1985pdf

14352_21985.pdf

Blood,Vol72.No6(December),1988:pp1985-19911985Lymphokine-ActivatedKillerCellsFromNormalandLymphomaSubjectsAre

CytotoxicforCellsCoatedWithAntibodyDerivativesDisplayingHumanFc'y ByR.J.Dearman,F.K.Stevenson,M.Wrightham,T.J.Hamblin,M.J.Glennie,andG.T.Stevenson Lymphokine-activatedkiller(LAK)cellsweresuccessfully generatedinallcasesfrombloodmononuclearcells obtainedfromsixpatientswithlymphoma.TheLAKcells fromthreeofthesepatientsandfromfivenormaladult donorsweretestedfortheireffectorabilitiesinantibody- dependentcellularcytotoxicity(ADCC)againstguineapig leukemiclymphocytescoatedwithvariousantiidiotype antibodies.Cellsfromallthedonorsbehavedsimilarly.

MousemonoclonalantibodiesofIgGi.lgG2a.andlgG2b

isotypes invokednoADCC.However.substantialADCC wasinvokedbythechimericantibodyFabFc.inwhich

HUMANLYMPHOCYTESactivatedincultureby

interleukin-2(IL-2)for48hoursorlongeracquirea broadcytotoxicpotentialthatsometimesencompassesautol- ogoustumorcells.'.2Strenuouseffortsarebeingmadeto exploitthisphenomenontherapeutically.3

Thelymphokine-activatedkiller(LAK)cellsappearpre-

dominantlytoarisefromlymphocytesdescribedvariouslyas null(ie,non-B,non-T),largegranular,ornaturalkiller (NK)-theclassesdefinedbyeachofthesethreedesigna- tionsshowingatleastanextensiveoverlap.
7Itshouldbe notedthatNKtargetsaremuchmorerestrictedthanLAK targets,2althoughthismightresultfromtheactivationby

IL-2raisingcytotoxicperformanceabovepreviouslylow

levels,ratherthanpermittingattacksoncompletelynew targets.7AminorcontributiontoLAKactivityappearsto derivefromTcellsrenderedbroadlycytotoxicinculture.5

Weshallaccepttheview(summarizedinreference7)that

largegranularlymphocytes(LGL)areadistinctclassdistin- guishedbymorphology,surfacemarkersthatincludethe Fc'y-receptorCD16,andthetwocytotoxicactivitiesNKand LAK.Athirdcytotoxicroleattributedtothemisthatofthe principallymphocyticeffector,orKcell,inantibody- dependentcell-mediatedcytotoxicity(ADCC).Heretarget cellsaremarkedfordestructionbyacoatingofIgG antibody,whichadherestotheFc-y-receptorofLGL.89 ConsistentwiththisactivityofLGL,murineLAKcellshave recentlybeenshowntobeactiveinADCC.'#{176}

Thecellsofnon-Hodgkin'slymphoma(NHL)presentan

attractivetargetforLAKcelltherapy,beingdisseminated andoftenhavingreadyaccesstotheblood.Intheclinical studycited,3twoof108patientsreceivingLAKcellshad

NHL.Bothunderwentremissions,onecompleteandone

partial.

WehavebeeninterestedforsomeyearsintreatingNHL

withantiidiotypeantibody(anti-Id)oritsderivatives."Early therapeuticeffortshavegivenvariableresults,'2'6commonly apartialremissioninducedbyinfusionofuptoseveralgrams ofantibody.Amongthedifficultiesthathaveemergedare modulationofsurfaceidiotype,'7mutationoftheidiotype,'8 andanuncertaintyabouttherolesofthevariouseffectorsin eliminatingtheantibody-coatedneoplasticcells."Thelastof theseproblemsimpedestherapeuticstrategy:itisnotclear whetheroneshouldbetryingtoinvokecomplementor ADCCorscavengingbymacrophagestoclearthecells.Fab"yfrommouseantiidiotypeisthioether-bondedto humannormalFc'y.Similarresultswereobtainedontest- ingLAKcellsfromanormaldonoragainstuncultured humanlymphomatargetscoatedwithnativeorchimeric antiidiotype.TheADCCinvokedbythemouse-human chimericantibodiesappearstodependonthehumanFcy theydisplayandnotontheunivalencyofthederivatives used.ThefindingsimplythatLAKtechnologycouldusefully augmentserotherapythatusesantibodyderivativesdis- playinghumanFcy.

S1988byGrune&Stratton.Inc.

Readymeansofenhancingtheactivitiesoftheseeffectors havenotbeenavailable.Now,however,LAKcelltechnology promisestoenhanceADCCandmighttherebyusefully complementtheinfusionofantibody.

ThisreportdescribestheabilityofLAKcells,from

normalorNHL-bearingsubjects,tokilllymphomacells coatedwithanti-Id.Amongtheantibodiesusedisthe chimericderivativeFabFc,whereFab"yfrommousemono- clonalanti-Idisthioether-bonded,hingetohinge,tohuman normalFc'y.Thisderivativeisunivalent,therebyavoiding rapidantigenicmodulation,andpossesseshumaninsteadof mouseFc'ysoastoprolongmetabolicsurvival,reduce immunogenicity,andbetterrecruithumaneffectors."'6We presentevidencethatLAKcellsarestronglyandconsistently effectiveinADCCwhenpresentedwithhumanFc
yonthe targetcells.

