Single-Domain Antibody–Based and Linker-Free Bispecific Antibodies Targeting FcgRIII Induce ...

12 Pages·2013·1.02 MB·English

Published OnlineFirst June 11, 2013; DOI: 10.1158/1535-7163.MCT-12-1012
Molecular
Cancer
Large Molecule Therapeutics
Therapeutics
Single-Domain Antibody–Based and Linker-Free Bispecific
Antibodies Targeting FcgRIII Induce Potent Antitumor
Activity without Recruiting Regulatory T Cells
CarolineRozan1,2,3,4,9,Ame(cid:1)lieCornillon1,2,3,4,9,CorinnePe(cid:1)tiard1,2,3,4,9,MartineChartier1,2,3,4,9,
GhislaineBehar1,2,3,4,9,CharlotteBoix5,6,7,9,BrigitteKerfelec1,2,3,4,9,BrunoRobert8,9,Andre(cid:1)Pe(cid:3)legrin8,9,
PatrickChames1,2,3,4,9,Jean-LucTeillaud5,6,7,9,andDanielBaty1,2,3,4,9,þ
Abstract
Antibody-dependentcell-mediatedcytotoxicity,oneofthemostprominentmodesofactionofantitumor
antibodies,suffersfromimportantlimitationsduetotheneedforoptimalinteractionswithFcgreceptors.In
thiswork,wereportthedesignofanewbispecificantibodyformat,compactandlinker-free,basedontheuseof
llamasingle-domainantibodiesthatarecapableofcircumventingmostoftheselimitations.Thisbispecific
antibody format was created by fusing single-domain antibodies directed against the carcinoembryonic
antigen and the activating FcgRIIIa receptor to human Ck and CH1 immunoglobulin G1 domains, acting
as a natural dimerization motif. In vitro and in vivo characterization of these Fab-like bispecific molecules
revealedfavorablefeaturesforfurtherdevelopmentasatherapeuticmolecule.Theyareeasytoproducein
Escherichiacoli,verystable,andelicitpotentlysisoftumorcellsbyhumannaturalkillercellsatpicomolar
concentrations. Unlike conventional antibodies, they do not engage inhibitory FcgRIIb receptor, do not
compete with serumimmunoglobulins G forreceptor binding, andtheircytotoxic activity isindependent
ofFcglycosylationandFcgRIIIapolymorphism.Asopposedtoanti-CD3bispecificantitumorantibodies,they
donotengageregulatoryTcellsastheselattercellsdonotexpressFcgRIII.Studiesinnonobesediabetic/severe
combined immunodeficient gammamice xenografted with carcinoembryonic antigen–positive tumorcells
showed that Fab-like bispecific molecules in the presence of human peripheral blood mononuclear cells
significantly slow down tumor growth. This new compact, linker-free bispecific antibody format offers a
promisingapproachforoptimizingantibody-basedtherapies.MolCancerTher;12(8);1481–91.(cid:2)2013AACR.
Introduction clinical trial results support an important role of anti-
body-dependent cell-mediated cytotoxicity (ADCC) for
Almost30monoclonalantibodies(mAb)arenowinthe
both lymphomas and solid tumors (3). The affinity
market,including15mAbsforcancertherapyinEurope
between the Fc portion of human IgG1 and FcgRIIIa
and the United States. Most of these mAbs (9/15) are
(CD16a),anactivatingreceptormostlyexpressedbynat-
human immunoglobulin G1 (IgG1) that trigger various
uralkiller(NK)cells,monocytes,andmacrophages,hasa
mechanisms,suchastargetsignalinginhibition,apoptosis,
profoundimpactonADCCexertedbyantibodies(4,5).
activation of the classical complement pathway, and/or
Correlation of clinical responses to mAb therapy with
ofimmuneeffectorcellsexpressingFcg receptors(FcgR;
FcgRIIIa polymorphism has been observed with more
refs.1,2).Althoughitisdifficulttoassessthecontribution
favorable response in patients homozygous for the
of each of these mechanisms in their in vivo efficacy,
high-affinityFcgRIIIa(V158;refs.6,7).Thefactthatabout
80%oftheCaucasianpopulationishomozygous(F/F)or
heterozygous (F/V) for low-affinity FcgRIIIa (F158) is
Authors'Affiliations:1INSERMU1068,CRCM;2CNRSUMR7258;3Insti-
tutPaoli-Calmettes;4Universite(cid:1)Aix-Marseille;5INSERMU872,Centre likely an important issue for antibody-based immuno-
deRecherchedesCordeliers;6Universite(cid:1)PierreetMarieCurie,UMR-S therapy. Moreover, the in vivo efficacy of therapeutic
872;7Universite(cid:1)ParisDescartes,UMR-S872,Paris;8IRCM,INSERM,
U896, Universite(cid:1) Montpellier 1, CRLC Val d'Aurelle, Montpellier; antibodiesishinderedbythepresenceoffucoseresidues
9GDR3260,CNRS,Tours,France in the N-glycosylation motif (N297 residue) of the Fc
regionthatmarkedlydecreasestheiraffinityforFcgRIIIa
Note:SupplementarydataforthisarticleareavailableatMolecularCancer
TherapeuticsOnline(http://mct.aacrjournals.org/). (8–10).Therapeuticantibodiesalsocompetewithserum
IgG for binding to the high-affinity FcgRI and to the
CorrespondingAuthor:Daniel Baty,INSERMU1068,163Avenuede
Luminy, Marseille 13009, France. Phone: [phone]; Fax: 33- intermediate-affinityFcgRIIIa.Strikingly,mostantibodies
491826083;E-mail:[email protected] have to be injected at high doses to reach a serum con-
doi:10.1158/1535-7163.MCT-12-1012 centrationbetween10and100mg/mLwhiletheyusually
(cid:2)2013AmericanAssociationforCancerResearch. elicitamaximalcytotoxicactivityat10ng/mLininvitro
www.aacrjournals.org 1481
Downloaded from mct.aacrjournals.org on April 10, 2019. © 2013 American Association for Cancer Research.
Published OnlineFirst June 11, 2013; DOI: 10.1158/1535-7163.MCT-12-1012
Rozanetal.
ADCC experiments. Competition with patient IgG has ingastrongcytotoxicityagainstCEA-positivetumorcells
been proposed to account for this large difference of by human NK cells in vitro, and of inhibiting tumor
concentration(11).Finally,theuseoftherapeuticantibo- growthinaninvivomousemodel.
dies may be hampered by their ability to engage the
inhibitoryFcgRIIbreceptor.FcgRIIbpossessesaninhibi- Materials and Methods
tory immunoreceptor tyrosine-based inhibitory motif in Expressionvectordesignandgenerationof
itscytoplasmicdomainandhasbeenshowntoimpactthe recombinantantibodies
antitumor efficacy of therapeutic antibodies in mouse
Bicistronicexpressionvectorsweredesignedtoexpress
models (12). Several approaches such as site-directed bsFabmoleculesinperiplasm(fordetailsseeSupplemen-
mutagenesis,computationalstructure–baseddesign,gly-
taryMethods).
cosylation engineering, and selection-based methods
havebeenusedtoincreaseFcbindingtoactivatingrecep- Productionandpurificationofantibodies
tors (FcgRI,FcgRIIa,and FcgRIIIa) andtodecreasetheir bsFabswerepurifiedfromperiplasmofEscherichiacoli
interaction with inhibitory FcgRIIb (3, 13, 14). Variants (E.coli)DH5aandsdAbC17-Fcfromculturesupernatant
possessingupto100-foldincreasedaffinityforFcgRIIIa, ofHEK293Tcells(fordetailsseeSupplementaryMethods).
