E X P E R I M E N T A L C E L L R E S E A R C H 3 1 6 ( 2 0 1 0 ) 9 0 7 – 9 1 4

ava i l ab l e a t www.sc i enced i r ec t . com

www.e l sev i e r . com/ loca te /yexc r

Research Article

Autocrine effects of VEGF-D on endothelial cells aftertransduction with AD-VEGF-DΔNΔC

Izabela Papiewska-Pajaka, Joanna Boncelab, Patrycja Przygodzkab, Czeslaw S. Cierniewskia,b,⁎aDepartment of Molecular and Medical Biophysics, Medical University of Lodz, PolandbInstitute of Medical Biology, Polish Academy of Science, Lodz, Poland

A R T I C L E I N F O R M A T I O N

⁎ Corresponding author. Department of MolecPoland. Fax: +48 42 6789433.

E-mail address: cciern@zdn.am.lodz.pl (C.S.

0014-4827/$ – see front matter © 2010 Elseviedoi:10.1016/j.yexcr.2010.01.014

A B S T R A C T

Article Chronology:

Received 5 May 2009Revised version received27 November 2009Accepted 8 January 2010Available online 22 January 2010

Endothelial cells in tumor vessels display unusual characteristics in terms of survival andangiogenic properties which result from the increased expression of VEGF-D and its autocrineeffect. To evaluate mechanisms by which VEGF-D leads to such abnormal phenotype, we searchedfor proteins with modified expression in HUVECs enriched in the recombinant mature VEGF-D(VEGFDΔNΔC) delivered by adenovirus. Expression of membrane proteins in endothelial cells wascharacterized by FACS using anti-human IT-Box-135 antibodies. HUVECs transduced with Ad-VEGF-DΔNΔC revealed markedly increased expression of proteins involved in adhesion andmigration such as (a) integrins (αVβ5, α2β1, α5β1, αMβ2, αLβ2), (b) matrix metalloproteinases(MMP-2, MMP-9, and MMP-14), (c) components of fibrinolytic system (PAI-1, u-PAR), and (d)

CD45, CD98, CD147. Interestingly, there also were numerous proteins with significantly reducedexpression, particularly among surface exposedmembrane proteins. Thus, it can be concluded thatto induce proangiogenic phenotype and facilitate migration of HUVECs, VEGF-DΔNΔC not onlyupregulates expression of proteins known to participate in the cell-matrix interactions but alsosilences some membrane proteins which could interfere with this process.

© 2010 Elsevier Inc. All rights reserved.

Keywords:

Endothelial cellsVEGF-DIntegrins

ProteomicsMigration assay

Introduction

Vascular endothelial growth factors (VEGF-C and VEGF-D) withintumors simulate endothelial cells to grow and generate newlymphatics and new blood vessels [1,2]. Like VEGF-C, VEGF-D isinitially synthesized as a disulfide-linked prepropeptide contain-ing N- and C-terminal extensions, not found in other VEGFpolypeptides, flanking a central receptor-binding VEGF homologydomain [3]. When it is not processed, it has low affinity for VEGFR-2 but preferentially binds to VEGFR-3 that signals for lymphan-giogenesis [4]. After cleavage of the N- and C-terminal domains, afully processed mature molecule is released. It forms noncovalent

ular and Medical Biophysi

Cierniewski).

r Inc. All rights reserved.

homodimers that bind with greatly increased affinity to VEGFR-2which signals for angiogenesis [5].

VEGF-D has been shown to be mitogenic for endothelial cellsand to promote angiogenesis in in vitro and in vivo models [6–8].When overexpressed in skin keratinocytes and tumors, itstimulates lymphoangiogenesis in a diversity of mouse tumormodels [9–12] and induces Akt activation, cell survival, andmigration in lymphatic endothelial cells [13]. Although VEGF-Dand VEGF-A exhibit some overlap in their endothelial biologicaleffects, they have marked differential effects on VEGFR-2 phos-phorylation, VEGFR-2-mediated signaling, and on a range ofbiological responses including PGI2 production, eNOS phosphor-

cs, Medical University of Lodz, 6/8 Mazowiecka Street, 92-215 Lodz,

908 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 6 ( 2 0 1 0 ) 9 0 7 – 9 1 4

ylation, gene expression, proliferation, survival, migration, tubu-logenesis, and angiogenesis in vivo [14].