MATERIALSANDMETHODS

PatientsandLAKcellgeneration.SixpatientswithNHLwere chosenforstudy(Table1).Theyhadallpresentedwithfollicular centercelllymphoma(FCC),andnonewasreceivingchemotherapy atthetimeofsamplingforLAKcellgeneration.Allgaveinformed consentaccordingtoguidelinesofthelocalEthicalCommitteeafter beinginformedoftheproceduresandattendantrisks.Normal donorswerehealthylaboratorypersonnelintheagerangeof21to55 years.Bloodwascollectedintopreservative-freeheparinandsepa- ratedinFicoll-Hypaque.Cellscollectedattheinterfacewere washedwithEagle'sminimalessentialmedium(MEM)containing heparin(20U/mL)andcounted.Theywerethenresuspendedin flasksat1.5xI0'/mLinRPMI1640mediumcontainingglutamine (200mmol/L),pyruvate(100mmol/L),penicillinandstreptomycin FromtheLymphomaResearchUnit,TenovusLaboratory,Gen- era!Hospital,Southampton,UK.

SubmittedMarch9,1988;acceptedJuly29,1988.

SupportedbyTenovusandtheCancerResearchCampaign.

AddressreprintrequeststoG.T.Stevenson.MD,Lymphoma

ResearchUnit,TenovusResearchLaboratory.SouthamptonGen- era!Hospital.TremonaRd.Southampton5094XY,England. Thepublicationcostsofthisarticleweredefrayedinpartbypage chargepayment.Thisarticlemustthereforebeherebymarked "advertisement"inaccordancewith18U.S.C.section1734solelyto indicatethisfact.

©/988byGrune&Stratton,Inc.

0006-4971/88/7206-0017$3.OO/ODownloaded from http://ashpublications.org/blood/article-pdf/72/6/1985/598624/1985.pdf by guest on 16 August 2023

1986DEARMANETAL

Table1.GenerationofLAKCellsFromPatientsWithB-CellLymphoma %SpecificLysis(±SD)ofIndicatedTargetsatIndicatedEffector:TargetRatios

TumorCellsinBlOOd

Patient(%ofIndicatedK562DaudiAutologousTumor

(Sex,Age[yr])DiseaseTypeandStatusMononuclearCount)40:120:140:120:140:1

S.A.CB/CC.nodular

M,72Partialremissionfor

2yrpostchemotherapy071±369±486±277±730±6

N.L.CC,diffuse

M.60Progressiveon

intermittentchemotherapy53%of7200/zL84±577±272±266±6-4±2

BR.CB/CC.nodular

F,55Partialremissionfor

2yrpostantibodyt60%of3,4004iL80±276±372±572±416±4

G.H.CC,diffuse

M,571mopostchemotherapy0666354510.5

JR.CB/CC.nodular

M,55Slowlyprogressiveon

intermittentchemotherapy075±572±7107±14100±7ND

SR.CB/CC,nodular

M,76Partialremissionfor

5yrpostchemotherapy067±254±393±472±3ND

Incubationtime,fourhours.

Abbreviations:CB,centroblastic;CC,centrocytic;ND.notdetermined. #{176}KieIclassification.26 tUnivalentchimeric(mouse/human)antiidiotype.1#{176}

(100IU/mL),fungizone(2zg/mL)(conditionedmedium[CM]),Monoclonalantibodies.Mousemonoclonalanti-Idspecificfor

and10%decomplementedhumanABserum,towhich10U/mLofsurfaceIgontheguineapigL2Cleukemiawereraisedaspreviously

recombinanthumanIL-2(donatedbyDuPont,Glenolden,PA)haddescribed.2'Anti-Id(I)andanti-Id(2)areofisotypesIgGiand

beenadded.Cellswereincubatedforthreetofivedaysat37#{176}CinanIgG2a,anddisplayKAforreactionwithsurfaceIgat37#{176}Cof1.Ix

incubatorwith5%CO2.Activatedcellswerecollectedbycentrifuga-i09mol/L'and2.3xI0#{176}mol/L',respectively.22Anti-Id(11),of

tionandsuspendedinmediumforcytotoxicityassays.isotypeIgG2b,wasraisedagainsttheA-chainofL2C1g.whichcan

Targetcells.ThehumanerythroleukemiacelllineK562andbeharvestedfromurineofleukemia-bearinganimals.23Itrecognizes

twohumanB-lymphoblastoidcelllines,DaudiandRaji,wereidiotypesexpressedonboththefreeA-chainsandtheIgMXsynthe-

culturedinCMcontaining10%fetalcalfserum(FCS).Theyweresizedbythetumor,anditfailstoreactwithnormal1gM.At

useddirectlyfromculture.Freshunculturedtumortargetsconsistedconcentrationsabove5
ig/mL,allthreeoftheseantibodiessaturate

ofpatients'tumorcellsobtainedfrombloodorinvolvedlymphnode,theL2CsurfaceIg(1.3to1.6xl0
molecules/cell
).

eitheratpresentationorduringdiseaseexacerbation(Table2).CellsMousemonoclonalantibodiesagainstidiotypicandisotypicdeter-

wereprepared,washed,andcryopreserved'9;theproportionoftumorminantsonpatients'tumorcells(N.L.andA.H.)wereraisedusing

cellsinthepreparationswasestimatedfromthesurfaceIgpheno-asantigenthe1gMrescuedfromthetumorcellsbyxenohybridiza-

typeobtainedbyimmunofluorescence.TargetBcellsfromthetion.24Anti-Idforeachpatient,bothofisotypeIgG1,wereselected

guineapigL2Cleukemiawereobtainedfreshfrombloodofterminalonthebasisofspecificbindingtotumor-derived1gMinthepresence

animalsasdescribedpreviously.20ofnormalhumanserum(50%).Specificbindingtothetumorcells

Table2.HumanLymphomaCellsUsedasUnculturedTargets

PatientSurfaceIgIsotype

(Sex.Age[wi)DiseaseTypeSourceofCells(Density,%Cells)

S.A4CB/CC.nodularLymphnodeI9MDX

M,69+(77)

N.L.CC.diffuseBloodIgMDA

M,60+(90)

BR.CB/CC.nodularBloodIgMDX

F,52+(85)

G.H.CC.diffuseBloodlgGx

M,57+(90)

A.H.CB/CC.terminalblasticBloodIgMA

M,72+++(90)

Abbreviations:CB.centroblastic;CC,centrocytic.