resultingin100-foldenhancedinvitroADCC(3,15),have
beenselected. Celllines
Attractive alternative to recruit and activate effector Jurkat(ATCCTIB-152),LS174T(ATCCCL-188),BXPC3
cells is to use bispecific antibodies (bsAb) capable of (ATCC CRL-1687), LAN-1 (16), HT29 (ATCC HTB-38),
simultaneously binding to a target antigen and to an and HEK293T (ATCC CRL-11268) cells purchased from
activating receptor such as FcgRI or FcgRIIIa (16, 17). AmericanTypeCultureCollection(ATCC)werecultured
Althoughtheinvitroandinvivoantitumorefficienciesof in RPMI-1640 þ GlutaMax-I medium (Invitrogen) sup-
bsAbshavebeenlargelyshownoverthelasttwodecades, plementedwith10%FBS(PAA).SKOV3(ATCCHTB-77)
theirdevelopmenthasbeenseverelyhinderedbydifferent andCO115(anestablishedCEA-negativesubclonefroma
factors, including immunogenicity and the difficulty to human colon carcinoma cell line; refs. 31, 32) were cul-
efficientlyproducelargeamountsofactivemolecules(18). turedinDulbecco’sModifiedEagleMedium(DMEM)þ
With the development of antibody engineering, several GlutaMax-I medium (Invitrogen) supplemented with
innovative recombinant formats have been proposed, 10%FBS(PAA).MC38andC15.4.3.AP(MC38-CEA)cells
including diabodies, tandem single-chain variable frag- are murine colorectal cancer cells (33). MC38-CEA cells
ments (scFvs), and minibodies (17, 19, 20), and some of areMC38cellstransfectedwithacDNAencodingCEA.
thesemoleculesarebeingtestedintheclinics(21).How- MC38 were cultured in DMEM þ GlutaMax-I medium
ever,mostdescribedrecombinantbsAbsheavilyrelyon supplementedwith10%FBSandMC38-CEAinDMEMþ
the use of peptide linkers. Although these linkers have GlutaMax-Imediumsupplementedwith10%FBSand0.5
obvious advantages in terms of antibody engineering, mg/mLofgeneticin.Jurkat-huFcgRIIIa/gcellsareJurkat
theirhydrophilicnaturemakesthempronetoproteolytic lymphomaTcellstransfectedwithacDNAencodingthe
cleavage, potentially leading to production issues, poor extracellulardomainofFcgRIIIafusedtothetransmem-
antibody stability, aggregation, and increased immuno- braneandintracellulardomainsoftheg chain(generous
genicity (22). Llama single-domain antibodies (sdAb), gift of Prof. E. Vivier, CIML, Marseille, France; ref. 34).
derivedfromheavy-chainantibodiesnaturallydevoidof These cells were cultured in RPMI-1640 þ GlutaMax-I
lightchains(23)aresmall(13kDa),wellexpressed,and mediumsupplementedwith10%FBSand0.5mg/mLof
extremelystablefragments.Theyarehighlyhomologous geneticin. SKOV3-CEA-Luc are SKOV3 human ovarian
totheVHIIIsubsetfamilyofhumanVH(24,25),which carcinoma transfected with a cDNA encoding the extra-
shouldresultinalowantigenicityinhuman,ifany,and cellulardomainofCEAandwithacDNAencodinglucif-
whichmakesthehumanizationprocesseasierifneeded erase;cellswereculturedinDMEMþGlutaMax-Imedium
(26).Becauseoftheirsmallsizeandsingle-domainnature, supplementedwith10%FBS,0.5mg/mLofgeneticin,and
sdAbs can recognize epitope usually not accessible to 0.4mg/mLhygromycinB.Allcelllinesweregrownina
conventionalantibodies.Thesefragmentsrepresentideal humidified37(cid:2)Cincubatorcontaining5%CO .
2
molecularbuildingunitsforbsAbconstruction(27).Ina AllcelllinespurchasedfromATCCwerenotcultured
previouswork,weisolatedtwosdAbscapableofbinding formorethan2months.Othercelllinesusedinthisstudy
and activating FcgRIIIa while showing no binding to havenotbeenauthenticated.
FcgRI,FcgRIIa,orinhibitoryreceptorFcgRIIb(28)aswell
asonesdAbabletobindcarcinoembryonicantigen(CEA) HumanNKcellpurification
with high specificity (29), a well-characterized tumor Human peripheral blood mononuclear cells (PBMC)
markerofinterestformAb-basedcancertherapy. were isolated from fresh peripheral blood of healthy
In the present work, we have exploited the natural donors (Etablissement Franc¸ais du Sang, Marseille and
affinityofhumanCH1andCkIgGdomainsasanhetero- Ho^tel-Dieu hospital, Paris, France) by Ficoll LSM 1077
dimerization motif (30) to produce Fab-like bispecific (PAA) gradient centrifugation. Cells were counted and
antibodies(bsFab),devoidoflinkerandcapableofinduc- tested for viability by Trypan blue exclusion assay. NK
1482 MolCancerTher;12(8)August2013 MolecularCancerTherapeutics
Downloaded from mct.aacrjournals.org on April 10, 2019. © 2013 American Association for Cancer Research.
Published OnlineFirst June 11, 2013; DOI: 10.1158/1535-7163.MCT-12-1012
Anti-FcgRIII(cid:3)CEABispecificFab-likeAntibodies
cells were isolated by depleting non-NK cells using the 4hours.Rosetteformationwasthenobservedbyoptical
NKcellisolationkit(MiltenyiBiotec)asdescribedbythe microscopy.
manufacturer.
Interleukin-2secretionassay
Flowcytometryassays Jurkat-huFcgRIIIa/gcells(106;ornontransfectedJurkat
IncreasingconcentrationsofbsFabs(1.2–100nmol/L) cellsasnegativecontrol)wereincubatedfor18hoursin
were used for staining 2 (cid:3) 105 Jurkat-huFcgRIIIa/g or RPMI-1640medium,containing5%FBS,10ng/mLphor-
2 (cid:3) 105 MC38-CEA cells in 50 mL PBS supplemented bol 12-myristate 13-acetate (PMA; Sigma Aldrich) and
with2%bovineserumalbuminat4(cid:2)Cfor1hour.Bound bsFabs (50 nmol/L) or anti-FcgRIII mAb 3G8 (35) used
bsFabsweredetectedbyincubatingcellswitheitheran at50nmol/Lasapositivecontrol.Insomeexperiments,
anti-flag mAb M2 (4 mg/mL; Sigma Aldrich) or anti-c- anti-flagM2mAbwasadded(4mg/mL)topromotebsFab
myc mAb 9E10 (10 mg/mL; Santa Cruz Biotechnology) dimerization. In some experiments, 105 LS174T cells (or
(cid:2)
for30minutesat4 Cfollowedbyafluoresceinisothio- LAN-1cellsusedasanegativecontrol)wereusedastarget
cyanate (FITC)-conjugated goat anti-mouse IgG F(ab’) cells.Afterincubation,cellswerecentrifugedfor5min-
2
(10mg/mL;JacksonImmunoResearch)incubated30min- utesat1,200(cid:3)gandhumaninterleukin(IL)-2presentin
(cid:2)
utesat4 C.Afterwashing,labeledcellswereanalyzedby the culture medium was measured by ELISA using
flowcytometryusingaFACSCalibur(BDBiosciences)or READY-SET-GOhumanIL-2Kit(eBioscience).
