Attenuation of colitis in the cotton top tamarin by anti alpha 4 integrin monoclonal antibody

Attenuation of colitis in the cotton-top tamarin

by anti-alpha 4 integrin monoclonal antibody.

D K Podolsky, … , P Sehgal, M deBeaumont

J Clin Invest. 1993;92(1):372-380. https://doi.org/10.1172/JCI116575.

Recent studies have demonstrated the induced expression of endothelial adhesion molecules including E-selectin (also called endothelial leukocyte adhesion molecule-1), vascular cell adhesion molecule and intercellular adhesion molecule in actively involved mucosa of patients with ulcerative colitis and Crohn's disease. Similar induction has been demonstrated in the colon of the Cotton-top tamarin (CTT), a New World primate that experiences a spontaneous acute and chronic colitis resembling ulcerative colitis. To assess the potential importance of leukocyte adhesion as a necessary step in acute colitis, the effect of parenteral mAb directed against adhesion molecules on CTT colitis was evaluated in placebo-controlled blinded trials. Serial administration of either of two anti-E-selectin mAb designated BB11 and EH8 effectively coated endothelial surfaces expressing this vascular adhesion molecule. Although colitis activity was slightly diminished after the 10-d treatment period in CTT receiving either BB11 or EH8, this reduction was not

significantly different than that seen in animals given a placebo control when assessed by a previously validated standardized scale of inflammatory activity: mean histologic activity grade 2.2 +/- 0.2 pretreatment vs 1.5 +/- 0.5 posttreatment in group receiving mAb and 2.1 +/- 0.1 pretreatment vs 1.3 +/- 0.5 posttreatment in the placebo group (P > 0.2). In contrast, administration of an anti-alpha 4 integrin mAb designated HP1/2 that binds VLA4 (alpha 4 beta 1) and presumably alpha […]

Research Article

Find the latest version:

http://jci.me/116575/pdf

Attenuation

of

Colitis

in the Cotton-top Tamarin

by Anti-a4 integrin Monoclonal Antibody

Daniel K. Podolsky,** Roy Lobb, NorvalKing,*11Christopher D.Benjamin,'

BlakePepinsky,IPrebhat Sehgal,'* and Michelle deBeaumont**

*Gastrointestinal Unit, Massachusetts General Hospital and Department ofMedicine, Harvard Medical School; MassachusettsGeneral

Hospital/New England Regional Primate ResearchCenterfortheStudyofInflammatoryBowelDisease, BostonMassachusetts02114; 1Biogen, Inc., Cambridge, Massachusetts 02142; andIINewEngland Regional Primate Research Center, Southboro, Massachusetts01772

Abstract

Recent studies have demonstrated the induced expression of endothelial adhesion molecules including E-selectin (also called endothelial leukocyte adhesion molecule-i), vascular celladhesion molecule and intercellular adhesion molecule in actively involvedmucosaofpatientswith ulcerative colitis and Crohn's disease. Similar induction has been demonstrated in the colon of theCotton-toptamarin(CTU),aNew World

pri-matethat experiencesaspontaneous acuteandchroniccolitis resembling ulcerative colitis.

Toassessthepotentialimportance of leukocyte adhesionas anecessarystep inacutecolitis, theeffectof parenteral mAb directedagainst adhesionmoleculesonCIT colitiswas evalu-ated inplacebo-controlledblinded trials. Serial administration ofeither oftwoanti-E-selectinmAbdesignatedBB11andEH8 effectivelycoatedendothelial surfacesexpressingthis vascular adhesion molecule. Althoughcolitisactivitywasslightly dimin-ishedafter the 10-dtreatmentperiod inCITreceivingeither BB11 orEH8, this reduction was not significantly different thanthatseeninanimalsgivenaplacebocontrol when assessed bya previously validated standardized scale ofinflammatory activity:meanhistologic activity grade2.2±0.2 pretreatmentvs

1.5±0.5 posttreatment in group receiving mAb and 2.1±0.1 pretreatmentvs1.3±0.5 posttreatment in theplacebogroup(P

>0.2).In contrast,administrationofananti-a4integrinmAb designated HP1/2 that binds VLA4

(a4f31)

and presumably

a4187

integrinsresulted insignificantattenuation ofacutecolitis

whencomparedtoboth pretreatmentactivityindex(P=_0.005)

andtheplacebocontrolgroup(P<0.01

):

mean

histologic

activ-itygrade 1.6±0.3pretreatmentvs0.2±0.1posttreatment in the

groupreceivingHP1/2and 1.8±0.5pretreatmentand 1.2±0.2 posttreatment in theplacebocontrolgroup.

These studies usinga model ofspontaneous colitis in the CMT demonstrate the

feasibility

of modulation of

leukocyte-vascularadhesionand/orother

integrin-mediated

events

possi-bly including T cell

aggregation

and T cell-stromal interac-tions, as well as lymphocyte

homing.

These results suggest both that theseprocessesare

important

and

possibly

essential

elements insustaining acute colitis and that their

disruption

may result in therapeutic benefit. (J. Clin. Invest. 1993. 92:

Addresscorrespondence to Daniel K. Podolsky, M.D., GI Unit,

Jack-son 7, Massachusetts GeneralHospital, 50 FruitStreet, Boston, MA 02114.

Receivedfor publication30October 1992 and in revisedform20 January 1993.

372-380.)Key words: adhesion * colitis*Cotton-top tamarine inflammatoryboweldisease

Introduction

Although theinitiatingeventsin the major forms of

inflamma-toryboweldisease

(IBD)'

remain unclear, considerable pro-gresshas been made in delineating some ofthepresumably

common,nonspecific inflammatoryprocesses that result in tis-sueinjuryandclinical manifestationsin association with the

periodic increases ofinflammatoryactivitythat are the

hall-marks of these disorders.Manyoftheseinflammatoryfactors arefoundinactively affectedtissue from patientswith either

ulcerativecolitis(UC) orCrohn's disease(CD). The various

mediatorsorsubstancesproduced by the diversepopulations ofcellular constituents present atthe sites of diseaseactivity

includeproducts of arachidonic acidmetabolism, both

prosta-glandins and leukotrienes, a complex mixture of cytokines,

and oxygen-free radicals. The relative importance ofeach of

these substancesorwhetheranysingleoneofthesemediators is

essential to sustaining or amplifying the inflammatory re-sponseandinjuryisnotclear. Theeffect ofagentsthat specifi-callyantagonize individual mediatorsmayprovide important insights in thiscontext.