#{176}Kielclassification.25

tSurfaceIgwasdetectedbyimmunofluorescence.Levelsoffluorescenceareindicatedbyplussigns.andthepercentagesofpositivecellsinthe

preparationsareinparentheses.

ThefirstfourpatientsalsoappearinTable1asdonorsofLAKcells.AgesgivenareatthetimeofdonatingtumororLAKcells.Downloaded from http://ashpublications.org/blood/article-pdf/72/6/1985/598624/1985.pdf by guest on 16 August 2023

Fab'V 'S R ,S
R:
b

R':O,,sN

2

Cy3#{149}Cy2

y3(AKCELLSANDANTIBODYDISPLAYINGHUMANFc'y1987 wasconfirmedbyimmunofluorescence.AnIgG1antibodytohuman C
wasobtainedduringthegenerationoftheanti-Id,beingselected byspecificreactivitywith1gMpreparationsonELISA. Polyclonalantibodies.RabbitantiserumspecificforL2Csur- faceidiotypewasraisedagainstFab
sisolatedfromapapaindigest ofthetumorcellsurfaces.26TheIgGfraction,preparedbyprecipita- tionwithammoniumsulphateandDEAE-cellulosechromatogra- phy,wasabsorbedwithguineapignormalIgtoyieldpolyclonal anti-Id.FromthistheFab/cderivativewaspreparedbylimited digestionwithpapain.27 Chimericantibodies.Thesewereconstructedchemically. FabIgG,inwhichFab"yfrommouseantibodyisthioether-bondedto humannormalIgG,waspreparedasdescribedpreviously.
FabFc derivatives(Fig1)werepreparedusingFab'7frommousemono- clonalIgG1antibodyandFcfromhumannormalIgG.F(ab"y)2was derivedbypepticdigestion29withIgGI18mg/mL,pepsin(product P6887,SigmaChemicalCo.StLouis)0.3mg/mL,pH4.2,37#{176}C. ProgresswasmonitoredbypassingaliquotsthroughZorbaxGF250 gel(DuPont,Wilmington,DE),generallyrevealingthatapproxi- mately90%oftheIgG1hadbeensplitwithineighthours.The F(ab'-y)2wasthenseparatedbyrecyclingchromatographyonSepha- cryl5200gel(PharmaciaInc.Piscataway,Ni).ToprepareFc-y,the basicfractionofhumannormalIgG(unretardedonDEAE- Trisacryl[LKBInstruments,Gaithersburg,MD]in40mmol/L TrisHCl,pH8.0)wasdigestedwithpapain:IgG18mg/mL,papain (productP3125,Sigma)0.1mg/mL,2-mercaptoethanol(British DrugHouses,Poole,Dorset,UK)1mmol/L,pH7.4,37#{176}C,60 minutes.The50-Kdfractionofthedigest(Fab'yplusFc-y)was separatedonSephacryl5200andthenpassedontoDEAE-Trisacryl in20mmol/LTrisHCl,pH8.0.AllFab-ypassesthroughtheDEAE column,sincethestartingbasicIgGlacksthemoreacidicFab'y fractions,whiletheretardedFc'yremainstobeelutedwith0.5 human Fig1.Postulatedstructureofthechimericunivalentantibody FabFc.formedbyunionofFab"yfrommouselgGlantiidiotypeand Fc'yfromhumannormallgG.Mostofthey-lightSSbondsinthe

FabareknowntosurvivethemildreductionofF(ab"y)2by

2-mercaptoethanol;thosethatdonotbecomethioetherbridged.

Majorsetsofnoncovalentbondsaredepictedby(#{149}#{149}#{149}).V.

variable;C.constant;Hi,hinge;H,heavychain;Llightchain.mol/LTrisHCl,pH8.0.TheFc'ythuspreparedisessentially

entirelyfromIgG1:IgG2isrelativelyresistanttocleavageunderthe digestionconditionsused,IgG3yieldsalargeFcfragmentseparated ontheSephacryl,andIgG4isnotpresentinsignificantamountin thebasicIgG.

FabFccannowbepreparedby(1)reducingF(ab"y)2tomono-

mericformwithfreesulfhydryl(SH)groups;(2)allowingtheFab"y toreactwithalargemolarsurplusofo-phenylenedimaleimide (Sigma)toyieldFab"ywithfreemaleimidegroups;(3)allowingthe

FabmaleimidegroupstoreactwithSHofreducedFc.Inmore

detail,theF(ab"y)2wasreducedwith10mmol/L2-mercaptoethanol atpH8.0andthenseparatedonSephadexG25(Pharmacia)in10 mmol/LdisodiumEDTA,pH4.6.Totheelutedproteinwasadded anequalvolumeof4mmol/Lo-phenylenedimaleimidein40% (vol/vol)dimethylformamide,60%50mmol/Lsodiumacetate,pH

5.3.Thereactionwasallowedtoproceedfor30minutesat5#{176}C.The

proteinwasthenseparatedbybindingtothecation-exchangeresin Phospho-Ultrogel(LKB),fromwhichitwaselutedwith0.5mol/L sodiumacetate,pH5.3.Thenumberofmaleimidegroupsper molecule(obtainedbyaddingastandardsolutionof2-mercaptoe- thanolandback-titratingSHgroupswith2,2'-dithiopyridine
#{176})was

0.84to1.05,suggestingthatwheretwovicinalSHgroupsoccur,the

bismaleimidecompoundlinksthemintramolecularly.TheFabwas thenaddedtoFc'ythathadearlierbeenreduced(10mmol/L dithiothreitol,pH8.0),separatedonSephacryl5200,concentrated toabout10mg/mLinanAmiconUltrafiltrationcell(PMIO membrane),andkeptinanitrogenatmosphereatpH5.3.Themolar ratioFab:FcwasI:2.5.Theproteinmixturewasincubatedat20#{176}C fortwohoursbeforeseparatingtheFabFcproductonSephacryl