MACSQuant(MiltenyiBiotec)flowcytometers.
Forserumstabilityassays,bsFabswereincubatedat1 ADCCassay
mmol/Lin90%humanserum(notheat-inactivated;Sig- Targetcells (BxPC3,HT29, LS174T,SKOV3-CEA-Luc,
ma Aldrich) at 37(cid:2)C for various periods of time (7–168 or SKOV3;5(cid:3)103cells/well) weremixedwith5(cid:3)104
hours). Samples were collected at different time points, freshly isolated human NK cells (effector/target ratio:
frozen,andkeptat(cid:4)20(cid:2)C.Theremainingbindingactivity 10:1). Variable concentrations of bsFabs or sdAb C17-Fc
of the samples after storage was determined by flow were addedto the cells in a final volume of200 mL. All
cytometryusingnonsaturatingbsFabconcentrations(10 proceduresweredoneintriplicatewithdifferentdonors.
(cid:2)
nmol/L for bsFab C21 and 100 nmol/L for bsFab C28 Following overnight incubation at 37 C, target cell via-
against Jurkat-huFcgRIIIa/g, and 100 nmol/L for both bility was quantified with CellTiter-Glo viability assay
againstMC38-CEA). (Promega) according to manufacturer’s protocol. The
ForbindingcompetitionassayswithserumIgG,Jur- formula used for % lysis calculation was: % lysis ¼ [T-
kat-huFcgRIIIa/gcellswerepreincubatedwithincreas- (T -E)]/[T-T ](cid:3)100(T¼target,E¼effector,T ¼
EAb dead dead
ingconcentrationsofhumanserum(12.5%–100%)at4(cid:2)C target lysed with 1% Triton solution, T ¼ target þ
EAb
for 1 hour then incubated with a constant amount of effectorþantibodyluminescentsignal).
bsFab C21 (4 nmol/L) or bsFab C28 (36 nmol/L) or To investigate the effect of soluble CEA, target and
biotinylatedsdAbC17-Fc(200nmol/L).Boundantibo- effector cells were incubated as previously described
dies were detected using either mouse anti-flag mAb withvariableconcentrationsofbsFabandconstantcon-
M2 (4 mg/mL; Sigma Aldrich) followed by incubation centration (0.1 mg/mL or 1 mg/mL) of soluble CEA
with FITC-conjugated goat anti-mouse IgG F(ab0) (10 (Chemicon).FcgRIIIagenotypingofdonorswascarried
2
mg/mL; Jackson ImmunoResearch Laboratories) for outasdescribed(36).Dose–responsecurvesweretreated
bsFab labeling or streptavidin-phycoerythrin diluted by nonlinear regression analysis using Prism software
1:10 (Beckman Coulter) for biotinylated sdAb C17-Fc (GraphPad Software). Data were expressed as mean (cid:5)
labeling. In another setting, Jurkat-huFcgRIIIa/g cells SEM.
were preincubated with constant concentrations of
human serum (90%) or PBS 1(cid:3) at 4(cid:2)C for 1 hour and Biodistributionassay
then incubated with various amounts of bsFab C21 or Swiss female nude mice (Charles River Laboratories,)
bsFabC28.Boundantibodiesweredetectedasdescribed bearinghumancoloncarcinomaLS174T(CEApositive)or
previously. CO115(forCEA-negativecontrol)tumorsontherightand
ForCEAdetection,cellswerepreincubatedwithanti- leftflanks,respectively,wereinjectedintravenouslywith
CEA sdAb C17 (10 mg/mL) for 1 hour at 4(cid:2)C. Bound 125I-labeled bsFab C21 (2.8 (cid:3) 105 Bq, 3 mg) 14 days after
antibody was detected using mouse anti-6his mAb (1 subcutaneoustumorcellinjection.Fourgroupsof3mice
mg/mL; Novagen) followed by incubation with FITC- weresacrificedat3,6,15,or24hoursafterbsFabinjection.
conjugatedgoatanti-mouseIgGF(ab0) (10mg/mL;Jack- Bloodwascollectedandtissuesandorganswereweighted
2
sonImmunoResearch). anddissectedateachtimepoint.Radioactivitywasquan-
tified using a g- counter COBRAII (Packard) and results
Rosetteformationassay wereexpressedas%injecteddose/gtissue(%ID/g).Mice
A total of 5 (cid:3) 104 MC38-CEA and 25 (cid:3) 104 Jurkat- housed in individual cages with laboratory chow and
huFcgRIIIa/gwerecoculturedinRPMI-1640þGlutaMax- water.Allexperimentswerecarriedoutaccordingtothe
Imediumsupplementedwith10%FBSand0.5mg/mL FrenchAnimalProtectionLawwiththepermissionfrom
ofgeneticinwithorwithoutbsFabs(2mg/mL)at37(cid:2)Cfor thelocalauthorities.
www.aacrjournals.org MolCancerTher;12(8)August2013 1483
Downloaded from mct.aacrjournals.org on April 10, 2019. © 2013 American Association for Cancer Research.
Published OnlineFirst June 11, 2013; DOI: 10.1158/1535-7163.MCT-12-1012
Rozanetal.
Results
Tumorgrowthinhibition
Seven-week-oldfemalenonobesediabetic/severecom- bsFabdesignandproduction
bined immunodeficient gamma (NSG) mice (Charles ForconstructionofbsFabs,wedesignedabicistronic
River Laboratories; n ¼ 8/group) were xenografted expression vector allowing the coexpression of one
subcutaneously with human pancreatic carcinoma sdAbdirectlyfusedtothehumanCkchainandanother
CEA-positiveBxPC3(2(cid:3)106cells/mouse)ontheright sdAb directly fused to the human CH1 chain in the
flank and were injected with freshly purified human periplasm of E. coli (Fig. 1A). We used the anti-CEA
PBMCs [30 (cid:3) 106/mouse, intraperitoneally (i.p.)] and sdAb C17 previously characterized (K : 8 nmol/L;
d
bsFab C21 (100 mg/mouse) or PBS (i.p.) at day 0. At ref.29)fortargetingCEA-positivecellsandtwodiffer-
days1and2,micewereinjectedintraperitoneallywith ent anti-FcgRIII sdAbs for targeting FcgRIIIa-positive
bsFab C21 (50 mg/mouse) or PBS. Tumor growth was effectorcells.Theanti-FcgRIIIsdAbC28waspreviously
measured with a Vernier caliper during 28 days to shown (28) to display an apparent affinity for FcgRIII
permit calculation of tumor volumes [V ¼ (L(cid:3)W2)/2, (K :82nmol/L)intherangeoftheFcportionofhuman
d
where L and W were length and width, respectively]. IgG1(>100nmol/L),whereastheanti-FcgRIIIsdAbC21
Allanimalsweresacrificedattheendofexperimentin displays a higher affinity of K : 10 nmol/L for FcgRIII
d
accordancewiththeInstitutionalguidelines.Micewere (28). The resulting bispecific Fab-like fragments were
housed in individual cages with laboratory chow and namedbsFabC21andbsFabC28,respectively(Fig.1B).
