aproangiogenicpeptidederivedfromvascularendothelialgrowthf ... · hemostasis, thrombosis,and...

HEMOSTASIS, THROMBOSIS, AND VASCULAR BIOLOGY

A proangiogenic peptide derived from vascular endothelial growth factorreceptor-1 acts through �5�1 integrinSimonetta Soro,1 Angela Orecchia,2 Lucia Morbidelli,3 Pedro Miguel Lacal,4 Veronica Morea,5 Kurt Ballmer-Hofer,6

Federica Ruffini,4 Marina Ziche,3 Stefania D’Atri,4 Giovanna Zambruno,2 Anna Tramontano,1,7 and Cristina Maria Failla2

1Department of Biochemical Sciences A. Rossi Fanelli, University La Sapienza, Rome, Italy; 2Molecular and Cell Biology Laboratory, Istituto Dermopaticodell’Immacolata-Istituto di Ricovero e Cura a Carattere Scientifico (IDI-IRCCS), Rome, Italy; 3Department of Molecular Biology, University of Siena, Siena, Italy;4Molecular Oncology Laboratory, IDI-IRCCS, Rome, Italy; 5National Reseach Council (CNR) Institute of Molecular Biology and Pathology, University LaSapienza, Rome, Italy; 6Paul Scherrer Institut (PSI), Biomolecular Research, Molecular Cell Biology, Villigen-PSI, Switzerland; and 7Istituto Pasteur-FondazioneCenci Bolognetti, University La Sapienza, Rome, Italy

Vascular endothelial growth factorreceptor-1 (VEGFR-1) is a tyrosine kinasereceptor for growth factors of the VEGFfamily. Endothelial cells express amembrane-bound and a soluble variant ofthis protein, the latter being mainly con-sidered as a negative regulator of VEGF-Asignaling. We previously reported thatthe soluble form is deposited in the extra-cellular matrix produced by endothelialcells in culture and is able to promote celladhesion and migration through bindingto �5�1 integrin. In this study, we demon-

strate that the Ig-like domain II ofVEGFR-1, which contains the binding de-terminants for the growth factors, is in-volved in the interaction with �5�1 inte-grin. To identify domain regions involvedin integrin binding, we designed 12 pep-tides putatively mimicking the domain IIsurface and tested their ability to inhibit�5�1-mediated endothelial cell adhesionto soluble VEGFR-1 and directly supportcell adhesion. One peptide endowed withboth these properties was identified andshown to inhibit endothelial cell migra-

tion toward soluble VEGFR-1 as well. Thispeptide directly binds �5�1 integrin, butnot VEGF-A, inducing endothelial cell tu-bule formation in vitro and neoangiogen-esis in vivo. Alanine scanning mutagen-esis of the peptide defined which residueswere responsible for its biologic activityand integrin binding. (Blood. 2008;111:3479-3488)

© 2008 by The American Society of Hematology

Introduction

Vascular endothelial growth factor receptor-1 (VEGFR-1) is atyrosine kinase receptor for members of the vascular endothelialgrowth factor (VEGF) family.1 VEGFR-1 is closely related to thetype III tyrosine kinase-Fms/Kit/platelet-derived growth factorreceptors, but it is classified into a distinct class of receptors whoseextracellular region is composed of 7 immunoglobulin (Ig)-likedomains.2 Ig-like domain II (dII) comprises the region involved inVEGF-A and placenta growth factor (PlGF) binding,3,4 whereasIg-like domains III and IV are involved in heparin binding anddomain III is also responsible for interaction with neuropilin-1.5

Studies carried out with PlGF, which is a VEGFR-1-selectiveligand, indicated that the tyrosine kinase activity of this receptor ismostly required for adult pathologic angiogenesis.6 VEGFR-1 doesnot play a significant role in endothelial cell proliferation, althoughits activation is greatly involved in chemotaxis of monocyte-macrophages7 and in mobilization of endothelial and hematopoieticstem cells from the bone marrow.8,9

Gene knockout studies demonstrated that VEGFR-1 is essentialfor the development and differentiation of embryonic vasculatureand that its inactivation results in increased hemangioblast commit-ment and disorganization of blood vessels.10,11 However, micecarrying a homozygous deletion of the intracellular kinase domainshow a correct blood vessel development, and a substantialimpairment of monocyte-macrophages migration was the only

reported phenotype.12 This result suggests that the extracellularregion of VEGFR-1 is sufficient to support embryonic angiogenesis.

In addition to the transmembrane isoform of VEGFR-1 (tm-VEGFR-1), endothelial cells (ECs) produce a soluble form of thereceptor, arising from alternative splicing of the same gene.13

Soluble VEGFR-1 (sVEGFR-1) comprises the first 6 Ig-likedomains of the extracellular region of the receptor and a specific30 amino acid tail at its C-terminus. Both isoforms bind VEGF-Awith high affinity, and sVEGFR-1 acts as a potent antagonist ofVEGF-A signaling by sequestering the growth factor and therebyinhibiting its interaction with the transmembrane receptors.14 Wepreviously demonstrated that sVEGFR-1 secreted by ECs inculture becomes part of the extracellular matrix and is able tomediate EC adhesion through direct binding to �5�1 integrin.15

A key role of this integrin in promoting vasculogenesis andangiogenesis has been previously highlighted by the defectivevascular phenotype of �5 knockout mice16 and by the ability ofeither antibody raised against this integrin or antagonist peptides toblock in vitro and in vivo angiogenesis.17

Integrin binding to their ligands usually occurs through recogni-tion of short amino acid sequences. Many integrins, including�5�1, bind to arginine-glycine-aspartic acid (RGD)–containingpeptides, although other binding motifs have also been identi-fied.18,19 Because none of the classic �5�1 integrin binding motifs

Submitted March 2, 2007; accepted December 18, 2007. Prepublished online asBlood First Edition paper, January 9, 2008; DOI 10.1182/blood-2007-03-077537.

S.S. and A.O. contributed equally to this work.

The online version of this article contains a data supplement.

The publication costs of this article were defrayed in part by page chargepayment. Therefore, and solely to indicate this fact, this article is herebymarked ‘‘advertisement’’ in accordance with 18 USC section 1734.

© 2008 by The American Society of Hematology

3479BLOOD, 1 APRIL 2008 � VOLUME 111, NUMBER 7

For personal use only.on May 30, 2017. by guest www.bloodjournal.orgFrom

http://www.bloodjournal.org/

http://www.bloodjournal.org/site/subscriptions/ToS.xhtml

is present in the sequence of sVEGFR-1, we investigated themolecular determinants responsible for sVEGFR-1/integrin interac-tion. We found that Ig-like dII of VEGFR-1 directly interacts with�5�1 integrin. Because this domain is also involved in growthfactor binding, we tried to identify the domain regions specificallyinvolved in integrin binding. We designed and tested 12 peptidesputatively mimicking the domain surface and found one peptideable to inhibit �5�1-mediated endothelial cell adhesion to VEGFR-1and directly support cell adhesion. This peptide also inhibitedendothelial cell migration toward sVEGFR-1 and induced endothe-lial cell tubule formation in vitro and neoangiogenesis in vivo.Finally, we performed alanine scanning mutagenesis of the peptideto define which residues were responsible for its biologic activitiesand integrin binding.

