Therapeutics Context-Dependent Role of Angiopoietin-1 ......AMG 386 is an investigational first-in-class peptide-Fc fusion protein (peptibody) that inhibits angiogenesis by preventing

Ther

ConSupfor

AngelJi-RonDavidEunjuDonal

Abst

Intro

Angegy toon necertaiage-recondi

AuthorBiologyand Dr5CardioDepartCaliforn

Note: SCancer

CurrentYongin

Current

CurrentDiego,

www.a

Do

Molecular
Cancer
apeutics

apeutic Discovery

text-Dependent Role of Angiopoietin-1 Inhibition in thepression of Angiogenesis and Tumor Growth: Implications

Ther

AMG 386, an Angiopoietin-1/2–Neutralizing Peptibody

a Coxon1, James Bready1, Hosung Min3, Stephen Kaufman2, Juan Leal1, Dongyin Yu1, Tani Ann Lee1,g Sun1, Juan Estrada1, Brad Bolon2, James McCabe2, Ling Wang1, Karen Rex1, Sean Caenepeel1, Paul Hughes1,Cordover2, Haejin Kim3, Seog Joon Han3, Mark L. Michaels3, Eric Hsu3, Grant Shimamoto3, Russell Cattley2,
Hurh4, Linh Nguyen4, Shao Xiong Wang4, Anthony Ndifor2, Isaac J. Hayward2, Beverly L. Falcón5, d M. McDonald5, Luke Li3, Tom Boone3, Richard Kendall1, Robert Radinsky1, and Jonathan D. Oliner1
ractAM

by preapeutipotentwe hator, TiediseasovariahibitocontexAMGgraft geffectretinaldual A

lated mtions (e.g

s' Affiliationand Safety

ug Metabovascular Rement of Aia

upplementTherapeuti

address fo-si, Gyeong

address for

addressCA.

acrjourna

wnloade

G 386 is an investigational first-in-class peptide-Fc fusion protein (peptibody) that inhibits angiogenesisventing the interaction of angiopoietin-1 (Ang1) and Ang2 with their receptor, Tie2. Although the ther-c value of blocking Ang2 has been shown in several models of tumorigenesis and angiogenesis, theial benefit of Ang1 antagonism is less clear. To investigate the consequences of Ang1 neutralization,ve developed potent and selective peptibodies that inhibit the interaction between Ang1 and its recep-2. Although selective Ang1 antagonism has no independent effect in models of angiogenesis-associatedes (cancer and diabetic retinopathy), it induces ovarian atrophy in normal juvenile rats and inhibitsn follicular angiogenesis in a hormone-induced ovulation model. Surprisingly, the activity of Ang1 in-rs seems to be unmasked in some disease models when combined with Ang2 inhibitors, even in thet of concurrent vascular endothelial growth factor inhibition. Dual inhibition of Ang1 and Ang2 using386 or a combination of Ang1- and Ang2-selective peptibodies cooperatively suppresses tumor xeno-rowth and ovarian follicular angiogenesis; however, Ang1 inhibition fails to augment the suppressiveof Ang2 inhibition on tumor endothelial cell proliferation, corneal angiogenesis, and oxygen-inducedangiogenesis. In no case was Ang1 inhibition shown to (a) confer superior activity to Ang2 inhibition orng1/2 inhibition or (b) antagonize the efficacy of Ang2 inhibition. These results imply that Ang1 plays at-dependent role in promoting postnatal angiogenesis and that dual Ang1/2 inhibition is superior to
contex
selective Ang2 inhibition for suppression of angiogenesis in some postnatal settings. Mol Cancer Ther; 9(10);

2641–51. ©2010 AACR.

psoriacurrenistratisunittreatmfactor

duction

iogenesis inhibition holds great promise as a strat-treat diseases in which progression is dependent

ovascularization. These illnesses include cancer,n ocular diseases (e.g., diabetic retinopathy and

acular degeneration), and inflammatory., rheumatoid arthritis, osteoarthritis, and

sistanbeen

s: Departments of 1Oncology Research, 2ComparativeSciences, 3Protein Science, and 4Pharmacokinetics

lism, Amgen, Inc., Thousand Oaks, California; andsearch Institute, Comprehensive Cancer Center, andnatomy, University of California, San Francisco,

ary material for this article is available at Molecularcs Online (http://mct.aacrjournals.org/).

r H. Min: Samsung Advanced Institute of Technology,gi-do, South Korea.

J. Leal: Biopharmacology Consulting, Minneapolis, MN.

for T.A. Lee: Prometheus Laboratories, Inc., San

Current

Current

Current

Current

CurrentChina.

CorresCenter805-44

doi: 10

©2010

ls.org

on August 27, 2021. © 20mct.aacrjournals.org d from

sis; refs. 1, 2). The four antiangiogenic therapiestly approved by the U.S. Food and Drug Admin-on [bevacizumab (Avastin), sorafenib (Nexavar),inib (Sutent), and pazopanib (Votrient)] for theent of cancer all target vascular endothelial growth(VEGF) or its receptors. Despite clinical benefit, re-
ce is common, and class-specific side effects haveseen (3). These side effects include hypertension,
address for B. Bolon: GEMpath, Inc., Longmont, CO.

address for L. Nguyen: Exelixis, Inc., South San Francisco, CA.

address for A. Ndifor: Johnson and Johnson, San Diego, CA.

address for I.J. Hayward: HaywardHistopathology, PC, Conyers, GA.

address for S.X. Wang: Genor Biopharma Co. Ltd., Shanghai,

ponding Author: Jonathan D. Oliner, Amgen, Inc., One AmgenDrive, Thousand Oaks, CA 91320. Phone: 805-447-3066; Fax:7-1982. E-mail: [email protected]

.1158/1535-7163.MCT-10-0213

American Association for Cancer Research.

2641

10 American Association for Cancer Research.

http://mct.aacrjournals.org/

thromtions,encinsafetydrugstientsbe usthe syat the(4), an(intramacujectionand reThe

be facof thibiochreplacneovanifica(Ang)involvof whincludknowAng4genestumoAng2tion (7humahavepooreAng

genestion iproanpostnpostnwidespromrodenleakatumohanceaccomlatedmeasuvesselcell pand ptransfgraft(22).suppr(23). Lin mo

ThechallesionsTo cirinconsuprato invenousneutrcombof poDet

has imtargetationCVX-hibito(http:well aonly oVEGF386, aproteneutrconceref. 7]2xConand ton cucornegethemechtientsof AMtoleraassocidid notestinmentreducized ppaclitdepen125 reinhibiisingattrib

Mate

PhagAng1A f

109 into selewas c

Coxon et al.