water. All experiments were carried out according to bsFabs were produced in E. coli periplasm and affinity
theFrenchAnimalProtectionLawwiththepermission purified by anti-CH1 followed by anti-Ck columns. As
from the local authorities. showninFig.1C,SDS-PAGEanalysesofthesecondstep
of bsFab C21 purification reveal, under nonreducing
Statisticalanalysis conditions,onebandintheexpectedrangeofsize(50–
Data are presented as mean (cid:5) SEM. Cytotoxicity 55 kDa) due to heterodimerization.
assays were statistically analyzed using a one-way Inparallel,wedevelopedahumanFcfusionprotein,
ANOVA test. In vivo studies were analyzed using Stu- sdAbC17-Fc,byfusinganti-CEAsdAbC17tothehinge,
dentttest. Forall tests,P<0.05wasconsideredstatis- CH2 and CH3 domains of IgG1 (Fig. 1B). Anti-CEA
tically significant. sdAbC17–Fcwasproducedbytransienttransfectionin
A
Ptac RRBBSS1 55 sdAb αCEA Cκ F RRBBSS2 wt sdAb αFcγRIII CH1 C6H
Figure1. Antibodyformats.
A,abiscistronicconstructwas
PΒ-act KKoozzaacc µp sdAb αCEA H CH2 CH3 designedtoallowtheproduction
ofbsFabintheperiplasmofE.coli.
B Darkgray,portionofhinge(H)and
CkandCH1-3constantdomains
ofhumanIgG1;lightgray,sdAbs;
Ptac,tacpromoter;RBS1and
H H RBS2,ribosome-bindingsites;55,
Cκ
optimizedPelBsignalsequence;
CH1 CH2 WT,PelBwild-typesignal
sequence;F,Flagtag;C,c-myc
CH2
tag;6H,hexahistag;Pb-act,
C F CH3 chickenb-actinpromoter;Kozac,
6H ribosome-bindingsite;mp,
CH3
microphosphatasesignal
bsFab sequence.B,schemeofantibody
formatsusedinthisstudy.C,
sdAb C17-Fc analysisofpurifiedbsFabsby
Coomassieblue–stained
C LoadFTWashMW Elution kDa SchDrSom-PaAtGogEraapftheyroafnfinIgitGy–CH1
184 matrixfollowedonLC–kmatrix.
98
64 MW,molecularweightladder;
50 FT,flowthrough.
36
22
1484 MolCancerTher;12(8)August2013 MolecularCancerTherapeutics
Downloaded from mct.aacrjournals.org on April 10, 2019. © 2013 American Association for Cancer Research.
Published OnlineFirst June 11, 2013; DOI: 10.1158/1535-7163.MCT-12-1012
Anti-FcgRIII(cid:3)CEABispecificFab-likeAntibodies
HEK293T cells and purified from culture supernatant MC38 and Jurkat cells was observed with any of these
byaffinitychromatographyonaG-proteincolumn.This bsFabs(datanotshown).
construction was designed to provide a molecule with TheabilityofbsFabstosimultaneouslybindtoMC38-
the same epitope specificity of bsFabs, but triggering CEAandJurkat-huFcgRIIIa/gcellswasshowedbyrosette
ADCC via a conventional Fc/FcgR interaction. formationbetweenthetwotypesofcells.Theadditionof
bsFabsinMC38-CEAandJurket-huFcgRIIIa/g coculture
BindinganalysisofbsFabs ledtorosetteformationbetweenJurkat-huFcgRIIIa/gand
BindingspecificitiesofbsFabstoFcgRIIIAandCEAwere MC38-CEAcells(upto9Jurkat-huFcgRIIIa/gcellsbound
analyzed by flow cytometry using Jurkat-huFcgRIIIA/g tooneMC38-CEAcellhavebeenobserved;Fig.2B).Inthe
and MC38-CEA cells, respectively. As shown in Fig. 2A, absenceofbsFab,norosetteformationwasobserved.
bothbsFabsefficientlyboundtothesecellsinadose-depen-
dent manner. To ensure that signals were generated by InvitrostabilityofbsFabsinhumanserum
heterodimers,bsFabbindingtoMC38-CEAcellswasass- In vitro stability was analyzed by incubation of each
(cid:2)
essedusingthec-myctagfusedtothechainbearingtheanti- bsFab in human serum at 37 C for up to168 hoursand
FcgRIIIA sdAb and vice-versa. A marked shift in mean subsequentexaminationofCEAandFcgRIIIabindingby
fluorescenceintensitywasobservedwhencomparingbsFab flow cytometry. Figure 2C shows that bsFab C21 and
C21 and bsFab C28 binding with Jurkat-huFcgRIIIA/g, bsFab C28 retain their full binding on both target and
likely related to the 8-fold difference of affinity between effectorcells,evenafter168hoursofincubation,showing
theseanti-FcgRIIIasdAbs(28).Asexpected,bsFabC21and aremarkablestability.SDS-PAGEandWesternblotanal-
C28,bothbearingthesameanti-CEAsdAbC17exhibiteda yses showed the absence of breakdown products of
similarbindingprofiletoMC38-CEAcells.Nobindingto both bsFabs (data not shown), further confirming that
A
FAi,gbusrFea2b.sbwseFraebi-nbciunbdaintegdawnaitlhysis. cells MC38-CEA
antigen-positivecellsatvarious of
concentrations.BindingonMC38- er Jurkat-huFcγγRIIIa/γ
b
CEAandJurkat-huFcgRIIIa/gwas m
detectedusinganti-c-mycoranti- Nu 0 1.2 3.7 11 33 100
flagantibodies,respectively,
bsFab (nmol/L)
followedbygoatanti-mouse Fluorescence intensity
labeledantibodies.Filledblack,
isotypecontrol;filledgray,bsFab B
C21;openblack,bsFabC28.B,
MC38-CEA(blackarrow)and
Jurkat-huFcgRIIIa/g(dottedarrow)
werecoculturedwithorwithout
bsFabsC21orbsFabC28.
Rosetteformationwasobserved
byopticalmicroscopy.C,bsFab
MC38-CEA + Jurkat-huFcγRIIIa/γ MC38-CEA + Jurkat-huFcγRIIIa/γ MC38-CEA + Jurkat-huFcγRIIIa/γ
stabilityinhumanserum.bsFabs
+ bsFab C21 + bsFab C28
wereincubatedat1mmol/Lin90%
serumforupto168hours.Binding Time (h)
C
experimentswerenextcarriedout
byflowcytometryusing 7
nonsaturatingconcentrations 24
(10nmol/LforbsFabC21and
100nmol/LforbsFabC28against 31
Jurkat-huFcgRIIIa/g,and100
48
nmol/LforbothagainstMC38-
CEA).Filledblack,isotypecontrol; 72
openblack,bsFabincubatedfor
varioustimesat37(cid:2)CinPBS;filled 96
gray,bsFabincubatedforvarious s 168
timesat37(cid:2)Cin90%human ell
sFlinecergu.RmII.IaC:EJuAr:kMatC-h3u8F-cCgERAIIIcae/glllcineell. ber of c FcγRIIIa CEA FcγRIIIa CEA Control
m bsFab C21 bsFab C28
u
N
Fluorescence intensity
www.aacrjournals.org MolCancerTher;12(8)August2013 1485
Downloaded from mct.aacrjournals.org on April 10, 2019. © 2013 American Association for Cancer Research.