Methods

Reagents

Recombinant human VEGFR-1/Fc, VEGF-A, and biotinylated VEGF-Awere purchased from R&D Systems (Minneapolis, MN). The purified �5�1integrin, octyl-�-D-glucopyranoside formulation, was from Millipore (Bil-lerica, MA). Peptide synthesis was carried out by Primm (Milan, Italy), andthe GRGDTP and GRGESP peptides were from Invitrogen (Paisley, UnitedKingdom). Antibodies used in this paper are detailed in Document S1(available on the Blood website; see the Supplemental Materials link at thetop of the online article).

Cell culture, cell adhesion, and endothelial tubule formation

Human umbilical vein endothelial cells (HUVECs) were obtained andcultured as described.15 A pool obtained from different individuals was usedin all the experiments. Adhesion assays to recombinant proteins wereperformed as previously described.15 Adhesion to peptides was carried outby using anhydride maleic activated microplates (Reacti-Bind; Pierce,Rockford, IL) coated with 500 �g/mL of each peptide. In competitionexperiments, HUVECs were preincubated with 10 �g/mL antibody or20 �g/mL peptide for 10 minutes at room temperature before plating andthen allowed to adhere for 30 minutes at 37°C in the presence of thecompounds. Where indicated, ethylenediaminetetraacetic acid (EDTA) orethyleneglycoltetraacetic acid (EGTA) (10 mM) was added at the time ofplating. To assay for tubule formation, HUVECs were examined in athree-dimensional collagen gel in the presence or absence of proteins,peptides, or antibodies as previously described.20 Wells were viewed withan Axiovert 135 inverted microscope (Zeiss, Oberkochen, Germany) usingACHROPLAN at 10�/0.25 PH1 and 20�/0.40 PH2. Images were acquiredusing a Canon TWAIN60 camera (Tokyo, Japan), and were processed withZoomBrowser EX version 5.5.0.190 software (Canon Information SystemsResearch Australia, North Ryde, Australia) and Adobe Photoshop version7.0 software (Adobe Systems, San Jose, CA).

Generation of recombinant Ig-like dII

The tmVEGFR-1 cDNA was kindly provided by Dr M. Shibuya (Institute ofMedical Science, University of Tokyo, Japan). The cDNA was modified tointroduce a Kpn I site followed by a Kozak sequence, a start codon at the 5� end, aBsi WI site followed by a HATag, and a stop codon at the 3� end. The polylinkerof the Baculovirus vector pFastBac 1 (Invitrogen) was replaced by a linkercontaining a Kpn I and a Bsi WI site together with a His tag. The tmVEGFR-1cDNA was first cloned into the pFastBac 1 vector using the Kpn I and Bsi WIsites. To construct the recombinant Ig-like dII, the signal sequence of VEGFR-1was amplified by polymerase chain reaction using the oligonucleotides: 5�-GCGCGGATCCGGTACCGC-3� and 5�-GTGATGCGTACGTGAACCT-GAACTAGATCCTGTG-3� and cloned into the modified pFastBac vector. Thenthe dII sequence was amplified with the oligonucleotides 5�-GTGATGCGTAC-GAGTGATACAGGTAGACCTTTC-3� and 5�-GTGATGCGTACGTCCGATT-GTATTGGTTTGTCGATG-3�. The polymerase chain reaction fragment was

cloned into the modified pFastBac vector carrying the VEGFR-1 signalsequence. Recombinant Baculovirus was obtained using the Bac-to-Bac system(Invitrogen). SF9 (Spodoptera frugiperda) cell line was purchased from Invitro-gen and maintained in culture at 27°C in TC100 medium (Invitrogen) containing10% heat inactivated fetal calf serum. Infected SF9 cells were cultured inserum-free SF-900 medium (Invitrogen). For protein expression, SF9 cells weregrown up to 60% confluence and infected with the corresponding recombinantvirus at a multiplicity of infection of 1. Forty-eight hours postinfection cells werecollected, and the recombinant protein was purified by affinity chromatographyusing the B-PER His Fusion Protein purification kit (Pierce). Cell lysates wereprepared using the B-PER extraction reagent and loaded on a nickel-chelatedcolumn, previously equilibrated with B-PER reagent. The column was washedwith wash buffer I and II, and protein eluted with elution buffer containing excessimidazole. The recombinant protein had the expected molecular weight onsodium dodecyl sulfate-polyacrylamide gel electrophoresis and was recognizedby an anti-HIS antibody (Santa Cruz Biotechnologies, Santa Cruz, CA).

Solid-phase binding assays

Integrin binding to recombinant proteins was assayed as previouslydescribed.15 Interaction of peptides (500 �g/mL, coated on Reacti-Bindmicroplates) with the purified integrin was evaluated with a 1:100dilution of the NKI-SAM 1 anti-�5 monoclonal antibody (mAb)followed by plate incubation with a biotinylated anti-mouse secondaryantibody (Vector Laboratories, Burlingame, CA). Streptavidin-alkalinephosphatase-conjugated and the appropriate substrate (4-nitrophenylphos-phate, Roche Diagnostics, Basel, Switzerland) were then used fordetection. Absorbance was determined at 405 nm using a Microplatereader 3550-UV (Bio-Rad, Hercules, CA). In a different experiment,500 �g/mL peptide were coated on the Reacti-Bind microplates andinteraction with VEGF-A was evaluated by incubating 200 ng biotinyl-ated VEGF-A on the plates. The amount of retained biotinylatedVEGF-A was then determined as described above.

Rabbit cornea assay

Angiogenesis was studied in the cornea of albino rabbits (Charles River,Calco, Como, Italy). Experiments were performed in accordance withguidelines of the European Economic Community for animal care andwelfare (EEC Law No. 86/609). Slow-release pellet preparation, surgicalprocedure, and quantification of angiogenesis (angiogenic score) wereperformed as previously reported.21 Multiple comparisons were performedby one-way analysis of variance and individual differences tested byStudent-Newman-Keuls test after the demonstration of significant inter-group differences.

Immunofluorescence studies

Aliquots of Dynabeads M-270 Carboxylic Acid (Dynal Biotech, Oslo,Norway) were coated with the selected peptides as recommended by themanufacturer. HUVECs were plated on fibronectin-coated coverslips andallowed to adhere for 2 hours in serum-free medium. Peptide-coated beadswere then added for 30 minutes. Cells were washed with phosphate-buffered saline (PBS), fixed with 3% formaldehyde in PBS, and permeabil-ized with 0.5% Triton X-100/PBS. Nonspecific staining was blocked with3% bovine serum albumin (BSA), and coverslips were incubated withprimary antibody against �5 subunit NKI-SAM 1 diluted 1:100 in 1%BSA/PBS. Coverslips were washed with PBS and stained with fluoresceinisothiocyanate-conjugated rabbit anti–mouse IgG (Dako Denmark A/S,Glostrup, Denmark) in 1% BSA/PBS for 45 minutes. Coverslips werewashed in PBS, mounted in Vectashield, and observed by confocalmicroscopy (Zeiss LSM 510 Meta, Oberkochen, Germany).