Mol C2642

Do

boemboli, hemorrhage, gastrointestinal perfora-and wound dehiscence, with some patients experi-g serious or life-threatening events. Although theprofile of systemically administered VEGF-targetedis acceptable in selected populations of cancer pa-, these characteristics are unfavorable for agents toed in chronic nononcology indications. To avoidstemic safety issues and maximize drug exposuresite of action, two VEGF antagonists, ranibizumabd pegaptanib (5), have been developed for local

vitreal) administration in patients with age-relatedlar degeneration. However, repeated intraocular in-s are required, conferring a risk for endophthalmitistinal detachment.expanded use of antiangiogenic therapies wouldilitated by improvements in the efficacy and safetys class of agents. To this end, intervention in otheremical pathways may be necessary to augment ore VEGF pathway inhibition in the suppression ofscularization. One pathway that has received sig-nt attention in this regard is the angiopoietinaxis. Like the VEGF pathway, the Ang pathwayes a receptor tyrosine kinase, Tie2, the expressionich is restricted to a limited number of cell types,ing the vascular endothelium (6). Tie2 binds threen secreted ligands in humans: Ang1, Ang2, and. Ang2 is upregulated at sites of postnatal angio-is, and Ang2 inhibitors suppress angiogenesis andr growth in preclinical models, suggesting thatis a key regulator or mediator of neovessel forma-). Ang2 is upregulated in the vasculature of manyn tumors, and higher levels of Ang2 upregulationbeen associated with more advanced disease andr prognosis (8–10).1 plays an important role in developmental angio-is (11), but its function in postnatal neovasculariza-s less clear. Ang1 has been shown to mediategiogenic and antiangiogenic effects in variousatal settings (12). In the first few weeks of rodentatal life, systemically administered Ang1 inducedpread circumferential venous enlargement byoting endothelial cell proliferation (13). In adultts, Ang1 made blood vessels resistant to plasmage and induced lymphatic sprouting (14, 15). Inr xenografts, ectopic overexpression of Ang1 en-d (16, 17) and retarded (18–21) tumor growth. Inpanying tumor angiogenesis studies, Ang1 stimu-(16, 17) and inhibited (18) neovascularization asred by assessments of vessel numbers (16–18),diameter (19), vessel branching (17), endothelialroliferation (18), pericyte coverage (16, 17, 21),erfusion (17). In a loss-of-function study, stablyected Ang1 antisense RNA reduced tumor xeno-growth and decreased tumor microvessel densityIn conditional transgenic mice, Ang1 inductionessed choroidal and retinal neovascularization
astly, Ang1 overexpression increased angiogenesisdels of ischemia (24, 25).
96-weon en

ancer Ther; 9(10) October 2010

on August 27, 2021. © 20mct.aacrjournals.org wnloaded from

conflicting results of these studies highlight thenges encountered in trying to draw general conclu-about the role of Ang1 in postnatal angiogenesis.cumvent the possibility that some of the observedsistencies may be a consequence of administeringphysiologic levels of exogenous Ang1, we choseestigate the function of Ang1 by inhibiting endog-Ang1. To that end, we have developed Ang1-

alizing peptibodies and tested them alone or inination with Ang2 inhibitors in preclinical modelsstnatal angiogenesis.ermining the consequences of Ang1 neutralizationportant therapeutic implications. Four biologics

ing the Tie2 axis are currently under clinical evalu-in oncology (AMG 386, AMG 780, CVX-060, and241), but only AMG 386 and AMG 780 are dual in-rs of Ang1 and Ang2. CVX-060 (26) and CVX-241//clinicaltrials.gov/ct2/show/NCT01004822), ass a preclinical-stage biologic called 3.19.3 (27), havene Ang target, Ang2 (although CVX-241 also targets). Of these five agents, the most advanced is AMGn investigational recombinant peptide-Fc fusionin (peptibody) that reduces tumor angiogenesis byalizing Ang1 and Ang2 [Half maximal inhibitoryntration (IC50) of 0.9 and 0.023 nmol/L, respectively;. In preclinical studies, AMG 386 [previously named4(C)] blocked tumor endothelial cell proliferationumor xenograft growth in mice but had no effectltured tumor cells. In addition, AMG 386 suppressedal angiogenesis in a VEGF-driven rat model (7). To-r, these data are consistent with an antiangiogenicanism of action (7). In a first-in-human study in pa-with advanced solid tumors, weekly administrationG 386 seemed to be well tolerated (a maximum-ted dosewas not reached; ref. 28). Toxicities typicallyatedwith inhibitors of the VEGF pathway (29) eithert occur (bleeding, thromboembolic events, gastroin-al perforations) or were not considered to be treat-related. AMG 386 mediated antitumor efficacy anded tumor blood flow/permeability (28). In a random-hase 2 study of AMG 386 combined with weeklyaxel in patients with recurrent ovarian cancer, dose-dent increases in progression-free survival and CA-sponse were seen (30). These data suggest that dualtion of Ang1 and Ang2 by AMG 386 confers prom-clinical activity, but the component of this activityutable to Ang1 inhibition remains to be evaluated.

rials and Methods

e display selection and characterization of-binding peptidesilamentous phage peptide library containing 2.3 ×dependent transformants (Dyax Corp.) was usedct for Ang1-binding phage. Phage display selectionarried out as described using Ang1 protein-coated
ll plates (7). Peptide sequences were selected basedzyme-linked immunosorbent assay (ELISA) results
Molecular Cancer Therapeutics



and DpeptibmogeweregenerlibrarindepsimilaPep

descrifollowsequeDAMTEEMMe

mentsare dpharm386 (7each amolarpharm

Ang:Ttime-The

usedthe sudetermeach(no A(no pewereXLfit4

Ang:TThe

plates1 nmocal deMax pAng:Tparisoof seriXLfit.

AnimAll

Care amentdardsLabor

TumoFem

or 2 ×one-thwere

treatmTumotwicefashioheighrepeaDunnmentBonfetweeninhibanaly(termfrompointbeforeFor thvolumANOAng2For thlyzedday, aas fixrandothe ni

HistoTum

werein colby stawereCD31usingly couvessecalcufractioVia

termiref. 33oldingNikonor uspixelexpremor,fractiotological anfollowBas

membBriefl1% pa

Ang1 Mediates Context-Dependent Postnatal Angiogenesis

www.a

Do

NA sequencing of phage clones and expressed in aody format. Peptibodies were evaluated in a ho-neous time-resolved fluorescence assay, and severalchosen for affinity maturation, which was done byating and panning nucleotide-doped phage displayies (31). These focused libraries, with >1 × 109

endent transformants, were panned by a procedurer to that used for panning the primary library.tibody mL4-3 was expressed and purified asbed (7). The amino acid sequence of mL4-3 is ass, in which Fc denotes the human IgG1 Fc

nce as described previously (7): MREWTEEMQVIF-MFGPRNDRGGSGSATGSGSTASSGSGSATHREW-QVIFDAMMFGPRNDRGGGGG-Fc.thodology and results for pharmacokinetic assess-of mL4-3 in CD-1 mice and Sprague-Dawley rats

escribed in the supplementary data. Because theacokinetic properties of mL4-3, L1-7, and AMG) were dissimilar, the dose levels and schedules ofgent were chosen, where possible, to achieve equi-serum steady-state Cmin concentrations withinacology studies.

ie2 neutralization homogeneousresolved fluorescence assayhomogeneous time-resolved fluorescence assay

to evaluate Ang:Tie2 neutralization is detailed inpplementary data. Neutralization potency wasined by calculating the percentage inhibition of

peptibody dilution in reference to the maximumng in the assay mixture) and minimum inhibitionptibody in the assay mixture) controls. IC50 valuescalculated by plotting percentage inhibition using, inwhich fit =A+ {(B −A) / [1 + (C /X)^D]} (IDBS).

ie2 neutralization ELISAELISAwas done as described (7) using Ang-coatedand mL4-3 or Fc serially diluted in a solution ofl/L Tie2-Fc/1% bovine serum albumin/PBS. Opti-nsity was measured at 650 nmol/L using Spectra-late reader (Molecular Devices), and the degree ofie2 neutralization (IC50) was determined by com-n against a Tie2 standard curve (the binding activityally diluted Tie2 in the absence of competitor) using

al studiesprocedures were approved by the Amgen Animalnd Use Committee and met Association for Assess-and Accreditation of Laboratory Animal Care stan-. All animals were purchased from Charles Riveratories.