Normally, endothelial cells do not synthesize VEGF-D but undersome conditions they show increased expression of this cytokine.This is the case, for example, (a) in tumor vessels in whichendothelium expresses VEGF-D and possesses distinct and uniquephenotype maintained by the functional autocrine pathwayrelated to this cytokine [1]. Tumor-derived endothelial cells, incontrast to normal endothelial cells, are resistant to apoptosis anddo not undergo senescence, show constant expression of markersof endothelial activation and angiogenesis, are proadhesive forrenal carcinoma cells, and able to grow and organize in the absenceof serum in persistent capillary-like structures [2]. (b) There arepotential clinical approaches currently being tested in the settingof cardiovascular diseases that involve the use of adenovirusesexpressing VEGF-D—these adenoviruses will undoubtedly infectendothelial cells in vivo. Such adenoviruses were reported to havetherapeutic potential due to inducing angiogenesis in varioustissues [15,16] and found to be useful in the catheter-mediatedapproach for the treatment of restenosis [17,18]. Therefore, itwould be beneficial to model the likely effects of such agents onthe endothelium.

In the present study, we attempted to characterize a role of theautocrine pathway related to VEGF-D in abnormal phenotype ofendothelial cells. For this purpose, we analyzed expression of theselected proteins in endothelial cells that were enriched inrecombinant mature VEGF-D (VEGF-DΔNΔC) delivered byadenovirus.

Materials and methods

Materials

HUVECs and Medium 200 were purchased from HUVEC CascadeBiologics. All cell culture media and recombinant human VEGF165were from Sigma. EGF was from Upstate, Lake Placid, NY.Adenoviruses, expressing human VEGF-D polypeptide with resi-dues from 93 to 201 corresponding to the fully processed VEGF-D(Ad-VEGF-DΔNΔC) or green fluorescence protein (Ad-GFP) werekind gifts from Dr. Tatiana Byzova (The Cleveland Clinic Founda-tion, Cleveland, USA). The interleukin 3 signal sequence was usedto drive secretion of the recombinant VEGF-DΔNΔC from cells [19].A cocktail of protease inhibitors (Complete) was from Roche.Antibodies to CD antigens (anti-human IT-Box-135), were fromImmunoTools. Antibodies to αvβ3, αvβ5, α2β1, α5β1 andsecondary antibodies anti-mouse IgG conjugated with FITC werefrom Chemicon International Inc., anti-α2 and anti-α4 from BDPharmingen, anti-α5 and anti-α6 from Serotec and anti-β1 fromDAKO.Monoclonal antibody to VEGF-Dwas from R&D Systems Inc.(Minneapolis, USA), and horseradish peroxidase-conjugated anti-murine IgG rabbit antibody or anti-rabbit IgG goat antibody wasfrom CHEMICON International.

Cell culture

HUVECs cultured in Medium 200 supplemented with Low SerumGrowth Supplement containing 2% FBS and antibiotics (strepto-mycin and penicillin) were grown in monolayer in flasks or dishesat 37 °C in 5% CO2. For experiments, HUVECs were harvested at

confluence with trypsin/EDTA and suspended in Medium 200containing 1% FBS and antibiotics. Human Embryonic Kidney Cells(HEK 293) were grown in Minimum Essential Medium (MEM)with 10% FBS, streptomycin (100 μg/ml), penicillin (100 U/ml),amphotericin B (1.5 μg/ml) and MEM nonessential amino acids.

Adenovirus purification and cell transduction

Adenoviruses were propagated in HEK293. For this purpose, flaskswith subconfluent monolayers of HEK293 in MEM containing 2%FBS were exposed to adenovirus vector until moderate cytopathiceffect was observed. Then, cell pellets were suspended in MEMwith 2% FBS, exposed to three freeze-thaw cycles and centrifugedat 4000 rpm for 10 min. Adenovirus vectors were purified using acesium chloride concentration gradient according to the standardprotocol. For experiments, HUVECs were washed with PBS andtransduced with Ad-VEGF-DΔNΔC or Ad-GFP at appropriatemultiplicities of infection (MOI). Incubation with adenoviruseswas led for 1.5 h in Medium 200 with 2% FBS, then medium waschanged into fresh one containing 1% FBS and antibiotics.Expression of VEGF-DΔNΔC in HUVECs was analyzed at the levelof mRNA and protein synthesis by RTPCR and Western immuno-blotting, respectively. In addition, secretion of VEGF-DΔNΔC fromtransduced HUVECs was evaluated by ELISA using Human VEGF-DImmunoassay kit (R&D Systems Inc., USA).