Western blotting analysis

HUVECs were starved for 24 hours in EBM (Clonetics, Lonza, Valais,Switzerland) complemented with 0.4% fetal calf serum, treated for 2 hourswith cycloheximide (20 �M), and plated on peptide 12- or fibronectin-coated dishes for 1 hour in the presence of cycloheximide (20 �M;Sigma-Aldrich, St Louis, MO), monensin (1 �M; Sigma-Aldrich), or

3480 SORO et al BLOOD, 1 APRIL 2008 � VOLUME 111, NUMBER 7




VEGF where indicated. Afterward, cells were directly lysed and boiled inSDS sample buffer, and 50 �L of the lysate run in a 10% sodium dodecylsulfate-polyacrylamide gel. Proteins were transferred to supported nitrocel-lulose membranes (Hybond-C; GE Healthcare, Chalfont St Giles, UnitedKingdom), using a Trans-Blot Cell from Bio-Rad. Membranes wereblocked in blocking solution (5% BSA, 0.1% Tween 20, 0.9% NaCl, 20 mMTris-HCl, pH 7.4) overnight at 4°C, washed with TBST buffer (0.1% Tween20, 0.9% NaCl, 20 mM Tris-HCl, pH 7.4), and incubated with the primaryantibody diluted 1:1000 in 3% BSA/TBST for 4 hours at room temperature.After washing with TBST, membranes were incubated for 1 hourat room temperature with the secondary antibody diluted 1:5000 in3% BSA/TBST. Membranes were then washed with TBST, and detectionwas carried out using the ECL Western blotting detection reagents andHyperfilms (both from GE Healthcare).

Structure analysis

Analysis of the three-dimensional structures of VEGFR-1 dII, availablefrom the Protein Data Bank (PDB),22 was performed using the interactivegraphics software InsightII23 and its module Biopolymer (Accelrys, SanDiego, CA).

Results

Ig-like dII of VEGFR-1 is involved in the interaction with�5�1 integrin

We previously showed that purified �5�1 integrin binds toVEGFR-1/Fc protein, comprising the soluble receptor fused to theFc domain of human IgG.15 We also demonstrated that an antibodyrecognizing a peptide derived from Ig-like dII of VEGFR-1competed with �5�1 integrin-mediated binding of EC to VEGFR-1.15 In the present study, we analyzed the effect of this antibody onVEGFR-1/integrin interaction. In a solid-phase binding assay, thisantibody completely inhibited �5�1 binding to VEGFR-1/Fc(Figure 1A), whereas another pAb (Flt-1 H225) recognizing theextracellular domain of VEGFR-1 did not. As expected, noinhibition was observed using an antibody (Flt-1 C17) against theintracellular domain of VEGFR-1 (Figure 1A). We concluded thatdeterminants for integrin binding were probably to be presentwithin Ig-like dII.

Based on domain boundaries observed in the experimentallydetermined structures available from the PDB database (PDB IDs:1flt, 1qty, 1rv6, 1qsv, 1qsz), we designed a recombinant proteincorresponding to dII (residues 129-229). This protein was producedin the SF9 insect cell line using the baculovirus-mediated expres-sion system and was purified by affinity chromatography. Recombi-nant dII was able to directly support HUVEC adhesion (Figure 1B)and to bind purified �5�1 integrin in a concentration-dependentmanner (Figure 1C).

Design and experimental testing of peptides mapping on dII

Because VEGFR-1 Ig-like dII contains binding determinants forboth growth factors of the VEGF family and �5�1 integrin, butneither VEGF-A nor PlGF compete with the interaction betweenVEGFR-1 and �5�1 integrin,15 dII regions responsible for integrinand growth factor binding are expected not to be overlapping. Toidentify those regions specifically responsible for integrin binding,we took advantage of the domain structures determined by x-raycrystallography and available from the PDB database. The three-dimensional structures were examined to identify contiguousamino acid sequences that might mimic exposed regions of dII.Twelve peptides (Table 1) were designed according to the follow-

ing criteria: (1) they are on the protein surface (ie, they have highoverall solvent-accessible surface area); (2) they correspond todifferent regions of the domain surface, covering it almost entirely;(3) they are shorter than 15 residues; (4) they contain polar residuesand, whenever possible, residues charged at physiologic pH values(Asp, Glu, Lys and Arg); (5) they contain a low percentage ofhydrophobic residues (ie, Leu, Val, Ile, Met, Phe, and Trp), tominimize solubility problems; and (6) they do not contain Cys

Figure 1. VEGFR-1 dII mediates cell adhesion and interaction with �5�1integrin. (A) In a solid-phase binding assay, plates coated with VEGFR-1/Fc weretreated with anti–VEGFR-1 antibodies and incubated with purified �5�1 integrin.The amount of bound integrin was quantified by incubation with an anti-�5�1antibody and colorimetric detection of this antibody. (B) HUVECs were incubatedon wells coated with recombinant dII. VEGFR-1/Fc and fibronectin (FN) wereused as positive controls, bovine serum albumin (BSA) as a negative control. Therelative number of attached cells was assessed by staining with crystal violet anddetermining A540 1 hour after plating. Absorbance resulting from nonspecific celladhesion was measured on BSA-coated wells. (C) Purified VEGFR-1/Fc or dIIwere added at the indicated concentrations to wells coated with �5�1 integrin.Bound molecules were detected using either an anti-Fc alkaline phosphatase(AP)-conjugate or an anti–VEGFR-1 antibody and a secondary anti-rabbitantibody AP-conjugate. Representative experiments performed in triplicate areshown; data are mean plus or minus SEM.

A VEGFR-1 PROANGIOGENIC PEPTIDE 3481BLOOD, 1 APRIL 2008 � VOLUME 111, NUMBER 7




residues, which might lead to formation of dimeric peptides. Tosatisfy the last requirement, VEGFR-1 residue Cys158, which isinvolved in the formation of a disulfide bond within dII structures,was replaced by Ala in peptide 4. In addition, to explore whetherthe presence of 2 branched hydrophobic residues next to each othermight cause solubility problems, the adjacent Ile145 and Ile146were both replaced by Thr in peptide 3. The sequence of peptide 5has been previously published and reported to act as an angiogen-esis inhibitor without binding to or inhibiting VEGF-A binding toits receptors.24 The location of the different peptides on dIIstructure is reported in Figure 2A. To test the ability of the designedpeptides to inhibit EC adhesion to VEGFR-1, each peptide wasadded to a HUVEC suspension before cell platingon VEGFR-1/Fc coated wells. As shown in Figure 2B, peptides4, 5, 7, and 12 effectively blocked EC adhesion to VEGFR-1/Fc.Moreover, peptides 4, 7, and 12 were able to partially inhibit ECadhesion to fibronectin, the canonical �5�1 integrin ligand,suggesting that they might interfere with ligand binding sites of�5�1 integrin. However, when linked to microtiter plates, onlypeptide 12 directly supported EC adhesion (Figure 2C).

