1998;58:3209-3214. Cancer Res   Zhenping Zhu, Patricia Rockwell, Dan Lu, et al.   Receptor Single-Chain Antibodies from a Phage Display LibraryReceptor Activation with Anti-Kinase Insert Domain-containing Inhibition of Vascular Endothelial Growth Factor-induced

  Updated version

  http://cancerres.aacrjournals.org/content/58/15/3209

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  .permissions@aacr.orgDepartment at

To request permission to re-use all or part of this article, contact the AACR Publications

Research. on September 28, 2014. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

Research. on September 28, 2014. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

[CANCER RESEARCH 58, 3209-3214, August I, 1998]

Advances in Brief

Inhibition of Vascular Endothelial Growth Factor-induced Receptor Activation withAnti-Kinase Insert Domain-containing Receptor Single-Chain Antibodies from a

Phage Display Library

Zhenping Zhu,1 Patricia Rockwell,2 Dan Lu, Helen Kotanides, Bronislaw Pytowski, Daniel J. Hicklin, Peter Bohlen,

and Larry Witte

Departments of Molecular and Cell Biology [Z. Z, D. L, H. K., B. P., L W.¡Immunology ¡P.R., D. J. H.¡,and Research ¡P.B.¡.ImClone Systems Inc.. New York, New York 10014

Abstract

A single-chain antibody phage display library was constructed from

spleen cells of mice immunized with a soluble form of a human vascularendothelial growth factor (VEGF) receptor, kinase insert domain-contain

ing receptor (KDR). After two rounds of biopanning, >90% of the clonesrecovered were specifically reactive to KDR. Subsequent selection identified two clones that blocked VEGF binding to KDR. The clones wereexpressed in Escherichia coli and purified as soluble single-chain Fv (scFv)

antibodies. The affinities of the scFv for binding to KDR were determinedby BIAcore analysis (2.1 x 10~'-5.9 x 10"' M). One scFv, plCll, was

shown to inhibit VEGF-induced KDR phosphorylation and VEGF-stim-

ulated DNA synthesis in human umbilical vein endothelial cells. There ismuch experimental evidence to suggest that the VEGF/KDR/Flk-1 path

way plays an important role in tumor angiogenesis, a process that isessential for tumor growth and metastasis. The antibodies discussed here,which block VEGF binding to KDR, have potential clinical application inthe treatment of cancer and other diseases where pathological angiogenesis is involved.

Introduction

Angiogenesis, a multistep process that results in new blood vesselformation from preexisting vasculature (see Ref. l for review), isessential for both the growth of solid tumors beyond the size of 2 mm3

and for tumor metastasis (2, 3). A number of growth factors, e.g.,VEGF,3 bFGF, and angiogenin have been implicated as possible

regulators in this process (4, 5). VEGF is distinct among these factorsin that it acts as an endothelial cell-specific mitogen during angiogenesis (6-8). Its receptor, KDR/Flk-1 (also known as VEGF receptor 2),

was shown to be expressed exclusively on endothelial cells andhematopoietic cells (9, 10). The level of the receptor has been shownto be up-regulated by VEGF in several tumors and confined to the

vascular endothelial cells in close proximity to the tumors (11).Endothelial cells transfected with a dominant-negative mutant of theKDR/Flk-1 receptor inhibited glioblastoma growth in vivo (12). Takentogether, these findings suggest that the VEGF/KDR/Flk-1 pathway

plays an important role in tumor angiogenesis (6, 7, 13). This role isfurther supported by studies showing inhibition of tumor growth in

Received 3/4/98; accepted 6/16/98.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1To whom requests for reprints should be addressed, at Department of Molecular and

Cell Biology, ImClone Systems, Inc., 180 Varick Street, New York, NY 10014. Phone:(212)645-1405; Fax: (212)645-2054; E-mail: zhenping@imclone.com.

2 Present address: Department of Biological Science, Hunter College, 695 Park Ave

nue, New York, NY 10021.3 The abbreviations used are: VEGF, vascular endothelial growth factor; bFGF, basic

fibroblast growth factor; Flk-1, fetal liver kinase 1; KDR, kinase insert domain-containingreceptor; scFv, single-chain Fv; HUVEC, human umbilical vein endothelial cell; AP,alkaline phosphatase; RT, room temperature; PBST, PBS containing 0.1% Tween 20;HRP. horseradish peroxidase; MAP, mitogen-activated protein; Mab, monoclonal antibody.

nude mice by anti-VEGF antibodies (14), anti-Flk-1 antibodies (15,16), anti-VEGF antisense RNA expression (17), and VEGF-diphtheria

toxin conjugate (18).We previously reported the production of a rat anti-Flk-1 Mab,

DC 101, by hybridoma technology and demonstrated that this antibodycould specifically inhibit VEGF stimulation of Flk-1 receptor ex

pressed on transfected 3T3 cells (19). Furthermore, DC101 stronglyinhibited growth of human tumor xenografts in nude mice by blockingthe tumor-associated angiogenesis (15, 16). DC101 does not, however, react with KDR, the human homologue of murine Flk-1: there

fore, it is not a suitable candidate for potential human therapy. Here,we describe the production and characterization of scFv anti-KDR

antibodies isolated from an antibody phage display library. One scFv,plCll, binds to KDR (but not to Flk-1) with high affinity (2.1 nM),blocks VEGF binding to KDR, and shows potent inhibition of VEGF-

induced receptor phosphorylation and DNA synthesis in HUVECs.