r xenograft modelsale nude mice were injected s.c. with 5 × 106 HT-29106 Colo205 human colon cancer cells mixed with
ird volumeMatrigel (BD Biosciences). Once tumorsestablished, animals were randomly assigned to
Cuttinstat s

acrjournals.org


ent groups (n = 10 per group except where noted).r measurements and body weights were recordedper week. All tumor studies were done in a blindedn. Tumor volumewas calculated as length ×width ×t in cubic millimeters. Data were analyzed usingted measures ANOVA followed by a Scheffé orett's post hoc test. Terminal tumor volume assess-s were analyzed using ANOVA followed by arroni-Dunn post hoc test. To assess differences be-selective Ang2 inhibition and combined Ang1/2

ition across multiple Colo205 studies, two meta-ses were done using either terminal tumor volumesinal-time-point meta-analysis) or all tumor volumesthe initiation to the termination of dosing (all-time-meta-analysis). The data were log transformedthe statistical analysis for each individual study.e terminal-time-point meta-analysis, final tumores for each study were analyzed using a one-wayVA with group (dual Ang1/2 inhibition, selectiveinhibition, and negative control) as a fixed effect.e all-time-point meta-analysis, the data were ana-using a mixed-effect model that included group,nd the two-way interaction between day and timeed effects and subject as a random effect. Them-effects model was used to pool the results fromne studies in a meta-analysis.

logic tumor analysesor blood vessel area. Blood vessel area assessmentsdone on Colo205 tumors that were immersion-fixedd zinc Tris solution (32) and then paraffin embeddedndard methods (n = 9 or 10 per group). Sectionsimmunostained for vascular endothelium (anti-antibody MEC 13.3; BD Biosciences Pharmingen)3,3'-diaminobenzidine as the chromogen and light-nterstained with hematoxylin. The total bloodl area (square millimeter) for every section waslated (viable tumor area × respective vessel arean). Details are outlined in the supplementary data.ble tumor burden. Colo205 tumor viability was de-ned histologically as described (n = 10 per group;). Viable tumor area was analyzed by RGB thresh-(using a Nikon DXM1200 camera mounted on aFXA compound microscope with a 1× objective

ing Aperio virtual slide scanning) and automatedcounting (Visiopharm Integrator System) and wasssed as a fraction of total tumor area. For each tu-the tumor burden (gram) was calculated as viablen × corresponding terminal tumor weight. All his-c analyses were done in a blinded fashion. Statisti-alyses of histologic data were done by ANOVA,ed by a Bonferroni/Dunn post hoc test.ement membrane histology. Empty basementrane sleeves were evaluated as described (34).

y, tissues were collected from mice perfused withraformaldehyde and frozen in Tissue-Tek Optimal
g Temperature compound (Sakura Finetek). Cryo-ections (80 μm) were stained for endothelial cells
Mol Cancer Ther; 9(10) October 2010 2643



(anti-basemCosm510 co

TumoTum

descrisize, Ctreate386 (6mL4-mL4-singlean un

CornVE

inducweek-(35).(5 mg7(N) agrouption awas tdescrivesseldisc adoneassess

RetinIsch

as desper gdata.(N) ((100 mstartinStatisFisher

OvarSup

standmentaeithervascuclonaldoneanalypost h

EvalutreateSpr

L1-7(N

imalsand etoxyliAll

softw

Resu

IdentAng1An

mL4-3ited sdisplaSuppmeterTableAng2(C)],and ahad aturedmenta

mL4-Ang1To

questcontrochallephospan effL1-7(NNex

neutrAng1Devethat Afoldinsystehad rbut leos (Supplementary Fig. S3B; Supplementary Table S3).Unlike

Table b yTie2 n

Agent 1 2

L1-7(NmL4AMFc

Coxon et al.

Mol C2644

Do

CD31; Clone 2H8, 1:500; Thermo Scientific) andent membrane (anti-type IV collagen; 1:8,000;o Bio Co. Ltd.) and examined using a Zeiss LSMnfocal microscope.

r endothelial cell proliferation assayor endothelial cell proliferation was assayed asbed (7). Briefly, when tumors were ∼500 mm3 inolo205 tumor-bearing mice (n = 3 per group) wered s.c. for 3 days with Fc (5.7 mg/kg daily), AMGmg/kg single dose), L1-7(N) (2.2 mg/kg daily),

3 (3.5 mg/kg daily), or L1-7(N) combined with3 (at the same doses and schedules used in the-agent groups). Statistical analysis was done usingpaired t test.

eal angiogenesis modelGF- and basic fibroblast growth factor (bFGF)-ed angiogenesis studies were done in 8- to 12-old female CD rats (n = 8 per group) as describedTreatment (i.v.) with Fc (60 mg/kg), L1-7(N)/kg), mL4-3 (60 mg/kg), or the combination of L1-ndmL4-3 (at the same doses used in the single-agents) was initiated on the day before corneal implanta-nd continued on days 3 and 6. On day 8, the studyerminated, and the corneas were photographed asbed (7). For each corneal image, the number of bloods intersecting the midpoint between the implantednd the limbus was counted. All evaluations werein a blinded fashion. Statistical significance wased by ANOVA followed by Fisher's post hoc test.

al neovascularizationemic retinopathy was induced in C57BL/6J micecribed (35) using postnatal day 7 (P7) pups (n = 7roup). Details are outlined in the supplementaryFc control (200 mg/kg), mL4-3 (100 mg/kg), L1-7100 mg/kg), or mL4-3/L1-7(N) combinationg/kg each) was administered s.c. daily for 9 daysg on P8. All counts were done in a blinded fashion.tical analysis was done by ANOVA followed by's post hoc test.

ian follicular angiogenesiserovulation was induced in C57BL/6J mice usingardmethodology; details are provided in the supple-ry data. Fixed sections of ovaries (32) were stainedwith hematoxylin and eosin or immunostained forlar endothelium (anti-CD31; rat anti-mouse mono-MEC 13.3). Two replicates of this experiment wereon different days (n = 8 to 10 per group). Statisticalsis was done by ANOVA, followed by Dunnett'soc test.

ation of ovaries and epiphyseal plates ind rats
ague-Dawley rats received 300 mg/kg of AMG 386,), or mL4-3 i.v. twice weekly for 28 days (n = 10 an-
Abb



per sex per group). At scheduled necropsy, ovariespiphyseal plates were sectioned, stained with hema-n and eosin, and observed for microscopic changes.statistical analyses were done using StatView 5.0.1are (SAS Institute, Inc.).

lts

ification and characterization of-selective inhibitorsg1-neutralizing peptibodieswere generated and one,, was chosen for use in these studies. mL4-3 exhib-imilar potency against several Ang1 orthologs andyed >40,000-fold selectivity over Ang2 (Table 1;lementary Table S1). The pharmacokinetic para-s of mL4-3 in rodents are shown in SupplementaryS2. Peptibodies L1-7(N), a potent and selectiveinhibitor, and AMG 386 [also known as 2xCon4a dual Ang1/2 inhibitor, have been described (7)re shown in Table 1. None of these Ang inhibitorsny effect on the proliferation or permeability of cul-human umbilical vein endothelial cells (Supple-ry Figs. S1 and S2).

