nM (Papapetropoulos et al., 1997). Ouresults demonstrate that at 0.034 nM, VEGF does notnhance permeability over baseline in vitro. At con-entrations above this threshold, VEGF caused hyper-ermeability in a dose-dependent manner andeached a peak at 1 nM. Because of the probableathophysiologic relevance of VEGF at this concentra-

ion, we performed our examination of hyperperme-bility signaling pathways with VEGF at 1 nM.

Signaling Pathways

PKB/akt. VEGF stimulates several signal trans-duction pathways after it is internalized by its R1 andR2 receptors. VEGF stimulates PI 3-K activity to in-duce proliferation of bovine aortic endothelial cells(Guo et al., 1995) and promote EC survival (Gerber et

259

al., 1998) by increasing PI 3-K phosphorylation. Addi-tionally, overexpression of PI 3-K or PKB/Akt resultsin angiogenesis in the chick chorioallantoic membrane(Jiang et al., 2000). The VEGF-stimulated pathways


opahlpaa

ota

(

ancuie

h

have been extensively investigated in relation to an-giogenesis, cell survival, and apoptosis. However, thesignaling pathways involved in hyperpermeability re-main largely unexplored.

Does VEGF stimulate PKB/akt to induce hyper-permeability? The studies aimed at answering thisquestion are designed with the understanding thatcurrently there are no direct antagonists for PKB/akt. Thus, the approach is based on inhibition ofPI-3K that serves to block PKB. For this reason, weused two structurally different agents to block PI-3K, namely wortmannin and LY294002. Both effec-tively and significantly inhibited VEGF-induced hy-perpermeability in HUVEC. Taken together, ourwork demonstrates that inhibition of PI-3K–PKBsignificantly reduces the hyperpermeability re-sponse to VEGF and supports the involvement ofPKB/akt in the signaling cascade leading to VEGF-stimulated hyperpermeability.

Our results regarding PKB differ from an earlierreport in which the authors blocked PI 3-K with asingle dose of wortmannin (10 nM) and subse-quently stimulated a HUVEC monolayer with a sin-gle dose of VEGF, 660 pM (Kevil et al., 1998). Withthis approach, they observed no inhibition of VEGF-induced increase in permeability. There are a fewpossible explanations for this difference. We studiedFITC– dextran transport in HUVEC demonstrating abaseline permeability of 3.9 3 1026 6 0.7 3 1026

cm/s, while the HUVEC permeability to albumin inKevil’s study ranged from 14.5 3 1026 to 19.3 3 1026

cm/s. Baseline HUVEC permeability coefficients toalbumin of 3.09 3 1026 6 1.27 3 1026 cm/s were

btained by Chang et al. (2000). The lower baselineermeability coefficient in our work and in Chang etl.’s study, relative to Kevil’s, may be the result of aigher degree of confluence of the HUVEC mono-

ayers. In addition, we performed a more robustharmacologic study to probe the role of PKB. Thispproach included a dose response for VEGF as wells for wortmannin and LY294002.Nitric oxide synthase. We and others have dem-

260

nstrated that NOS activity is important as a signal inhe events leading to agonist-induced hyperperme-bility in vivo (Kobayashi et al., 1994; Ramırez et al.,

1995, 1996; Mayhan, 1994; Yuan et al., 1993). Our in

imer


vitro results confirm that inhibition of NOS in ECblocks the permeability response induced by VEGF. Itis important to highlight that the only cells present inthis permeability assay are ECs. Additionally, VEGF-induced hyperpermeability is observed within 5–15min. Thus, it is reasonable to conclude that eNOSplays a significant role in the modulation of enhancedpermeability in vivo.

Our data on the concerted activation of PKB andNOS are in agreement with reports that stimulation ofPKB induces NOS phosphorylation and leads to in-creased NO production (Fulton et al., 1999; Dimmeleret al., 1999). Our results suggest that this sequence ofactivation may be of significance for the regulation ofmicrovascular permeability by EC. VEGF-inducedphosphorylation of eNOS in HUVEC and its associa-tion with VEGF-stimulated hyperpermeability in iso-lated coronary venules have been documented previ-ously (Wu et al., 1999; Yuan et al., 1993). We have alsoobtained supportive evidence for the role of NOSphosphorylation and NO production in PAF-inducedhyperpermeability in vivo in the hamster cheek pouchDuran et al., 2000).

MAP kinases. VEGF has the ability to promotengiogenesis and to increase permeability. MAP ki-ases appear to play a role in both biological pro-esses. Our report, and the study by Kevil et al. (1998)sing AG126, confirms a role for MAPK in VEGF-

nduced hyperpermeability. In addition, we providevidence for the role of both p38MAPK and p42/44MAPK

in the control of endothelial permeability.The signaling pathways activated in VEGF-in-

duced hyperpermeability include the same key mol-ecules as the pathway stimulated by VEGF-inducedangiogenesis and cell survival. The similarities be-tween the VEGF biochemical pathway for perme-ability and the VEGF pathway for angiogenesis andcell survival, as well as the hyperpermeability sig-naling pathways for PAF (Kobayashi et al., 1994),histamine (Yuan et al., 1993), or bradykinin (May-

an, 1992), highlight the concept that cells use sim-

Lal et al.

lar or identical signaling pathways to evoke re-arkably different responses. The particular

lements that code and provide specificity to theseesponses still remain elusive.