Whilethe absolute dependenceoftheinflammatory activ-ityonvariousindividual solublemediatorsremains uncertain, it isapparentthatproduction ofthesemediatorsisintrinsically relatedtorecruitment of inflammatorycellsfromthe vascular

compartmenttothesiteof diseaseactivity (1). Theseinclude neutrophils, lymphocytes, and monocytes, as well as (to a

lesserextent) eosinophils andmast cells. Conversely, disrup-tion ofprocessesresponsible for themigration of leukocytes fromthe vascularspace tosites ofinflammatory activity might result inattenuation ofthegeneration of inflammatory

media-torsthataretheproximatecauseoftissueinjuryandthe mani-festations ofIBD.

Some ofthe mechanisms

responsible

forthe

targeted

mi-gration of leukocytes to various sites of inflammation in the bodyhaverecentlybeen delineated(2-7).

Specific

interactions between

proteins

presentonleukocyte surfacemembranes and proteins expressed on endothelial surfaces as

receptor-coun-terreceptor

pairs

appeartobe centraltotheseprocesses.

Many

of these proteins have been identified and they include members of integrnn, selectin, and

immunoglobulin protein

superfamilies.

Itis notable that

expression

ofsome

key

adhe-sion moleculespresentonthe endothelium are

specifically

in-duced bycytokines producedat

inflammatory

foci

including

interleukin-1, tumornecrosisfactor,and interferon(2, 8-11

).

1.Abbreviations usedinthispaper: CD, Crohn's disease;CTT, Cotton-toptamarin;IBD,inflammatoryboweldisease;ICAM1,intracellular adhesion

molecule-i;

UC,ulcerative colitis;VCAM1,vascular adhe-sionmolecule-l;VLA4, very lateantigen-4.

372 Podolsky, Lobb, King, Benjamin, Pepinsky, Sehgal,and deBeaumont J.Clin. Invest.

© The American Society for Clinical Investigation, Inc.

Amongthe various adhesion molecules identified, endothelial leukocyte adhesion molecule- 1 (E-selectin), vascular cell ad-hesion molecule- 1 (VCAM 1), and intracellular cell adhesion

molecule (ICAM1) on the endothelial surface appear to be particularly important in adhesion of neutrophils,

lympho-cytes, and monocytes (2-12). E-selectin plays an important role in neutrophil-endothelial adhesion but may also bind

monocytes. Inaddition, E-selectin bindssome Tcell subsets includingmemorycells, andmaybeimportantinearlyT cell recruitment (13-16). VCAM1 appears toplayapredominant role in lymphocyte and monocyte adhesion to endothelium and ICAM-1contributesto awidevarietyof cellular adhesion

events.Itshould be noted that VCAM1-and ICAM 1-mediated interactionsmayplayarole in immune cell activationaswell

asrecruitment. Byimplicationthe "counterreceptors"onthe leukocytes' surfacesarealsoimportantandincludethe

puta-tivesialylLewis bloodgroupbearingE-selectin ligand,verylate antigen-4(VLA4) and CDl1 / 18(17-23).

Recentstudies in this laboratory and elsewhere have dem-onstrated theexpression of theadhesionmoleculesE-selectin and ICAM1 at sites of active inflammation in patients with

IBD(24, 25).Asexpectedthese adhesion moleculesare pres-entin both UC and CD.Inaddition,wehavedemonstrated the

presenceofVCAM1in bothinflamedandnoninflamedbowel, suggesting itmayplayarole in theconstitutive recruitment of lymphocytes andmonocytes tothemucosa.Theexpressionof leukocyte-endothelial adhesion molecules in association with

IBD areconsistent with thehypothesisthatrecruitmentof leu-kocytes from the vascular space contribute to the sustained inflammatory response in these disorders. Conversely, it im-pliesthatdisruption ofthese processescouldattenuatethe in-flammatory infiltrate. Indeed, one report has indicated that monoclonal antibodies targetedto

CDl

1 /18 ameliorated in-flammation in a rodent model of colonic inflammation in-ducedbyanexogenous agent(26).

TheCotton-toptamarin (CTT) offersaparticularly useful modeltoexplore theimportance of adhesionmolecules in coli-tis. The CTT isaNewWorldprimatethat

experiences

a sponta-neouschronic colitis marked byperiodicflares ofacute inflam-mation that closely mimics human UC in its histologic and

clinicalfeatures,aswellasitsresponse topharmacologicagents

(27, 28). Previous studies confirmed theinductionof E-selec-tin inassociation with active colitis in the CTT. Although the lackofreagents suitable foruseinthisprimate precluded as-sessment ofVCAM1, presumably this ligand might also be expressed in the CTTcolonic mucosainamannersimilarto

thatfound inman. Inthisreport, we assesstheeffect ofadminis-tration of monoclonal antibodiesdirectedagainst either

E-se-lectin,whichcontributestoneutrophil-endothelial adhesionor

theintegrina4, oneofthesubunits ofVLA4 thatcontributesto

lymphocyteand monocyteadhesion, on the activity of acute

flaresof CTT colitis. While coating ofE-selectin by mAb did nothaveasignificant effect,the anti-a4 mAb led to significant

reductionin the acute colonicinflammation. Methods

Monoclonal antibodies and antisera

BB 1Iand EH8, two anti-E-selectin mAbs of IgG2b and

_IgG,

isotypes, respectively, prepared against the human ligand were demonstrated in earlier studies to recognize the Cotton-top tamarin counterpart (24).

PreparationofBB 1 has been

previously

reported(29)while EH8was

prepared by immunization with pure recombinant E-selectin and

screening for blockade of neutrophil adhesion to rE-selectin (13). HP1/2,anIgGImonoclonalantibody directed against the a4 chain of the human VLA4 heterodimer (a4fl) has been previously described (30) and proventorecognizea cross-reacting determinantonCTT leukocytesbyFACS® analysis (Becton Dickinson Immunocytometry Systems, Mountain View, CA) and staining ofCTT colonicmucosa.In addition, HP1/2wasproventoblock adherenceofCTT leukocytesto

recombinant soluble VCAM 1-coated plates in vitro,asdemonstrated below. For in vivo studies mAbs BB1 1, EH8, and HP1 /2werepurified from ascitesby proteinAchromatography, followed by gel filtration chromatography in endotoxin-free physiologic saline. All mAbs

con-tained< 1endotoxin unit/mg.