Materials and Methods

Cell Lines and Proteins. Primary-cultured HUVECs were obtained from

Dr. S. Rafii at Cornell Medical Center (New York, NY) and maintained inEBM-2 medium (Clonetics, Walkersville, MD) at 37°Cand 5% CO2. Cells

were used between passages 2 and 5 for all assays. VEGFI65 protein was

expressed in baculovirus and purified following the procedures described (20).cDNA encoding the extracellular domain of KDR was isolated by RT-PCR

from human fetal kidney mRNA (Clontech, Palo Alto, CA) and subcloned intothe Bgtn and BspEl sites of the vector AP-Tag (21). In this plasmid, the cDNAfor KDR extracellular domain is fused in-frame with the cDNA for human

placenta! AP. The plasmid was electroporated into NIH 3T3 cells together withthe neomycin expression vector pSV-Neo, and stable cell clones were selected

with G418. The soluble fusion protein KDR-AP was purified from cell culture

supernatant by affinity chromatography using immobilized Mab to AP (MedixBiotech, Inc., Foster City, CA).

Immunization of Mice and Construction of a scFv Antibody PhageDisplay Library. Female BALB/C mice (HaríanSprague Dawley, Indianapolis, IN) were given two i.p. injections of 10 /j.g of KDR-AP in 200 /xl of RIBI

Adjuvant System (RIBI Immunochem Research, Inc., Hamilton, MT), fol

lowed by one i.p. injection without RIBI adjuvant over a period of 2 months.The mice were also given a s.c. injection of 10 fj.g of KDR-AP in 200 /¿Iof

RIBI solution at the time of the first immunization. The mice were boosted i.p.with 20 ¿igof KDR-AP 3 days before euthanasia. Spleens from donor mice

were removed and the cells were isolated. Total splenocyte RNA was firstextracted with a kit from Ambion (Austin, TX), and mRNA was purified fromtotal RNA with the Quickprep mRNA kit (Pharmacia Biotech, Piscataway,NJ). A scFv phage display library was constructed using the mRNA with theuse of the Recombinant Phage Antibody System (Pharmacia Biotech) following the manufacturer's protocol. The assembled scFv DNA was ligated into the

pCANTAB 5E vector using SfiUNoll restriction sites and electroporated intoTGI cells. The transformed TGI cells were plated onto 2YTAG plates (16 gof tryptone, 10 g of yeast extract, and 5 g of NaCl per liter medium, plus 100ju,g/ml ampicillin and 2% glucose) and incubated overnight at 30°C.The

3209

Research. on September 28, 2014. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

INHIBITION OF VEOF ACTIVITIES BV scFv ANTIBODIES

colonies were scraped into 10 ml of 2YT medium, mixed with 5 ml 50%glycerol, and stored at -70°C as the library stock.

Biopanning. The library stock was grown to logarithmic phase, rescuedwith M13K07 helper phage, and amplified overnight in 2YTAK medium (2YTcontaining 100 ng/ml ampicillin and 50 jug/ml kanamycin) at 30°C.The phage

preparation was precipitated in 4% PEG/0.5 M NaCl, resuspended in 3%fat-free milk/PBS containing 500 fig/ml AP protein, and incubated at 37°Cfor

l h to capture phage displaying anti-AP scFv and to block other nonspecific

binding.KDR-AP (10 |U.g/ml)-coated Maxisorp Star tubes (Nunc, Roskilde, Den

mark) were first blocked with 3% milk-PBS at 37°Cfor l h and then incubated

with the phage preparation at RT for 1 h. The tubes were washed 10 times withPBST, followed by 10 washes with PBS. The bound phage was eluted at RTfor 10 min with 1 ml of a freshly prepared solution of 100 mM triethylamine(Sigma Chemical Co., St. Louis, MO). The eluted phage were incubated with10 ml of mid-logarithmic-phase TGI cells at 37°C,first stationary (30 min)

and then shaking (30 min). The infected TGI cells were then plated onto2YTAG plates and incubated overnight at 30°C.

Phage ELISA. Individual TGI clones were grown at 37°Cin 96-well

plates and rescued with M13K07 helper phage as described above. Theamplified phage preparation was blocked with 1/6 volume of 18% milk-PBS atRT for l h and added to Maxi-sorp 96-well microtiter plates (Nunc) coatedwith KDR-AP or AP (1 ¿tg/mlX 100 ^1). After incubation at RT for 1 h, the

plates were washed three times with PBST and incubated with a rabbitanti-Mi3 phage antibody-HRP conjugate (Pharmacia Biotech). The plates

were washed 5 times, TMB peroxidase substrate (KPL, Gaithersburg, MD)was added, and the /1450nmread using a microplate reader (Molecular Devices,Sunnyvale, CA).

DNA BstNl Pattern Analysis and Nucleotide Sequencing. The diversityof the KDR binders after each round of selection was analyzed by restrictionenzyme digestion patterns (i.e., DNA fingerprinting). The scFv insert ofindividual clones was amplified with PCR using the following primers: PUC19reverse, AGCGGATAACAATTTCACACAGG; and fdtet seq, GTCGTCTT-TCCAGACGTTAGT. The amplified product was digested with a frequent-

cutting enzyme, BstNl, and analyzed on a 3% agarose gel. DNA sequences ofrepresentative clones from each digestion pattern were determined bydideoxynucleotide sequencing.

Preparation of Soluble scFv. Phage of individual clones were used toinfect a nonsuppressor Escherichia coli host HB2151, and the infectant wasselected on 2YTAG-N plates (2YTAG containing 100 /xg/ml nalidixic acid;

Sigma). Expression of scFv in HB2151 cells was induced by culturing the cellsin 2YTA medium containing l mM isopropyl-1-thio-ß-D-galactopyranoside(Sigma) at 30°C.A periplasmic extract of the cells was prepared by resus-

pending the cell pellet in 25 mM Tris (pH 7.5) containing 20% (w/v) sucrose,200 mM NaCl, 1 mM EDTA, and 0.1 mM phenylmethylsulfonyl fluoride,followed by incubation at 4°Cwith gentle shaking for 1 h. After centrifugation

at 15,000 rpm for 15 min, the soluble scFv was purified from the supernatantby affinity chromatography using the RPAS Purification Module (PharmaciaBiotech).