Endothelial cell migration and proliferation assays

HUVECs were grown to confluence in DMEM/F12 mediumsupplemented with 15% FCS, 150 μg/ml endothelial growth factor(Clonetics, San Diego, CA), and 90 μg/ml heparin (Sigma). Cellculture was maintained in 1% FCS for 20 h before experiments.Migration assay was conducted in Transwell® chambers (pores8 μm, Costar®). Filters of the chambers were precoated withcollagen type I or fibrinogen (15 μg/ml in PBS, pH 7.4) for 3 h at 37°C. Cells transduced with Ad- VEGF-DΔNΔC or Ad-GFP wereharvested with trypsin/EDTA and suspended in Medium 200 with0.1% BSA to the final concentration of 1×106/ml. Before placementin Transwells, 100 μl aliquots of cell suspensionswere preincubatedfor 30 min with blocking antibodies to ανβ3, α5β1 (20 μg/ml) orα2β1 (10 μg/ml). The chambers were incubated for 6 h at 37 °C in5% CO2. Then, the cells on the upper face of the filterswere removedusing a cotton swab. The filters were fixed in methanol and stainedwith Mayer's hematoxylin (AQUA-MEDICA) and eosin (PPH PochS.A.). Cells present on the lower surface of the filter were counted.Nonspecific migration was assessed by subtraction of the cellsmigrated in 0.5% BSA precoated filters. Each experiment wasperformed in triplicate and repeated at least 3 times.

“Wound healing”-like cell migration and cell proliferationassay were performed as previously described [20].

Flow cytometric analysis

HUVECs were transduced with either Ad VEGF-DΔNΔC or controladenovirus for 24 h and harvested using trypsin/EDTA. Afterresuspension in Medium 200, they were incubated with FITC or PEconjugated antibodies. In some experiments, unlabelled antibo-dies to α2β1 or α5β1 were used followed by anti-mouse IgGconjugated with FITC. FACS analysis was done using Calibur flowcytometer (Becton Dickinson).

909E X P E R I M E N T A L C E L L R E S E A R C H 3 1 6 ( 2 0 1 0 ) 9 0 7 – 9 1 4

HUVECs cultured in 3D fibrin gel

HUVECs cultured in Medium 200 supplemented with Low SerumGrowth Supplement containing 2% FBS and antibiotics (strepto-mycin and penicillin) were grown in dishes coveredwith fibrin gel.For this purpose, wells were coated with fibrinogen (2 μg/ml) for1 h and washed with 0.1 M Tris buffered saline, pH 7.3 (TBS). Thenfibrinogen bound was converted to fibrin by incubation withthrombin (10 min, 2.5 U/ml). After extensive washing with TBS,fibrinogen solution (2mg/ml)was added again and converted intofibrin with thrombin (2 h, 2.5 U/ml). To detect proteolyticdegradation of fibrin in the vicinity of cells, anti D-D antibodieslabeled with FITC were used and cells analyzed by confocalmicroscopy.

SDS-PAGE and Western blotting analysis

HUVECs transduced with either Ad VEGF-DΔNΔC or Ad-GFP weresolubilized in the lysis buffer (200 mM NaCl, 20 mM Tris–HCl, pH

Fig. 1 – Expression of VEGF-D in human endothelial cells (HUVECs)incubatedwith Ad-GFP for 2 hwere transduced in 70%. Such cells we20 h after transduction was completed (B). Panels C and D show expafter 24 h byWestern immunoblotting and its release fromHUVECsrespectively. Data are shown as a mean of at least three determinatrespectively.

8.3 containing 1 mM EDTA, 5% glycerol, and 1% Triton X-100) for10 min on ice, scraped and centrifuged at 12,000×g for 30 min at4 °C. After dilution with the sample buffer, aliquots were separatedby SDS-PAGE under reducing conditions and transferred to anitrocellulose membrane. Membranes were blocked for 1 h with3% BSA in TBS, pH 8.3 containing 0.04% Tween-20 and incubatedovernight at 4 °C with monoclonal antibody to VEGF-D or rabbitpolyclonal antibody to MMP-14. After washing, blots were stainedwith horseradish peroxidase-conjugated anti-murine IgG rabbitantibody or anti-rabbit IgG goat antibody, respectively. Afterextensive washing, the bands were detected using enhancedchemiluminescence substrate (Pierce, USA).