Rabbit anti-human VCAM 1 serum (rabbit no. 349) was prepared by immunization of rabbits with recombinant soluble human VCAM 1 purified byimmunoaffinity chromatography using immobilized anti-VCAM1 mAb 4B9asdescribed (31 ). Anti-VCAM 1 serum was affinity purified by sequentialchromatographyonproteinA-Sepharose andan

affinity resin prepared by immobilization of rsVCAM 1 on Affigel (8 mg/ml resin)followed by elution with pH 3.0 buffer, immediate

neu-tralization, and dialysis against PBS. The affinity-purified antiserum completely blocked thebinding of human Ramos cellstorsVCAM 1-coated plates and binds human, mouseand ratVCAM 1 stably ex-pressed in Chinese hamster ovary cells as determined by FACS® analysis.

Evaluation of E-selectin, VCAMI, and a4 integrin expression inCTT

Pinch mucosal biopsieswereobtained fromrectosigmoid of CTT in parallel with samples obtained for histologic evaluation of inflamma-tion. Tissue sampleswereplaced in OCT compound(Tissue Tek, Elk-hart, IN)andfrozensections prepared ( -44m). Immunohistochemis-try was performed usinga commercially available peroxidase-based technique (Vectastain; Vector Labs, Inc., Burlingame, CA) after incu-bation withspecific first antibodyaspreviously described (24).

Evaluation of

serum

half-life of

anti-E-selectin

and

anti-a4 integrin

mAbs

mAbs BBl 1, EH8,and HPl/2 were each administeredassingle intra-muscular doses (2 mg/kg) to CTT (n=2for each antibody) and serum was collected at 0, 1, 2, 6, 12, 24, and48 h. Levelsof mouse Ig were determined by ELISA assay.

Effect

of anti-E-selectin and

anti-a4 integrin

mAbs

on CTTcolitis

CTT maintained at the New England Regional Primate Research Centerwerescreenedfor signs of activecolitis. Pinchcolonicmucosal biopsies obtained after sedation with ketamine (25 mg/kg) by routine

techniqueswereevaluatedforactive inflammationusing a previously validated scoring system (27). CTT with active colitis grades 2+ or 3+ (grading scale range 0-3+) were selected for study to be initiated within 3 wkof the screening biopsy. CTT were rebiopsied at the initiation of the studyperiod. Animals were randomized to receive mAb or placebo

(saline) injection.

Anti-E-selectin studies. In separate studies CTTreceived either

BBll orEH8 (- 2 mg/kg) on days 0, 2, 4, 6, and 8; n = 5/mAb; placebo controls (n=5for each mAb trial) were injected intramuscu-larlywith saline on the same schedule. Blood (- 0.5 ml) and colonic mucosalbiopsies were obtained from mAb and control CTT on days 0, 2, 4, 6, 8, and 10. Serum prepared from peripheral blood was assessed for mouse IgG as noted above. Colonic mucosal biopsies on days 0 and 10wereevaluatedfor acute colitis activity in a blinded fashion using thevalidated index noted above. Colonic biopsies were also assessed for the presenceof E-selectin using the previously described histochemical technique (24), as well as the presence of mouse IgG. The latter was accomplished by omission of the first antibody in the peroxidase-based staining process.

Anti-a4integrinstudies. CTT received eitherHPl/2mAb(2mg/

3,4,5,6, and 7. Colonic mucosal biopsies, as well as peripheral blood, were obtained on days 0, 2, 4, 6, 8, and10.Colitis activity was assessed in ablinded fashion on biopsies obtained on days0and 10.Peripheral

blood was used todetermineserum mouse IgG levels, as well as com-plete blood counts.

Results

Expressionofvascular adhesionmolecules in CTT. Inprevious

studies, E-selectin an inducibleleukocyte adhesion molecule expressedonendothelial surfaceswas foundtobepresentin association with active colitis in patients with IBD (24). In contrast, VCAM 1,adistinct vascular adhesion molecule that bindspredominantlylymphocytes and monocyteswaspresent in substantialamountsin association withtissuelymphoid ag-gregates incolonic mucosaof patients withIBDinamanner independent of activeinflammation, as well as normal

con-trols. Itshould be noted that VCAM1could also be detectedon

endothelial surfacesinmanyofthe IBDtissues.Parallelstudies demonstrated a similarpattern of expression ofE-selectin in colonic mucosa of CTT, a New World monkey that

experi-ences a spontaneousdiffuse colitis withcloseclinicaland histo-pathological similarity to humanUJC. Thus, anti-E-selectin-stained endothelial surfaces in mucosal biopsies exhibiting signs of active colitisbutnotthose inwhichthediseaseprocess

was quiescent. However, itwas not possible in those earlier studiestoassesstheexpressionof VCAM1 in theprimate

be-cause of lack of cross-reactivity ofthe monoclonal reagents

preparedagainst human VCAM1thenavailable. The

develop-ment ofa conventional antisera against VCAM1 with broad speciesrecognition, as describedin Methods, allowed

exten-sion ofthoseearlier studiestodefineVCAM1expressionin the CTT. AsillustratedinFig. 1, VCAM 1 isindeedpresentin the colonicmucosaofCTTasobservedin humantissueinearlier studies.AsillustratedinFig.1, VCAM 1ispresentinsignificant

amountsinareasof lymphoidaggregatesindependent ofactive colitisina mannersimilartothatobserved inman.

In addition to studies definingthe presence ofadhesion moleculesontheendothelialsurface,theexpression ofoneof the cognate counter receptors present on leukocyte surfaces

wasalsoexamined.Specifically, colonicmucosalbiopsies from CTTwereassessedforthepresenceofoneof thetwosubunits thatcompriseVLA4by

immunohistochemistry

usingan

anti-a4integrinmAb.Asillustrated inFig. 2, leukocytes,both lym-phocytes and monocytes, present in the

inflammatory

infil-trateinmucosaof CTT with active colitisexhibit thea4 inte-grin subunit oftheVLA4ligand.

Staining

was notobserved in the absenceoftheacutecellular

inflammatory

infiltrate.

