Radiolabeled Chimeric Anti-CEA Monoclonal Monoclonal Antibody for ...

Radiolabeled Chimeric Anti-CEA Monoclonal Antibody Compared with the Original Mouse Monoclonal Antibody for Surgically Treated Colorectal Carcinoma Franz Buchegger, Jean-Pierre Mach, AndréPèlegrin,Michel Gullet, Charles-AndréVogel, Thierry Buclin, Jean-Etienne Ryser, Bernard Delaloye, and Angelika Bischof Delaloye DivL@ionofNuclear Medicine, Depamnentr ofSwgety and Clinical Pharmacolo@, Centre Hospitalier Unive,@itaire VaudoLr and I,t@titute ofBiochemist,y, Univei@y ofLausanne, Lausanne, Switzerland; and L@epa,iment of Nuclear Medicine,

HOpital Cantonal

Universitaire,

Genève, Switzerland

radiosensitive tumorssuch as B cell lyinphomas in patients Biodistribution and tumoruptakeof a chimerichuman-mouse (9,10), they havethe additionaladvantagethat they can monoclonalantibody(MAb)andthe originalmouseMAbhave yield precise information concerning biodistribution and beencomparativelystudied.Methods: Eighteenpatientswith tumor dosimetry. These data can be useful for the selection suspectedcolorectalcancerscheduledfor surgeryunderwent of the most appropriatetype of antibody for radioimmuno immunoscintigraphy with msWabaledthimedc anti-CEAMAb. scintigraphy and radioimmunotherapy (9—19). Iodine-125and 1311 trace-labeledchimericand onginalmouse Monoclonal antibodies used for tumortargetingare gen MAb were simultaneousiyinjectedfor b@disthbudon studies. erally ofmurine origin and can elicit human anti-mouse IgG Results:Similarserumkineticsanda lowimmunogenlcfty were observedfor both antibodies.Mean bindingcapadtyto CEA antibodies (HAMA) in many patients (20—22).High titers measuredin PBSafterradiolabeling wasidenticalforbothMAbs of HAMA are frequentlyobserved after repeatedinjections andit wasslightlydecreasedwhenmeasuredin serum1-4 hr of MAbs or when they are coupled to other immunogenic of anti after injection.Radiochromatograms of patientssera showed proteins such as toxins or enzymes. Chimerization immunecomplexformation relatedtotheamountofcirculating bodies representsa first step toward reducingthe immuno CEA.Postoperative exvivoradioa@tMty countingintissuesam genicity of MAbs for applicationin patients (1Z20,23,24). piesrevealedsimilarantibodydistributionswith notal@y similar Immunogenicity is, however, still a problem for certain chi antibodyuptakesintumors.Hightumoruptakes(between0.02 mericantibodies(25). Furtherhumanizationby graftingonly to 0.06%injecteddoseperg) wereobservedin 3 of 13patients the complementaiydeterminantregions(CDR)of the mouse operatedfor pnmaryor metastaticcokxectalcancer.Conclu MAb DNA into human IgO DNA as described for the re don: In this dual-labeltechnique,the radiolodinated anti-CEA shaped MAbs (26,27) or production of human antibodies 19G4chimencMAb and the o@iginaI mouse lgG1MAb were shownto havevery similarbehavlOrin colorectalcancerpa (28-30) might be necessaiy to further reduce immunogem city. tients. Chimeric antibodies allow the comparison of the biolog Key Words: chimencmonodonalantibody;carcinoembryonic ical behavior of selected human IgO subclasses and verifi antigen;tumoruptake;colorectalcarcinoma cation of their potential for tumor targeting.They can also be used to study their effector functions and the possibility J Nuci Med 1995; 36:420-429 of obtaining fragments for immunoscintigraphyand radio immunotherapy (1Z23,31,32). We have used such chi meric anti-CEA MAbs of different IgG subclasses in ex he concept of using monoclonal antibodies (MAbs) as perimental animal models and could show that both the carriers to deliver cytotoxic drugs, radioisotopes or toxins intact Ig02 MAb and its F(ab')2 fragment demonstrated more selectively into tumors continues to stimulate exper excellent tumor targeting and in vivo stability (31). The tumor localization capacity of the intact Ig04 chimeric imental and clinical research (1—3). While radiolabeled an tibodies have been shown to be useful for therapy of hu MAb has been shown to be identical to the origmal mouse man carcinomas in nude mice (4—8),and of more MAb in tumor bearing nude mice (23). However, the in vivo behavior of chimeric 1g04 F(ab')2 fragments was un satisfactory in mice (32) and this fragment was therefore RecolvedMay31,1994;revisionacce@*ed Sept20.1994. Forcorrespondence andrepdnts cont@FranzBuchegger, MD,D@scnof not used in patients. NudearMedicine.CHIN, CH-1011Lausanne, Switzerland. Here we compare the intact chimeric anti-CEA MAb