5200(FigI).ResidualSHgroupswithintheFchinge(theoretically,

threeoftheoriginalfourremain)wereencouragedtoreformatleast oneofthetwooriginalhingeregiondisulfidebondsbyundergoing disulfideexchangewithImmol/L2,2'-dithiopyridine(Aldrich, Gillingham,Dorset,UK)atpH5,3,30Acycleofabsorptionelution onproteinA-Sepharose(Pharmacia)gotridofasmallamountof contaminatingFabdimer.Theonlycontaminantthendetectable wasFcdimer(<10%).RecoveryofantibodyFab"yintheFabFcis usually50%to60%. Measurementsofcytotoxicity.Targetcellswereuseddirect fromculture(K562,Raji),freshfromthebloodofleukemia-bearing guineapigs(L2C),orimmediatelyafterthawing(lymphomacells frompatients).Cells(-3x106)wereresuspendedin100
Lof phosphate-bufferedsaline(PBS),and100
tLofNa25'Cr04(Amer- shamInternational,UK)wasadded.Incubationwasforonehourat

37#{176}C,followedbythreewasheswithPBSandresuspensionat

105/mLinCMwith10%FCS.

ForLAKcytotoxicity(intheabsenceofantibody),targetcells wereadded,atl0
cellsin100
L,towellscontaining100
L CM/FCSandeffectorcellsinnumbersspecifiedforindividual experiments.Cultureswereintriplicateinround-bottommicrotiter platesunlessindicatedotherwise.Afterincubationforfourhoursat

37#{176}Cin5%CO2.theplateswerecentrifugedat175gforfive

minutesandthesupernatantsharvested.Releasedisotopewas countedinay-counterwithmaximumreleaseestimatedfromcells treatedwith0.5%NonidetP40.Supernatantreleasewasdetermined byincubationoftargetcellswithCM/FCSonly.Thepercent specificlysiswascalculatedas experimentalcpm-spontaneouscpm .xIOO%maximumcpm-spontaneouscpm

ForADCC,afixedantibodyconcentration(20zg/mL)and

effector:targetratio(25:1)wereused,afterpreliminarytitrations hadshownthesevaluestoachieveplateaulevelsofcelllysis(Fig2).

Labeledtargetcells(10
)in50j
LCM/FCSweretreatedwith100Downloaded from http://ashpublications.org/blood/article-pdf/72/6/1985/598624/1985.pdf by guest on 16 August 2023

0a, CsCs 6
C.) U U Csa a,40 20 0

A1006.30.390.024

Antibody(pg/mI)

50
40
30U)
Cl) >- -J 0 IL 0uJ 0
U)20 10

B0255070100achieved.1988DEARMANETAL

EFFECTORTOTARGETRATIO

Fig2.TitrationsofantibodyandeffectorcellsforLAK-

cell-mediatedADCC.(A)LysisofL
Ccellsinvokedbyanti-Id FabFc(mouse-human).Eachstepalongthex-axisrepresentsa fourfolddilution.(B)LysisofL
CcellsmediatedbyLAKeffectors fromasingledonor(male,age30)inthepresenceofanti-Id antibodyat25
g/mL.Targetcellsweremaintainedat104/well.

PointsrepresentthemeansoftriplicateswithSD<5%.(A)

mouse-humanFabFc;(
)mouselgG;(#{149})noantibody.
Lantibody(50
g/mLinCM/FCS)orcontrolIgG,andincubated thusfor15minutesat4#{176}C.Effectorcellsin50
iLmediumwerethen addedandthesuspensiongentlycentrifuged(100g,fiveminutes)in thecoldtopromoteeffector-targetcontact.(Thesestepswere carriedoutinthecoldsoastoachieveeffector-targetapposition beforebivalentantibodiesinducerapidmodulationofsurfaceIg'7). Finally,thesystemwaswarmedto37#{176}Cforathree-hourincubation period.SpecificlysiswasassessedasforLAKcytotoxicity.

ToexaminetheeffectonADCCofblockingtheFc'yreceptor

CDI6,theeffectorcellswerepreincubatedfor30minutesat0#{176}C

withmousemonoclonalanti-CD16IgGiantibody(3G83')orcontrolIgG1antibodyatindicatedconcentrations,washedinthecold,and

thensuspendedincoldmediumforuseasabove.

RESULTS

LAKcellsfrompatientswithlymphoma.Thedisease

statusforeachofihesixpatientsisindicatedinTable1:none hadrapidlyprogressivetumoratthetimeofstudy.Asjudged bythekillingofeitherK562orDaudicells,bloodmononu- clearcellsfromallthepatientsgeneratedgoodLAKactivity (Table1).ItisnotablethatthisabilitypersistedinN.L.and

BR.,bothofwhomhadasubstantialproportionoftumor

cellsamongthebloodmononuclears.Normaladultdonors generatedLAKactivityagainstK562cellsatacomparable level:74%±10%(SD)specificlysiswithaneffector:target ratioof20:1.

LAKcellsfromfourofthepatientswerealsotested

againstcryopreservedautologoustumorcells.Intwoofthe fourcases,amodestdegreeoflysiswaseffected(Table1).