3 can be used as a reagent for interrogatingfunction in vivo

assess whether mL4-3 was capable of selectively se-ering Ang1 in vivo, mL4-3, L1-7(N), and Fc (negativel) were administered s.c. to mice, followed by an i.v.nge with recombinant Ang1. Ang1 induced Tie2horylation in mouse lung endothelium (∼5-fold),ect that could be prevented by mL4-3, but not by) or Fc (Supplementary Fig. S3A).t, we wanted to determine whether mL4-3 couldalize endogenous Ang1 in a setting in whichwas known to play a physiologically relevant role.lopmental genetic knockout studies have shownng1 deletion reduces cardiac size and endocardialg in embryos (11). E12.5 mouse embryos exposedmically to mL4-3 at early and middle gestationeduced cardiac size and trabeculation similar toss dramatic than that observed in Ang1-null embry-

re

10

Ang1 knockout embryos, which display a less

1. Peptiinteractio

viation: h, hu

American A

odies competitivels

man.

Molecular Cance

ssociation for Cancer

inhibit Ang:

hAng
IC50 (nmol/L) hAng IC50 (nmol/L)
)
>10,000 0.064 0.022 3,085 -3
G 386 6.2 0.029>10,000 >10,000

r Therapeutics

Research.


comppartsserve(datamL4-3subopAng1embrAng1mL4-3

Ang1supprIn a

minis

growtmiceand Lmedia(P = 0fers btion agroupand Lgroupgroupboth gby eit

FigureAng1 agrowthxenogrColo20either Aper weeL1-7(N)(3.2 mgof L1-7same dsingle-anecessaddedof protecombinViable tblood vdetermpeptibomean vversus§, P <versusmL4-3.images(CD31;membrAfter 26mL4-3,were acmembrabnormvesselscompa(regresstreated(N) wasof vessendothempty b(arrow).


www.a

Do

lex vascular network than their wild-type counter-(11), no extracardiac vascular changes were ob-d by light microscopy in mL4-3 treated embryosnot shown). The less pronounced phenotype of thetreated embryos may have been a consequence oftimal embryonic mL4-3 exposures and incompletesequestration. Nonetheless, mL4-3 clearly inducedyonic cardiac defects that phenocopy those ofgenetic knockout mice, confirming the utility ofas a reagent for investigating Ang1 function in vivo.

antagonism augments Ang2 antagonism inessing tumor growth
previous report, we showed that systemically ad-tered L1-7(N) and AMG 386
had n

Scale bar, 15 μm.

acrjournals.org


h of Colo205 tumor xenografts implanted into nude(7). In that study, the antitumor effects of AMG 3861-7(N) seemed to be similar, although AMG 386ted modestly superior efficacy to that of L1-7(N).006). To confirm that dual Ang1/2 inhibition con-etter tumor growth suppression than Ang2 inhibi-lone, a similar experiment was done, but this time,s treated with mL4-3 or a combination of mL4-31-7(N) were also tested. The AMG 386 treatmentand the mL4-3/L1-7(N) combination treatmentshowed comparable antitumor efficacy; moreover,roups exhibited efficacy superior to that mediatedher L1-7(N) or mL4-3 alone (Fig. 1A). In fact, mL4-3
o discernable single-agent effect on tumor growth,
could inhibit the implying that combining Ang2 antagonism with Ang1

1. The effect of combinednd Ang2 inhibition on theof Colo205 tumorafts. A, mice implanted with5 cells were treated withMG 386 (5.6 mg/kg twicek), Fc control (5.2 mg/kg),(2.0 mg/kg), or mL4-3/kg) daily; or the combination(N) and mL4-3 (at theosing regimens used in thegent groups). Whereary, Fc control protein wasto match the total amountin delivered in theation group (5.2 mg/kg).umor burden (B) and totalessel area (C) wereined following 7 days ofdy administration. Data arealues ± SE. *, P < 0.05L1-7(N), mL4-3, and Fc;0.05 versus Fc; †, P < 0.01Fc; ‡, P = 0.02 versusD, confocal microscopicof tumor endothelial cellsgreen) and basementane (type IV collagen; red).days of treatment with Fc,or L1-7(N), tumor vesselscompanied by basementane with varyingalities (arrowheads), but theare continuous. By

rison, a discontinuousing) vessel in a tumorwith both mL4-3 and L1-7observed (right). The regionel regression, which lackselial cells, is spanned by anasement membrane sleeve




antagAng1dual Aadditithese,ited smted wshowntisticatermixenogbetweaverainhibimeta-volumof dotweenwhethmodeinhibihibitiorepresFig. S4periortwo ein theno acwas teAd

donehibitiocomb(Fig.tendebitorseithersion o(CD31treatmstaininIn thetumosleeveL1-7(Nemptytive ofindinfluoreagainsstainidetecnot shand dTumofeatur(37) aAng1

tumopericyWe

Ang aor bevgrowtdose land Bmoreconteof L1-subopBothAMGactivitimement(Fig. 2tion ising Cactivi

Ang2inhibangioWe

supprin vivothroution owithwith tsectiothelia(Fig. 3fect oaloneeffectsgrowtof endderlyiantagThe

in thecular,secontranspels ofof AnAs obangioon Andrawnbodie(39) inin am

Coxon et al.

Mol C2646

Do

onism may have unmasked the antitumor effect ofinhibition (Fig. 1A). Selective Ang2 inhibitors andng1/2 inhibitors were tested head-to-head in sevenonal Colo205 tumor xenograft studies, and in six ofthe groups treated with the dual inhibitors exhib-aller terminal tumor volumes than the groups trea-

ith the selective Ang2 inhibitors (ref. 36 and data not). In some cases, these differences did not reach sta-l significance. However, a meta-analysis comparingnal tumor volumes over all nine Colo205 tumorraft studies showed a highly significant differenceen these two treatments (P = 0.0018; 26% greaterge tumor growth suppression with dual Ang1/2tion than with Ang2 inhibition alone). A separateanalysis of these studies, comparing all tumores from the initiation of dosing to the terminationsing, yielded similarly significant differences be-these two treatments (P < 0.0002). To determineer these findings were specific to the Colo205l, we did experiments comparing selective Ang1tion, selective Ang2 inhibition, and dual Ang1/2 in-n in the HT-29 colon tumor xenograft model. Twoentative experiments are shown (Supplementary). Dual Ang1/2 inhibition mediated statistically su-efficacy to selective Ang2 inhibition in one of thexperiments and similar efficacy to Ang2 inhibitionother experiment. Selective Ang1 inhibition had

tivity in any of the HT-29 experiments in which itsted.ditional Colo205 tumor xenograft studies wereto evaluate the histologic consequences of Ang in-n. Treatment with AMG 386 or the mL4-3/L1-7(N)ination induced reductions in viable tumor burden1B) and total blood vessel area (Fig. 1C). L1-7(N)d to show less activity than the dual Ang1/2 inhi-, and mL4-3 had no statistically significant effects inassay. To determine whether blood vessel regres-ccurred, tumors were stained for blood vessels) and basement membrane (type IV collagen). Allent groups had some scattered type IV collageng not associated with the tumor vessels (Fig. 1D).groups treated with Fc, mL4-3, or L1-7(N), the

r vessels were covered by basement membranes. After combination treatment with mL4-3 and), however, tumor vascularity was reduced, andbasement membrane sleeves were present, indica-