Statistical analysis

We have analyzed statistical significance of our data using pairedStudent t-test or ANOVA. Statistically significant data bearingP value <0.05 or <0.001 are annotated by ⁎ and ⁎⁎ symbols,respectively. Data are expressed as a mean±SD.

transduced with Ad-EGF-DΔNΔC. Panel A shows that HUVECsre enriched inmRNAVEGF-D as analyzed by real time PCR, 8 andression of recombinant VEGF-DΔNΔC, detected in HUVECs extracttransducedwith Ad-VEGF-DΔNΔC analyzed by ELISA in cellmedia,ions±SD. P values <0.05 and <0.001 are annotated by ⁎ and ⁎⁎,

Fig. 2 – Migration and proliferation human endothelial cells transduced by Ad-VEGF-DΔNΔC. Confluent HUVEC culture transducedwith Ad-GFP or Ad-VEGF-DΔNΔC was starved for 4 h. After the wounding, the cells were maintained in M199 containing 1% bovineserum albumin. Cell culture images were recorded immediately (0 h), and after 24 h and 48 h. Migration of wounded cells wasestimated by quantification of % of recovery as described in Materials and methods (B). Panel C shows proliferation rate of cellstransduced with Ad-GFP (circles) or Ad-VEGF-DΔNΔC (squares), respectively. Cells were counted manually in the Boyden chamber.Cell viability was determined microscopically by Trypan blue exclusion. Data are shown as a mean of at least three determinations±SD. P values <0.001 are annotated by ⁎⁎.

Fig. 3 – Increased migration of HUVECs transduced withAd-VEGF-DΔNΔC results from upregulation of integrinexpression. Migration of control HUVECs (transduced withAd-GFP) and HUVECs transduced with Ad-VEGF-DΔNΔC wasanalyzed using Transwell policarbonate filters coated withfibrinogen or collagen. Migration of cells was tested in thepresence or absence of blocking antibodies specific to αVβ3 orα2β1, respectively. Migrant cells were quantified at the bottomsurface of policarbonate filters after 16 h. Data are shown as amean of at least three determinations±SD. P values <0.001 areannotated by ⁎⁎.

910 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 6 ( 2 0 1 0 ) 9 0 7 – 9 1 4

Results

In preliminary experiments, we focused on setting up the optimalconditions to transduce HUVECs with Ad-VEGF-DΔNΔC or Ad-GFPand to operate its autocrine pathway. Fluorescent images ofHUVECs treated with Ad-GFP for 8 h showed that GFP protein wasexpressed by most cells indicating high yield of transduction(Fig. 1A). Similarly, HUVECs transduced with Ad-VEGF-DΔNΔC

expressed the recombinant VEGF-DΔNΔC detectable at the level ofmRNA, particularly after 8 h since transduction (Fig. 1B). Extensivesynthesis of VEGF by HUVECs could be seen after 24 h (Fig. 1C) andits secretion reached a plateau 10 h later (Fig. 1D). Conditioned-medium from Ad-VEGF-infected HUVEC cells contained largeamounts of VEGF-DΔNΔC protein (8000 pg/ml) compared with nodetectable VEGF in the conditioned-media from Ad-GFP trans-duced or control cells. As expected, transduction of HUVECs withAd-VEGF-DΔNΔC resulted in markedly increased cellular migrationwhen analyzed by both the “Wound Healing”-like assay and thetranswell migration assay. There was more rapid migration acrossthe wound in cells expressing VEGF-DΔNΔC than in controls(HUVECs transduced with Ad-GFP) (P<0.05; Figs. 2A, B). There

911E X P E R I M E N T A L C E L L R E S E A R C H 3 1 6 ( 2 0 1 0 ) 9 0 7 – 9 1 4

was significantly increased cell proliferation evaluated by countingof cells transduced with VEGF-DΔNΔC (Fig. 2C). Similarly, whentested using the transwell filters coated on the lower side withfibrinogen or collagen type I, such cells showed increasedmigration when compared to those transduced with Ad-GFPalone. Migration was significantly inhibited by blocking antibodiesto αVβ3 or α2β1, respectively (Fig. 3).

Since HUVECs exposed to autocrine loop of VEGF-DΔNΔC

showed profoundly changed adhesive properties, we turned toflow cytometry to characterize surface expression of membraneproteins. Almost all subunits of integrin receptors showedtendency to be upregulated in HUVECs transduced with Ad-VEGF-DΔNΔC (Fig. 4) Interestingly, αL, αM, and β2 subunits thatnormally are not expressed in endothelial cells were alsodetectable. Main integrin receptors indicated tendency to bemarkedly upregulated in such cells. On the other hand, majority ofthe tested membrane receptors but CD45RB, CD98, and CD147,

Fig. 4 – Differential changes in expression of membrane proteins inshows expression of integrin subunits or entire integrinmolecules aanalyzed by FACS. Panel B shows the effect of transduction with VEChanges in expression are plotted as a % of surface expression of c100%. Data are shown as a mean of at least three determinations±respectively.