EC adhesion on peptide 12 was completely abolished bytreatment with either EDTA or EGTA (Figure 3A), indicating thatthis interaction required divalent cations, as it is generally the casefor integrin binding to extracellular matrix proteins.18 To demon-strate whether �5�1 integrin was indeed the protein mediating ECadhesion to peptide 12, we tested the effect of integrin-blockingantibodies. Preincubation of HUVECs with a mAb against �5�1integrin inhibited cell adhesion to peptide 12, whereas a mAbagainst a different integrin subunit (�6) was ineffective (Figure3A). In particular, when an anti-�5 subunit antibody was used,complete inhibition of EC adhesion was observed (Figure 3A).

Unligated integrins are diffusely distributed in the membrane,but after ligand binding they are activated and form aggregates.25

To examine whether peptide 12 was able to induce �5�1 integrinclustering on binding, adherent HUVECs were incubated withmagnetic beads coated with peptide 12, or with RGD or RGEpeptides as positive and negative controls, respectively, andimmunostained for �5�1 integrin (Figure 3B). A higher number ofRGD-coated (Figure 3Bii) than of peptide 12-coated beads (Figure3Biii) adhered to HUVECs, probably because of their interactionwith other RGD-binding proteins on the cell surface. Indeed, adistinct staining for �5�1 was only observed in approximately 66%of RGD-coated beads (Figure 3Biv), supporting the idea thatdifferent ligands were involved in cell binding of these RGD-coated beads. On the other hand, 98% of peptide 12-coated beads

(Figure 3Bv) were immunostained with the anti-�5�1 integrinantibody, indicating that �5�1 integrin is the main cellular ligandfor this peptide. As expected, no RGE-coated beads attached toHUVECs in this assay (Figure 3Bi).

Preincubation of HUVECs with peptide 12 completely abro-gated the migration response toward VEGFR-1/Fc (Figure S1), aprocess mediated by �5�1 integrin.15 The inhibitory effect ofpeptide 12 on cell migration was specific for VEGFR-1 because nosignificant inhibition of fibronectin-induced migration was ob-served (Figure S1). On the other hand, peptide 12 did not affect cellproliferation (data not shown).

Table 1. Amino acid sequence of peptides mapping on domain II ofVEGFR-1 and of the scrambled peptide 12 (sP12)

Peptide Sequence VEGFR-1 region

P1 GRPFVEMYSE 132-141

P2 YSEIPEIIH 139-147

P3 ETTHMTEGR 144-152

P4 TEGRELVIPARVT 149-161

P5 NITVTLKKFPL 164-174

P6 KKFPLD 170-175

P7 KFPLDTLIPDG 171-181

P8 DTLIPDGKRII 175-185

P9 DSRKGFIISNAT 187-198

P10 TYKEIGL 198-204

P11 EATVNGHLYKT 208-218

P12 NYLTHRQ 219-225

sP12 LTQNYRH —

— indicates not applicable.

Figure 2. Peptide 12 inhibits endothelial cell adhesion to VEGFR-1 and directlysupports cell adhesion. (A) Ribbon representation of the alpha carbon trace ofVEGFR-1 dII structure determined by X-ray crystallography (PDB ID: 1flt).Peptides 1, 3, 5, 7, 9, and 11 (left panel) and 2, 4, 6, 8, 10, and 12 (right panel) arecolored red, orange, yellow, blue, purple and magenta, respectively. Residuesbelonging to both peptide 5 and 7 (left panel) or 6 and 8 (right panel) are coloredgreen. (B) Microtiter plates were coated with VEGFR-1/Fc or fibronectin (FN).HUVECs were preincubated with each peptide and then plated on the coatedwells. Data are reported as percentage of adhesion, calculated relative to thenumber of cells attached on the same substrate in the absence of the peptides.(C) Different VEGFR-1 peptides and RGD peptide, as a positive control, werecovalently linked to microtiter plates. HUVECs were then added to the coatedplates and allowed to adhere. Nonspecific cell adhesion was measured by theabsorbance in BSA-coated wells. Representative experiments performed intriplicate are shown; data are mean plus or minus SEM.





Identification of peptide 12 residues involved in integrinbinding

To assess the relative contribution given by peptide 12 residues tothe interaction with �5�1 integrin, 7 different peptides carrying asingle substitution to Ala were synthesized. As shown in Figure 4A,the ability of peptide 12 to support EC adhesion was significantlyaffected by single substitution of Leu221, His223 and, above all,Tyr220, with Ala. A “scrambled” peptide containing the sameresidues as peptide 12 in a different order (Table 1) was used as acontrol and did not sustain cell adhesion.

As a direct confirmation of peptide 12/integrin interaction, weused a solid phase binding assay with wild-type and mutatedpeptides. In this assay, Tyr220, Leu221, and His223 were con-firmed to be critical residues mediating peptide binding to purified�5�1 integrin (Figure 4B). In addition, Arg224 was also shown toplay an important role. The specific contribution of Arg224 insustaining peptide binding to purified integrin might be the result ofa different conformation of the regions involved in peptide bindingbetween purified and cell membrane integrin, possibly caused byinteractions of the latter with other membrane proteins.

No significant interference with cell attachment or integrinbinding was observed after substitution of Gln225, a VEGFR-1residue known to be involved in VEGF-A binding.26 However,

because other residues of peptide 12 are in contact with VEGF-A inthe known structures of complexes with VEGFR-1, we analyzedwhether peptide 12 was able to directly interact with VEGF-A. We alsotested peptide 7, containing 6 residues of VEGFR-1 that interact with thegrowth factor, and peptide 1, mapping to a receptor region not involvedin interactions with VEGF-A. The VEGFR-1/Fc protein was used as apositive control. As shown in Figure 4C, VEGF-A did not bind topeptide 12, 7, or 1, indicating that none of these peptides was sufficientto mediate VEGF-A binding.

Peptide 12 supports in vitro and in vivo angiogenesis

The ability of peptide 12 to induce EC tubule formation was testedon a collagen I matrix and found to be similar to that of VEGF-A(Figure 5B,C, respectively) compared with unstimulated mono-layer (Figure 5A). The Q225A peptide, which binds to �5�1integrin and supports EC adhesion-like peptide 12, also induced ECtubule formation (Figure 5D). Conversely, treatment with theY220A peptide, which does not bind �5�1 integrin, did not inducevessel shaping (Figure 5E). As a control of specificity, tubuleformation assay was performed using peptide 12, previouslyincubated with an antibody specifically recognizing its amino acidsequence. In this case, formation of capillary structures was notobserved (Figure 5F).