Published OnlineFirst June 11, 2013; DOI: 10.1158/1535-7163.MCT-12-1012
Rozanetal.
40,000
L)
m 30,000 Figure3. IL-2releaseassays.
g/ JurkatorJurkat-huFcgRIIIa/gcells
p
2 ( wereprestimulatedwithPMAand
L- 20,000 incubatedwith50nmol/LofbsFab
d I oranti-FcgRIIIabivalentmAb3G8.
e Insomecase,anti-FlagtagmAb
cret 10,000 wasaddedtoinducebsFab
e dimerization.TheadditionofCEA-
S
positiveLS174Ttumorcellscould
induceIL-2secretion,whereas
0 CEA-negativeLAN-1human
Jurkat + + + tumorcellsdidnot.IL-2secretedin
Jurkat-huFcγγRIIIa/γ + + + + + + + + + + + + themediumwasquantifiedby
bsFab C21 + + + + + ELISA.ErrorbarsrepresenttheSD
bsFab C28 + + + + + ofexperimentscarriedoutin
triplicates.
Anti-flag mAb + +
3G8 +
LS174T (CEA+) + + + + + +
LAN-1 (CEA-) + + +
both bsFabs are stable up to 168 hours in physiologic absence of target cells (Fig. 3), unlike bivalent anti-
conditions. FcgRIIIa mAb 3G8. The low difference in IL-2 secretion
observedbetweenthetwobsFabscouldbeexplainedby
CompetitionwithhumanIgGforFcgRIIIabinding their different affinityfor FcgRIIIaand/or different epi-
Thepotentialimpactofserum,containinghighamount topes. The addition of a monoclonal antibody targeting
of endogenous IgG, on FcgRIIIa binding by bsFabs was the Flag tag fused to the C-terminus of Ck domain and
explored. Jurkat-huFcgRIIIa/g cells were preincubated thusleadingtobsFabdimerizationledtoIL-2secretionas
inthepresenceofvariousvolumesofhumanserumand expected (Fig. 3). These results show that monovalent
thenstainedwithbsFabs.SdAbC17-Fcwasusedascon- bsFabdonotinduceeffectorcellactivationintheabsence
trol. As shown in Supplementary Fig. S1A, both bsFabs oftargetcells.
retainbindinginthepresenceofhumanserum,whereas However, the coculture of Jurkat-huFcgRIIIa/g with
sdAb C17-Fc binding is, as expected, strongly inhibited. CEA-positiveLS174TtumorcellsinthepresenceofbsFab
Competition assays were also conducted on Jurkat- C21 or C28 could lead to IL-2 secretion. This activation
huFcgRIIIa/g cellsinthepresenceof90% humanserum was not seen in the absence of bsFab, or when using
andvariousconcentrationsofbsFabs(SupplementaryFig. nontransfected Jurkat cells or CEA-negative LAN-1
S1B). The interaction of bsFabs with FcgRIIIa was not humantumorcellsinthepresenceofbsFab,asnegative
hinderedbythepresenceofhumanIgG,evenatlowbsFab controls (Fig. 3). These results show in vitro that the
concentrations.Thesameresultswereobservedwithpuri- efficientactivationofeffectorcellsbymonovalentbsFabs
fiedhumanpolyclonalIgGinacompetitionassay(datanot isdependentonthepresenceofcellsexpressingthetumor
shown).Theseresultsareinagreementwithourprevious antigen,anddoesnotleadtounwantedcytokinerelease
data (28) showing that sdAbs C21 and C28 recognize byFcgRIII-positiveeffectorcellsintheabsenceoftumor
FcgRIIIa epitopes different from those bound by human cells.
IgGFcportion.
bsFab-dependentNKcell-mediatedcytotoxicity
IL-2releaseassays The cell-mediated cytotoxicity of bsFabs was investi-
Onepossiblesideeffectofantibodiestargetingactivat- gatedusingapaneloffourdifferenthumantumorcells
ing receptors is the activation of effector cells in the expressing various levels of CEA, as shown in Fig. 4A.
absence of target cells, potentially leading to a systemic Freshly purified human NK cells were used as effector
inflammatoryresponse(cytokinestorm).Toevaluatethis cellsatanE:Tratioof10:1.Dose–responsecurvesallowed
possibility, Jurkat and Jurkat-huFcgRIIIa/g cells were thedeterminationofEC values(Fig.4B)andpercentages
50
cultured in the presence of monovalent bsFab. Bivalent ofmaximallysis.BothbsFabstriggeredastrongcytotox-
anti-FcgRIIIamAb3G8,abletocrosslinkthereceptorand icityagainstSKOV3-CEAcellsinadose-dependentman-
triggeractivation,wasusedaspositivecontrol.Monova- nerbutwereineffectivewhenCEA-negativeSKOV3cells
lentbsFabsonlyledtoamarginalIL-2secretion,indicat- weretested(Fig.4B).CellsfromallCEA-positivetumor
ingthatFcgRIIIa-positivecellscannotbeactivatedinthe celllineswerekilledwithasimilarefficiency,leadingto
1486 MolCancerTher;12(8)August2013 MolecularCancerTherapeutics
Downloaded from mct.aacrjournals.org on April 10, 2019. © 2013 American Association for Cancer Research.
Published OnlineFirst June 11, 2013; DOI: 10.1158/1535-7163.MCT-12-1012
Anti-FcgRIII(cid:3)CEABispecificFab-likeAntibodies
Figure4. bsFab-dependentNK
cytotoxicity.A,CEAexpressionof
varioustumorcelllineswas
assessedbyflowcytometry(open
black,isotypecontrol;filledgray,
anti-CEAsdAbC17)onSKOV3,
SKOV3-CEA,BxPC3,LS174T,or
HT29cells.B,humanNKcellswere
mixedatratio10:1withCEA-
positivecellsinthepresenceof
indicatedconcentrationsof
bsFabs.Targetcellviabilitywas
measuredbyCellTiter-Gloviability
assayafterovernightincubationat
37(cid:2)Cinthepresenceofindicated
concentrationsofbsFabC21(open
circle),bsFabC28(opensquare),
exceptforSKOV3cellsforwhich
bsFabC21isrepresentedbyopen
triangleandbsFabC28byopen
diamond.Table1summarizes
EC50(pmol/L)ofeachbsFab
determinedforeverytumorcell
line.ErrorbarsrepresenttheSDof
experimentscarriedoutin
triplicates.
EC inthe10pmol/Lrangeusingthreedifferentdonors. cytotoxic activity triggered by sdAb C17-Fc was lower
50
Inallthreecases,aslightlyhigherefficiencywasobserved thanthatofbsFabC21,bothintermsofEC (valuesinthe
50
withbsFabC21,potentiallyduetoitshigheraffinityfor nmol/Lrange)andofmaximallysis(<50%),despiteits
FcgRIIIa(Table1). bivalentbindingtoCEA.This resultislikelyduetothe
weakerinteractionoftheFcportionofsdAbC17-Fcwith
SensitivitytoFcgRIIIapolymorphism FcgRIIIa,ascomparedwithbsFabC21.Moreimportantly,
BecausetheefficiencyofADCCisaffectedbyFcgRIIIa the lytic activity of sdAb C17-Fc varied according to
polymorphismatposition158,experimentswerecarried FcgRIIIa polymorphism, NK cells from 158F/F donor
out using SKOV3-CEA cells as target cells and purified being significantly less efficient than NK cells from
NKcellsfromFcgRIIIa158V/V,F/F,orV/Fdonorsand 158V/V and 158V/F donors. In contrast, the cytotoxic
bsFab C21 or sdAb C17-Fc. As shown in Fig. 5A, the activity of bsFab C21 was very similar whatever the
FCGR3genotypeofthedonor.Theseresultsconfirmthat
theepitoperecognizedbysdAbC21isnotaffectedbythe
Table 1. Half maximal effective concentration
nature of residue 158, and show that bsFab can trigger
(EC ) of bsFabs inADCC assay
50 highly potenttumorkilling atverylowdoses,indepen-
dentlyofFcgRIIIapolymorphism.