Effect of antiadhesion moleculemAbs on CTTcolitis. To

assesstherelative importanceofvascularadhesion molecules

insustainingactive colitis in the CTT, the effect ofparenterally administered mAbsspecificforadhesion moleculeson active CTTcolitiswasdetermined. Initial studies focusedontheuse

of mAbs directed toward E-selectin

designated

BB 1 andEH8.

Inpilotstudies,theindividual antibodieswereadministeredas

asingleintramuscularinjectiontodefinetheserumhalf-life for

more prolonged therapeutic trials. Administration of these mAbs in concentrations predicted to leadto adequate tissue concentrationbypreviousstudies ledtoextendedserum

IgG

levelsinCTT,witha calculatedT.1/2of 20-24hforBBl1 and EH8. Mucosal biopsies obtainedover the course of 48 hafter administration of the mAbs led to detectablemouse IgG on the

CTT colonic mucosal endothelial surface. Importantly, the presenceofIgGwascorrelated with the lack of ability to detect

E-selectin afteraddition of exogenous anti-E-selectinmAbsto

the mucosal biopsies. E-selectin became undetectable by this approach within 6 h of the _{parenteral administration of the} mAb andpersisted upto24h.However, E-selectin could again

bedetected to some extentby 24handhadreturned to essen-tially pretreatmentlevelsby48 h(Fig. 3).

On the basis of these pilot studies, cohorts of CTT with

active colitis as definedby previously validated histological

cri-teria wereenteredintotrialstoreceive repeateddosesof BB 11,

EH8,or aplaceboonalternate days for8 d. Colonicmucosal biopsieswereobtained atthe

beginning

and2dafterthelast

mAbinjectiontoascertain thepresenceofmAb in thetissue,

theexpressionofE-selectin,andthehistologicalgrade ofcolitis activity. Administration oftheanti-E-selectin mAbs led, as

anticipated,todemonstrable levels ofmouseIgGon endothe-lial surfacesofCTTreceiving theseagents. Mostimportantly, the presence ofIgG on endothelial surfaceswas again asso-ciated withapparentblocking of E-selectin. Despite the

appar-entcoatingofE-selectinonendothelial surfaces, neitherBB 11 orEH8anti-E-selectinmAb led to asignificant improvement in colitis activity when comparedtoCTT receivingsaline pla-cebo(Table I).

The lack ofananti-VCAM1mAb that recognizes the CTT species precluded directassessmentof theeffect of blockade of this endothelial ligand on colitis activity. However,

prelimi-narystudies described above demonstrated thepresenceofa4

integrin,oneof thesubunits ofVLA4thecognateligand for VCAM1 onleukocytesurfaces inassociation withactive coli-tis. The ability ofananti-a4 integrin mAb, HPl /2,toblock

adhesion of CTT leukocytestorecombinant VCAM1 was

con-firmed by incubation of isolatedleukocyteswith HPl /2 before

exposure to a surface coated with the recombinant protein (Fig. 4). Adhesion of CTT peripheral blood leukocytes to

rsVCAM1 was completely blocked at concentrations of 10

,ug/ml.

Subsequently, the effect of the anti-a4 integrin mAb on

active CTT colitiswasassessed ina mannersimilartothat used in thestudy ofthe anti-E-selectin mAbs. Initial administration of HPl /2 demonstrateda somewhat shorter serum half-life

(T,,2

= 16 h) than that observed for BB11 and EH8. Initial circulatingvalues after HP1/2 (2mg/kg) rangedfrom 10-20 ,ugfallingtothe 3-5 ,ug/mlrangeat24 h. This indicated the feasibilityofachievingconcentrations that affectblockingof CTT

VLA4-dependent

adhesionevents.Because of the shorter half-life, HPl/2 (orsalineplacebo)wasgiven by daily injec-tion for 8 d and the effectoncolitis monitoredbyhistological

assessment_{of colonic mucosal biopsies.}

In contrast to theanti-E-selectin mAbs, HPl/2 ledtoa

highly significant improvementinacutecolitis when compared

toplacebo,P<0.01 (Fig.5 andTableII).Of12CTTreceiving anti-a4integrin mAb,all butoneshowedareduction in colitis activityatthe end of the I0-d study

period;

eightshowed

com-pletequiescenceoftheacuteinflammation.Incontrast, six of the CTTreceiving placeboshowed reduction in coliticactivity, andonlythree had resolution ofacuteinflammation.Thus,the

mean level ofacute

inflammatory activity

was reduced from 1.6±0.3to 0.2±0.1 after administration oftheHPl/2,while activitydiminished from 1.8±0.5to1.2±0.2 in CTTreceiving

saline placebo during thesameinterval (P<0.01). Not surpris-ingly, no change wasobserved inhistological features that

-:p

.4 tf,

k

<I

'SF+T SI

we. It:4

L.p;' t "

*: . <siA

A~~~~~

2 N

At

V

I

4 .a' .;;

0*

;F' . I

P'

P.

a

A

'V.

_I

11%~~ ~

I')o~

,-0

_~~~~ai,r 9L

sit

'ii .

8 ~ f ; ^4b :t

4sl

a t f

+

~~~~~~~~~~~~~~~~.

4i

Figure1. Expressionof VCAM 1 in colonicmucosaof

cotton-top tamarins. Colonic mucosalpinch biopsieswereobtained from CTT with active(A)andinactive(B)colitis. Frozen

sec-tions(--4pm)wereincubated withaffinity-purifiedrabbit anti-human rsVCAM1.Afterwashing,sectionswerestained

us-ingaperoxidase-baseddetection system(Vectastain;Vector

Labs)and thencounterstained withmethylblue.Control

sec-* ~~~tions (C)wereincubated withanonimmune rabbit antiserum rather than theanti-rsVCAM1.X450.

A t

V~~~V

f

bSt

~~~~~~~~~.