Figure 3. �5�1 integrin mediates the biologic activ-ity of peptide 12. (A) Peptide 12 or BSA, as anonspecific adhesion control, was coated onto microti-ter plates, and HUVECs were added in the presence ofchelating agents (EDTA or EGTA). Alternatively,HUVECs were preincubated with an antibody against�5�1 integrin, against �5 or �1 integrin subunits, oragainst �6 integrin subunit and allowed to adhere oncoated wells. Data are mean plus or minus SEM; arepresentative experiment performed in triplicate isshown. (B) Fibronectin-adherent HUVECs wereincubated with magnetic beads coated with RGEpeptide (i), RGD peptide (ii,iv), or peptide 12 (iii,v).HUVECs were then immunostained for �5�1 integrin.Integrin clusters around the beads are indicated (arrow-heads). Bars represent 10 �m.





To evaluate a possible role of VEGF-A in peptide 12-mediatedvessel shaping, tubule formation assay was also performed in thepresence of function blocking antibodies against VEGFR-1 or ofVEGFR-1/Fc chimera. In these experiments, addition of peptide 12partially restored EC growth and vessel shaping (Figure S2),suggesting that this peptide acted through a mechanism parallel tothat of VEGF-A.

To address the significance of these findings in vivo, pelletsreleasing peptide 12 were implanted into avascular rabbit corneasand neovascularization was monitored for 17 days. As shown inFigure 6A, peptide 12 (100 �g/pellet) strongly promoted angiogen-esis. Extent, growth rate, and pattern of capillary growth, reported

as angiogenic score in Figure 6B, overlapped with that obtainedwith VEGF-A (400 ng/pellet). Consistent with the in vitro data, thecontrol Y220A peptide lacked angiogenic activity. Results ob-tained with 2 different concentrations of peptide 12, 10 �g and100 �g/pellet, indicated that angiogenesis was sustained in adose-dependent manner (Table 2). Histologic and immunohisto-chemical analysis of the corneal tissues are shown in Figure S3.

The ability of peptide 12 to induce angiogenesis in the rabbitcornea neovascularization assay was also tested in the presence ofVEGF-A, resulting in additional angiogenesis (Figure S4). Tofurther evaluate the relevance of VEGF-A signaling on peptide12-induced angiogenesis in vivo, assays were also performed in thepresence of bevacizumab, a recombinant humanized monoclonalantibody that prevents VEGF-A interaction with its tmVEGFRs, orsunitinib, a tyrosine kinase inhibitor also acting on VEGFRs.27 Theangiogenic effect of peptide 12 was significantly inhibited by bothbevacizumab and sunitinib treatment (Figure S5), indicating thatVEGF-A signaling played a role in peptide 12-induced cornealneovascularization. Similar inhibition was obtained on VEGF-A-induced angiogenesis (Figure S5).

Peptide 12 does not activate FAK signaling pathway

The mechanism of action of peptide 12 through �5�1 integrinactivation was analyzed. No phosphorylation of integrin�1 subunit could be observed after HUVEC adhesion on peptide12 (data not shown).

Figure 4. Alanine substitution in peptide 12. (A) Peptides whose sequence differsfrom that of peptide 12 for the substitution of one residue with Ala were covalentlylinked to microtiter plates. A “scrambled” version of peptide 12 (sP12) and BSA, as anegative control, were also tested. HUVECs were added to the plates and allowed toadhere. Results are expressed as percentage of cell adhesion observed on peptide12-coated plates. (B) Purified �5�1 integrin was added to wells coated with the samepeptides as in panel A, and bound integrin was detected by using an anti-�5�1antibody and a colorimetric assay. Data are expressed as percentage of proteinbinding on peptide 12-coated plates. (C) Biotinylated VEGF-A was added to wells coatedwith peptide 12, 1, 7, VEGFR-1/Fc, as a positive control, or BSA, as a negative control. Theamount of attached VEGF-A was quantified by incubation with streptavidin alkalinephosphatase-conjugated and a colorimetric assay. Representative experiments performedin triplicate are shown; data are mean plus or minus SEM.

Figure 5. Peptide 12 sustains vessel formation in vitro. HUVECs were incubatedon 24-well plates precoated with a type I collagen gel mixture, in the absence (A) orpresence of peptide 12 (B), peptide Q225A (D), or peptide Y220A (E). VEGF-A wasused as a reference control (C). To assess the specificity of peptide 12 activity,an antibody against peptide 12 was also tested in combination with peptide12 (F). Results are representative of 3 independent experiments. Bars represent25 �m in C, 50 �m in the others.





After binding to fibronectin, �5�1 integrin activates 2 majortyrosine kinase-dependent pathways, the focal adhesion kinase(FAK)28 pathway and the adaptor protein Shc28 one. To analyzesignals activated by EC adhesion on peptide 12, HUVECs wereplated on peptide 12- or fibronectin-coated dishes for 30 minutes or1 hour in the presence or absence of VEGF-A. As shown in Figure7A,B, adhesion on peptide 12 did not promote FAK phosphoryla-tion and did not result in significant extracellular signal-regulatedkinase (ERK1/2) phosphorylation, the latter being a downstreameffector of both integrins and growth factor receptors. ERK1/2phosphorylation was observed in EC adhering on peptide 12 in thepresence of VEGF-A. On the other hand, phosphorylation of Shccould be observed in HUVECs adhering on fibronectin and onpeptide 12 in the absence of exogenous VEGF-A (Figure 7C).

Formation of membrane macromolecular complexes betweenintegrins and growth factor receptors has been previously high-lighted.29 In particular, a recent study on acute myeloid leukemiacell lines revealed an association between tmVEGFR-1, integrin�1, and human ERG K� channel.30 We investigated whether also inHUVECs �1 integrin could form a complex with tmVEGFR-1, butcoimmunoprecipitation analysis could not reveal any interaction

(Figure S6A). Moreover, expression of the mRNA coding forhuman ERG-1 protein was not observed in HUVECs (Figure S6B).

Discussion

Soluble VEGFR-1 can play 2 opposite biologic roles. It can actboth as a negative regulator of VEGF-related growth factorsignaling13 and as an extracellular matrix protein involved inendothelial cell adhesion and migration.15 To completely under-stand the angiogenesis and vasculogenesis processes, it is ofprimary importance to define when and where this moleculebehaves as a positive or a negative regulator of angiogenesis.

We previously reported that the involvement of sVEGFR-1 incell adhesion and migration relies on its interaction with �5�1integrin. In this study, we show that a peptide, isolated fromVEGFR-1, is able to bind and activate �5�1 integrin.