showed a reduced expression (Fig. 4B). To address in vivo-likemigration within a 3D substrate that provides both adhesion sitesand physical resistance towardsmoving cells, we next used a fibrinnetwork. Endothelial cells were grown on 3D fibrin and treatedwith VEGF-D. The extracellular proteolytic activity in the closevicinity of cells was monitored by confocal microscopy using FITC-labeled monoclonal antibodies specific to D-dimer, a degradationproduct of cross-linked fibrin. As it is seen in Fig. 5, there wasalready slight but visible degradation of fibrin by control cells.Treatment of cells with VEGF-D resulted in strongly enhanced localdegradation of fibrin which was abolished by MMPs inhibitorGM6001. Transduction of HUVECswith Ad-VEGF-DΔNΔC resulted insignificant upregulation of MMP-2, MMP-9 and MMP-14 detect-able at the level of their mRNA and protein expression (Fig. 6).MMPswere secreted as zymogens from cells to the cell surface andinto the extracellular environment where they were able todegrade both ECM and non-ECM proteins. Interestingly, increased

duced upon transduction of HUVECs with VEGF-DΔNΔC. Panel Analyzed 24 h after transduction of HUVECs with Ad-VEGF-DΔNΔC

GF-DΔNΔC on expression of the selected membrane proteins.orresponding membrane proteins in control HUVECs taken asS.D. P values <0.05 and <0.01 are annotated by ⁎ and ⁎⁎,

Fig. 5 – Enhanced fibrinolytic activity of CRC in 3D fibrin gels after exposure to VEGF-D. Endothelial cells were grown on fibrin gelsthen they were treated with 50 ng/ml VEGF-D for 24 h in the presence or absence of 25 μM GM6001. Pericellular fibrinolysis wasestimated with monoclonal antibody reacting with D-dimer epitope (green). Nuclear DNA was stained with DAPI (cyan).

912 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 6 ( 2 0 1 0 ) 9 0 7 – 9 1 4

fibrinolytic activity observed under these conditions was notaffected by PAI-1 which together with u-PAR was also markedlyupregulated by VEGF-D in endothelial cells at the level of its mRNAand protein synthesis (Figs. 6B, E).

Discussion

Recently, endothelial cell lines obtained from human renalcarcinomas were found to differ from normal human microvascu-lar endothelial cells in terms of survival and proangiogenic

Fig. 6 – Expression of matrix metaloproteinases and PAI-1 in endotfor MMP-2, MMP-9, MMP-14, PAI-1 and u-PAR in HUVECs transduceRTPCR, 8 and 20 h after transduction (A and B). In the same cells 20 hzymography (C) whereas protein synthesis of MMP-14 and PAI-1 wshow representative data obtained during several experiments (3–5<0.05 and <0.01 are annotated by ⁎ and ⁎⁎, respectively.

properties. They were characterized by increased expression ofVEGF-D and its receptors, VEGFR-2 and VEGFR-3, and showed theantiapoptotic phenotype probably due to persistent activation bythe tumor secreted angiogenic growth factors [1,2]. Consistently,present data indicate that HUVECs, enriched in recombinant VEGF-DΔNΔC delivered by adenoviral vector, behave similarly to thetumor-derived endothelial cell lines as far as proadhesive,prosurvival, and proangiogenic properties are concerned. Theyshowed significantly increased migration, probably due to theenhanced expression or activation of adhesion molecules andintegrins. Their adhesion and migration on fibrinogen or collagen

helial cells transduced by Ad-VEGF-DΔNΔC. Expression of mRNAsd with Ad-GFP (control cells) or Ad-VEGFDΔNΔC was analyzed byafter transduction, activity of MMP-2 andMMP-9was tested byas evaluated by Western immunoblotting (D and E). All figures). Amounts of mRNA are expressed as a mean±SD and P values

913E X P E R I M E N T A L C E L L R E S E A R C H 3 1 6 ( 2 0 1 0 ) 9 0 7 – 9 1 4

type I could be abrogated by anti-αVβ3 or anti-α2β1 function-blocking antibodies, respectively. Surface concentration of bothintegrin receptors in the Ad-VEGF-DΔNΔC transduced HUVECs wasnot significantly upregulated, confirming previous observationsindicating that Ad-VEGF-DΔNΔC influences the integrins activationwithout increasing their number of copies [7]. Similar to others, weshowed that such cells have an increased ability to migrate oncollagen type I that can be abolished by anti β1 blocking antibody[21,22]. These findings confirm that VEGF-D functions as anautocrine growth factor and stimulator of proangiogenic pheno-type of endothelial cells. Accordingly, HUVECs exposed to VEGF-Dshow increased expression of proteins that indicate proangiogenicactivity, such as (a) MMPs (MMP-2, MMP-9, and MMP-14), (b)PAI-1 and u-PAR, (c) integrins, (d) proteins that modulate MMPsactivity (CD147, EMMPRIN) or integrins activity (CD98). Thus,VEGF-D similar to VEGF-A increased expression of MMP-2, andMMP-14 [23].