Ax~

Figure 2. Expressionof A4 integrin in colonic mucosa of cotton-top tamarins. Colonic mucosalpinch biopsieswereobtained from CTT with active (A) and inactive (B) colitis. Frozen sections ( - 4gm)wereincubated withanti-a4integrinmAbHPl/2.Specificallyboundantibody

wasvisualized with the peroxidase-basedtechnique usingaconjugated rabbit anti-mouse IgGasnoted inlegendtoFig. 1 and Methods. Control section (C)wasincubated with supernatant from the parent NS 1myelomacellline.

flect chronic mucosalinjuryusingaseparatevalidated scale of chronicinjury(24),meanchronicactivityremained thesame

beforeandafter the study periodin both CTTreceivingHPI/2

and thosereceivingtheplacebo.Thehistologicalimprovement in acute colitis activity was associated with an increase in

weight.MeanweightofCTTreceivingmAb HPI/2increased

from 394to422 gduringthetreatmentperiod(+5.4%),while

CTTreceivingsalineplaceboshoweda meanchangeof-2.7% (409-399g).Serialleukocytecountsobtainedduringthe

treat-mentperiod didnotshow anyconsistentchangesinthe

C

I

Figure2(Continued) eral bloodduringthe first 2 d ofthe trial.However,asindicated

in Fig. 6 A, moderate increases in the peripheral leukocyte

counts were observed by day four. Subsequently, levels fell

somewhatbut remained above initial levels at the conclusion

of the study. Paralleldeterminations showedconcomitant

re-ductions inthe hematocrit of both CTTreceiving mAb and

ELAM-Expresson1

+ I

o

rn

0 1 12 24 48

TIME(Hours)

Figure 3.Effect of systemic anti-E-selectin mAbs ondetectable

co-lonic mucosal E-selectin.CTTswithactivecolitis(n=_{6 pergroup)}

weregivenBB11 or EH8anti-E-selectin mAbs intramuscularly(2

mg/kg)orsaline placebobyintramuscular injection. Colonic

muco-salbiopsieswereobtainedatvarying intervals. E-selectinwasdetected infrozen sectionsbyincubationwith BB 11 mAb andvisualization

usingaperoxidase-baseddetection systemasdescribedinMethods.

thosereceiving placebo injection (Fig.6 B). Thismay reflect the effect ofrepeated phlebotomy; nosignsof overtbleeding diathesiswerenoted.

Discussion

Recent advanceshave ledtotheidentificationofmanyofthe receptors and counter-receptors present on endothelial and

TableL. Effect ofanti-E-selectinmAbsonCTT Colitis

Treatment group Pretreatment Posttreatment A

I. mAb-BB l1 (n= 5) 2.2±0.2 1.8±0.5 0.4

P> 0.1 Control(n=5) 2.0±0 1.4±0.5 0.6

II. mAb-EH8 (n= 5) 2.2±0.2 1.4±0.5 0.8

P>0.1 Control (n= 5) 2.2±0.2 1.2±0.5 1.0

CTT with active colitis received saline or either of twoanti-E-selectin

mAbs,BB 1,orEH8(2 mg/kgi.m. ondays 0, 2,4, 6,and8).Colitis

wasassessedby blinded evaluation of colonic mucosal biopsiesusing

astandardized gradingscaleranging from3+(severe)to0(inactive).

Separate placebo (saline) controlgroup weretreated in parallelwith eachmAbtreatmentgroup.

p0

. _

c %

1000- Figure 4. Effect of

anti-Cells a4integrinmAbs on 750 - VCAM 1-mediated

bound/ adhesion of CTT leuko-500.- cytes. Peripheral

mm2

_T

leukocytes obtained

250- from CTTwere

incu-batedwith rs VCAM

1-0 - coatedglassplates

TamarinPBLs aloneorinthepresence

Tamarin PB~s

of anti-humanVCAM

1

mAb 4B9 F(ab)2 or

anti-a4integrinmAb.HPl/2 Adhesionwasassessed as the number of adherent cells per squaremillimeter.*, rsVCAM; 1, 4B9(Fab')2;

l,

HP1/2.

leukocytesurfaces that participate in theadhesion of these cell populationsto vascularsurfacesatsites ofactiveinflammation. Key endothelial surface proteins including E-selectin and VCAM 1 are notconstitutively present but are induced by in-flammatory cytokines including interleukin-1 and tumor ne-crosisfactor (2, 4, 32-36). Previousstudies by ourselves and others haveconfirmedthe presenceofE-selectinonendothelial surfaces ofcolonicmucosa inassociation with active inflamma-tioninpatients withIBD(24, 25). E-sgelectinappearstobea

key ligandrecognized byneutrophilsand anecessaryfactor in recruitmentoftheseleukocytes to sitesof inflammation.

Ac-tivecolitisinpatients withIBD,particularly UC, ismarkedby ahigh concentration ofneutrophils presumably recruited from the vascular space.

Inthestudiespresented above, theimportance of E-selec-tin-mediated neutrophil adhesionto sustaining active colitis was explored using the CTT as a primate model of colitis closely resembling UC in man. Earlier studies demonstrated that activecolitiswasassociatedwith endothelial

expression

of E-selectinina mannersimilartothat observed in IBD. Paren-teraladministration ofanti-E-selectinmAbs

apparently

coated E-selectin present on endothelial surfaces.

However,

despite

thepresumed central role of E-selectin-mediated

neutrophil

TableII. Effect ofanti-a4Integrin mAbonCTTColitis

I. Acuteactivity (mean±SD)

Pre-Tx Post-Tx A

HP 1/2 1.6±0.3 0.2±0.1 1.4

P<0.01 Placebo 1.8±0.5 1.2±0.2 0.6

II. Chronicactivity(mean±SD)

Pre-Tx Post-Tx

HP 1/2 1.6±0.5 1.6±0.5 Placebo 2.0±0.5 2.0±0.5

CTTwithactive colitis received salineoranti-a4integrinmAb HP1/2

(2mg/kgondays 0, 1,2,3, 4, 5,6,and7).Colitiswasassessedby

blinded evaluation of colonic mucosalbiopsies usingstandardized

gradingscaleforacuteandchronicinflammation,eachranging3+

(severe)to0(noabnormality).Tx,treatment.

2+

AcuteCONSl 1+

Pre Post Pro Post HP 1/2 Placebo (Anti-VLA4) Control

Figure 5. Effect of anti-a4 integrin mAb on CTT Colitis. 24 CTT with acute colitic activity (grade 1+ to 2+ according to a previously validated scale [ 23])wererandomized (n= 12 per group) to receive

eitheranti-a4integrinmAb HPI/2 or saline placebo by intramus-cularinjectionlasting for7d. Acutecolitis was evaluated in a blinded

fashiononbiopsies obtainedonday 0 (pre) and day 10 (post).

adhesion, neither of two anti-E-selectin mAbs studied had a demonstrableeffectonCTT colitis. The lack of effect suggests eitherthat E-selectin-dependent neutrophil adherence is not necessary to sustain the inflammatory process or that despite blockingdetection of E-selectin by immunohistochemical ap-proaches,functional ligand is still accessible to neutrophils. If, as these results would suggest, E-selectin is not necessary to sustain neutrophil migration into the colonic mucosa, other pathways of neutrophil adhesion must be present in associa-tionwith thecolitis.However, it is also possible that a benefi-cial effectof E-selectin blockade could have been obscured by complement-mediated inflammation as an indirect effect of immune complex formation on the vessel wall. It is possible thatF(ab)2fragmentsmay be needed to demonstrate the im-portanceof E-selectinmodulation in CTT colitis in a manner similartothat observed in a model of lung inflammation in the rat(16).