C

D

D

D

K

K

K

K

M

M

M

M

M

M

M

ACKNOWLEDGMENTS

This work was supported by Grants KO8 HL03354, KO7HL03437, and HL 43146 from the National Institutes of Health.

REFERENCES

Bernatchez, P. N., Soker, S., and Sirois, M. G. (1999). Vascularendothelial growth factor effect on endothelial cell proliferation,migration, and platelet-activating factor synthesis is Flk-1-depen-dent. J. Biol. Chem. 274, 31047–31054.

hang, Y. S., Munn, L. L., Hillsley, M. V., Dull, R. O., Yuan, J.,Lakshminarayanan, S., Gardner, T. W., Jain, R. K., and Tarbell,J. M. (2000). Effect of vascular endothelial growth factor on cul-tured endothelial cell monolayer transport properties. Microvasc.Res. 59, 265–277.

immeler, S., Fleming, I., Fisslthaler, B., Hermann, C., Busse, R., andZeiher, A. M. (1999). Activation of nitric oxide synthase in endo-thelial cells by Akt-dependent phosphorylation. Nature 399, 601–605.

uran, W. N., Seyama, A., Yoshimura, K., Gonzalez, D. R., Jara, P. I.,Figueroa, X. F., and Boric, M. P. (2000). Stimulation of NO pro-duction and of eNOS phosphorylation in the microcirculation invivo. Microvasc. Res. 60, 104–111.

vorak, H. F., Nagy, J. A., Feng, D., Brown, L. F., and Dvorak, A. M.(1999). Vascular permeability factor/vascular endothelial growthfactor and the significance of microvascular hyperpermeability inangiogenesis. Curr. Topics Microbiol. Immunol. 237, 97–132.

Franke, T. F., Kaplan, D. R., Cantley, L. C., and Toker, A. (1997).Direct regulation of the Akt proto-oncogene product by phospha-tidylinositol-3,4-bisphosphate. Science 275, 665–668.

Fulton, D., Gratton, J. P., McCabe, T. J., Fontana, J., Fujio, Y., Walsh,K., Franke, T. F., Papapetropoulos, A., and Sessa, W. C. (1999).Regulation of endothelium-derived nitric oxide production bythe protein kinase Akt. Nature 399, 597–601.

Gawlowski, D. M., and Duran, W. N. (1986). Dose-related effects ofadenosine and bradykinin on microvascular permselectivity tomacromolecules in the hamster cheek pouch. Circ. Res. 58, 348–355.

Gerber, H. P., McMurtrey, A., Kowalski, J., Yan, M., Keyt, B. A.,Dixit, V., and Ferrara, N. (1998). Vascular endothelial growthfactor regulates endothelial cell survival through the phosphati-dylinositol 39-kinase/Akt signal transduction pathway. Require-ment for Flk-1/KDR activation. J. Biol. Chem. 273, 30336–30343.

Gluzman-Poltorak, Z., Cohen, T., Herzog, Y., and Neufeld, G.(2000). Neuropilin-2 is a receptor for the vascular endothelial


growth factor (VEGF) forms VEGF-145 and VEGF-165. J. Biol.Chem. 275, 18040–18045.

Guo, D., Jia, Q., Song, H. Y., Warren, R. S., and Donner, D. B. (1995).Vascular endothelial cell growth factor promotes tyrosine phos-

M

phorylation of mediators of signal transduction that contain SH2domains. Association with endothelial cell proliferation. J. Biol.Chem. 270, 6729–6733.

Hood, J. D., Meininger, C. J., Ziche, M., and Granger, H. J. (1998).VEGF upregulates ecNOS message, protein, and NO productionin human endothelial cells. Am. J. Physiol. 274, H1054–H1058.

Jiang, B. H., Zheng, J. Z., Aoki, M., and Vogt, P. K. (2000). Phospha-tidylinositol 3-kinase signaling mediates angiogenesis and ex-pression of vascular endothelial growth factor in endothelialcells. Proc. Natl. Acad. Sci. USA 97, 1749–1753.