ADCCmediatedbyLAKcellsagainstaxenogeneic

target.TheguineapigL2Cleukemia,maintainedbycon- tinualpassageinvivo,provedaconvenientmodelforassess- ingtheADCCcapabilityofLAKcells:theL2Ccellslabel well,arerelativelyinsensitivetohumanLAKeffectorsinthe absenceofantibody,andhaveavailableagainstthemawell characterizedpanelofmousemonoclonalanti-Id.22

Resultsobtainedusingcellsfromfivenormaldonorsand

threesubjectswithlymphomaareshowninTable3.Fourof thefivenormalsubjectsyieldedasmalldegreeofLAK activity,ie,nonantibody-dependentlysis,whiletwoofthe threelymphomasubjectsyieldedsubstantiallymore.There wasneveranysignificantincreaseinlysiswhennativemouse antibodycoatedthecells;indeed,inthemajorityofdetermi- nationsthelysisachievedwasless,andonoccasionnotably so.Thepicturewasentirelydifferentwhenthechimeric FabFc,displayinghumanFc'yl,coatedthecells:hereinall instancestherewasastrikingincreaseintheextentoflysis

ToassesswhetheraminormouseIgGsubclass,ora

combinationofsubclasses,couldbeeffectiveinmediating

ADCC,IgGfromapooledmouseantiserumtoL2Csurface

IgwasincludedamongtheantibodiesinTable3.Nolysis

abovethatoccurringintheabsenceofantibodywas observed.Itshouldbenoted,inviewofrecentreportsthat mouseIgG3canmediateADCCwithhumanLGL,3235that wedonotknowwhatproportionofourIgGpreparationwas representedbythisminorsubclass. Inadditiontousingchimericantibodiescontaininghuman

Fc'yl,wehavefoundthathumanLAKcellpreparations

regularlymediateagoodlevelofADCCagainstL2Ccellsin thepresenceofrabbitpolyclonalanti-Id(Table4),rabbit polyclonalanti-it(datanotshown),andguineapigpolyclonal anti-majorhistocompatibilitycomplex(MHC)classII (raisedinstrain13guineapigsagainststrain2lymphocytes; datanotshown).

ADCCmediatedbyLAKcellsagainsthumantumor

targets.LAKcellsfromanormaldonorwerearrayed againstallogeneiclymphomacellsfromtwopatientsinthe

presenceofvariousantibodies(Table5).Negligiblecytotox-Downloaded from http://ashpublications.org/blood/article-pdf/72/6/1985/598624/1985.pdf by guest on 16 August 2023

Table3.ActivityofLAKCellsAgainstAntibody-CoatedL
CLeukemicLymphocytes AntibodyIsotype%SpecificLysis(±5D)GivenbyLAKCellsFromIndicatedDonor

NiN2N3N4N5S.A.N.L.BR.

Nil

Anti-ld(1)

Anti-Id(2)

Anti-ld(11)

Potyclonal

anti-Id

FabFcanti-IdMouselgGl

MouselgG2a

MouselgG2b

MouseIgG

(subclasses

1.2a.2b.3)

MouseFab"yl-

humanFc'yl0.6±0.4

2.9±0.7

0.1±0.5

ND ND

32±26.8±0.5

7.4±2.3

10±1

2.1±1.6

ND

32±36.4±2.68.8±1.08.2±2.636±7

2.4±1.83.8±1.07.2±0.526±2

2.8±5.23.7±0.98.5±0.7ND

3.1±3.42.9±1.62.9±1.6ND

5.4±1.8ND8.3±1.0ND

59±346±639±364±40.3±1.6

-1.6±1.1 -0.8±0.8 -0.2±0.9 -0.6±1.1

16±0.325±2

9.9±2.0

13±2

4.7±1.6

13±2

79±4

Table4.ActivityofLAKCellsAgainstL
CLeukemic

LymphocytesCoatedwithBivalentorUnivalentAnti-Id

Bivalent(rabbitlgG)39±333±345±645±4

Univalent(rabbitFab/c)44±0.444±464±257±1 DesignatedasinTable3.wherethelevelsofLAKcytotoxicity effectedbycellsfromthesedonorsintheabsenceofantibodyarealsoset

Out.LAKCELLSAWDANTIBODYDISPLAYINGHUMANFc'y1989

Effector:targetratiois25:1.Incubationtimeisthreehours.

Abbreviation:ND:notdetermined.

N1throughN5arenormalsubjects;S.A.,N.L..BR.aresubjectswithIymphoma. icitywasseenintheabsenceofspecificantibodyandinthe presenceofspecificmonoclonalanti-Idofmouseisotype IgG1.However,inthepresenceofunivalentchimericderiva- tivescontainingFab"yfromspecificanti-Idanddisplaying humanFc-y,theLAKcellseffectedmodestbutsignificant killingoverthethree-hourincubationperiod.

RoleofantibodyvalencyinmediatingADCC.Thefact

thattheunivalentchimericantibodiesexaminedherecan invokesignificantADCC,whiletheparentmouseIgGl antibodiesdonot,couldbeduetothechimericderivatives possessingahumanFc-yand/ortotheirbeingunivalent.

Previouslywehavereportedthatunivalencyofantibody

enhancesbothcomplementlysisandADCCbecauserapid antigenicmodulationisavoided.'7'
Thecontributionof univalencytotheLAK-mediatedlysisinvokedbyFabFcand

FabIgGcouldnotbeexploreddirectly,aswedidnothave

%SpecificLysisGivenbyLAKCeUsFrom

AntibodyN3IndicatedDonor

N4N5BR.bivalentcounterpartscontaininghumanFc'y.However,it waspossibletocomparethelysisinvokedbyrabbitIgG antibodyanditsFc-containingderivativeFab/c,27inthe knowledgethatrabbitFc-ycanrecruithumanlymphocytic effectors.36

InTable4itisseenthathumanLAKcellsachievedgood

levelsoflysisofL2Cleukemiccellsduringathree-hour incubationinthepresenceofeitherIgGorFab/canti-Id. Despitethefactthattheunivalentderivativewasconsis- tentlythemoreeffective,theresultsleavenodoubtthat bivalencyalonedoesnotruleoutsubstantialADCCbyLAK cells.