f vessel regression (34). We attempted to bolster thisg by measuring tumor endothelial cell apoptosis byscence-activated cell sorting (using an antibodyt activated caspase-3) and histologically (by TUNELng), but there were too few events to reliablyt differences between the treatment groups (dataown). Tumor vessels are typically leaky, tortuous,ilated, with loose or absent pericyte associations.r vessel normalization, a reversal of these aberrantes, has been observed with inhibition of the VEGF
nd Ang (36) pathways. Ang2 inhibition, but notinhibition or dual Ang1/2 inhibition, resulted in
clinicangio



r vessel normalization as evidenced by enhancedte coverage (Supplementary Fig. S5).next evaluated the combined effects of inhibitingnd VEGF activity. Suboptimal doses of AMG 386acizumab alone failed to confer significant tumorh inhibition, but combining the two agents at theseevels yielded cooperative antitumor activity (Fig. 2A). To assess whether dual Ang1/2 inhibition wasefficacious than selective Ang2 inhibition in thext of concurrent VEGF antagonism, optimal doses7(N) and AMG 386 were combined with the sametimal dose of bevacizumab used in Fig. 2A and B.combination treatments reduced tumor growth.386 plus bevacizumab mediated better antitumorty than L1-7(N) plus bevacizumab over the latterpoints of the study (Fig. 2C), and terminal assess-of viable tumor burden substantiated this findingD). These results confirm that dual Ang1/2 inhibi-s superior to selective Ang2 inhibition for suppres-olo205 tumor growth, regardless of whether VEGFty is simultaneously antagonized.

antagonism, but not Ang1 antagonism,its tumor endothelial cell proliferation, cornealgenesis, and retinal angiogenesispreviously showed that dual Ang1/2 inhibitionessed Colo205 tumor endothelial cell proliferation(7). To investigate whether this effect was conferred

gh Ang1 inhibition, Ang2 inhibition, or a combina-f the two, Colo205 tumor-bearing mice were treatedmL4-3, L1-7(N), mL4-3/L1-7(N), or AMG 386. Ashe tumor volume results described in the previousn, mL4-3 had no single-agent effect on tumor endo-l cell proliferation, whereas L1-7(N) was inhibitoryA). Dual Ang1/2 inhibition conferred no greater ef-n endothelial cell proliferation than Ang2 inhibition(Fig. 3A) in contrast to the apparently cooperativeof combined Ang1/2 inhibition on Colo205 tumorh (Fig. 1). This dissimilarity implies that repressionothelial cell proliferation is only one component un-ng the tumor growth inhibition mediated by Angonism.se agents were next tested inmodels of angiogenesiscornea and the retina. The cornea is normally avas-but pathologic angiogenesis can occur in the corneadary to conditions such as keratitis and corneallant rejection (38). VEGF- and bFGF-induced mod-corneal angiogenesis (35) were used to test the rolesg1 and Ang2 antagonism in blood vessel formation.served with endothelial cell proliferation, cornealgenesis seemed to be dependent on Ang2, but notg1 (Fig. 3B and C). The same conclusion could befrom evaluation of these Ang-antagonizing pepti-

s in a Tie2-dependent retinal model of angiogenesiswhich neovascularization was induced by changesbient oxygen tension (Fig. 3D). Thus, in three pre-
al settings (endothelial cell proliferation, cornealgenesis, and retinal angiogenesis), Ang2 inhibition



dramAng1with A

SelecatropTo

mals,(N), otagonthickemechaapy (epiphwhere

epiphmentathickAng2duceding thto ind

Ang1ovariTo

the ovfollicu

Figuretreated(at the(n = 10)with eitthe tota*, P < 0L1-7(N)


www.a

Do

atically suppressed neovessel formation, whereasinhibition had no effect alone or in combinationng2 inhibition.

tive inhibition of Ang1 or Ang2 induces ovarianhy, but not epiphyseal plate thickeningassess the effects of Ang inhibition in normal ani-rats were treated systemically with mL4-3, L1-7r AMG 386 for 1 month. AMG 386, like VEGF an-ists, has been observed to induce epiphyseal platening and ovarian atrophy, effects considered to benism-based consequences of antiangiogenic ther-7, 40). In the present study, AMG 386 provoked

at terminal measurement.

yseal plate thickening in all treated animals,as remarkably, L1-7(N) and mL4-3 failed to alter

mentously

acrjournals.org


yseal morphology in any of the animals (Supple-ry Table S4). Thus, induction of epiphyseal plateening seems to require inhibition of Ang1 and. In striking contrast, all three peptibodies pro-ovarian atrophy at similar incidence rates, indicat-at selective inhibition of Ang1 or Ang2 is sufficientuce ovarian atrophy.

and Ang2 inhibitors cooperatively suppressan follicular angiogenesisbetter understand the effects of Ang inhibition onary, we used a hormone-induced model of ovarianlar angiogenesis that allowed controlled assess-

2. The effect of combined Ang1 and Ang2 inhibition in the presence of VEGF pathway antagonism. A, mice implanted with Colo205 cells werewith human IgG1 (0.4 mg/kg, i.p.), bevacizumab (0.4 mg/kg, i.p.), AMG 386 (0.11 mg/kg, s.c.), or the combination of bevacizumab and AMG 386same dose and schedule as the single-agent groups) twice per week. B, viable tumor burden was determined at the end of the experiment. Bev, bevacizumab. C, Colo205 tumor-bearing mice were treated with IgG1 (0.4 mg/kg, i.p.; n = 10) or bevacizumab (0.4 mg/kg, i.p.) in combinationher AMG 386 (2.8 mg/kg, s.c.) or L1-7(N) (2.8 mg/kg, s.c.) twice per week (n = 25 per group). Fc or IgG1 control proteins were added to matchl amount of protein delivered in the combination group in both experiments. D, viable tumor burden (n = 23-25). Data are mean values ± SE..0001 versus bevacizumab or AMG 386; §, P < 0.05 versus bevacizumab or AMG 386; †, P ≤ 0.05 versus IgG1; ‡, P < 0.05 versus bevacizumab +

of neovascularization in mice that had never previ-ovulated. Two replicates of this experiment yielded




almosage rereplicmL4-3nation(N) grcant i(P < 0genes

citedthe nofailed

Discu

We

Figureincorpoendothand VE†, P <Pups w

Coxon et al.

Mol C2648

Do

t identical activity profiles with respect to percent-duction in blood vessel area fraction (replicate 1,ate 2): L1-7(N) (8%,11%), mL4-3 (15%, 14%), and/L1-7(N) (24%, 26%). All single-agent and combi-peptibody groups, with the exception of the L1-7oup in experiment 1, mediated statistically signifi-nhibition of angiogenesis relative to the Fc control
.05; Fig. 4). Thus, the inhibition of ovarian angio-is and induction of ovarian atrophy could be eli-
first spable

0.002 versus Fc + bFGF; ‡, P < 0.01 versus Fc + VEGF. D, inhibition of Ang2 preere treated with Fc, L1-7(N), mL4-3, or L1-7(N) plus mL4-3. Data are mean value



by inhibiting Ang1, Ang2, or both, consistent withtion that the ovarian atrophy was a consequence ofneovessel development.

ssion

describe here the generation and evaluation of the
ystemically administrable pharmacologic agents ca-of selectively inhibiting Ang1 in vivo. We show that
3. The effect of Ang1 and Ang2 antagonism on tumor endothelial cell proliferation, corneal angiogenesis, and retinal angiogenesis. A, BrdUrdration in endothelial cells. Colo205 tumor-bearing mice were treated with Fc, AMG 386, L1-7(N), mL4-3, or L1-7(N) plus mL4-3. Bar, meanelial:total mouse cell BrdUrd ratios. Data are mean values ± SE. *, P < 0.05 versus Fc. The effect of inhibition of Ang1 and Ang2 on bFGF-induced (B)GF-induced (C) corneal angiogenesis following treatment with Fc, L1-7(N), mL4-3, and L1-7(N) plus mL4-3. Data are mean values ± SE.

vents oxygen-induced neovascularization in the mouse retina.s ± SE. §, P < 0.0001 versus Fc.