Quantitative KDR Binding Assay. Two assays were used to examinequantitatively the binding of purified soluble scFv to KDR. In the directbinding assay, various amounts of soluble scFv were added to KDR-coated96-well Maxi-sorp microtiter plates and incubated at RT for 1 h, after which

the plates were washed three times with PBST. The plates were then incubatedat RT for 1 h with 100 j^l of mouse anti-E tag antibody (Pharmacia Biotech),followed by incubation with 100 ju.1of rabbit antimouse antibody-HRP con

jugate (Biosource, Camarillo, CA). The plates were washed and developedfollowing the procedure described above for the phage ELISA. In anotherassay, the competitive VEGF blocking assay, various amounts of soluble scFvwere mixed with a fixed amount of KDR-AP (50 ng) and incubated at RT for1 h. The mixture were then transferred to 96-well microtiter plates coated with

VEGF165 (200 ng/well) and incubated at RT for an additional 2 h, after whichthe plates were washed five times and the substrate for AP (p-nitrophenylphosphate; Sigma) was added to quantify the bound KDR-AP molecules. IC50,

i.e., the scFv concentration required for 50% inhibition of KDR binding toVEGF, was then calculated.

BIAcore Analysis of the Soluble scFv. The binding kinetics of solublescFv to KDR were measured using BIAcore biosensor (Pharmacia Biosensor).KDR-AP fusion protein was immobilized onto a sensor chip, and soluble scFv

antibodies were injected at concentrations ranging from 62.5 to 1000 nM.Sensorgrams were obtained at each concentration and were evaluated using theBIA Evaluation 2.0 program to determine the rate constants fconand koff. Kdwas calculated from the ratio of rate constants kofl/kon.

Phosphorylation Assay. Subconfluent HUVECs were grown in growthfactor-depleted EBM-2 medium for 24-48 h prior to experimentation. After

pretreatment with 50 nM sodium orthovanadate for 30 min, the cells wereincubated in the presence or absence of antibodies (5 fig/ml) for 15 min,followed by stimulation with 20 ng/ml VEGF 165 or 10 ng/ml bFGF (R&DSystems, Minneapolis, MN) at RT for an additional 15 min. The cells werethen lysed in lysis buffer (50 nM Tris, 150 mM NaCl, 1% Nonidet P-40, 2 mM

EDTA, 0.25% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride, 1fig/ml leupeptin, 1 fig/ml pepstatin, and 10 fig/mi aprotinin; pH 7.5), and thecell lysate was used for both the KDR and MAP kinase phosphorylationassays. The KDR was immunoprecipitated from the cell lysates with ProteinA-Sepharose beads (Santa Cruz Biotechnology, Santa Cruz, CA) coupled to ananti-KDR antibody, Mab 4.13 (ImClone Systems). Proteins were resolved withSDS-PAGE and subjected to Western blot analysis. To detect KDR phospho

rylation, blots were probed with an antiphosphotyrosine Mab, PY20 (ICNBiomedicals, Aurora, OH). For the MAP kinase activity assay, cell lysateswere resolved with SDS-PAGE followed by Western blot analysis using a

phosphospecific MAP kinase antibody (New England Biolabs, Beverly, MA).All signals were detected using enhanced chemiluminescence (Amersham,Arlington Heights, IL). In both assays, the blots were reprobed with a poly-clonal anti-KDR antibody (ImClone Systems) to assure that equal amount ofprotein was loaded in each lane of SDS-polyacrylamide gels.

Antimitogenic Assay. HUVECs (5 X IO3 cells/well) were plated onto

96-well tissue culture plates (Wallac, Gaithersburg, MD) in 200 ju.1of EBM-2

medium without VEGF, bFGF, or epidermal growth factor and incubated at37°Cfor 72 h. Various amounts of soluble scFv were added to duplicate wellsand preincubated at 37°Cfor 1 h, after which VEGF165 was added to a final

concentration of 16 ng/ml. In a parallel assay, bFGF (4 ng/ml) was added to thecells to evaluate the specificity of anti-KDR scFv on ligand-induced DNAsynthesis in HUVECs. After 18 h of incubation, 0.25 /¿Ciof [3H]thymidine

(Amersham) was added to each well and incubated for an additional 4 h. Thecells were placed on ice and washed twice with serum-containing medium,followed by a 10-min incubation at 4°Cwith 10% trichloroacetic acid. The

cells were then washed once with water and solubilized in 25 /^l of 2% SDS.Scintillation fluid (150 p,l/well) was added, and DNA-incorporated radioactivity was determined on a scintillation counter (Wallac; model 1450 Micro-

beta Scintillation Counter).