Kevil, C. G., Payne, D. K., Mire, E., and Alexander, J. S. (1998).Vascular permeability factor/vascular endothelial cell growthfactor-mediated permeability occurs through disorganization ofendothelial junctional proteins. J. Biol. Chem. 273, 15099–15103.

im, D., Armenante, P. M., and Duran, W. N. (1993). Transientanalysis of macromolecular transport across the microvascularwall and into the interstitium. Am. J. Physiol. 265, H993–H999.

obayashi, I., Kim, D., Hobson, R. W., II, and Duran, W. N. (1994).Platelet-activating factor modulates microvascular transport bystimulation of protein kinase C. Am. J. Physiol. 266, H1214–1220.

raft, A., Weindel, K., Ochs, A., Marth, C., Zmija, J., Schumacher, P.,Unger, C., Marme, D., and Gastl, G. (1999). Vascular endothelialgrowth factor in the sera and effusions of patients with malignantand nonmalignant disease. Cancer 85, 178–187.

ranz, A., Rau, C., Kochs, M., and Waltenberger, J. (2000). Elevationof vascular endothelial growth factor-A serum levels followingacute myocardial infarction. Evidence for its origin and func-tional significance. J. Mol. Cell. Cardiol. 32, 65–72.arshall, C. J. (1995). Specificity of receptor tyrosine kinase signal-ing: Transient versus sustained extracellular signal-regulated ki-nase activation. Cell 80, 179–185.ayhan, W. G. (1994). Nitric oxide accounts for histamine-inducedincreases in macromolecular extravasation. Am. J. Physiol. 266,H2369–H2373.ayhan, W. G. (1992). Role of nitric oxide in modulating perme-ability of hamster cheek pouch in response to adenosine 59-diphosphate and bradykinin. Inflammation 16, 295–305.ayhan, W. G. (1993). Role of nitric oxide in leukotriene C4-inducedincreases in microvascular transport. Am. J. Physiol. 265, H409–H414.ayhan, W. G. (1999). VEGF increases permeability of the blood–brain barrier via a nitric oxide synthase/cGMP-dependent path-way. Am. J. Physiol. 276, C1148–C1153.illauer, B., Wizigmann-Voos, S., Schnurch, H., Martinez, R.,Moller, N. P., Risau, W., and Ullrich, A. (1993). High affinityVEGF binding and developmental expression suggest Flk-1 as amajor regulator of vasculogenesis and angiogenesis. Cell 72, 835–846.urray, M. A., Heistad, D. D., and Mayhan, W. G. (1991). Role of

261

protein kinase C in bradykinin-induced increases in microvascu-lar permeability. Circ. Res. 68, 1340–1348.yoken, Y., Kayada, Y., Okamoto, T., Kan, M., Sato, G. H., and Sato,J. D. (1991). Vascular endothelial cell growth factor (VEGF) pro-


duced by A-431 human epidermoid carcinoma cells and identi-fication of VEGF membrane binding sites. Proc. Natl. Acad. Sci.USA 88, 5819–5823.

Papapetropoulos, A., Garcia-Cardena, G., Madri, J. A., and Sessa,W. C. (1997). Nitric oxide production contributes to the angio-genic properties of vascular endothelial growth factor in humanendothelial cells. J. Clin. Invest. 100, 3131–3139.

Quinn, T. P., Peters, K. G., De Vries, C., Ferrara, N., and Williams,L. T. (1993). Fetal liver kinase 1 is a receptor for vascular endo-thelial growth factor and is selectively expressed in vascularendothelium. Proc. Natl. Acad. Sci. USA 90, 7533–7537.

Ramırez, M. M., Quardt, S. M., Kim, D., Oshiro, H., Minnicozzi, M.,and Duran, W. N. (1995). Platelet activating factor modulatesmicrovascular permeability through nitric oxide synthesis. Micro-

262

vasc. Res. 50, 223–234.Ramırez, M. M., Kim, D. D., and Duran, W. N. (1996). Protein kinase

C modulates microvascular permeability through nitric oxidesynthase. Am. J. Physiol. 271, H1702–H1705.


Risau, W. (1997). Mechanisms of angiogenesis. Nature 386, 671–674.Rousseau, S., Houle, F., Landry, J., and Huot, J. (1997). p38 MAP

kinase activation by vascular endothelial growth factor mediatesactin reorganization and cell migration in human endothelialcells. Oncogene 15, 2169–2177.

Shasby, D. M., Shasby, S. S., Sullivan, J. M., and Peach, M. J. (1982).Role of endothelial cell cytoskeleton in control of endothelialpermeability. Circ. Res. 51, 657–661.

Toker, A., and Cantley, L. C. (1997). Signalling through the lipidproducts of phosphoinositide-3-OH kinase. Nature 387, 673–676.

Wu, H. M., Yuan, Y., Zawieja, D. C., Tinsley, J., and Granger, H. J.(1999). Role of phospholipase C, protein kinase C, and calcium inVEGF-induced venular hyperpermeability. Am. J. Physiol. 276,H535–H542.

Lal et al.

Yuan, Y., Granger, H. J., Zawieja, D. C., DeFily, D. V., and Chilian,W. M. (1993). Histamine increases venular permeability via aphospholipase C-NO synthase-guanylate cyclase cascade. Am. J.Physiol. 264, H1734–H1739.

vegf increases permeability of the endothelial cell monolayer by activation of pkb/akt, endothelial...

Documents