Ang1supprand inhibitioAng1develAng1ducedby inHowehad liseemewith Adiffereto beIt is

whichtion wrole inunderextenAng1Ang2phasecles alicles,of thesequecorpuopposAng1angiothat toustsescencfromproan

In tof Anbitorspressiindividentthrouerativtifiedinhibiin thetion oAng1binativentewithoinhibregrestors frtent wvesselnot. Ttion mselectmaineffectsIn c

of Aneffectsconfeimplyin ponant AgestinarounAn

oppossystemconclquencences include (a) in vitro versus in vivo systems, (b) prenatalversu

FigurefolliculacombinexperimmL4-3/versus


www.a

Do

inhibition plays a context-dependent role in theession of angiogenesis in preclinical disease modelsnormal animals. In utero, pharmacologic Ang1 in-n substantially phenocopied the genetic ablation of, consistent with the important role of Ang1 inopmental angiogenesis (11). Postnatally, selectiveantagonism inhibited ovarian angiogenesis and in-ovarian atrophy, effects that could also be achieved

hibiting Ang2 alone or Ang1 plus Ang2 together.ver, in postnatal disease models, Ang1 inhibitionttle effect on its own, although its biological activityd to be unmasked in some settings when combinedng2 suppression. The mechanism underlying thential dependency on Ang1 in these settings remainsdetermined.intriguing that the ovary was the only organ ina biological consequence of selective Ang1 inhibi-as observed postnatally. The ovary, by virtue of itsreproductive cycling, is one of the few organs thatgo normal angiogenesis in adults. Based upon ansive analysis of ovarian expression patterns ofand Ang2 in hormone-induced ovulating rats,is thought to play two distinct roles at separates of follicular development: (a) preovulatory folli-nd (b) aging corpora lutea (41). In preovulatory fol-Ang2 expression is increased around blood vesselstheca interna. Because Ang1 expression is sub-

ntly increased around blood vessels in the earlys luteum, Ang2 and Ang1 are hypothesized to haveing functions, in which Ang2 initially displacesfrom Tie2, resulting in vessel destabilization andgenesis. The temporal expression patterns implyhis state of plasticity is later reversed when Ang1Ang2 from the receptor to re-establish vascular qui-e and stability. In conflict with this model, the data

L1-7(N) combination versus either single-agent alone; #, P < 0.05Fc.

the current study imply that Ang1 and Ang2 playgiogenic roles in the ovary.

FigureColo20

acrjournals.org


he Colo205 tumor xenograft model, the antagonismg1 and Ang2 either by combining individual inhi-or by using AMG 386 mediated greater tumor sup-on than was achieved by inhibiting Ang1 or Ang2dually, indicating that this tumor model is depen-on both angiopoietins. A potential mechanismgh which inhibition of Ang1 and Ang2 could coop-ely suppress tumor growth has recently been iden-(36). In the Colo205 tumor xenograft model, Ang2tion mediated two vascular outcomes: (a) reductiontotal number of tumor vessels and (b) normaliza-f the remaining tumor vessels (Fig. 5). Selectiveinhibition had no vascular effect. However, in com-on with Ang2 inhibition, Ang1 neutralization pre-d Ang2 inhibitor-mediated vessel normalizationut antagonizing the antivascular effects of Ang2ition. Normalization is thought to hinder vesselsion (42), which may prevent selective Ang2 inhibi-om reaching their full antitumor potential. Consis-ith this concept, dual Ang1/2 inhibition inducedregression, whereas selective Ang2 inhibition didhus, combining Ang1 inhibition with Ang2 inhibi-ay counteract a disadvantageous consequence of

ive Ang2 suppression (vessel normalization) whiletaining or enhancing the beneficial antivascularof Ang2 antagonism.ontrast to the subtle and context-dependent effectsg1 inhibition, Ang2 inhibition frequently mediatedthat were equivalent or nearly equivalent to those

rred by combined antagonism of Ang1 and Ang2,ing that Ang2 may be the dominant Ang involvedstnatal angiogenesis. Ang1 seems to be the domi-ng involved in prenatal angiogenesis (11, 43), sug-g a shift in the dependency on these two factorsd the time of birth.g1 and Ang2 have been shown to play similar anding functional roles in various in vitro and in vivos (18, 41, 44–47). The inability to draw consistent

usions from these studies may be in part a conse-e of differing experimental conditions. These differ-

4. Ang1 and Ang2 inhibitors cooperatively suppress ovarianr angiogenesis. Fc, mL4-3, L1-7(N), or an mL4-3/L1-7(N)ation were administered to superovulated mice (two independentents). Data are mean values ± SE. *, P = 0.005 comparing

s postnatal angiogenesis, (c) varying vascular beds,

5. Mechanistic consequences of Ang2 and Ang1 inhibition in the5 tumor model.




(d) paof-funThis fcausemay bfunctimostare thdeleteies prrolesgenetand Aout syin uterstraintype oby resing phthe pocircumstudyand, iAng2Pat

levelsretinoosteoationscal beantagsomeas achpeuticFurththe prIn ear

antituIn thesolida toxiadverhibitoal phVEGFregimevidestudywith rcrease(30). Aand oinform

Discl

All acurrent

Ackn

We ttheir crand Htechnic

Grant

ComMcDonLung atute (D

Thepaymenadvertisthis fac

Refe1. Pe

the2. Pe

MEEy

3. vantorFro

4. Rocu141

5. GraPeEn

6. Ekcon200

7. Olitum20

8. Bu

Coxon et al.

Mol C2650

Do

thologic versus normal angiogenesis, and (e) gain-ction versus loss-of-function experimental designs.inal difference may be particularly important be-the addition of exogenous factors to a model systeme a less physiologically relevant means to elucidateon than removal of endogenous factors. Perhaps theinformative published experiments in this regardose in which Ang1 and Ang2 have been geneticallyd in the germline of rodents (11, 41, 48). These stud-ovide significant insight into the developmentalof Ang1 and Ang2. However, it is more difficult toically examine the postnatal in vivo function of Ang1ng2 without the availability of conditional knock-stems; the constitutive Ang1 knockout mouse dieso (as does the constitutive Ang2 knockout on somebackgrounds; refs. 11, 41), and the postnatal pheno-f surviving Ang2 knockout mice may be influencedidual effects of developmental gene deletion. By us-armacologic Ang1 and Ang2 inhibitors to examinestnatal roles of Ang1 and Ang2 in vivo, we havevented these issues. The results of the currentshow that Ang1 neutralization does not counteractn some cases augments, the efficacy mediated byneutralization.hologic angiogenesis is associated with altered Angin a number of diseases, including cancer, diabeticpathy, macular degeneration, rheumatoid arthritis,rthritis, and psoriasis (1). Ang-targeted interven-in these therapeutic indications may provide clini-nefit alone or in combination with VEGF pathwayonists. The data presented here suggest that, insettings, combined inhibition of Ang1 and Ang2,ieved with AMG 386, may provide superior thera-efficacy to that mediated by targeting Ang2 alone.