EC (pmol/L)
50
Celllines bsFabC21 bsFabC28 SensitivitytothepresenceofsolubleCEA
SKOV3-CEA 4(cid:5)0.2 6(cid:5)0.5 It is well known that CEA can be released by phos-
LS174T 3(cid:5)1.5 4(cid:5)0.4 pholipasesfromthecellsurfacethroughcleavageofits
BxPC3 8(cid:5)3 17(cid:5)6 glycosyl-phosphatidyl-inositol linkage, leading to the
HT29 5(cid:5)1.5 27(cid:5)10 appearanceofasolubleformintheblood(37).Itrepre-
sents a valuable tumor marker because its serum level
www.aacrjournals.org MolCancerTher;12(8)August2013 1487
Downloaded from mct.aacrjournals.org on April 10, 2019. © 2013 American Association for Cancer Research.
Published OnlineFirst June 11, 2013; DOI: 10.1158/1535-7163.MCT-12-1012
Rozanetal.
therapy,inparticularinclinicalsettingswherelevelsof
solubleCEAarelow.
BiodistributionofbsFabinxenograftedmice
ThebiodistributionofbsFabC21,whichexhibitedthe
most favorable in vitro characteristics in terms of cell
bindingandcytotoxicity,wasstudiedusingxenografted
nude mice. bsFab C21 was labeled with 125I and was
injectedintravenouslyinmicebearingbothLS174Tcells
(CEA-positive colorectal tumor; right flank) and CO115
cells(CEA-negativecolorectaltumor;leftflank,asnega-
tivecontrol).Miceweresacrificed3,6,15,24hoursafter
bsFabinjectiontoquantifythepresenceofbsFabinblood,
variousorgans,andtumors.Besttumor-to-normalorgan
ratioswereobserved15hoursafterinjection(Fig.6A)with
A
25
20
e
u
s 15
s
Ti
% ID/g 10 n.*s***
*
5
Fdmiegixpueerdneda5et.nrtaInNtifloKu1ec0ne:lc1lecwyotiftohFtocSxgKiRcOiItIVyIa.3A-pC,ohElyuAmm–oparonpshFiticsivgmeRILoIIurac-s1-op5luo8bsVil/teFivCNeEKcAeclolesnllsinbwstheFeraeb- 0BlCoEodA+ tuCEmoAr- tumorLiverKidney LungSpleenHeartMuscle Bone SkIinntestineColon
presenceofvariousconcentrationsofbsFabC21orsdAbC17-Fc.Target
cellviabilitywasmeasuredbyCellTiter-Gloviabilityassayafterovernight B
incubationat37(cid:2)C.bsFabC21(blackcircle)orsdAbC17-Fc(opencircle) 1,000
with158V/VNKcells,bsFabC21(blacksquare)orsdAbC17-Fc(open 3m) 800
square)with158F/VNKcells,andbsFabC21(blacktriangle)orsdAb m
C17-Fc(opentriangle)with158F/FNKcells.(cid:6),asignificantdifference me ( 600
betweensdAbC17-Fcwith158V/VNKcellsgroupandsdAbC17-Fcwith u
E15rr8oFr/bFaNrsKrceeplrlsesgeronutpth(e(cid:6),SsiDgnoififceaxnptedriimffeernetnscceabreritewdeoeuntginroturpipsliPca<te0s.0.5B),. or vol 400 *
humanNKcellsweremixedataratioof10:1withBxPC3cellsinthe um 200
presenceofvariousconcentrationsofbsFabC21andofsolubleCEA. T
TargetcellviabilitywasmeasuredbyCellTtiter-Gloviabilityassayafter 0
0 5 10 15 20 25 30
overnightincubationat37(cid:2)C.bsFabC21(opencircle)orbsFabC21with Days
1mg/mL(opensquare)or0.1mg/mL(opentriangle)ofsolubleCEA.No
significantdifferencewasfoundbetweenbsFabC21with1mg/mLof
solubleCEAgrouporbetweenbsFabC21with0.1mg/mLofsolubleCEA Figure6. bsFabinxenograftedmice.A,biodistribution:125I-labeledbsFab
groupversuscontrolbsFabC21group.ErrorbarsrepresenttheSDof
C21wasinjectedintravenouslyintoLS174T(CEApositive)andCo115
experimentscarriedoutintriplicates. (CEAnegative)xenograftednudemice.Mice(n¼3/group)weresacrificed
atindicatedtimes3hours(black),6hours(darkgray),15hours(lightgray),
and24hours(white)andradioactivityofeachorganwasmeasured
correlates disease progression. However, this soluble
andplottedas%injecteddose/goftissue.ThesignificanceofbsFab
form of CEA could also interfere with antibody-based
retentioninCEA-positivetumorversusCEA-negativetumorwas
immunotherapy by competing with membrane-bound comparedateachtime.(cid:6),significantdifferencebetweengroups:
CEA for antibody binding. Thus, we conducted ADCC (cid:6)(cid:6),P<0.005,(cid:6),P<0.05;andn.s,nonsignificant.Errorbarsrepresentthe
assays using BxPC3 tumor cells in presence of interme- SDofexperimentscarriedoutintriplicates.B,tumorgrowthinhibition
assay:atday0,NSGmice(n¼8/group)werexenografted(s.c.)with
diate (0.1 mg/mL) or high (1 mg/mL) concentrations of
BxPC3cellsandinjectedintraperitoneallywithhumanPBMCsandbsFab
soluble CEA. As shown in Fig. 5B, soluble CEA has a C21(100mg)orPBS,followedatdays1and2by50mgofbsFabC21
detectablebutmoderateeffectontheantitumorefficacyof orPBS(i.p.).TumorgrowthwasfollowedusingaVerniercaliper.PBS
bsFabC21.EC valuesincreasedtoaround100pmol/Lat (opencircle),bsFabC21(opensquare),PBMCsandPBS(opentriangle),
50 andPBMCsandbsFabC21(opendiamond).(cid:6),significantdifference
thehighestCEAconcentration,althoughitdidnotaffect
betweenPBMCsandbsFabC21groupandPBMCsandPBSgroup
themaximallysisvalue.Theseresultssuggestthatcircu- (P<0.05).ErrorbarsrepresenttheSDofexperimentscarriedoutwith8
latingCEAshouldnotcriticallyimpacttheeffectofbsFab micepergroup.
1488 MolCancerTher;12(8)August2013 MolecularCancerTherapeutics
Downloaded from mct.aacrjournals.org on April 10, 2019. © 2013 American Association for Cancer Research.