RoleofCDI6.ThefactthatLAK-cellADCCisdepen-

dentonuseoftheFc7-receptorCD16(FcRIII)wascon- firmedbytheinhibitionshownbyantibody3G8,directed againstthisreceptor.3'Table6showsthediminishedeffi- ciencyofathree-hourADCCoccurringafterpreincubating theeffectorswithanti-CD16.Preincubationwithcontrol antibodyhadnosignificanteffect.Theinhibitionwas obviousbutrequiredhigherconcentrationsofanti-CD16 thanwereneededforcomparableinhibitionsoftheADCC mediatedbyunstimulatedbloodlymphocytes.

DISCUSSION

NostrictdefinitionofLAKcellshasbeenproffered.

However,itappearsthatthebroad,non-MHC-restricted,

nonantibody-dependentcytotoxicitythatistheLAKhall- markrequiresthatmononuclearcellsbeactivatedbyIL-2in Table5.ActivityofLAKCellsFromaNormalDonorAgainstAntibody-CoatedHumanLymphomaCells Ant
ody
Isotype%SpecificLysis(±50)AgainstIndicatedTarget

N.L.CellsA.H.Cells

Anti-Id(N.L.)

Anti-ID(A.H.)

FabFcanti-Id(N.L.)

FablgGanti-Id(A.H.)MouselgGl

MouseIgGi

MouseFab"yl-humanFc'yl

MouseFab"yl-humanlgG3.6±3.8ND

0.6±4.0
2.0±6.4

14±4.62.8±4.2

2.0±2.8
28±4

Abbreviation:ND,notdetermined.

Male,age29years.

tMousemonoclonalanti-IdpreparedagainstidiotypicgMfromeitherN.L.orA.H.2eandchimericantibodiespreparedfromthem.IntheFablgG

derivative,thehumanlgGisusedsimplybecauseoftheFc'yitdisplays.28

Anti-Id(N.L.)anditschimericderivativeshowednobindingwhatevertoA.H.cells,andviceverse(examinationsbyflowcytofluorimetry).sothatthe

indicatedtestsrepresentnegativeresultsforLAK-typecytotoxicity.Downloaded from http://ashpublications.org/blood/article-pdf/72/6/1985/598624/1985.pdf by guest on 16 August 2023

REFERENCES1990DEARMANETAL

Table6.ADCCExhibitedbyLAKCellsAgainstAntibody-Coated

L2CLeukemicLymphocytes:InhibitionbyAnti-CD16

LAK-Cell

Donor(Sex.Age[yr])%Specific'
Cr-Re1easebyEffectorsPreincubated

WithAnti-CD16atIndicatedConcentration(pg/mL)

00.050.5550500

F,24433629ii1210

M,30555151332431

M,26585556553628

Anti-CD16antibody:mouseIgGi3G8.3'Anti-L2Cantibody:anti-Id FabFc(mouseFab"yl-human-Fc'yl)at25sg/mL.LAK-cellculture:three days.Preincubationwithanti-CD16:30minutesat0#{176}C.Incubationfor cytotoxicity:threehours. cultureforatleast48hours2;anditseemsthatthesoleor principalLAKprogenitorsareLGL.
7MurineLAKcellsso definedareefficientinADCC,'#{176}'37andinthisreportwe describesimilarbehaviourbytheirhumancounterparts.

TherearetohandconsistentreportsthatADCCdueto

humanLGLcanbeaugmentedbybriefexposureofthe effectorstoIL-2,32-34andthatADCCcanbeeffectedby

IL-2--dependentclonedLGLlines.33

Themajorfindingtoemergefromthepresentworkisthe

efficiencywithwhichLAKcellslysetargetscoatedwith antibodythatdisplayshumanFc'y,evenoverasshorta periodasthreehours(Table3).Theactualhumanisotype involvedisundoubtedlyFc'yl,whichaccountsformorethan

90%oftheFcinFabFc(seeMaterialsandMethods).

Chimericantibodiescontainingmousevariableregionsand humanFccanbesynthesizedeitherbychemicalmeans28or bygeneticengineering.38'39Amongtheadvantagesenvisaged forthemisefficientrecruitmentofhumaneffectors(comple- ment,macrophages,Kcells),anexpectationconfirmedhere forADCCbyLAKcells.

NoLAKdonorhasbeenfoundtoyieldcellsthatfailto

effectADCCwhenpresentedwithhumanFc'yontheir targets.ApparentlythevariabilityinADCCperformance amongLAKcellsfromdifferentdonorsislessthanthe well-knownvariabilityintheperformanceofunstimulated bloodlymphocytes'#{176}Itisalsoreassuringtonotethatallsix lymphomapatientstestedwereabletogenerateactiveLAK cells(TableI),evenwhenthepresenceoftumorcellsinthe bloodsuggestedextensivemarrowinfiltration.

AlthoughLAKcellshavecharacteristicallybeenused

aftercultureforthreetofivedaysinIL-2,thereisacasefor lookingattheirperformanceinADCCatearliertimes,as

1.LotzeMT.GrimmE,MazumderA,StrausseriL,Rosenberg

SA:InvitrogrowthofcytotoxichumanlymphocytesculturedinT cellgrowthfactor(TCGF).CancerRes41:4420,1981

2.GrimmEA,WilsonDi:Thehumanlymphokine-activated

killersystem.CellImmunol94:568,1985

3.RosenbergSA,LotzeMT,MuulLM,ChangAE,AvisFP,

Leitman5,LinehamWM,RobertsonCN,LeeRE,RubiniT,Seipp CA,SimpsonCG,WhiteDE:Aprogressreportonthetreatmentof

157patientswithadvancedcancerusinglymphokine-activatedkiller

cellsandinterleukin-2orhighdoseinterleukin-2alone.NEngIi

Med316:889,1987

4.PhillipsiH,LanierLL:Dissectionofthelymphokine-

activatedkillerphenomenon.iExpMed164:814,1986theLAKactivityofCDl6-positivecellshasbeenreportedto peakaftercultureforonlyonetotwodays,5althoughwith somevariationfordifferentcelltargets.Areductioninthe necessaryperiodofculturewouldclearlybeofgreattechni- calconvenience.