ermore, this enhanced activity is also observed in
esence of concurrent VEGF pathway antagonism. Rece
or growth by selective inhibition of angiopoietin-2. Cancer Cell04;6:507–16.none G, Vigneri P, Mariani L, et al. Expression of angiogenesis

stiwit

9. Eto2 iin21

10. SfibreInt

11. Suliga19

12. YuFu

13. Thenme

14. Talym

15. Mo1-p



mor activity, as well as pharmacodynamic activity.first-in-human study in patients with advanced

tumors, AMG 386 was well tolerated and showedcity profile that did not include treatment-relatedse events typically associated with angiogenesis in-rs targeting the VEGF pathway (28). Two addition-ase 1 studies of AMG 386 in combination withpathway inhibitors (49) or various chemotherapyens (50) showed AMG 386 to be well tolerated withnce of antitumor efficacy. A randomized phase 2of AMG 386 combined with paclitaxel in patientsecurrent ovarian cancer yielded dose-dependent in-s in progression-free survival and CA-125 responsedditional controlled studies of AMG 386 in cancerther angiogenesis-associated diseases will furtherthe potential clinical benefit to patients.

osure of Potential Conflicts of Interest

uthors (with the exception of B.L. Falcón and D.M. McDonald) areor former employees of Amgen, Inc.

owledgments

hank Drs. Glenn Begley, David Weinreich, and Beate Quednau foritical reading of the manuscript; Dr. Julia Gage, Jennifer Keysor,eather Hartley for manuscript preparation; Michael Mann foral assistance; and Lei Zhou for statistical analysis.

Support

mercial research grant from Amgen, Inc. (B.L. Falcón and D.M.ald), NIH grants HL24136 and HL59157 from the National Heart,nd Blood Institute and CA82923 from the National Cancer Insti-.M. McDonald).costs of publication of this article were defrayed in part by thet of page charges. This article must therefore be hereby markedement in accordance with 18 U.S.C. Section 1734 solely to indicatet.

ived 03/03/2010; revised 08/11/2010; accepted 08/12/2010;
ly clinical studies, AMG 3 86 has shown promising published 10/11/2010.
rencesrry BN, Arbiser JL. The duality of angiogenesis: implications forrapy of human disease. J Invest Dermatol 2006;126:2160–6.nn JS, Madan A, Caldwell RB, Bartoli M, Caldwell RW, Hartnett. Vascular endothelial growth factor in eye disease. Prog Retine Res 2008;27:331–71.Cruijsen H, van der Veldt A, Hoekman K. Tyrosine kinase inhibi-

s of VEGF receptors: clinical issues and remaining questions.nt Biosci 2009;14:2248–68.senfeld PJ, Brown DM, Heier JS, et al. Ranibizumab for neovas-lar age-related macular degeneration. N Engl J Med 2006;355:9–31.goudas ES, Adamis AP, Cunningham ET, Feinsod M, Guyer DR.gaptanib for neovascular age-related macular degeneration. Ngl J Med 2004;351:2805–16.lund L, Olsen BR. Tie receptors and their angiopoietin ligands aretext-dependent regulators of vascular remodeling. Exp Cell Res6;312:630–41.ner J, Min HS, Leal J, et al. Suppression of angiogenesis and

mulators and inhibitors in human thyroid tumors and correlationh clinical pathological features. Am J Pathol 1999;155:1967–76.h T, Inoue H, Tanaka S, Barnard GF, Kitano S, Mori M. Angiopoietin-s related to tumor angiogenesis in gastric carcinoma: possiblevivo regulation via induction of proteases. Cancer Res 2001;61:45–53, [print].ligoi C, de Luca A, Cascone I, et al. Angiopoietin-2 expression inast cancer correlates with lymph node invasion and short survival.J Cancer 2003;103:466–74.ri C, Jones PF, Patan S, et al. Requisite role of angiopoietin-1, and for the TIE2 receptor, during embryonic angiogenesis. Cell96;87:1171–80.Q. The dynamic roles of angiopoietins in tumor angiogenesis.ture Oncol 2005;1:475–84.urston G, Wang Q, Baffert F, et al. Angiopoietin 1 causes vessellargement, without angiogenic sprouting, during a critical develop-ntal period. Development 2005;132:3317–26.mmela T, Saaristo A, Lohela M, et al. Angiopoietin-1 promotes
phatic sprouting and hyperplasia. Blood 2005;105:4642–8.risada T, Oike Y, Yamada Y, et al. Angiopoietin-1 promotes LYVE-ositive lymphatic vessel formation. Blood 2005;105:4649–56.



16. Shangmic

17. MaAnAm

18. Ah-2cer

19. TiacanCa

20. Haan

21. Hatorsqu

22. Shsiotum

23. NaneoTh

24. Shintbuisc

25. ZhlatangCa

26. Hutuming16;

27. BroANVECa

28. Hetumpa

29. EsklialcepCa

30. Kaplapa50

31. Lofor

32. Besech

33. Cocanvasand

34. Ina

thetheba

35. Cotumco26

36. Fatioiza

37. Jaian

38. CuU.str20

39. DaAnne

40. Gecoge19

41. Maan19

42. BicvasBio

43. Gaforlat

44. Momuca11

45. KimattheOn

46. DaligaCe

47. Visanthe82

48. Naaffbu20

49. Hoty,ne(mAn

50. Mi


www.a

Do

im WSN, Teh M, Bapna A, et al. Angiopoietin 1 promotes tumoriogenesis and tumor vessel plasticity of human cervical cancer ine. Exp Cell Res 2002;279:299–309.chein MR, Knedla A, Knoth R, Wagner S, Neuschl E, Plate KH.giopoietin-1 promotes tumor angiogenesis in a rat glioma model.J Pathol 2004;165:1557–70.

mad SA, Liu WB, Jung YD, et al. The effects of angiopoietin-1 andon tumor growth and angiogenesis in human colon cancer. Can-Res 2001;61:1255–9.n S, Hayes AJ, Metheny-Barlow LJ, Li LY. Stabilization of breastcer xenograft tumour neovasculature by angiopoietin-1. Br Jncer 2002;86:645–51.yes AJ, Huang WQ, Yu J, et al. Expression and function ofgiopoietin-1 in breast cancer. Br J Cancer 2000;83:1154–60.wighorst T, Skobe M, Streit M, et al. Activation of the Tie2 recep-by angiopoietin-1 enhances tumor vessel maturation and impairsamous cell carcinoma growth. Am J Pathol 2002;160:1381–92.imWS, Teh M, Mack PO, Ge R. Inhibition of angiopoietin-1 expres-n in tumor cells by an antisense RNA approach inhibited xenograftor growth in immunodeficient mice. Int J Cancer 2001;94:6–15.mbu H, Nambu R, Oshima Y, et al. Angiopoietin 1 inhibits ocularvascularization and breakdown of the blood-retinal barrier. Gene