Published OnlineFirst June 11, 2013; DOI: 10.1158/1535-7163.MCT-12-1012
Anti-FcgRIII(cid:3)CEABispecificFab-likeAntibodies
morethan4.5%ID/gofbsFabbeinglocalizedintheCEA- produce.Thus,wehavestudiedaformatrelyingonthe
positivetumor,whereaslessthan0.8%ID/gremainedin useofllamasdAbsthatpossessvaluablefunctionaland
the blood or other organs and CEA-negative tumor, structuralpropertiesandofhumanCH1/Ckdomainsas
except for kidneys (around 1%). Nontumor infiltrated heterodimerization motif, to produce Fab-like bispecific
organswithnoticeablelevelofbsFabwerekidneys,prob- antibodyfragments.
ablyduetobsFabclearance. Asaproof-of-concept,weconstructedbsFabstargeting
CEA-positive tumor cells and human FcgRIIIa–positive
bsFabinvivoefficacyinxenograftedmice immune effector cells, using previously isolated anti-
To determine whether the potent in vitro activity of FcgRIIIa sdAbs–binding epitopes located outside the
bsFabwouldtranslateintoinhibitionoftumorgrowthin Fc-binding site, and an anti-CEA sdAb chosen for its
vivo,anadoptivetransfermodelwasused.Atday0,NSG affinity comparable with conventional anti-CEA mAbs,
micewereengraftedsubcutaneouslywith CEA-positive despiteitslackofbivalency.
human pancreatic cancer BxPC3 cells and with human These molecules were easily produced in E. coli with
PBMCsfromhealthydonorsandtreatedwith100mgof routine yield ranging from 0.5 to 2 mg/L of culture in
bsFabC21byintraperitonealinjection.InjectionsofbsFab nonoptimalshakerflaskconditionsinanacademiclabo-
C21(50mg)weredoneforthe2followingdays(cumula- ratory,whichcomparesfavorablywithotherformatssuch
tivedose/mouse¼200mg).Micetreatedonlywitheither as tandem scFv (the BiTE format) requiring eukaryotic
PBS, bsFab C21, or PBMC þ PBS were used as control cellsforproduction(41).ThehighstabilityofthesebsFabs
groups (n ¼ 8). As shown in Fig. 6B, significant tumor conferredbytheabsenceoflinkersisillustratedbythefact
growthinhibitionwasvisibleinmicetreatedwithPBMCs thattheyretainedtheirfullbindingactivityatleastfor7
andbsFabC21ascomparedwithPBSorbsFabtreatment daysinhumanserumat37(cid:2)C,evenatlowconcentration.
alone.Aslightreductionoftumorgrowthwasobserved One of the key issues with bsAbs is their capacity to
withhumanPBMCstreatmentalone. triggeranefficientcell-mediatedcytotoxicity.Ourinvitro
datashowedastrongandspecificNKcell–mediatedlysis
Discussion
ofCEA-expressingtumorcells(frompancreaticandcolo-
Redirectingthecytotoxicpotentialofleukocytestoelim- rectalcancers)atpicomolarconcentrations,thatis,orders
inatetumorcellshasbeenamajorimpulseforthedevel- of magnitude lower than conventional mAbs (10, 42).
opment of bsAbs for cancer immunotherapy. FcgRIIIa These results suggest that a high affinity for FcgRIIIa
is an attractive candidate to recruit effector cells. It is translatesintoimprovedcytotoxicactivityofeffectorcells
stronglyexpressedbyNKcellsthatplayacriticalrolein in vitro, a finding also described with mutated and gly-
theinductionofADCC,oneofthemajormodesofaction coengineeredantibodies(43,44).Asopposedtoclassical
ofantitumorantibodies.Inaddition,FcgRIIIaisalsoexp- mAbs such as rituximab, target antigen density did not
ressedonmonocytesandmacrophagesthatareimportant significantlyimpactbsFab-mediatedADCCasshownby
actorsofantitumorimmunity.Anti-FcgRIIIabsAbshave theuseoftumorcelllinesexpressingdifferentamountsof
the potential to bypass several important limitations CEA.Thesignificantinhibitionoftumorgrowthobserved
facedbytherapeuticantibodies.However,difficultiesin upon human PBMCs adoptive transfer to CEA-positive
producing sufficient amounts of functional and stable tumor–bearingNSGmicesupportstheinvivoefficacyof
bsAbsatreasonablecostshavestronglyhamperedtheir thebsFabformat.ThelowlevelofINF-gsecretionbyNK
developmentformanyyears,althoughearlyclinicaltrials cellsinducedbybsFabintheabsenceoftargetcellsinvitro
hadbeenpromising(38,39).Molecularconstructs,such suggests in addition that bsFabs should not induce any
as bispecific diabodies, single-chain diabodies, tandem systemic activation of NK cells in vivo and should not
0
scFvsandF(ab)obtainedbyrecombinantDNAtechnol- provokecytokinestorm.
ogy,mightcircumventtheabovelimits.Theseformatscan Importantly, by carefully choosing the anti-FcgRIII
be expressed in bacteria or mammalian cells and may sdAb,wehaveconstructedbsFabswhoseactionisinsen-
showbenefitsduetotheirsmallsize,althoughtheycan sitive to FcgRIIIa polymorphism and that do not bind
presentmanufacturingchallengesrelatedtotheirproduc- inhibitoryFcgRIIb.
tionandinvitroandinvivostability.Notably,thevariable Because of their high efficiency and the absence of
andunpredictableexpressionyieldsobservedfordiffer- competitionwithserumIgG,thesemoleculesshouldbe
ent scFvs hamper their widespread use. Moreover, the clinicallyactiveatmuchlowerconcentrationsthanthose
linker connecting the VH and VL in the scFv and that usedforconventionalmAbsasalreadyreportedforthe
between different scFvs can provoke aggregation of the bispecific T-cell engager format (21). These properties,
molecule. Shortening the linkers to produce diabodies addedtothepossibilitytoproduceeasilythesefragments
does not always guarantee success, as the linkers can inE.coliinlargeamounts,havethepotentialtosubstan-
induceanerroneousanglebetweentheVH–VLpair,thus tiallyreducethehighmanufacturingcostsassociatedwith
formingpoorlyfunctionalantibodies(40). mAbtherapy.
Animportantmotivationoftheworkreportedherein TheretargetingofFcgRIII-expressingcellshasanother
þ
wasthereforetodesignanewlinker-freebispecificformat advantage over CD3 T-cell retargeting by bsAb. Anti-
to avoid aggregation and stability issues and easy to CD3 bsAbshave been shown tolead to the recruitment
www.aacrjournals.org MolCancerTher;12(8)August2013 1489
Downloaded from mct.aacrjournals.org on April 10, 2019. © 2013 American Association for Cancer Research.
Published OnlineFirst June 11, 2013; DOI: 10.1158/1535-7163.MCT-12-1012
Rozanetal.
and the activation of regulatory T cells (T ; ref. 45) activityirrespectiveofFcgRIIIapolymorphism,glycosyl-
reg
possibly leading to a downregulation of the antitumor ation,andcompetingendogenousIgGissues.Altogether,
response. FcgRIII retargeting does not activate the T thesepropertiesassociatedtoasignificantcytotoxicactiv-
reg
subset,anddoesnottriggerimmunosuppression. ityinapreclinicalinvivomodelconfertothisbispecific
A possible limitation of anti-CEA bsFabs might be Fab-like antibody format the potential to lead to a new
thepresenceofshedextracellulardomainsofCEAinthe generation of highly active therapeutic antibodies for
serumofpatientswithcancer,possiblyblockingtheanti- tumortherapy.