First, we demonstrated that Ig-like dII of sVEGFR-1 is involvedin the interaction with the integrin. To identify dII regionsmediating this interaction, we tested 12 different peptides mimick-ing most of the dII surface and found that 4 of them inhibited celladhesion to sVEGFR-1. Of these, peptides 4 and 12 are situated onopposite sides of the domain, each on one of the 2 �-sheets packedface-to-face, whereas peptide 7 lies on a plane roughly perpendicu-lar to the 2 �-sheets. Only peptide 5 is relatively close to bothpeptide 7, partially overlapping with it, and peptide 12, which issituated on the same �-sheet. The relatively distant position ofthese peptides suggests that their interference with cell adhesion tosVEGFR-1 may be achieved through interaction with differentbinding partners and diverse mechanisms. Accordingly, peptide 12is the only one in our series that directly supported EC adhesion,providing us with the first indication that this peptide may directlyinteract with �5�1 integrin. EC binding to peptide 12 was divalentcation-sensitive and antibodies against �5�1 integrin specificallyblocked this interaction, supporting the assumption that EC adhe-sion to peptide 12 was �5�1 integrin-mediated. In particular, ourresults suggest that the �5 subunit gives the most relevantcontribution to peptide 12 binding. Moreover, �5�1 integrinclustering was observed with peptide 12-coated beads, indicatingthat a direct interaction between peptide 12 and integrin did occurleading to integrin activation. Finally, peptide 12 inhibited cellmigration toward VEGFR-1, a process that is mediated by�5�1 integrin.15

In addition to these indirect observations, we also detected adirect interaction between �5�1 integrin and peptide 12 in a solidphase binding assay. Alanine-scanning mutagenesis identified4 residues of peptide 12, namely, Tyr220 (giving the highestcontribution), Leu221, His223, and Arg224, which, if mutated,significantly reduced peptide ability to support direct interactionwith �5�1 integrin and, with the exception of Arg224, cell

Figure 6. Peptide 12 sustains neoangiogenesis in vivo. (A) The effect of100 �g/pellet of peptide 12 or Y220A peptide was compared with that of 400 ng/pelletVEGF-A. *Pellet implants. (B) Quantification of the data (means � SEM) shown inpanel A reported as angiogenic score during time (days). Numbers are the means of6 implants for each experimental point. One-way analysis of variance (P � .001).*P � .05 peptide 12 versus Y220Apeptide (Student-Newman-Keuls multiple compari-son test).

Table 2. Peptide 12 promotes in vivo angiogenesis in a dose-dependent manner

10 �g/pellet 100 �g/pellet

Peptide 12 2.0 � 0.5* 8.2 � 2.0*

Peptide 12 Y220A 0.3 � 0.3 1.4 � 0.8

The effect of 10 and 100 �g/pellet of peptide 12 or Y220A peptide was tested inthe rabbit cornea assay. Data (mean � SEM) are reported as angiogenic score at day17 and are the means of 6 implants for each experimental point. (One-way analysis ofvariance P � .001.)

*P � .05 peptide 12 versus Y220A peptide (Student-Newman-Keuls multiplecomparison test).





adhesion. This YLXHR motif shows some similarities but alsonoticeable differences with previously described �5�1 integrinbinding sequences.18,19 Similarly to the RGD motif, the fibronectinPHSRN and VKNEED synergy sites, and the majority of peptidesequences identified from phage library screenings,31,32 the YLXHRmotif bears a positive charge, which can be contributed either bythe Arg or His residue. Moreover, similarly to many peptidesequences selected from phage libraries, it contains an aromaticresidue (Tyr) separated by 2 or 3 amino acids from the positivelycharged one. However, this motif differs from those previouslyidentified in the position of the aromatic and positive residues(RXXXW being the most typical arrangement) and in the roleplayed by the hydrophobic residue Leu. In addition, this is the firstmotif for which an aromatic residue, rather than a charged one,plays the most important role for �5�1 integrin binding.

We demonstrated that peptide 12 induces tubule-like structuresformation in an in vitro angiogenic assay and supports neoangiogen-esis in vivo.

Interestingly, addition of peptide 12 in the in vitro vessel formationassay circumvents the effects of either sVEGFR-1 or anti–VEGFR-1function-blocking antibodies, partially restoring EC survival and tubulestructuring. These data suggest that peptide 12 activates a signal

pathway parallel to that of VEGF-A. Accordingly, peptide 12 enhancesthe angiogenic effect of VEGF-A in vivo.

Peptide 12 plausibly exerts these proangiogenic propertiesthrough integrin binding and activation of an �5�1-mediatedsignaling pathway because this peptide is not able to directly bindVEGF-A, its tyrosine kinase receptors, or other molecules involvedin VEGFR signaling, such as neuropilins (A.O., P.M.L., C.M.F.,unpublished data). We analyzed whether �5�1 binding to peptide12 induced FAK phosphorylation, the main signaling mechanismof �5�1 integrin after fibronectin binding, but no FAK phosphory-lation could be observed. At the same time point, ERK phosphory-lation was detected in EC adhering on peptide 12 only in thepresence of VEGF-A. On the other hand, we could show a slightphosphorylation of the adaptor protein Shc28 in the absence ofexogenous VEGF-A. It remains to be clarified whether Shcphosphorylation induced by peptide 12 requires further VEGF-induced signaling to result in ERK phosphorylation.

Interestingly, peptide 12 angiogenic activity in vivo was signifi-cantly impaired by both bevacizumab, a monoclonal antibody thatblocks VEGF-A binding to its receptors, and sunitinib, a tyrosinekinase inhibitor acting also on VEGFRs. These results suggest thatpeptide 12 interactions with �5�1 integrin require simultaneous

Figure 7. Effect of peptide 12 on �5�1 integrin-mediated down-stream signaling. HUVECs were starved and treated with cyclohexi-mide as described in “Western blotting analysis” and then plated onpeptide 12-, or fibronectin- (FN)-coated dishes in the presence orabsence of VEGF-A. (A) Immunoblotting analysis for the presence ofphosphorylated-FAK or phosphorylated-ERK1/2 polypeptides after1-hour adhesion. Actin staining was performed to normalize the results(ACT). (B) Histograms showing the relative OD of the bands in panel A,normalized using absorbance of the actin bands for each sample, aftersubtraction of the background (phosphorylation levels in cells plated onBSA-coated dishes). (C) Immunoblotting analysis of phosphorylated-Shcand total Shc polypeptide after 30 minutes of adhesion. The histogramshows the OD ratio between the phosphorylated band and the totalprotein. Data represent the results of a representative experiment.





VEGF-A activation of its transmembrane receptors to give neoan-giogenesis. Alternatively, peptide 12 requires “active” endothelialcells, as those stimulated by VEGF-A, to operate through �5�1integrin. It should be consider that VEGF-A elicits differentfunctions in endothelium homeostasis, several of which we are justbeginning to understand.33

It has been previously reported that cross talk between integrinsand growth factors represents an important modulator of vasculo-genesis and angiogenesis.29 Significant enhancement of the angio-genic response can be specifically observed when both �5�1integrin and VEGF-A receptors are engaged.34 In this perspective,we also investigated whether �5�1 integrin could directly interactwith tmVEGFR-1, but we did not detect any interaction. Moreover,previous studies could not detect the formation of a macromolecu-lar complex involving �5�1 integrin and VEGFR-2.35 Neverthe-less, an additional transmembrane receptor could participate in theformation of the complex between �5�1 integrin and VEGFRs,and further investigation is needed to clarify this aspect.