Results

Library Construction and Biopanning. niRNA was isolatedfrom mice immunized with KDR-AP soluble protein and used to

construct a combinatorial scFv library, which was displayed on thesurface of the filamentous phage Ml3. In displaying the scFv onfilamentous phage surface, antibody VH and VL domains are joinedtogether by a 15-amino acid linker (GGGGS)3 and fused to the NH2terminus of phage protein III. A 15-amino acid E tag, which is

followed by an amber codon (TAG), was inserted between the COOHterminus of VL and the protein III for detection and other analyticpurposes. The amber codon, positioned between the E tag and theprotein III, enables the construct to make scFv in surface-displaying

format when transformed into a suppressor host (such as TGI cells)and to make scFv in soluble form when transformed into a nonsu-

pressor host (such as HB2151 cells).After ligation of the scFv fragments into the pCANTAB 5 E vector

and electroporation into TGI cells, a library of 2.7 X IO8 clones was

obtained. Anti-KDR scFv were selected from the library by two to

three rounds of biopanning. A saturating concentration of AP proteinwas added to phage preparation in all rounds of panning as a coun-terselection to remove phage-displaying anti-AP scFv. A total of 270

clones were randomly picked after the second round of biopanningand screened for binding to KDR-AP or AP, and 243 clones (90%) ofthem were found to bind to KDR-AP. None of the 270 clones bound

3210

Research. on September 28, 2014. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

INHIBITION OF VEGF ACTIVITIES BY scFv ANTIBODIF.S

to AP protein, indicating the high efficiency of the counter selectionused during the biopanning process. Further selection was achievedwith a competitive VEGF-binding assay in which the binding of

soluble KDR to immobilized VEGF in the presence or absence ofscFv was determined (see "Materials and Methods"). The assay iden

tified 53 clones (20% of all KDR binders) as being capable ofblocking the binding between VEGF and its receptor, KDR. DNABstNl fingerprinting of these 53 clones indicated the presence of fourdifferent digestion patterns (with one pattern dominant, 79%),whereas 26 randomly picked VEGF nonblockers yielded six differentpatterns (also with one pattern dominant, 46%).

Ninety-nine % (185 of 186) of clones screened after the third round

of panning were found to be specific KDR binders. However, only 15(8%) of these binders could block KDR binding to immobilizedVEGF. DNA ßj-?NIfingerprinting of these 15 clones indicated the

presence of two different digestion patterns, whereas 21 randomlypicked VEGF nonblockers yielded four different patterns. All of thedigestion patterns were also seen in clones identified after the secondround of panning. Representative clones of each digestion patternwere picked from clones recovered after the second round of panningand subjected to DNA sequencing. Of 15 clones sequenced. 2 uniqueVEGF blockers and 3 nonblockers were identified (Fig. 1). One scFv,p2A7, which neither binds to KDR nor blocks VEGF binding to KDR,was selected as a negative control for all studies.

KDR Binding Assay of Purified scFv. Four different clones,including the two VEGF blockers (p 1C11 and p 1F12), one nonblocker(the dominant clone p2A6), and the nonbinder p2A7, were expressedin shaker flasks using a nonsuppressor host E. coli HB2151 cells. Thesoluble scFv was purified from the periplasmic extracts of E. coli byanti-E tag affinity chromatography. The yield of purified scFv of theseclones ranged from 100-400 /ng/liter culture.

Fig. 1A shows the dose-dependent binding of the KDR-bindingscFv to immobilized KDR, as assayed by a direct-binding ELISA.

Clones plCll and plF12 but not p2A6 also block KDR binding to

Heavy chain

scFv CDR1 CDR2 CDR3

p1C11 GFNIKDFYMH WIDPENGDSDYAPKFQG YYGDYEGY

p1F12 GYSFTGYNMN IIDPYYGGTYYNQKFKG SGNVGGYFDY

p2A6 GYTFTEYTMH GINPNTGGTSYNQKFKG YRYDAMDY

p3A1 GSAFSSYWMN QIYPGDGDTKYNGKFKG EYGSDFDY

p2A11 GYTFTEYTMH GINPNTGGTGYNPKFKG SKYGSLDY

p2A7 GYPFSSYWMN QIYPGDGDTNYNGNFRN FYGNYDPYTMNY

0.001 0.01 0.1 1

scFv Concentration (ng/ml)

1.5

—¿�1.0 -

cIO

IO0 0.5^

0.0

p2A7

p1C11

0.01 0.1

scFv Concentration (ng/ml)

10

Fig. 2. Binding of scFv antibodies lo KDR. A. direct binding of scFv to immobilizedKDR. Various amounts of scFv antibodies were added to 96-well plates coated withKDR-AP and incubated at RT for I h, after which the plates were incubated with a mouseanti-E tag antibody followed by a rabbit antimouse antibody-HRP conjugate. The plateswere washed, peroxidase was substrate added, and ¿445Unmwas read (see "Materials andMethods" for details). B, inhibition of binding of KDR to immobilized VEGF,,,, by scFv.Various amounts of scFv antibodies were incubated with a fixed amount of KDR-AP insolution at RT for 1 h, after which the mixtures were transferred to 96-well plates coatedwith VEGF165.The amounts of KDR-AP that bound to VEGFI65 were quantified byincubation of the plates with AP substrate and reading of Am nm.Data points, means oftriplicate determinations; bars, SD.

Light chain

scFv CDR1 CDR2

p1C11 SASSSVSYMH STSNLAS

p1F12 SASSSVSSSYLH RTSNLAS

p2A6 RASSSVSWH DTSKLAS

p3A1 SASSSVSYMH STSNLAS

p2A11 SASSSVNYMH DTSNLAS

p2A7 RATSSVNYMS DTSNLAS

CDR3

QQRSSYPFT

QQWSDYPFT

QQWSDYPFT

FQGSGYPLT

QQWSSNPWT

QQHGESPLT

Fig. 1. Deduced amino acid sequences [according to the hypervariabledefinition of Kabat el al. (39)] of complementarity-determining regions (CDKIthe VH and VL of the scFv antibodies isolated from the phage display library.

sequenceCDR3) of

Table 1 KDR binding analysis of anti-KDR scFv antibodies

Bindingkinetics'scFv

cloneplCll

plF12p2A6p2A7KDR

binding"

(ED,nM)Yes

(0.3)Yes (1.0)Yes (5.0)No (NA)''VEGF

blocking*(1C,

DM)(1Yes

(0.3)Yes (15)No(>300)No (>300)0>M^.S-'>0.24

4.1NAV-,2.31.446.1

NA(10~9M)2.1

5.911.2NA

' Determined by direct-binding ELISA. Numbers in the parenthesis represent the scFv

concentrations thai give 50% of maximum binding (see also Fig. 2A).b Determined by competitive VEGF-blocking ELISA. Numbers in the parenthesis

represent the scFv concentrations required for 50% inhibition of KDR binding to immobilized VEGF (also see Fig. 2B).