CEA–bindingsite.Here,weshowthatbsFab-dependent
cytotoxicityisonlyslightlyimpactedbytheconcentration DisclosureofPotentialConflictsofInterest
of soluble CEA exceeding concentrations found in the J.-L.Teillaudisaconsultant/advisoryboardmemberofBiotechCom-
pany.Nopotentialconflictsofinterestweredisclosedbytheotherauthors.
majority of patients with cancer (46). Thus, as already
suggestedbysomeclinicalstudiesusinganti-CEAanti-
Authors'Contributions
bodies(47,48),solubleCEAshouldnothaveasignificant Conceptionanddesign:C.Rozan,A.P(cid:3)elegrin,P.Chames,J.-L.Teillaud,
impactontheclinicalefficacyofthebsAb. D.Baty
Anotherimportantissuefacedbytherapeuticantibody Development of methodology: C. Rozan, A. Cornillon, C. P(cid:1)etiard,
M.Chartier,G.Behar,C.Boix,P.Chames,D.Baty
fragmentsdevoidofFcportionistheirpharmacokinetic Acquisitionofdata(providedanimals,acquiredandmanagedpatients,
property.Asalikelyconsequenceoftheirsmallsizeandof providedfacilities,etc.):C.Rozan,A.Cornillon,C.P(cid:1)etiard,B.Robert
Analysisandinterpretationofdata(e.g.,statisticalanalysis,biostatis-
theirlackofinteractionwithFcRn,Fabfragmentshavea tics,computationalanalysis):C.Rozan,A.Cornillon,C.P(cid:1)etiard,G.Behar,
meanretentiontimeinthebody35-foldlowerthanfull- B.Kerfelec,B.Robert,P.Chames,J.-L.Teillaud
length IgG (49). Despite this disadvantage, it has been Writing,review,and/orrevisionofthemanuscript:C.Rozan,A.Cornil-
lon,B.Kerfelec,P.Chames,J.-L.Teillaud,D.Baty
shownthattheirsmallersizeascomparedwithwholeIgG
Administrative,technical,ormaterialsupport(i.e.,reportingororga-
limitstheirretentioninnontargetedorgansandincreases nizingdata,constructingdatabases):G.Behar,P.Chames
theirtumorpenetration.Consistentwiththosedata,bio- Studysupervision:A.P(cid:3)elegrin,D.Baty
distributionexperimentscarriedoutinnudemicebearing
Acknowledgments
CEA-positivetumorxenograftsrevealedarapidincrease TheauthorsthankSabrinaSeddikandSt(cid:1)ephanieCharlesforskilled
ofCEA-positivetoCEA-negativetumorratiouptoalmost technicalassistanceandIsabelleTeulon,JacquesBarbet,Herv(cid:1)eWatier,
7-foldat15hours,showingaspecificandefficienttumor S(cid:1)everineBarth,andMathieuCollinforhelpfuldiscussion.
targeting of bsFabs. Nevertheless, the bsFab format
GrantSupport
authorizesthelinker-freeadditionofsingle-domainanti-
ThisworkwassupportedbyINSERM,INSERM-Transfert(ProCop).C.
bodiesattheC-terminusofCH1orCkdomains(datanot BoixwassupportedbyagrantfromthePo^ledeComp(cid:1)etitivit(cid:1)eM(cid:1)editech
shown).Itcanthereforebeenvisagedtoconstructtrispe- Sant(cid:1)e,Ile-de-France(Immucanproject).C.P(cid:1)etiardwassupportedbya
cific formats by adding an anti-human serum albumin grantfromInserm-Transfert(ProCop).
Thecostsofpublicationofthisarticleweredefrayedinpartbythe
sdAb, a method that has been shown to significantly paymentofpagecharges.Thisarticlemustthereforebeherebymarked
increasetumorretentionandhalf-life(50,51). advertisementinaccordancewith18U.S.C.Section1734solelytoindicate
thisfact.
Inconclusion,wehavedesignedanewclassofsingle-
domain–based bispecific antibodies that combines easy
ReceivedOctober17,2012;revisedMay14,2013;acceptedJune5,2013;
production, high stability, and potent in vitro ADCC publishedOnlineFirstJune11,2013.
References
1. ChamesP,VanRegenmortelM,WeissE,BatyD.Therapeuticanti- HER-2/neu-positivemetastaticbreastcancer.JClinOncol2008;
bodies:successes,limitationsandhopesforthefuture.BrJPharmacol 26:1789–96.
2009;157:220–33. 8. SiberilS,DutertreCA,FridmanWH,TeillaudJL.FcgammaR:thekeyto
2. BorossP,LeusenJH.MechanismsofactionofCD20antibodies.AmJ optimize therapeutic antibodies?CritRevOncolHematol2007;62:
CancerRes2012;2:676–90. 26–33.
3. DesjarlaisJR,LazarGA,ZhukovskyEA,ChuSY.Optimizingengage- 9. Shinkawa T,NakamuraK, Yamane N, Shoji-HosakaE, Kanda Y,
mentoftheimmunesystembyanti-tumorantibodies:anengineer's SakuradaM,etal.Theabsenceoffucosebutnotthepresenceof
perspective.DrugDiscovToday2007;12:898–910. galactoseorbisectingN-acetylglucosamineofhumanIgG1com-
4. GloriusP,BaerenwaldtA,KellnerC,StaudingerM,DechantM,Stauch plex-type oligosaccharides shows the critical role of enhancing
M,etal.Thenoveltribody[(CD20)(2)xCD16]efficientlytriggerseffector antibody-dependent cellular cytotoxicity. J Biol Chem 2003;278:
cell-mediatedlysisofmalignantBcells.Leukemia2013;27:190–201. 3466–73.
5. UmanaP,Jean-MairetJ,MoudryR,AmstutzH,BaileyJE.Engineered 10. deRomeufC,DutertreCA,LeGarff-TavernierM,FournierN,Gaucher
glycoformsofanantineuroblastomaIgG1withoptimizedantibody- C,GlacetA,etal.Chroniclymphocyticleukaemiacellsareefficiently
dependentcellularcytotoxicactivity.NatBiotechnol1999;17:176–80. killedbyananti-CD20monoclonalantibodyselectedforimproved
6. CartronG,DacheuxL,SallesG,Solal-CelignyP,BardosP,Colombat engagement of FcgammaRIIIA/CD16. Br J Haematol 2008;140:
P, et al. Therapeutic activity of humanized anti-CD20 monoclonal 635–43.
antibodyandpolymorphisminIgGFcreceptorFcgammaRIIIagene. 11. PreithnerS,ElmS,LippoldS,LocherM,WolfA,daSilvaAJ,etal.
Blood2002;99:754–8. HighconcentrationsoftherapeuticIgG1antibodiesareneededto
7. MusolinoA,NaldiN,BortesiB,PezzuoloD,CapellettiM,Missale compensateforinhibitionofantibody-dependentcellularcytotox-
G, etal.Immunoglobulin GfragmentC receptor polymorphisms icity by excess endogenous immunoglobulin G. Mol Immunol
andclinicalefficacyoftrastuzumab-basedtherapyinpatientswith 2006;43:1183–93.
1490 MolCancerTher;12(8)August2013 MolecularCancerTherapeutics
Downloaded from mct.aacrjournals.org on April 10, 2019. © 2013 American Association for Cancer Research.