Altogether, the properties of peptide 12, located within thesVEGFR-1 sequence, support the hypothesis that sVEGFR-1 itselfis implicated in vessel shaping and angiogenesis as well, throughits interaction with �5�1 integrin. Interestingly, sVEGFR-1 hasbeen shown to act as a positive modulator of vascular sproutformation and vessel morphogenesis.36 The addition of VEGFR-1/Fc protein to in vitro cultures of VEGFR-1–lacking embryonicstem cells was also shown to partially rescue the mutant phenotype(ie, vascular overgrowth and disorganization).37 In those studies,sVEGFR-1 has been proposed to act on vascular structuring throughmodulation of VEGF-A receptor signaling by sequestration and neutral-ization of VEGF-A. Our results suggest that additional pathwaysrequired for vessel formation may be regulated through interactionbetween sVEGFR-1 and �5�1 integrin.

In our in vitro angiogenesis assay, addition of sVEGFR-1impaired vessel structure formation and EC growth is in agreementwith the concept that, in certain circumstances, sVEGFR-1 prefer-entially acts as a decoy receptor for VEGF growth factor family andnot as an activator of �5�1 integrin.38,39 It is possible that the location ofsVEGFR-1 within the extracellular matrix favors its interaction with�5�1 integrin over growth factors. It is also possible that matrix-associated proteases process sVEGFR-1 producing matrix-embedded

smaller polypeptides not able to bind the growth factors anymore butstill capable of interacting with �5�1 integrin.

The identification of a �5�1 integrin-binding small peptideable to activate a proangiogenic signaling in EC provides arational basis for the development of new therapeutic tools.Molecules able to sustain neoangiogenesis have been success-fully used for the treatment of limb ischemia by inducingcollateral vessel formation, and for tissue repair in pathologiesassociated with impaired wound healing, such as diabetes.40 Ourresults suggest that peptide 12 is a promising candidate for thedesign of additional proangiogenic drugs.

Acknowledgments

The authors thank Dr Shibuya for the gift of the human VEGFR-1cDNA plasmid, Dr Ragone for acquisition of the confocal micro-scope pictures, Dr Avitabile for the analysis of Shc phosphoryla-tion, Prof Defilippi and Prof Tarone for discussion of phosphoryla-tion assays, and Mr Inzillo for artwork.

This work was supported by Compagnia di San Paolo, Prat23 734/Ente 11 587, and the European Community (LSHB-CT-2005 512102 and LSHG-CT-2003-503265), the National ResearchCouncil (grant SV.P08.009.003 to V.M.), AIRC Regional Grant2005 (L.M.), and the Fondazione Monte dei Paschi di Siena (M.Z.).

Authorship

Contribution: S.S., A.O., L.M. designed and performed research;P.M.L. and V.M. designed the experiments, assessed the results,and revised the manuscript; K.B.-H. provided fundamental re-agents and technology and revised the manuscript; F.R. performedthe experiments; M.Z., S.D., G.Z. analyzed data and revised themanuscript; A.T. and C.M.F. designed research, assessed theresults, and wrote the paper.

Conflict-of-interest disclosure: A patent related to this peptidehas been deposited. The authors declare no other competingfinancial interests.

Correspondence: Cristina M. Failla, Molecular and Cell Biol-ogy Laboratory, IDI-IRCCS, via Monti di Creta, 104, 00167 Rome,Italy; e-mail: [email protected].

References

1. Korpelainen E, Alitalo K. Signaling angiogenesisand lymphangiogenesis. Curr Opin Cell Biol.1998;10:159-164.

2. Shibuya M, Yamaguchi S, Yamane A, et al. Nu-cleotide sequence and expression of a novel hu-man receptor-type tyrosine kinase (flt) closelyrelated to the fms family. Oncogene. 1990;8:519-527.

3. Wiesmann C, Fuh G, Christinger HW, EigenbrotC, Wells JA, de Vos AM. Crystal structure at 1.7 Aresolution of VEGF in complex with domain 2 ofthe Flt-1 receptor. Cell. 1997;91:695-704.

4. Christinger HW, Fuh G, de Vos AM, Wiesmann C.The crystal structure of placental growth factor incomplex with domain 2 of vascular endothelialgrowth factor receptor-1. J Biol Chem. 2004;279:10382-10388.

5. Fuh G, Garcia KC, de Vos AM. The interaction ofneuropilin-1 with vascular endothelial growth fac-tor and its receptor Flt-1. J Biol Chem. 2000;275:26690-26695.

6. Carmeliet P, Moons L, Luttun A, et al. Synergismbetween vascular endothelial growth factor andplacental growth factor contributes to angiogen-

esis and plasma extravasation in pathologic con-ditions. Nat Med. 2001;7:575-583.

7. Sawano A, Iwai S, Sakurai Y, et al. Flt-1, vascularendothelial receptor 1, is a novel cell surfacemarker for the lineage of monocyte-macrophagesin humans. Blood. 2001;97:785-791.

8. Lyden D, Hattori K, Dias S, et al. Impaired recruit-ment of bone-marrow-derived endothelial andhematopoietic precursor cells blocks tumor an-giogenesis and growth. Nat Med. 2001;7:1194-1201.

9. Luttun A, Tjwa M, Moons L, et al. Revasculariza-tion of ischemic tissues by PlGF treatment, andinhibition of tumor angiogenesis, arthritis and ath-erosclerosis by anti-Flt-1. Nat Med. 2002;8:831-840.

10. Fong G, Rossant J, Gertsenstein M, Breitman M.Role of the Flt-1 receptor tyrosine kinase in regu-lating the assembly of vascular endothelium. Na-ture. 1995;376:66-70.

11. Fong G, Zhang L, Bryce D, Peng J. Increasedhemangioblast commitment, not vascular disor-ganization, is the primary defect in flt-1 knock-outmice. Development. 1999;126:3015-3025.

12. Hiratsuka S, Minowa O, Kuno J, Noda T, ShibuyaM. Flt-1 lacking the tyrosine kinase domain is suf-ficient for normal development and angiogenesisin mice. Proc Natl Acad Sci U S A. 1998;95:9349-9354.

13. Kendall RL, Thomas KA. Inhibition of vascularendothelial cell growth factor activity by an endo-genously encoded soluble receptor. Proc NatlAcad Sci U S A. 1993;90:10705-10709.

14. Inoue T, Kibata K, Suzuki M, Nakamura S, Mo-toda R, Orita K. Identification of a vascular endo-thelial growth factor (VEGF) antagonist, sFlt-1,from a human hematopoietic cell line NALM-16.FEBS Lett. 2000;469:14-18.

15. Orecchia A, Lacal PM, Schietroma C, Morea V,Zambruno G, Failla CM. Vascular endothelialgrowth factor receptor-1 is deposited in the extra-cellular matrix by endothelial cells and is a ligandfor the �5�1 integrin. J Cell Sci. 2003;116:3479-3489.