Secreted enzymes influence proteolytic balance associatedwith endothelial cell membranes and facilitate their invasion of afibrin scaffold thus determining the progression of angiogenesis.The increased secretion of active PAI-1 from endothelial cellsexplains why incorporation of vitronectin into fibrin furtherimproves the facilitation of cell invasion by the fibrin matrix.PAI-1 binds to vitronectin with high affinity, but also both,vitronectin and PAI-1 bind directly to fibrin [24]. Thus, PAI-1concentrated in fibrin, particularly at sites of injury and inflam-mation, may account for the recent observations that both activeand latent forms of this protein stimulate cell migration inchemotaxis, haptotaxis, chemokinesis, and wound-healing assays[25]. PAI-1, via formation of multimolecular complexes on the cellmembranes or in extravascular areas, can have profound effects oncells, including their attachment, detachment, andmigration in theextracellular matrix [26,27]. Thus, two potential mechanismsexplaining angiogenic activity that depends upon PAI-1 concen-tration were proposed. First, by protecting the extracellular matrixagainst excessive degradation, PAI-1 may serve to maintain thematrix scaffold required for endothelial cell migration and tubeformation [28]. Second, due to specific interaction with uPAR,integrins, and vitronectin [29], alterations in PAI-1 expression andactivity would therefore be expected to alter the adhesive,migratory, and growth properties of endothelial cells, which inturn would regulate angiogenesis.

In addition to proteins that are directly involved in adhesionand migration of endothelial cells, VEGF-D appears to modulateexpression of membrane proteins which can act as cofactors.Among them, there are CD98 and CD147 controlling receptoractivity of integrins (β1 and β3 subfamilies) and activation ofMMPs, respectively. CD147 is thought to promote tumor angio-genesis mostly through its protease-inducing function and morerecently by its ability to increase HIF-2α, VEGFR-2 and the solubleforms of VEGF in endothelial [30]. The CD98hc binds to β1 and β3integrins and stimulates adhesion when crosslinkedwith antibody[31]. The extracellular portion of CD98 heavy chain is necessaryand sufficient for amino acid transport function of the hetero-dimer, while the cytoplasmic and transmembrane domains arenecessary and sufficient for integrin binding and complementationof dominant suppression.

In summary, exposure of endothelial cells to VEGF-DΔNΔC

causes major changes in the global protein expression pattern.Taken together, upregulation of so many different proteins by

VEGF-D in endothelial cells indicates that inducing proangiogenicphenotype is associated with extensive modulation of cellhomeostasis to ensure their survival under different conditions.

Acknowledgment

This work was supported by Projects N301 012 31/0230 and 343/N-INCA/2008/0 from the PolishMinistry of Scientific Research andHigher Education.

R E F E R E N C E S

[1] B. Bussolati, I. Deambrosis, S. Russo, M.C. Deregibus, G. Camussi,Altered angiogenesis and survival in human tumor-derivedendothelial cells, FASEB J. 17 (2003) 1159–1161.

[2] B. Bussolati, M.C. Deregibus, G. Camussi, Characterization ofmolecular and functional alterations of tumor endothelial cells todesign anti-angiogenic strategies, Curr. Vasc. Pharmacol. (2010)Jan 1. [Electronic publication ahead of print].

[3] M.G. Achen, M. Jeltsch, E. Kukk, T. Makinen, A. Vitali, A.F. Wilks, K.Alitalo, S.A. Stacker, Vascular endothelial growth factor D (VEGF-D) is a ligand for the tyrosine kinases VEGF receptor 2 (Flk1) andVEGF receptor 3 (Flt4), Proc. Natl. Acad. Sci. U. S. A. 95 (1998)548–553.

[4] B.K. McColl, K. Paavonen, T. Karnezis, N.C. Harris, N. Davydova,J. Rothacker, E.C. Nice, K.W. Harder, S. Roufail, M.L. Hibbs, P.A.W. Rogers, K. Alitalo, S.A. Stacker, M.G. Achen, Proproteinconvertases promote processing of VEGF-D, a critical step forbinding the angiogenic receptor VEGFR-2, FASEB J. 21 (2007)1088–1098.

[5] S.A. Stacker, K. Stenvers, C. Caesar, A. Vitali, T. Domagala, E. Nice, S.Roufail, R.J. Simpson, R. Moritz, T. Karpanen, K. Alitalo, M.G.Achen, Biosynthesis of vascular endothelial growth factor-Dinvolves proteolytic processing which generates non-covalenthomodimers, J. Biol. Chem. 274 (1999) 32127–32136.