WehavenotobservedanysignificantADCCeffectedby

LAKcellsagainsttargetscoatedwiththecommonlyoccur-

ringmousemonoclonalisotypesIgGl,IgG2a,andIgG2b.

Therearecontradictoryreportsastotheabilitiesofthe

progenitorlymphocytes,unstimulatedorexposedbrieflyto

IL-2oraninterferon,toeffectADCCviathevariousmouse

IgGisotypes.32'35'4143Thediscrepanciesaredifficulttoevalu- atebecauseofvariablesadditionaltoantibodyisotype:the naturesoftargetcellandsurfaceantigen,theantibody affinity,andthepossiblepresenceofmonocytesamong mononucleareffectorskeptonlybrieflyinculture.However, apersuasivecasehasnowbeenmadefortheabilityofhuman LGLtokilltargetscoatedwiththerarestmouseIgGisotype, IgG3.3235ItisareasonableexpectationthenthatLAKcells willalsousethisisotype.WesawnoADCCwhenusinga mousepolyclonalIgGpreparation,butitscontentofIgG3is likelytohavebeen<5%."

Usingratmonoclonalantibodies,theisotypeIgG2bstood

outastheonlyonecapableofinvokingADCCwithhuman lymphocyticeffectors,
#{176}soagainthereisanexpectationthat thisisotypewillinvokeADCCbyLAKcells.Fromourown andothers'experiencewecannowsummarizetheminimum

ADCCcapabilityofhumanLAKcells:(I)Ademonstrated

abilitytolysecellscoatedwithantibodydisplayinghuman Fc
yl;(2)Anexpectedabilitytolysecellscoatedwithmouse

IgG3orratIgG2bantibody.Thereisclearlyaneedeventu-

allytoassesswhetherthesegeneralizationsapplyovera rangeofneoplastictargets,includingcarcinomacells.

TheadditionofautologousLAKcellstoantibodytherapy,

providedanappropriateFc-yispresented,holdspromiseof enhancingtheattackontumorcellsbybothantibody- independentandantibody-dependentmeans.Itcanbeseen inTable3thattwoofthedonors-both,asithappens, subjectswithlymphoma-yieldedcellsthatdisplayedappre- ciableantibody-independentcytotoxicity,andinthepres- enceofchimericantibodyincreasedtheirkillingtothe highestobserved.Themajordrawbackweseeistheadded difficultyinevaluationwhentwotypesoftreatmentare combined.

5.OrtaldoJR.MasonA,OvertonR:Lymphokine-activated

killercells.Analysisofprogenitorsandeffectors.JExpMed

164:1193,1986

6.RobertsK,LotzeMT.RosenbergSA:Separationandfunc-

tionalstudiesofhumanlymphokine-activatedkillercells.Cancer

Res47:4366,1987

7.LotzovaE,HerbermanRB:ReassessmentofLAKphenome-

nology:Areview.NatIImmunCellGrowthRegul6:109,1987

8.OjoE,WigzellH:Naturalkillercellsmaybetheonlycellsin

normalmouselymphoidcellpopulationsendowedwithcytolytic abilityforantibody-coatedtumourtargetcells.ScandJImmunol

7:297,1978

9.TimonenT,OrtaldoiR,HerbermanRB:CharacteristicsofDownloaded from http://ashpublications.org/blood/article-pdf/72/6/1985/598624/1985.pdf by guest on 16 August 2023

LAKCELLSANDANTIBODYDISPLAYINGHUMANFc'y1991

humanlargegranularlymphocytesandrelationshiptonaturalkiller andKcells.JExpMed153:569,1981

10.ShiloniE,EisenthalA,SachsD,RosenbergSA:Antibody-

dependentcellularcytotoxicitymediatedbymurinelymphocytes activatedinrecombinantinterleukin2.iImmunol138:1992,1987

11.StevensonGT,GlennieMi:Surfaceimmunoglobulinof

B-lymphocytictumoursasatherapeutictarget.CancerSurveys

4:213,1985

12.HamblinTi,Abdul-AhadAK,GordonJ,StevensonFK,

StevensonGT:Preliminaryexperienceintreatinglymphocyticleu- kaemiawithantibodytoimmunoglobulinidiotypesonthecell surfaces.BrJCancer42:495,1980

13.MillerRA,MaloneyDG,WarnkeR,LevyR:Treatmentof

B-celllymphomawithmonoclonalanti-idiotypeantibody.NEnglJ

Med306:517,1982

14.MeekerTC,Lowderi,MaloneyDG,MillerRA,Thielemans

K,WarnkeR,LevyR:Aclinicaltrialofanti-idiotypetherapyforB cellmalignancy.Blood65:1349,1985

15.RankinEM,HekmanA,SomersR,tenBokkelHuininkW:

TreatmentoftwopatientswithBcelllymphomawithmonoclonal anti-idiotypeantibodies.Blood65:1373,1985

16.HamblinTi,CattanAR,GlennieMi,MacKenzieMR,

StevensonFK,WattsHF,StevensonGT:Initialexperiencein treatinghumanlymphomawithachimericunivalentderivativeof monoclonalanti-idiotypeantibody.Blood69:790,1987

17.GordonJ,StevensonGT:Antigenicmodulationoflympho-

cyticsurfaceimmunoglobulinyieldingresistancetocomplement- mediatedlysis.Immunology42:13,1981

18.MeekerT,LowderJ,ClearyML,Stewart5,WarnkeR,

SklarJ,LevyR:Emergenceofidiotypevariantsduringtreatmentof B-celllymphomawithanti-idiotypeantibodies.NEnglJMed

312:1658,1985

19.TuttAL,StevensonFK,SmithJL,StevensonGT:Antibodies

againsturinarylightchainidiotypesasagentsfordetectionand destructionofhumanneoplasticBlymphocytes.JImmunol

131:3058,1983

20.StevensonGT,EadyRP,HoughDW,JurdRD,Stevenson

FK:Surfaceimmunoglobulinofguinea-pigleukaemiclymphocytes.