' Determined by BIAcore analysis (see "Materials and Methods").d NA, not applicable.

3211

Research. on September 28, 2014. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

INHIBITION OF VEOF ACTIVITIES BY scFv ANTIBODIES

VEGF

«ePv

+ + +

i i I !^••HBfll

VEGF

scFv

+

5

KDR

ptìo»ptiofyl«Uonp«-

B

bFGF

scFv

p44-p42-

Fig. 3. Inhibition of VEGF-induced phosphorylation of KDR (A and fi) and MAP kinases p44/p42 (C and D) by anti-KDR antibodies. Growth factor-starved early-passage HUVECswere incubated in EBM-2 medium at RT for 15 min in the presence or absence of scFv antibodies (5 f¿g/ml),followed by stimulation with 20 ng/ml VEGF165 or 10 ng/ml bFGF atRT for an additional 15 min. The cells were lysed and the cell lysate used for both the KDR and MAP tunase phosphorylaüon assays. The KDR was immunoprecipitated from thecell lysates with protein A-Sepharose beads coupled to an anti-KDR antibody, Mab 4.13. Proteins were resolved with SDS-PAGE and subjected to Western blot analysis. To detectKDR phosphorylation, blots were probed with an antiphosphotyrosine Mab, PY20. For the MAP kinase activity assay, cell lysates were resolved with SDS-PAGE followed by Westernblot analysis using a phospho-specific MAP kinase antibody. In both assays, the blots were reprobed with an appropriate polyclonal antibody to assure that equal amount of proteinwas loaded on each lane of the SDS-polyacrylamide gels. All signals were detected using enhanced chemiluminescence. A and B, inhibition of VEGF-induced KDR phosphorylation:KDR phosphorylation (A) and KDR protein loaded on each lane in A (B). C, inhibition of VEGF-induced activation of MAP kinases p44/p42. D, inhibition of bFGF-induced activation

of MAP kinases p44/p42.

immobilized VEGF (Fig. 25). The negative control clone, p2A7, didnot bind to KDR or block KDR binding to VEGF (Fig. 2). CloneplCll, the dominant clone after each round of panning, showed thehighest KDR binding capacity and the highest potency in blockingVEGF binding to KDR (Table 1). The antibody concentrations ofclone plCl 1 required for 50% of maximum binding to KDR (Fig. 14)and for 50% of inhibition of KDR binding to VEGF (Fig. 2B) were 0.3nM and 3 nM, respectively (Table 1). Fluorescence-activated cell

sorting analysis demonstrated that plCll,plF12, and p2A6 were alsoable to bind to cell surface expressed receptor on HUVECs (data notshown).

The kinetics of scFv binding to KDR was determined by surfaceplasmon resonance on a BIAcore instrument (Table 1). The VEGF-

blocking scFv antibodies, plCll and plF12, bound to immobilizedKDR with KDs of 2.1 and 5.9 nM, respectively. The nonblocking scFvp2A6 bound to KDR with an ~6-fold weaker affinity (Kd, 11.2 HM)

than the best binder plCl 1, mainly due to a much faster dissociationrate. As anticipated, p2A7 did not bind to the immobilized KDR onthe BIAcore.

Phosphorylation and Mitogenesis Assays. The biological effectsof the anti-KDR scFv antibodies on VEGF-induced KDR phospho

rylation (Fig. 3) and cell mitogenesis (Fig. 4) were determined onHUVECs. In the phosphorylation assay, the levels of KDR phosphorylaüonstimulated by VEGF were assayed in the presence and absence of scFv antibodies. Results showed that VEGF-blocking scFv

p 1C 11 but not the nonblocking scFv p2A6 was able to inhibit KDRphosphorylation stimulated by VEGF (Fig. 3, A and B). Furthermore,plCl 1 also effectively inhibited VEGF-stimulated activation of MAP

kinases p44/p42 (Fig. 3C). In contrast, neither plCll nor p2A6inhibited bFGF-stimulated activation of MAP kinases p44/p42

(Fig. 3D).The ability of scFv antibodies to block VEGF-stimulated mitogenic

activity on HUVECs is shown in Fig. 4. The VEGF-blocking scFv

plCl 1 strongly inhibited VEGF induced DNA synthesis in HUVECswith an EC50, i.e., the antibody concentration that inhibited 50% of

VEGF-stimulated mitogenesis of HUVECs, of ~5 nM. The nonblock

ing scFv p2A6 showed no inhibitory effect on the mitogenic activityof VEGF. Neither plCll nor p2A6 inhibited bFGF-induced DNA

synthesis in HUVECs (data not shown).