InadditiontoE-selectin, earlier studies had demonstrated theexpression ofanother keyligandVCAM 1 that participates inadhesionofcirculatinglymphocytes and monocytes in hu-man colonic mucosa. However, this ligand was found to be present inhighestconcentrations in mucosal lymphoid aggre-gates rather than endothelialsurfaces.Incontrast toE-selectin, VCAM 1 expression wasfound to be present independent of

acuteinflammation, makingdeterminationofits relative im-portance in theacuteinflammatoryprocess difficult.Although

40

-35

-30

-25

-20 -

1S-HCTM%

o(placebo) x(HP112)

WC(x 104mnPI

+(placebo) *(HP 1/2)

Day 0 2 4 6 a 10

Figure6.Effectofanti-a4integrinmAb HPI/2on_peripheralblood

countsin CTT.Peripheral leukocytecountsandhematocritwere de-terminedonserial bloodsamplesobtained from CTTreceiving

anti-A4integrinmAbHPl/2(2 mg/kg)every otherdayasdetailed in

text.

the reagents available atthe time of initial studiesprecluded

assessment ofthe expression of VCAM 1 inCTT,in thepresent

studies the similarityof expression of this adhesion molecule wasconfirmedwith anti-VCAM 1affinity purifiedfrom a

con-ventional antiserum. These studiesdemonstrate the presence of the ligand on endothelial surfaces in both CTT andhuman

colonic mucosa in addition to its expression in mucosal

lym-phoid aggregates.

Unfortunately,theaffinity-purifiedrabbitanti-VCAM 1

re-agent wasunsuitable to use in studies to determine the effect of blockade of VCAM 1-mediated adhesion on colitis in amanner

similar to anti-E-selectin mAbs.However,asdemonstrated in this report, peripheralleukocytesexpress a4integrin,one sub-unit of the cognateligandfor VCAM1,onleukocyte surfaces, offeringanother target for assessing these ligands on colitis

ac-tivity. Thus, an anti-a4 integrin mAb was demonstrated to

prevent adhesionofperipheral CTT leukocytesto recombinant

VCAM1 indicating its capacity to functionally disrupt the

VLA4-VCAM1mediated process. While anti-E-selectinmAbs had noappreciable effectson CTTcolitis,administrationofan

anti-a4integrinmAbsignificantly attenuatedacutecolitic ac-tivity. Theabilityofanti-a4integrinmAb to effectan

improve-mentin the coliticactivity in CTT provides evidence ofthe central role thatadhesion eventsmediated through

heterodi-meric integrins containing this subunit playinthis model of IBD. Among the several processes in which the A4 integrin

chainis known toparticipate,itisplausiblethattheeffectsof the HPl/2mAb are at least in part theresultofdisruptionof

leukocyte-endothelial adhesion. Thus, the HPl /2 antibody

wasfoundtoinhibit adherence of CTT leukocytesto VCAM l-coatedplatesinvitro,aprocessdependentonVLA4

_(a4f13).

It

is perhaps surprising thatblocking this recruitment pathway directed toward lymphocytes and monocytes should have a more profound effect than disruption ofthe E-selectin-me-diated pathway directed towardsneutrophils,the more

charac-teristic inflammatory cell hallmark ofacuteinflammation in IBD. Notably, the anti-A4 integrin led toattenuation ofthe

histologic signs ofacutecolitis but hadnodiscernable effecton

histologic features associated with chronic inflammatory in-jury inthismodel.

Although the attenuation of CTT colitis may be most

di-rectlyattributabletothedisruptionofleukocyte-vascular adhe-sion, it should be noted that VLA4 appears to have several functional properties, andit is notpossibleto know apriori which of these may have been most central to the observed beneficial effect of anti-A4oncolitis(30). Thus, the VCAM/

VLA4 pathway mayplay a roleinimmunecellactivationas well asrecruitment( 19, 21 ).VLA4isknowntocontributeto T

cell-T cell aggregation and, in addition, binds ligandsother

than VCAM1includingfibronectin(18). Itis also possiblethat

theeffect of the HPl /2,anmAbdirected against thea4chain of the

a4flB

dimerthat comprises VLA-4, may in fact reflect bindingto a4associated with f7, theligand for lymphocyte

Peyer's patchhoming receptors. Recent studieshaveindicated that anti-a4 not only blocks lymphocyte migration into

in-flamed tissue inthe rat,butalsostronglyinhibits normal

lym-phocyte recirculation into mesenteric nodes and intestinal Peyer's patches (37-39). Nonetheless,thereduction incolitis observed suggestsanessential role for the ongoing recruitment oflymphocytesand/or monocytesforsustainingactivecolitic

activity.

It isclearthat the

_impact

on CTT colitiswas nota

nonspecific

effectonconcentrations of

_circulating

_leukocytes;

levels were not decreased but were paradoxically increased

withinafew

days

of theinitiationofdaily injectionswith the

anti-a4

_integrin

mAb.

Asamodel ofulcerative

colitis,

thesestudiesusingthe CTT

underscorethe

potential therapeutic impact

ofdisruptingthe

recruitment ofinflammatorycells fromthe vascular spaceto

the site ofintestinal inflammation. Furtherstudywillbe

neces-sarytodefinethepathwaysessential to themigrationof

neutro-philsin thecontextof activecolitis.The presentfindings

sug-gestthat theongoingrecruitmentofotherleukocytesmust facil-itate theabundantneutrophil representationin themucosain this

setting.

More detailed delineation of the mechanism of

anti-a4 inhibitionof CTT colitis mayprovidefurther

_insight

into the_relationship betweenlymphocyteand monocyte

mi-grationand the recruitmentofneutrophilsin active colitis.