420

TheJournalofNudearMedicine • Vol.36 • No.3 • March1995

(human I@G4)with its originalmurine IgG@MAb by differ ential labeling, co-injection and cx vivo measurement of tumor and normal tissue radioactivity distributionsin pa tients. METhODS Patients Patients selected for the comparative biodistributionstudy of

chimericand mouse anti-CEAMAb (n = 18)were suspectedof colorectal cancer. Surgery was planned 1 to 5 days after MAb injection. Immunoscintigraphy was performed with the aim of staging more precise'y the disease. Definitive diagnosis in these patients was: primary colorectal

adenocarcinoma

in nine patients,

reasons)

does

not significantly

influence

the pharmacokinetic

analysisbecause both antibodieshave been injectedin the same perfusion and showed a very similar behavior in the hours follow

inginjection. T@* Sam@s Patientserumwascollectedimmediatelyfollowingand1, 4—6, 24 and 48 hr after injection;additionallyblood samplingwas performed during surgery (at the moment of tumor removal). Tumorsamplesof primarytumorsor localrecurrenceswerean alyzed together with normal colon and fat. Normal colon tissue was dissected into the normal mucosa, known to contain CEA

(37), and the rest of normalbowel wall. Liver metastaseswere analyzedtogetherwithlivertissuesurroundingthe tumoranda

one of whom had initialliver metastasis, one local recurrenceand

small biopsy of distant normalliver and fat. The tumorwas mac

five liver metastases.

roscopicallyseparatedfromnormaltissueand fromnecrosis.Tis

One patient had an ovarian carcinoma, one

patient a benign polyp and one patient with a suspicion

of liver

metastasis had no tumorat the time of surgery. For three of these patients, surgely was either performed too late after antibody

sue samples were weighed and counted in a triplechannelgamma counter togetherwith a sample ofthe injected material that served

as referencefor the totalinjecteddose. Samplesstillcontaining

injectionor was cancelled.For determinationof the serum half lifeof chimericMAb,nineadditionalpatientswereincludedwho

radioactivitymeasurementswerecorrectedforcrossoverof 1311 in

were only injectedwith chimericMAb andhada follow-uptime of

the 1@Ichannel.

1231were counted

again after complete

decay ofthis

isotope.

Final

2 to 6 days. Finally,one patientincludedherehadsurgeiyfora livermetastasis8 days afterinjectionof 2 mgof the IgG4chimeric In Vitro Testing of Radlolabeled MAbs MAb together with 30 @gof a chimeric MAb with the same Theinvitroimmunoreactive fractionof radiolabeledchimeric variableregionsbut with human IgG2constant domains(32). and original mouse MAb was determined in a binding assay on CEA insolubilizedon CNBr-Sepharose(Pharmacia,Uppsala, Monoclonal AntibOdies Themouse-human chimericmonoclonalantibodyusedhereis Sweden).Tento 50 nCiof radiolabeledMAbwereincubatedfor in PBSbuffercontaining1%normalmouseserum of humanI@G4 subclassandwas derivedfromthe murineMAb 16hrat 25°C and 1%normalhumanserumwith5 @d packedCEA-Sepharose CE25/B7(6,23)(CIBAGEIGY,Basel,Switzerland).ThisMAbis (containing about2 @&g purified CEA).Afterwashing,boundra directed against the epitope Gold 4 of CEA (33). It has a high specificity for CEA (34) and does not crossreact with NCA-55 or NCA-95 (35) or other granulocyte glycoproteins. Theoretically,