Discussion

High-affinity anti-KDR scFv antibodies, which block VEGF bind

ing to KDR, were isolated from a phage display library constructedfrom mice immunized with a soluble form of the human VEGFreceptor. The following are noteworthy in this study: first, over 90%of recovered clones after two rounds of selection are specific to KDR.The binding affinities for KDR of these scFv are in the nanomolarrange, which are as high as those of several bivalent anti-KDR Mabsproduced using hybridoma technology.4 These results indicate high

efficiency of the immunization and/or biopanning process, plus arelatively large size of the library. Second, we show that a scFvantibody, plCl 1, is able to block VEGF-KDR interaction and inhibitVEGF-stimulated receptor phosphorylation and mitogenesis of HU

VECs. To our knowledge, this is the first report on the production ofhigh-affinity anti-KDR antibodies and the first demonstration thatsuch antibodies can specifically inhibit VEGF-stimulated activation of

KDR and mitogenesis of endothelial cells.Several results presented in this study suggest that the mechanism

of endothelial cell growth inhibition mediated by plCll scFv is dueto a direct block of VEGF-KDR interaction, resulting in the inhibitionof VEGF-induced KDR activation. This is based on the observation

that plCll scFv specifically binds to both soluble KDR and cellsurface-expressed KDR on HUVECs; once bound, it blocks VEGFbinding to KDR (Fig. 2). Furthermore, plCll blocks VEGF-induced

KDR phosphorylation (Fig. 3A) at saturating levels of VEGF (19).The scFv also blocks VEGF (but not bFGF)-stimulated activation of

MAP kinases p44/42 (Fig. 3, C and D). These MAP kinases are

4 Unpublished observations.

3212

Research. on September 28, 2014. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

INHIBITION OF VEGF ACTIVITIES BY scFv ANTIBODIES

3000

_ 2800

l 2600

!_ 2400

£ 2200

|T 2000

r;1800l 1600

< 1400

£ 1200-O)S IODO-E

800

6000.01 0.1 1 10

Antibody Concentration ((ig/ml)

Fig. 4. Inhibition of VEGF-induced HUVEC proliferation by scFv antibodies. HU-VECs (5 X 10' cells/well) were plated into 96-well tissue culture plates in 200 ^,1 ofEBM-2 medium without VEGF, bFGF, or epidermal growth factor and incubated at 37°C

for 72 h. Various amounts of soluble scFv were added to duplicate wells and preincubatedat 37°Cfor 1 h, after which VEGF, 65 was added to a final concentration of 16 ng/ml. After18 h of incubation, 0.25 /¿Ciof [3H]thymidine was added to each well and incubated for

an additional 4 h. DNA incorporated radioactivity was determined with a scintillationcounter. •¿�,VEGF only; Q, no VEGF. Data shown are the representative of at least threeseparate experiments.

known to be associated with cell proliferation induced by a number ofmitogens, including VEGF and bFGF (22). Finally, plCll specifically inhibits VEGF-induced but not bFGF-induced DNA synthesis in

HUVECs. On the other hand, p2A6, which binds KDR but does notblock VEGF-KDR interaction, has no effect on VEGF-induced DNA

synthesis in HUVECs (Fig. 4).A growing list of inhibitors of angiogenesis are currently under

evaluation as potential cancer therapeutics (23), including small molecule inhibitors such as linomide (24), AGM-1470 (also known as

TNP 470; Ref. 25), and thalidomide (26). More recently, several smallmolecule Flk-1 inhibitors, including SU1433 and SU1498, were foundto inhibit VEGF-induced endothelial cell proliferation and angiogenesis in the chorioallantoic membrane of chicken embryo (27). An-

giostatin and endostatin were found to suppress tumor growth andmetastasis by inhibiting angiogenesis (28, 29), and reintroduction ofwild-type p53 into glioblastoma cells was found to release an inhibitor

of vessel formation (30). Additionally, Mabs against VEGF (14) andavß3integrin (31) have been shown to inhibit tumor neovasculariza-

tion. There are several theoretical advantages for using antibodiestargeting KDR/Flk-1 on endothelial cells as cancer therapeutics. First,KDR/Flk-1 is expressed exclusively on proliferating endothelial cells

at tumor sites (6, 7, 11, 13); therefore, antibodies against the receptormay offer high specificity compared to other agents. Because tumorvessel endothelial cells are in direct contact with the blood, antibodiesagainst KDR/Flk-1 have greater accessibility to their targets on en

dothelial cells compared to antibodies against markers expressed onindividual tumor cells. Furthermore, conventional targeting therapieswith antibodies and/or their conjugates requires targeting individualtumor cells, and their efficacy is greatly limited by tumor cell heterogeneity. On the contrary, local interruption of tumor vasculature bytargeting molecules expressed on endothelial cells may produce anavalanche of tumor cell death (32). Finally, because endothelial cellspossess a normal complement of chromosomes and are relativelygenetically stable, they should be far less prone to develop resistanceto therapy than tumor cells themselves (29).

We have previously shown that a rat Mab against the mouse Flk-1

receptor, DC 101, effectively inhibits human tumor growth in nudemice by blocking the tumor-associated angiogenesis (15, 16). It is

important to note that, despite high sequence homology betweenmouse Flk-1 and its human homologue KDR, none of the VEGF-blocking anti-KDR scFv antibodies produced in our laboratory cross-react with Flk-1. Consequently, human tumors grown in mice, which

recruit the mouse vasculature, are not appropriate models to evaluatethe antiangiogenesis therapy in vivo of the anti-KDR antibodies described here, because these antibodies do not react with mouse Flk-1

receptor.Apart from cancer, the VEGF/KDR/Flk-1 pathway has been shown

to also play a role in diabetic retinopathy (33), psoriasis (34), heman-gioma (35), rheumatoid arthritis (36), and Kaposi's sarcomas in AIDS

patients (37). Thus, identification of KDR/Flk-1 inhibitors may have

potential application in the treatment of a variety of human diseases inwhich pathological angiogenesis is involved. Here, we demonstratethat antibodies against KDR can specifically block VEGF-KDR interaction and inhibit VEGF-induced KDR activation and DNA syn

thesis in human endothelial cells. Our data, together with previousresults with anti-Flk-1 antibodies (15, 16), lend support to further

evaluation of these antibodies as antitumor agents (38).