Acknowledgments

The authors thank GloriaChi-Rosso,IreneDougas,KathleenKimball,

and Diane Leone for technical assistance.

These studiessupported by grantsfrom the National Institutes of Health (P30DK43351 and RO1DK36350 to D. K. Podolsky, and NERPRC 5P5IRROO 168)andBiogen,Inc.

References

1. Butcher, E. C. 1991. Leukocyte-endothelial cellrecognition:three(or more)stepstospecificityanddiversity.Cell. 67:1033-1036.

2.Bevilacqua,M._P.,S.Stengelin,M. A.Gimbrone, Jr.,and B. Seed. 1989. Endothelialleukocyteadhesion molecule 1:aninducible receptor forneutrophils

relatedtocomplement regulatory proteinsandlectins. Science(Wash. DC).

243:1160-1165.

3.Yasuyuki, I.,M.S.Singer,C.Fennie,L. A.Lasky,and S. D. Rosen. 1991. Identification ofacarbohydrate-basedendothelialligandforalymphocyte

hom-ingreceptor.J.CellBiol. 113(5):1213-1221.

4.Osborn, L.,C. Hession,R.Tizard,C. Vassallo,S.Luhowskyj,G.

Chi-Rosso,and R. Lobb. 1989. Direct expressioncloningof vascular cell adhesion. 1.

Molecule,acytokine-inducedendothelialproteinthat bindstolymphocytes.

Cell.59:1203-1211.

5.Picker,L.J.,A.Warnock,A.R.Burns,C.M.Doerschuk,E.L.Berg,and E. C. Butcher. 1991. Theneutrophilselectin LECAM-1 presentscarbohydrate ligandstothe vascularselectins ELAM-IandGMP-140.Cell.66(5 ):921-933.

6. Lobb,R.R.,G. Chi-Rosso,D. R. Leone, M. D. Rosa, S. Bixler, B. M.

Newman,S.Luhowskyj,C. D. Benjamin, I. G. Dougas, S. E. Goelz et al. 1991.

Expressionand functional characterization of a soluble form of

endothelial-leu-kocyteadhesion molecule 1. J. Immunol. 147:124-129.

7.Jonjic, N.,P.Jilek,S.Bernasconi, G. Peri, I. Martin-Padura, S.Cenzuales,

E.Dejana,and A. Montovani. 1992. Molecules involved in the adhesion and

cytotoxicityof activated monocytesonendothelial cells. J. Immunol. 148:2080-2083.

8.Montgomery,K.F.,L. Osborn, C. Hession, R. Tizard, D.Goff,C.Vassallo,

P. I.Tar,K.Bomsztyk,R.Lobb,and J. M. Harlan. 1991. Activation of endothe-lial-leukocyte adhesion molecule 1 (ELAM-1) gene transcription. Med. Sci. 88:6523-6527.

9. Carlos, T. M., B. R. Schwartz, N. L. Kovach, E. Yee, M. Rosso, L.Osborn,

G. Chi-Rosso, B. Newman,R. Lobb, and J. M. Harlan. 1990. Vascular cell adhesion molecule- 1 mediates lymphocyte adherence tocytokine-activated cul-tured human endothelial cells. Blood. 76:965-970.

10. Kuijpers, T. W., M. Hoogerwerf, L. J. W. Van der Laan, G. Nagel, C.E.

Van der Schoot, F. Grunert, and D. Roos. 1992. CD66 nonspecific cross-reacting

antigensareinvolved inneutrophiladherencetocytokine-activated endothelial cells. J. Cell Biol. 118:457-466.

11.Graber, N.,T.V.Gopal,D.Wilson, D. L. Beall, T. Polte, and W.

New-man.1990. T cells bind to cytokine-activated endothelial cells via a novel, induc-iblesialoglycoprotein and endothelial leukocyte adhesion molecule- 1. J. Im-munol. 145:819.

12. Shimizu,Y.,W. Newman, T. Venkat Gopal, K. J. Horgan, S. N.Graber,

L. Dawson Beall, G. A. Van Seventer, and S. Shaw. 1991. Four molecular

13.Lobb, R. R., G. Chi-Rosso, D. R. Leone, M. D. Rosa, S.Bixler,B. M. Newman,S.Luhowkyj,C. D.Benjamin, I. G. Dougas, S. E. Goelz, et al. 1991. Expression and functional characterization of a soluble form of endothelial-leu-kocyteadhesion molecule1.J.Immunol.147:124-129.

14.Picker,L.J., T. K.Kishimoto,C. W.Smith,R. A.Warnock,and E. C. Butcher.1991. ELAM-Iis an adhesion molecule for skin-homing T cells. Nature

(Lond.). 349:796-799.

15.Shimizu,Y.,S.Shaw, N.Graber,T. V.Gopal,K.J.Horgan,G. A. Van Seventer, and W. Newman. 1991. Activation-independent bindingof human memory T cells toadhesionmolecule ELAM-1.Nature(Lond.).349:799-802.

16.Mulligan,M.S.,J.Varani,M. K.Dame,C. L.Lane,C. W.Smith,D. C. Anderson, and P. A. Ward.1991.Roleofendothelial-leukocyteadhesion mole-cule 1(ELAM-1)inneutrophil-mediated lung injuryin rats. J. Clin. Invest.

88:1396-1406.

17.Phillips,M.L.,E.Nudelman,F. C. A.Gaeta,M. Perez, A. K.Singhal, S. I.

Hakomori,and J.C.Paulson. 1990. ELAM-1 mediates cell adhesionby

recogni-tionofacarbohydrateligand, Sialyl-Lex.Science(Wash.DC). 250:1130-1132. 18.Elices,M.J., L. Osborn, Y.Takada,C.Crouse, S.Luhowskyj,M. E. Hemler, and R. R. Lobb.1990. VCAMI onactivated endotheliuminteracts with theleukocyteintegrinVLA-4at asite distinct from theVLA-4/fibronectin bind-ing site.Cell. 60:577-584.

19.Pulido,R., M. J.Elices,M. R.Campanero,L.Osborn,S.Schiffer,A.

Garcia-Pardo,R.Lobb,M.E.Hemler,and F. Sanchez-Madrid. 1991. Functional

evidence forthree distinct andindependentlyinhibitable adhesion activities

me-diated bythe humanintegrinVLA-4.J. Biol. Chem. 266:10241-10245.