the murineandchimericantibodysubclassesbothhaveminimal effector functions: they should not react with Fc receptors on

monocytesand macrophagesand shouldnot activatethe comple ment cascade (36). The original mouse MAb has been used in patients both for immunoscintigraphy(14) and in a first trial of

rsdioimmunotherapy(1). MouseMAbwas preparedfromascites by ammonium sulfate precipitation and ion exchange chromatog

dioactivity was determined as percent of input radioactivity. Sim

ilarly, patients, serum samples obtained 1 to 4 hr after injection (containingsimilar amounts of radioactivity)were diluted 113in

PBS buffercontaining1%normalmouse serumandwere also incubatedwith 5 @1 packedCEA-Sepharose.Nonspecificbinding was measured by incubation with irrelevant protein also coupled

to CNBr-Sepharose. Itwas alwaysbelow2%andwas subtracted fromCEA bindingvalues. Trichloroaceticacid precipitationof radioactivityafterlabelingshowed that more than 95%of the radioactivity was bound to protein for all preparations. MalytiCaI

raphy (6). ChimericMAbwas producedin Sp2@cellstransfected size chromatographyof the radiolabeledMAbs was done on a witha singlevectorcontainingboththechimericheavyandlight SephadexG-200columnor on a Superdex200 FPLCcolumn chain genes (23). (bothPharmacia).Immediatelyafterlabeling,a sharppeakwas obtainedforbothMAbswithoutdetectableamountsof aggregates Radlolabeling Two to 4 mg of chimeric MAb were labeled by the iodogen andwithonlytraceamountsof freeiodine. methodwith 15 to 30 mCiof ‘@I for immunoscintigraphy (final specific activities were 2 to 4.5 mCi per mg antibody and 4 to 18 CEA and HAMA Assay Circulating CPA was determined in a serum sample of each mCi were injected per patient). Batches of 0.2 mg chimeric MAb and oforiginal mouse MAb were separatelylabeled using 250 @.tCi patient taken before injection of radiolabeledantibodies using a

@

of 1@Iandof ‘@‘i, respectively(finalspecificactivitieswere 0.8 previouslydescribedsolidphase enzymeimmunoassay(38). to 1.1 @&Ci per antibodyfor the trace labelings).For three HAMA and anti-idiotype antibodies were measured in three patients,thetracelabelingswerereversedandthechimericMAb sandwichassays.Briefly,forthethreetestsA, B andC, polysty was labeled with ‘@‘I and the mouse MAb with 1@I. Using the reneballswerecoatedwithA), irrelevantmouseIgG,B), mouse pairedlabelingmethod,it is possibleto analyzethe biodistribution MAb CE 25 or C), chimenc MAb, respectively. These balls were and tumor localizationcapacity of the two MAbs in the same incubated for 3 hr at 25°Cwith 10 p1 patients serum (taken 5—6 wk patient and thus compare results obtained underidenticalbiolog after antibody injection)diluted in 300 @l PBS and peroxidase icalconditions.The 1@I-labeled and the trace-labeledMAbswere coupledmouseMAbCE25(testA andB)or radiolabeled chimeric pooled, diluted in 100ml of0.9% NaCl and perfused intravenously

MAb in test C. A rabbit anti-mouse F(ab')2 antiserum (10 @.d serum

within 15 mm. Total amountof injectedantibody(mouseand diluted1:10'000in PBS) served as a positivecontrol.Normal chimericMAbtogetheror chimericMAbalone)rangedfrom2 to humanseraservedas negativecontrols.All patientsseratested 4 mgin all patients.The 15-mmperfusiontime(usedfor safety before any MAbinjectionwere also negative.