16. Francis SE, Goh KL, Hodivala-Dilke K, et al. Cen-tral role of alpha5beta1 integrin and fibronectin invascular development in mouse embryos and





embryoid bodies. Arterioscler Thromb Vasc Biol.2002;22:927-933.

17. Kim S, Bell K, Mausa SA, Varner JA. Regulationof angiogenesis in vivo by ligation of integrin�5�1 with the central cell-binding domain of fi-bronectin. Am J Pathol. 2000;156:1345-1362.

18. Humphries MJ. Integrin structure. Biochem SocTrans. 2000;28:311-339.

19. Shimaoka M, Springer TA. Therapeutic antago-nists and conformational regulation of integrinfunction. Nat Rev Drug Discov. 2003;2:703-716.

20. Leu S, Lam S, Lau L. Pro-angiogenic activities ofCYR61 (CCN1) mediated through integrins al-phavbeta3 and alpha6beta1 in human umbilicalvein endothelial cells. J Biol Chem. 2002;277:46248-46255.

21. Morbidelli L, Ziche M. The rabbit corneal pocketassay for the study of angiogenesis. Cancer TreatRes. 2004;117:147-151.

22. Berman HM, Westbrook J, Feng Z, et al. The Pro-tein Data Bank. Nucleic Acids Res. 2000;28:235-242.

23. Dayringer HE, Tramontano A, Spang SR, Fletter-ick RJ. Interactive program for visualization andmodelling of proteins, nucleic acids and smallmolecules. J Mol Graph. 1986;4:82-87.

24. Tan DCW, Manjunatha Kini R, Jois SDS, LimDKF, Xin L, Ge R. A small peptide derived fromFlt-1 (VEGFR-1) functions as an angiogenic in-hibitor. FEBS Lett. 2001;494:150-156.

25. Yamada KM, Geiger B. Molecular interactions in

cell adhesion complexes. Curr Opin Cell Biol.1997;9:76-85.

26. Davis-Smyth T, Presta LG, Ferrara N. Mappingthe charged residues in the second immuno-globulin-like domain of the vascular endothelialgrowth factor/placenta growth factor receptorFlt-1 required for binding and structural stability.J Biol Chem. 1998;273:3216-3222.

27. Roskoski RJ. Sunitinib: a VEGF and PDGF re-ceptor protein kinase and angiogenesis inhibitor.Biochem Biophys Res Commun. 2007;356:323-328.

28. Barberis L, Wary KK, Fiucci G, et al. Distinct rolesof the adaptor protein Shc and focal adhesionkinase in integrin signaling to ERK. J Biol Chem.2000;275:36532-36540.

29. Eliceri BP. Integrin and growth factor receptorcrosstalk. Circ Res. 2001;89:1104-1110.

30. Pillozzi S, Brizzi MF, Bernabei PA, et al. VEGFR-1(FLT-1), �1 integrin, and hERG k� channel for amacromolecular signaling complex in acute my-eloid leukemia: role in cell migration and clinicaloutcome. Blood. 2007;110:1238-1250.

31. Koivunen E, Gay DA, Ruoslahti E. Selection ofpeptides binding to the �5�1 integrin from phagedisplay library. J Biol Chem. 1993;268:20205-20210.

32. Koivunen E, Wang B, Ruoslahti E. Isolation of ahighly specific ligand for the alpha5 beta1 integrinfrom a phage display library. J Cell Biol. 1994;124:373-380.

33. Lee S, Chen TT, Barber CL, et al. AutocrineVEGF signaling is required for vascular ho-meostasis. Cell. 2007;130:691-703.

34. Wijelath ES, Rahman S, Namekata M, et al. Hep-arin-II domain of fibronectin is a vascular endo-thelial growth factor-binding domain. Circ Res.2006;99:853-860.

35. Soldi R, Mitola S, Stasly M, Defilippi P, Tarone G,Bussolino F. Role of �v�3 integrin in the activa-tion of vascular endothelial growth factor recep-tor-2. EMBO J. 1999;18:882-892.

36. Kearney JB, Kappas NC, Ellerstrom C, DiPaolaFW, Bautch VL. The VEGF receptor flt-1(VEGFR-1) is a positive modulator of vascularsprout formation and branching morphogenesis.Blood. 2004;103:4527-4535.

37. Roberts DM, Kearney JB, Johnson JH, Rosen-berg MP, Kumar R, Bautch VL. The VEGF recep-tor flt-1 (VEGFR-1) modulates flk-1 (VEGFR-2)signaling during blood vessel formation. Am JPathol. 2004;164:1531-1535.

38. Kong HL, Hecht D, Song W, et al. Regional sup-pression of tumor growth by in vivo transfer of acDNA encoding a secreted form of the extracellu-lar domain of the flt-1 vascular endothelial growthfactor receptor. Hum Gene Ther. 1998;9:823-833.

39. Ambati BK, Nozaki M, Singh N, et al. Cornealavascularity is due to soluble VEGF receptor-1.Nature. 2006;443:993-997.

40. Folkman J. Clinical applications of research onangiogenesis. N Engl J Med. 1995;333:1757-1763.





online January 9, 2008 originally publisheddoi:10.1182/blood-2007-03-077537

2008 111: 3479-3488

Tramontano and Cristina Maria FaillaBallmer-Hofer, Federica Ruffini, Marina Ziche, Stefania D'Atri, Giovanna Zambruno, Anna Simonetta Soro, Angela Orecchia, Lucia Morbidelli, Pedro Miguel Lacal, Veronica Morea, Kurt

1 integrinβ5αfactor receptor-1 acts through A proangiogenic peptide derived from vascular endothelial growth

http://www.bloodjournal.org/content/111/7/3479.full.htmlUpdated information and services can be found at:

(2485 articles)Hemostasis, Thrombosis, and Vascular Biology Articles on similar topics can be found in the following Blood collections

http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requestsInformation about reproducing this article in parts or in its entirety may be found online at:

http://www.bloodjournal.org/site/misc/rights.xhtml#reprintsInformation about ordering reprints may be found online at:

http://www.bloodjournal.org/site/subscriptions/index.xhtmlInformation about subscriptions and ASH membership may be found online at:

Copyright 2011 by The American Society of Hematology; all rights reserved.of Hematology, 2021 L St, NW, Suite 900, Washington DC 20036.Blood (print ISSN 0006-4971, online ISSN 1528-0020), is published weekly by the American Society


http://www.bloodjournal.org/content/111/7/3479.full.html

http://www.bloodjournal.org/cgi/collection/hemostasis_thrombosis_and_vascular_biology

http://www.bloodjournal.org/site/misc/rights.xhtml#repub_requests

http://www.bloodjournal.org/site/misc/rights.xhtml#reprints

http://www.bloodjournal.org/site/subscriptions/index.xhtml





aproangiogenicpeptidederivedfromvascularendothelialgrowthf ... · hemostasis, thrombosis,and...

Documents