20. Tyrrell, D., P.James, N. Rao, C. Foxall, S.Abbas,F. Dasgupta, M. Nashed, A. Hasegawa, M.Kiso, D. Asa, J.Kidd,and B. K. Brandley. 1991.

Structuralrequirementsfor thecarbohydrate ligandof E-selectin.Biochemistry. 88:10372-10376.

21. VanDinther-Janssen,A.C. H.M.,E.Horst,G.Koopman,W.Newmann,

R. J.Scheper, C. J. L. M.Meijer,and S. T. Pals. 1991. TheVLA-4/VCAM-I

pathwayis involved inlympthocyteadhesiontoendothelium in rheumatoid

synovium.J.Immunol. 147:4207-4210.

22.Berg, E.L.,M. K.Robinson,0.Mansson,E. C.Butcher,and J. L.

Mag-nani.1991. Acarbohydrate domaincommon tobothSialylLe'andSialylLeXis

recognizedby theendothelialcellleukocyteadhesion molecule ELAM- 1. J.Biol. Chem.266:14869-14872.

23. Berg, E.L.,J.Magnani, R.A.Warnock, M. K. Robinson,and E. C.

Butcher. 1992.Comparison of L-selectin andE-selectinligand specificities:the

L-selectincanbindtheE-selectinligands Sialyl Le,and SialylLe'. Biochem.

Biophys.Res.Commun. 184:1048-1055.

24.Koizumi, M.,N.King,R.Lobb,C.Benjamin,and D. K.Podolsky.1992.

Expression of vascular adhesionmolecules ininflammatoryboweldisease.

Gas-troenterology.103:840-847.

25.Malizia, G.,A.Calabrese,M.Cottone,M.Raimondo,L. K.

Trejdosie-wicz, C. J. Smart, L.Oliva,and L.Pagliaro.1991.Expressionofleukocyte

adhe-sionmoleculesbymucosalmononuclearphagocytesininflammatorybowel dis-ease.Gastroenterology. 100:150-159.

26.Wallace,J.L.,A.Higa,G. W.McKnight,and D. E.MacIntyre. 1992.

Preventionand reversalofexperimental colitis by a monoclonal antibody which inhibitsleukocyte adherence.Inflammation.16:343-354.

27. Madara, J. L., D. K. Podolsky, N. W.King,P. K.Seghal, R. Moore, and H. S.Winter. 1985. Characterization of spontaneous colitis in cotton-top tamar-ins (Saguinus oedipus) and its response to sulfasalazine. Gastroenterology.

88:13-9.

28. Podolsky, D. K., J. L. Madara, N. King, P. Sehgal, R. Moore, and H. S. Winter. 1985. Colonic mucin composition in primates. Selective alterations

asso-ciated withspontaneous colitis in the Cotton-top tamarin. Gastroenterology. 88:20-25.

29.Benjamin,C.,I.Dougas, G. Chi-Rosso, S.Luhowskyj,M. Rosa, B. New-man, L.Osborn, C. Bassallo, C. Hession, S.Goelz,K.McCarthy, and R. Lobb. 1990. A blocking monoclonal antibody to endothelial-leukocyte adhesion

mole-cule-l(E-selectin). Biochem.Biophys. Res. Commun. 171:348-353.

30. Pulido,R., M. J. Elices, M. R. Campanero,L. Osborn,S. Schiffer, A.

Garcia-Pardo,R. Lobb, M. E. Hemler, and F. Sanchez-Madrid. 1991. Functional

evidence forthreedistinctand independentlyinhibitableadhesion activities

me-diatedby the human integrin VLA-4.J.Biol. Chem. 266:10241-5.

31. Lobb, R., G. Chi-Rosso, D. Leone, M. Rosa, B. Newman, S.Luhowskyj,

L.Osborn, S.Schiffer,C.Benjamin, I. Dougas, C. Hession, and P. Chow. 1991. Expression and functional characterization of a soluble form of vascular cell

adhesionmolecule 1 (VCAM-1). Biochem. Biophys. Res. Commun. 178:1498-1504.

32. Kansas, G. S.,0. Spertini, L. M. Stoolman, and T. F. Tedder. 1991. Molecular mapping of functional domains ofthe leukocyte receptor for endothe-lium,LAM-l. J. CellBiol. 114:351-358.

33. Yamada, K. M. 1991. Adhesive recognition sequences. J. Biol. Chem.

266:12809-12812.

34.Hession,C. L., Osborn, D. Goff, G. Chi-Rosso, C. Vassallo, M. Pasek, C.

Pittack,R. Tizard, S.Goelz, K. McCarthy, S. Hopple, and R. Lobb. 1990. Endo-thelial leukocyte adhesion molecule 1: direct expression cloning and functional

interactions. Cell Biol.87:1673-1677.

35.Carlos,T. M., B. R. Schwartz, N. L. Kovach, E. Yee, M. Rosso, L. Osborn,

G.Chi-Rosso, B. Newman, R. Lobb, and J. M. Harlan. 1990. Vascular cell

adhesionmolecule- Imediates lymphocyteadherencetocytokine-activated cul-turedhumanendothelialcells.Blood.76:965-970.

36.Postigo,A.A., R.Garcia-Vicuna,F.Diaz-Gonzalez, A. G. Arroyo, M.0. DeLandazuri,G. Chi-Rosso, R. R. Lobb, A. Laffon, and F. Sanchez-Madrid.

1992.Increasedbindingof synovial T lymphocytes from rheumatoid arthritis to

endothelial-leukocyte adhesionmolecule- I (ELAM- I ) and vascular cell adhesion

molecule-l (VCAM-I).J.Immunol. 89:1445-1452.

37.Issekutz, T. B. 1991.Inhibitionofin vivolymphocytemigration to

inflam-mationand homing to lymphoid tissues by the TA-2 monoclonal antibody: a

likelyroleforVLA-4 in vivo.J.Immunol. 147:4178-4184.

38.Ruiegg, C.,A. A.Postigo,E. E.Sikorski,E.C. Butcher, R. Pytela, and D. J. Erle. 1992. Roleof integrina4,77/a4flPin lymphocyte adherence to fibronectin andVCAM-Iand inhomotypiccellclustering. J. Cell Biol. 117:179-189.

39.Holzmann,B., andI.L.Weissman.1989. Integrin molecules involved in lymphocytehomingtoPeyer's patches.Immunol.Rev.108:45-62.