Immunology28:807,1975

21.StevensonFK,GlennieMJ,JohnsonDMM,TuttAL,Stev-

ensonGT:Consumptionofmonoclonalanti-idiotypicantibodyby neoplasticBlymphocytes:Aguideforimmunotherapy.BriCancer

50:407,1984

22.ElliottTi,GlennieMi,McBrideHM,StevensonGT:Analy-

sisoftheinteractionofantibodieswithimmunoglobulinidiotypeson neoplasticBlymphocytes:Implicationsforimmunotherapy.J

Immunol138:981,1987

23.StevensonFK,MoleLE,RaymontCM,StevensonGT:AX

BenceJonesproteininguineapigs.BiochemJ151:751,1975

24.Brown5,DilleyJ,LevyR:Immunoglobulinsecretionby

mousexhumanhybridomas:Anapproachfortheproductionof anti-idiotypereagentsusefulinmonitoringpatientswithBcell lymphoma.JImmunol125:1037,1980

25.LennertK:Non-Hodgkin'sLymphomas.Berlin,Springer-

Verlag,1981

26.StevensonGT,SmithJL,HamblinTi:ImmunologicalInves-

tigationofLymphoidNeoplasms.Edinburgh,ChurchillLiving- stone,198327.GlennieMi,StevensonGT:Univalentantibodieskilltumour cellsinvitroandinvivo.Nature295:712,1982

28.StevensonGT,GlennieMi,PaulFE,StevensonFK,Watts

HF:PreparationandpropertiesofFabIgG,achimericunivalent antibodydesignedtoattacktumourcells.BiosciRep5:991,1985

29.LamoyiE,NisonoffA:PreparationofF(ab')2fragmentsfrom

mouseIgGofvarioussubclasses.JImmunolMethods56:235,1983

30.BrocklehurstK:Specificcovalentmodificationofthiols:

Applicationsinthestudyofenzymesandotherbiomolecules.mtJ

Biochem10:259,1979

31.FleitHB,WrightSD,UnkelessJC:HumanFcgamma

receptordistributionandstructure.ProcNatIAcadSciUSA

79:3275,1982

32.HonsikCi,JungG,ReisfeldRA:Lymphokine-activated

killercellstargetedbymonoclonalantibodiestothedisialoganglio- sidesGD2andGD3specificallylysehumantumorcellsofneurec- todermalorigin.ProcNatlAcadSciUSA83:7893,1986

33.AnasettiC,MartinPi,MorishitaY,BadgerCC,Berstein

ID,HanseniA:Humanlargegranularlymphocytesexpresshigh affinityreceptorsformurinemonoclonalofIgG3subclass.i

Immunol138:2979,1987

34.OrtaldoJR,WoodhouseC,MorganAC,HerbermanRB,

ChereshDA,ReisfeldR:Analysisofeffectorcellsinhuman antibody-dependentcellularcytotoxicitywithmurinemonoclonal antibodies.iImmunol138:3566,1987

35.HellstromI,BrankovanV,HellstromKE:Strongantitumour

activitiesofIgG3antibodiestoahumanmelanoma-associated ganglioside.ProcNatlAcadSciUSA82:1499,1985

36.StevensonGT,GlennieMi,GordonJ:Thekillingoflym-

phomacellsbyunivalentderivativesoftumor-specificantibody.

UCLASympMolecCellBiol24:459,1982

37.Hinuma5,NaruoK,Shiho0,TsukamotoK:Characteristics

ofmurinenonspecifickillercellsinducedinvivobyhuman recombinantIL-2.Immunol59:251,1986

38.LiuAY,RobinsonRR,HellstromKE,MurrayED,Chang

CP,HellstromI:Chimericmouse-humanIgG1antibodythatcan mediatelysisofcancercells.ProcNatlAcadSciUSA84:3439, 1987

39.ShawDR,KhazaeliMB,SunLK,GhrayebJ,DaddonaPE,

McKinney5,LoBuglioAF:Characterizationofamouse/human chimericmonoclonalantibody(17-lA)toacoloncancertumor- associatedantigen.JImmunol138:4534,1987

40.HaleG,ClarkM,WaldmannH:Therapeuticpotentialofrat

monoclonalantibodies:Isotypespecificityofantibody-dependent cell-mediatedcytotoxicitywithhumanlymphocytes.JImmunol

134:3056,1985

41.RalphP,NakoinzI:Cellmediatedlysisoftumortargets

directedbymurinemonoclonalantibodiesof1gMandallIgG isotypes.iImmunol131:1028,1983

42.ChristiaanseniE,SearsDW:Unusuallyefficienttumorcell

lysisofhumaneffectsofantibody-dependentcellularcytotoxicity mediatedbymonoclonalantibodies.CancerRes44:3712,1984

43.KippsTi,ParhapP,PuntJ,HerzenbergLA:Importanceof

immunoglobulinisotypeinhumanantibody-dependent,cell- mediatedcytotoxicitydirectedbymurinemonoclonalantibodies.J

ExpMed161:1,1985

44.GreyHM,HirstJW,CohnM:Anewmouseimmunoglobu-

lin:IgG3.iExpMed133:289,1971Downloaded from http://ashpublications.org/blood/article-pdf/72/6/1985/598624/1985.pdf by guest on 16 August 2023