[6] L. Marconcini, S. Marchio, L. Morbidelli, E. Cartocci, A. Albini, M.Ziche, F. Bussolino, S. Oliviero, c-fos-induced growth factor/vascular endothelial growth factor D induces angiogenesis in vivoand in vitro, Proc. Natl. Acad. Sci. U.S.A. 96 (1999) 9671–9676.

[7] T.V. Byzova, C.K. Goldman, J. Jankau, J. Chen, G. Cabrera, M.G.Achen, S.A. Stacker, K.A. Carnevale, M. Siemionow, S.R. Deitcher,P.E. DiCorleto, Adenovirus encoding vascular endothelial growthfactor-D induces tissue-specific vascular patterns in vivo, Blood99 (2002) 4434–4442.

[8] T.T. Rissanen, J.E. Markkanen, M. Gruchala, T. Heikura, A. Puranen,M.L. Kettunen, I. Kholova, R.A. Kauppinen, M.G. Achen, S.A.Stacker, K. Alitalo, S.Y. Herttuala, VEGF-D is the strongestangiogenic and lymphangiogenic effector among VEGFs deliveredinto skeletal muscle via adenoviruses, Circ. Res. 92 (2003)1098–1106.

[9] T. Veikkola, L. Jussila, T. Makinen, T. Karpanen, M. Jeltsch, T.V.Petrova, H. Kubo, G. Thurston, D.M. McDonald, M.G. Achen, S.A.Stacker, K. Alitalo, Signaling via vascular endothelial growthfactor receptor-3 is sufficient for lymphangiogenesis in transgenicmice, EMBO J 20 (2001) 1223–1231.

[10] S.A. Stacker, C. Caesar, M.E. Baldwin, G.E. Thornton, R.A. Williams,R. Prevo, D.G. Jackson, S. Nishikawa, H. Kubo, M.G. Achen, VEGF-Dpromotes the metastatic spread of tumor cells via the lymphatics,Nat. Med. 7 (2001) 186–191.

[11] Z. Von Marschall, A. Scholz, S.A Stacker, M.G. Achen, D.G. Jackson,F. Alves, M. Schirner, M. Haberey, K.H. Thierauch, B.Wiedenmann,S. Rosewicz, Vascular endothelial growth factor-D induceslymphangiogenesis and lymphatic metastasis in models of ductalpancreatic cancer, Int. J. Oncol. 27 (2005) 669–679.

914 E X P E R I M E N T A L C E L L R E S E A R C H 3 1 6 ( 2 0 1 0 ) 9 0 7 – 9 1 4

[12] L. Kopfstein, T. Veikkola, V.G. Djonov, V. Baeriswyl, T. Schomber,K. Strittmatter, S.A. Stacker, M.G. Achen, K. Alitalo, G. Christofori,Distinct roles of vascular endothelial growth factor-D in lym-phangiogenesis and metastasis, Am. J. Pathol. 170 (2007)1348–1361.

[13] T. Makinen, T. Veikkola, S. Mustjoki, T. Karpanen, B. Catimel, E.C.Nice, L. Wise, A. Mercer, H. Kowalski, D. Kerjaschki, S.A. Stacker,M.G. Achen, K. Alitalo, Isolated lymphatic endothelial cellstransduce growth, survival andmigratory signals via the VEGF-C/D receptor VEGFR-3, EMBO J 20 (2001) 4762–4773.

[14] H. Jia, A. Bagherzadeh, R. Bicknell, M.R. Duchen, D. Liu, I. Zachary,Vascular endothelial growth factor (VEGF)-D and VEGF-Adifferentially regulate KDR-mediated signaling and biologicalfunction in vascular endothelial cells, J. Biol. Chem. 279 (2004)36148–36157.

[15] T.T. Rissanen, J.E. Markkanen, M. Gruchala, T. Heikura, A. Puranen,M.L. Kettunen, I. Kholová, R.A. Kauppinen, M.G. Achen, S.A.Stacker, K. Alitalo, S. Ylä-Herttuala, VEGF-D is the strongestangiogenic and lymphangiogenic effector among VEGFs deliveredinto skeletal muscle via adenoviruses, Circ. Res. 92 (2003)1098–1106.

[16] S. Bhardwaj, H. Roy, M. Gruchala, H. Viita, I. Kholova, I. Kokina, M.G. Achen, S.A. Stacker, M. Hedman, K. Alitalo, S. Ylä-Herttuala,Angiogenic responses of vascular endothelial growth factorsin periadventitial tissue, Human Gene Therapy 14 (2003)1451–1462.