Acknowledgments

We thank X. Jimenez, K. King, and W. O'Connor for excellent technical

assistance and Q. Zhou and L. Jacoby for synthesizing oligonucleotides.

References

1. Folkman, J. Angiogenesis in cancer, vascular, rheumatoid and other diseases. Nat.Med., /: 27-31, 1995.

2. Weinstat-Saslow, D., and Steeg, P. S. Angiogenesis and colonization in the tumormetastatic process: basic and applied advances. FASEB J., 8: 401-407, 1994.

3. Liotta, L. A., Steep, P. S., and Stetler-Stevenson, W. G. Cancer metastasis andangiogenesis: an imbalance of negative and positive regulation. Cell, 64: 327-336,

1991.4. Fidler, I. J., and Ellis, L. M. The implications of angiogenesis for the biology and

therapy of cancer metastasis. Cell, 79: 185-188, 1994.5. Klagsbrun, M., and D'Amore, P. A. Regulators of angiogenesis. Annu. Rev. Physiol.,

53: 217-239, 1991.6. Ferrara, N. Vascular endothelial growth factor. Trends Cardiovasc. Med., 3: 244-

250, 1993.7. Brown, L. F., Detmar, M., Claffey, K., Nagy, J. A., Feng, D., Dvorak, A. M., and

Dvorak, H. F. Vascular permeability factor/vascular endothelial growth factor: amultifunctional angiogenic cytokine. In: I. D. Goldberg and E. M. Rosen (eds.).Regulation of Angiogenesis, pp. 233-269. Basel: Birkhauser Verlag, 1997.

8. Leung, D. W., Cachianes, G., Kuang, W. J., Goeddel, D. V., and Ferrara, N. Vascularendothelial growth factor is a secreted angiogenic mitogen. Science (WashingtonDC), 246: 1306-1309, 1989.

9. Millauer, B.. Wizigmann-Voos, S., Schnurch, H., Martinez, R., Moller, N. P. H.,Risau, W., and Ullrich, A. High-affinity VEGF binding and developmental expressionsuggest Flk-1 as a major regulator of vasculogenesis and angiogenesis. Cell, 72:835-846, 1993.

10. Quinn, T. P., Peters, K. G., De Vries, C, Ferrara, N., and Williams, L. T. Fetal liverkinase 1 is a receptor for vascular endothelial growth factor and is selectivelyexpressed in vascular endothelium. Proc. Nati. Acad. Sci. USA, 90: 7533-7537,1993.

11. Plate, K. H., Breier, G.. Millauer, B., Ullrich, A., and Risau, W. Up-regulation ofvascular endothelial growth factor and its cognate receptors in a rat glioma model oftumor angiogenesis. Cancer Res., 53: 5822-5827, 1993.

12. Millauer, B.. Shawver, L. K., Plate, K. H., Risau. W., and Ullrich. A. Glioblastomagrowth inhibited in vivo by a dominant-negative Flk-1 mutant. Nature (Lond.), 367:576-579, 1994.

13. Terman, B. I., Dougher-Vermazen, M., Carrion, M. E., Dimtrov, D., Armellino, D. C,

Gospodarowicz, D., and Bohlen, P. Identification of the KDR tyrosine kinase receptoras a receptor for vascular endothelial growth factor. Biochem. Biophys. Res. Commun., 187: 1579-1586, 1992.

14. Kim. J., Li, B., Winer, J., Armanini, M., Gillett, N., Phillip, H. S., and Ferrara, N.Inhibition of vascular endothelial growth factor-induced angiogenesis suppressestumor growth in vivo. Nature (Lond.), 362: 841-844, 1993.

15. Rockwell, P., and Goldstein, N. Role of protein tyrosine kinase receptors in cancer:possibilities for therapeutic intervention. Mol. Cell. Differ., 3: 315-335, 1995.

16. Skobe, M., Rockwell, P., Goldstein, N., Vosseler, S., and Fusening, N. E. Haltingangiogenesis suppresses carcinoma cell invasion. Nat. Med., 3: 1222-1227, 1997.

17. Saleh, M., Stacker, S. A., and Wilks, A. F. Inhibition of growth of C6 glioma cells invivo by expression of antisense vascular endothelial growth factor sequence. CancerRes., 56: 393-401, 1996.

3213

Research. on September 28, 2014. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

INHIBITION OF VEGF ACTIVITIES BY scFv ANTIBODIES

18. Ramakrishnan, S., Olson, T. A., Baulch, V. L., and Mohanraj, D. Vascular endothelialgrowth factor-toxin conjugate specifically inhibits KDR/flk-1-positive endothelialcell proliferation in vitro and angiogenesis in vivo. Cancer Res., 56: 1324-1330,

1996.19. Rockwell, P., Neufeld, G., Classman, A., Carón, D., and Goldstein, N. In vitro

neutralization of vascular endothelial growth factor activation of Flk-1 by a monoclonal antibody. Mol. Cell. Differ., 3: 91-109, 1995.

20. Tessler, S., Rockwell, P., Hicklin, D., Cohen, T., Levi, B. Z.. Witte, L., Lemischka,I. R., and Neufeld, G. Heparin modulates the interaction of VEGF 165 with solubleand cell associated Flk-t receptor. J. Biol. Chem., 269: 12456-12461, 1994.