er 2004;11:865–73.yu KG, Manor O, Magner M, Yancopoulos GD, Isner JM. Directramuscular injection of plasmid DNA encoding angiopoietin-1t not angiopoietin-2 augments revascularization in the rabbithemic hindlimb. Circulation 1998;98:2081–7.ou YF, Stabile E, Walker J, et al. Effects of gene delivery on col-eral development in chronic hypoperfusion: diverse effects ofiopoietin-1 versus vascular endothelial growth factor. J Am Collrdiol 2004;44:897–903.ang H, Do J, Li L, et al. Angiopoietin-2 specific CVX-060 inhibitsor growth cooperatively with chemotherapy. 99th Annual Meet-of the American Association for Cancer Research 2008 Apr 12-San Diego, CA. 2008. p. Abstract 2493.wn JL, Cao ZA, Pinzon-Ortiz M, et al. A human monoclonal anti-G2 antibody leads to broad antitumor activity in combination withGF inhibitors and chemotherapy agents in preclinical models. Molncer Ther 2010;9:145–56.rbst RS, Hong D, Chap L, et al. Safety, pharmacokinetics, and anti-or activity of AMG 386, a selective angiopoietin inhibitor, in adult

tients with advanced solid tumors. J Clin Oncol 2009;27:3557–65.ens FA, Verweij J. The clinical toxicity profile of vascular endothe-growth factor (VEGF) and vascular endothelial growth factor re-tor (VEGFR) targeting angiogenesis inhibitors; a review. Eur Jncer 2006;42:3127–39.rlan BY, Oza AM, Hansen VL, et al. Randomized, double-blind,cebo-controlled phase II study of AMG 386 combined with weeklyclitaxel in patients (pts) with recurrent ovarian carcinoma [abstract00]. J Clin Oncol 2010;28 Suppl:15s.wman HB. Bacteriophage display and discovery of peptide leadsdrug development. Annu Rev Biophys Biomol Struct 1997;26.ckstead JH. A simple technique for preservation of fixation-nsitive antigens in paraffin-embedded tissues. J Histochem Cyto-em 1994;42:1127–34.xon A, Bush T, Saffran D, et al. Broad antitumor activity in breastcer xenografts by motesanib, a highly selective, oral inhibitor of
cular endothelial growth factor, platelet-derived growth factor,Kit receptors. Clin Cancer Res 2009;15:110–8.
i T, Mancuso M, Hashizume H, et al. Inhibition of vascular endo-

popyAS

acrjournals.org


lial growth factor (VEGF) signaling in cancer causes loss of endo-lial fenestrations, regression of tumor vessels, and appearance ofsement membrane ghosts. Am J Pathol 2004;165:35–52.xon A, Bolon B, Estrada J, et al. Inhibition of interleukin-1 but notor necrosis factor suppresses neovascularization in rat models of

rneal angiogenesis and adjuvant arthritis. Arthritis Rheum 2002;46:04–12.lcón BL, Hashizume H, Koumoutsakos P, et al. Contrasting ac-ns of selective inhibitors of angiopoietins-1 and -2 on the normal-tion of tumor blood vessels. Am J Pathol 2009;175:2159–70.n RK. Normalization of tumor vasculature: an emerging concept intiangiogenic therapy. Science 2005;307:58–62.rsiefen C, Hofmann-Rummelt C, Kuchle M, Schlotzer-SchrehardtPericyte recruitment in human corneal angiogenesis: an ultra-uctural study with clinicopathological correlation. Br J Ophthalmol03;87:101–6.s A, Fanslow W, Cerretti D, Warren E, Talarico N, McGuire P.giopoietin/Tek interactions regulate MMP-9 expression and retinalovascularization. Lab Invest 2003;83:1637–45.rber HP, Vu TH, Ryan AM, Kowalski J, Werb Z, Ferrara N. VEGFuples hypertrophic cartilage remodeling, ossification and angio-nesis during endochondral bone formation [comment]. Nat Med99;5:623–8.isonpierre PC, Suri C, Jones PF, et al. Angiopoietin-2, a naturaltagonist for tie2 that disrupts in vivo angiogenesis. Science97;277:55–60.knell R, Harris AL. Anticancer strategies involving the vasculature:cular targeting and the inhibition of angiogenesis. Semin Cancerl 1992;3:399–407.le NW, Thurston G, Hackett SF, et al. Angiopoietin-2 is requiredpostnatal angiogenesis and lymphatic patterning, and only the

ter role is rescued by angiopoietin-1. Dev Cell 2002;3:411–23.chizuki Y, Nakamura T, Kanetake H, Kanda S. Angiopoietin 2 sti-lates migration and tube-like structure formation of murine brainpillary endothelial cells through c-Fes and c-Fyn. J Cell Sci 2002;5:175–83.I, Kim JH, Moon SO, Kwak HJ, Kim NG, Koh GY. Angiopoietin-2

high concentration can enhance endothelial cell survival throughphosphatidylinositol 3′-kinase/Akt signal transduction pathway.cogene 2000;19:4549–52.vis S, Aldrich TH, Jones PF, et al. Isolation of angiopoietin-1, and for the TIE2 receptor, by secretion-trap expression cloning.ll 1996;87:1161–9.conti RP, Richardson CD, Sato TN. Orchestration of angiogenesisd arteriovenous contribution by angiopoietins and vascular endo-lial growth factor (VEGF). Proc Natl Acad Sci U S A 2002;99:19–24.sarre P, Thomas M, Kruse K, et al. Host-derived angiopoietin-2ects early stages of tumor development and vessel maturationt is dispensable for later stages of tumor growth. Cancer Res09;69:1324–33.ng D, Gordon M, Appleman L, et al. From a phase 1b study of safe-pharmacokinetics and tumor response of the angiopoietin1/2-utralizing peptibody AMG 386 in combination with AMG 706otesanib), bevacizumab or sorafenib in advanced solid tumors.n of Oncol 2008;19:Abstract 462D.ta AC, Wang D, Takimoto CH, et al. AMG 386, a selective angio-ietin 1/2-neutralizing peptibody, in combination with chemothera-
in adult patients with advanced solid tumors. J Clin Oncol, 2007CO Annual Meeting Proceedings Part I 2007;25:Abstract 14033.



2010;9:2641-2651. Mol Cancer Ther Angela Coxon, James Bready, Hosung Min, et al.

Neutralizing Peptibody−for AMG 386, an Angiopoietin-1/2Suppression of Angiogenesis and Tumor Growth: Implications

Context-Dependent Role of Angiopoietin-1 Inhibition in the

Updated version

http://mct.aacrjournals.org/content/9/10/2641

Access the most recent version of this article at:

Material

Supplementary

http://mct.aacrjournals.org/content/suppl/2010/10/08/9.10.2641.DC1

Access the most recent supplemental material at:

Cited articles

http://mct.aacrjournals.org/content/9/10/2641.full#ref-list-1

This article cites 48 articles, 15 of which you can access for free at:

Citing articles

http://mct.aacrjournals.org/content/9/10/2641.full#related-urls

This article has been cited by 22 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected] at

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

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mct.aacrjournals.org/content/9/10/2641To request permission to re-use all or part of this article, use this link

on August 27, 2021. © 2010 American Association for Cancer Research. mct.aacrjournals.org Downloaded from


http://mct.aacrjournals.org/content/suppl/2010/10/08/9.10.2641.DC1

http://mct.aacrjournals.org/content/9/10/2641.full#ref-list-1

http://mct.aacrjournals.org/content/9/10/2641.full#related-urls

http://mct.aacrjournals.org/cgi/alerts

mailto:[email protected]



Documents

Therapeutics Context-Dependent Role of Angiopoietin-1 ......AMG 386 is an investigational first-in-class peptide-Fc fusion protein (peptibody) that inhibits angiogenesis by preventing