[17] J. Rutanen, A.M. Turunen, M. Teittinen, T.T. Rissanen, T. Heikura, J.K. Koponen, M. Gruchala, M. Inkala, S. Jauhiainen, M.O. Hiltunen,M.P. Turunen, S.A. Stacker, M.G. Achen, S. Ylä-Herttuala, Genetransfer using the mature form of VEGF-D reduces neointimalthickening through nitric oxide-dependent mechanism, GeneTherapy 12 (2005) 980–987.

[18] J. Rutanen, T.T. Rissanen, J.E. Markkanen, M. Gruchala, P.Silvennoinen, A. Kivelä, A. Hedman, M. Hedman, T. Heikura, M.R.Ordén, S.A. Stacker, M.G. Achen, J. Hartikainen, S. Ylä-Herttuala,Adenoviral catheter-mediated intramyocardial gene transferusing the mature form of vascular endothelial growth factor-Dinduces transmural angiogenesis in porcine heart, Circulation 109(2004) 1029–1035.

[19] S. De, O. Razorenova, N.P. McCabe, T. O'Toole, J. Qin, T.V. Byzova,VEGF-integrin interplay controls tumor growth andvascularization, Proc. Natl. Acad. Sci. U.S.A. 102 (2005)7589–7594.

[20] Z. Pawlowska, P. Baranska, H. Jerczynska, W. Koziolkiewicz, C.S.Cierniewski, Heat shock proteins and other components ofcellular machinery for protein synthesis are up-regulated invascular endothelial cell growth factor-activated human endo-thelial cells, Proteomics 5 (2005) 1217–1227.

[21] J.F. Wang, X.F. Zhang, J.E. Groopman, Stimulation of beta 1integrin induces tyrosine phosphorylation of vascular endothelialgrowth factor receptor-3 and modulates cell migration, J. Biol.Chem. 276 (2001) 41950–41957.

[22] M.J. Mountain, M. Singh, K. Singh, Interleukin-1β-mediatedinhibition of the processes of angiogenesis in cardiac microvas-cular endothelial cells, Life Sciences 82 (2008) 1224–1230.

[23] E.N. Unemori, E.N.N. Ferrara, E.A. Bauer, E. Amento, Vascularendothelial growth factor induces interstitial collagenase expres-sion inhumanendothelial cells, J. Cell. Physiol. 153 (1992) 557–562.

[24] K. Smolarczyk, J. Boncela, J. Szymanski, A. Gills, CS. Cierniewski,Fibrinogen contains cryptic PAI-1 binding sites that are exposedon binding to solid surfaces or limited proteolysis, ArteriosclerThromb Vasc Biol 25 (2005) 2679–2684.

[25] B. Degryse, J.G. Neels, R.P. Czekay, The low density lipoproteinreceptor-related protein is a motogenic receptor for plasminogenactivator inhibitor-1, J. Biol. Chem. 279 (2004) 22595–22604.

[26] K.N. Al-Nedawi, M. Czyz, R. Bednarek, J. Szemraj, M. Swiatkowska,A. Cierniewska-Cieslak, J. Wyczolkowska, C.S. Cierniewski,Thymosin β4 induces the synthesis of plasminogen activatorinhibitor-1 in cultured endothelial cells and increases itsextracellular expression, Blood 103 (2004) 1319–1324.

[27] J. Boncela, I. Papiewska, I. Fijałkowska, B. Walkowiak, C.S.Cierniewski, Acute phase protein α1-acid glycoprotein interactswith plasminogen activator inhibitor type 1 and stabilizes itsinhibitory activity, J. Biol. Chem. 276 (2001) 35305–35311.

[28] M.S. Pepper, R. Montesano, Proteolytic balance and capillarymorphogenesis, Cell Differ. Dev. 32 (1990) 319–327.

[29] D.J. Loskutoff, S.A. Curriden, G. Hu, G. Deng, Regulation of celladhesion by PAI-1, APMIS 107 (1999) 54–61.

[30] F. Bougaten, C. Quemener, S. Kellouche, B. Naimi, M.P. Podgorniak,G. Millot, E.F. Gabison, F. Calvo, C. Dosquet, C. Lebbe, S. Menashi, S.Mourah, EMMPRIN promotes angiogenesis through HIF-2αmediated regulation of soluble VEGF isoforms and their receptorVEGFR-2, Blood 116 (2009) 5547–5556.

[31] C.A. Fenczik, T. Sethi, J.W. Ramos, P.E. Hughes, M.H. Ginsberg,Complementation of dominant suppression implicates CD98 inintegrin activation, Nature 390 (1997) 81–85.