21. Flanagan, J. G., and Leder, P. The kit ligand: a cell surface molecule altered in steelmutant fibroblasts. Cell, 63: 185-194, 1990.

22. D'angelo, G., Struman, I., Martial, J., and Weiner, R. I. Activation of mitogen-

activated protein kinases by vascular endothelial growth factor and basic fibroblastgrowth factor in capillary endothelial cells is inhibited by the antiangiogenic factor16-kDa N-terminal fragment of prolactin. Proc. Nati. Acad. Sci. USA, 92: 6374-

6378, 1995.23. Fan, T-P. D., Jaggar, R., and Bicknell, R. Controlling the vasculature: angiogenesis,

anti-angiogenesis and vascular targeting of gene therapy. Trends Pharmacol. Sci., 16:57-66, 1995.

24. Vukanovic, J., Passantii, A., Hirata, T., Traystman, R. J., Hartley-Asp, B., and Isaacs,J. T. Antiangiogenesis effects of the quinoline-3-carboxamide linomide. Cancer Res.,53: 1833-1837, 1994.

25. Yamamoto, T., Sudo, K., and Fujita, T. Significant inhibition of endothelial cellgrowth in tumor vasculature by an angiogenesis inhibitor, TNP-470 (AGM-1470).Anticancer Res., 14: 1-4, 1994.

26. D'Amato, R. J., Loughnan, M. S., Flynn, E., and Folkman, J. Thalidomide is an

inhibitor of angiogenesis. Proc. Nati. Acad. Sci. USA, 91: 4082-4085, 1994.

27. Strawn, L. M., McMahon, G., App, H., Schreck, R., Kuchler, W. R., Longhi, M. P.,Hui, T. H., Tang, C., Levitzki, A., Gazit, A., Chen, I., Keri, G., Orfi, L., Risau, W.,Flamme, I., Ullrich, A., Hirth, K. P., and Shawver, L. K. Flk-1 as a target for tumorgrowth inhibition. Cancer Res., 56: 3540-3545, 1996.

28. O'Reilly, M. S., Holmgren, L., Shing, Y., Chen, C., Rosenthal, R. A., Moses, M.,

Lane, W. S., Cao, Y., Sage, E. H., and Folkman, J. Angiostatin: a novel angiogenesisinhibitor that mediates the suppression of metastasis by a Lewis lung carcinoma. Cell,79: 315-328, 1994.

29. Boehm, T., Folkman, J., Browder, T., and O'Reilly, M. S. Antiangiogenic therapy of

experimental cancer does not induce acquired drug resistance. Nature (Lond.), 390:404-407, 1997.

30. Van Meir, E. G., Polverini, P. J., Chazin, V. R., Huang, H-J. S., de Tribolet, N., and

Cavenee, W. K. Release of an inhibitor of angiogenesis upon induction of wild typep53 expression in glioblastoma cells. Nat. Genet., 8: 171-176, 1994.

31. Brooks, P. C., Montgomery, A. M. P., Rosenfeld, M., Reisfeld, R. A., Hu, T., Klier,G., and Cheresh, D. A. Integrin av/33 antagonists promote tumor regression byinducing apoptosis of angiogenic blood vessels. Cell, 79: 1157-1164, 1994.

32. Huang, X., Molema, G., King, S., Watkins, L., Edgington, T. S., and Thorpe, P. E.Tumor infarction in mice by antibody-directed targeting of tissue factor to tumorvasculature. Science (Washington DC), 275: 547-550, 1997.

33. Pierce, E. A., Avery, R. L., Foley, E. D., Aiello, L. P., and Smith, L. E. H. Vascularendothelial growth factor/vascular permeability factor expression in a mouse model ofretinal neovascularization. Proc. Nati. Acad. Sci. USA. 92: 905-909, 1995.

34. Detmar, M., Brown, L. F., Claffey, K. P., Yeo, K-T.. Kocher, O., Jackman, R. W.,

Berse, B., and Dvorak, H. F. Overexpression of vascular permeability factor/vascularendothelial growth factor and its receptors in psoriasis. J. Exp. Med., 180: 1141-1146,

1994.35. Wizigmann-Voos, S., Bieier, G., Risau, W., and Plate. K. H. Up-regulation of

vascular endothelial growth factor and its receptor in von Hippel-Lindau disease-associated and sporadic hemangioblastomas. Cancer Res., 55: 1358-1364, 1995.

36. Koch, A. E., Harlow, L. A., Haines, G. K., Amento, E. P., Unemori, E. N., Wong,W. L., Pope, R. M., and Ferrara, N. Vascular endothelial growth factor. A cytokinemodulating endothelial function in rheumatoid arthritis. J. Immunol., Õ52:4149-

4156, 1994.37. Albini, A., Soldi, R., Giunciuglio, D., Giraudo, E.. Benelli, R., Primo, L., Noonan, D..

Salio, M., Camussi, G., Rockl, W., and Bussolino, F. The angiogenesis induced byHIV-1 tat protein is mediated by the Flk-1/KDR receptor on endothelial cells. Nat.Med., 2: 1371-1375, 1996.

38. Witte, L., Hicklin, D. J., Zhu, Z., Pytowski, B., Kotanides, H., Rockwell, P., andBohlen, P. Monoclonal antibodies targeting the VEGF receptor-2 (Flk-1/KDR) as ananti-angiogenic therapeutic strategy. Cancer Metastasis Rev., in press, 1998.

39. Kabat, E. A., Wu, T. T., Perry, H. M., Gottesman, K. S., and Foeller, C. Sequencesof Proteins of Immunological Interest. Ed. 5. Bethesda, MD: NIH, 1991.

3214

Research. on September 28, 2014. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from