@himenc versusMouseMAbin Patients• Bucheggerat Si.

421

TABLE I PatientsInjectedwithChirnedcandOnginalMouseAnti-CEA MAbsand MajorClinicalParameters Patient no.Age (day)147SexDiagnosisDifferentIatiOnDukesSurgery

rectumIntermB2271 MPrimary

(Simed,Creteil,France)individualpatientdatawereanalyzedin atwo-compartment model.A weighted least-squares method with weights being the reciprocal of the predicted radioactivity was

FUver sigmoldIntermC1369

met.

FPrimary ascIntermB2478

cci and metcci

FPrim

Pharmacokinetics were analyzedby modelinga time (hr)— radioactivity (cpm/ml) curve for each patient. Time 0 immediately after injection was taken as 100%.Using the SIPHAR program

ascIntermD1570

usedto estimatetheparameters.A linearcorrelationanalysiswas usedto calculatethecorrelationcoefficientbetweenthechimeric and the original mouse MAb in serum.

FOvarlanadenocaLow5657

MHyperplastic Polyp1765 sigmoldIntermC2879 FLocal rec.

sigmoldInterm-wellBI944 MPrimary MPrlmatysigmoidLowC21073

sigmoidIntermB—1 FPrimary 187 MPrimary sigmoidmt-wellD11266 MPrimaiycoltrIntermB—1364

sigmokinaB—1463 MUver met rectumIntern,C21 met MLiver colonnaC21672 MLiver met 574 FLiver rectumIntermC51742 met FSuspected

mata@asin11 drIntermB2na FPrimary 858

Eighteen patients were injected with ‘@I-labeled chi meric anti-CEA MAb together with “SI and ‘@‘I trace labeled chimeric and original mouse MAb (Table 1). Al though immunoscintigraphy is not the objective of this study, two representative illustrations of a primaiy and a metastatic tumorimmunoscintigraphyare shown in Figure

1. SerumCEA of thesetwo patientswaslow (1.7and1.3 liver

colon

ng/ml) and immune complex formation was almost nega tive. Both immunoscintigraphsclearly show antibody up take in the tumor 6 and 24 hr after injection. At these times,

blood radioactivity was still high as it has been observed earlier after injection of radioiodinatedintact MAbs.

=not available.

@

RESULTS

-:@‘

A

B

FiGURE 1. (A) Uptakeof thimeiic anti-CEAMAbIn a primarytumorof the colonascendens(Patient3) and (B)in a livermetastasisof

acoloncarcinoma (Patient19).InPanelA,a6-gprimarytumorofthe colonascendens iscleedyvisible ontheplanarview 6hrafter1@l-MAb injection.PanelB showsa coronalsectionoftheabdomenofa patientwithan8-9livermetastasis intheleftliverlobe24hrafter1@IMAb injection.Uptakeinthe metastasis(arrow)is ofsimilarintensitythan that of heartand blgvessela,witha hypoactivecentercorresponding tonecrotic tissue. Activity innormal liverislow.

422

TheJournalofNuclearMedicine • Vol.36 • No.3 • March1995

FIGURE2. (@4@ Radk,actMtyconcentra tionsof the chimericMAb (0) andof the orig inSimouseMAb 4). Arrowslnd@atepercent

A

Insaraof Patients2 and4 andshowexcep tionallyshortcirculationtimesfor bothchi merleMAbandoriginalmouseMAb.(B)Se rumdisappearance of mouseMAbIsplotted aginstthstofct@medc MAbofallserumsam

100

a p

•

•0

—InPanel A(Patients 2and4arenot

g60 g

a

@(

•

U

:-‘

hours aft.r Injic010n

20

Sirum Kinetics The serum half-Me of the two MAb forms was very similar in all patients with a tendency of marginallylonger retentionof the chimeric MAb later after injection (after24 to 48 hr, Fig. 2A). Many of these patients were followed for only one to two days since serum collection was discontinued after Frequently,

this

short

observation

time

Y=-1.54+1.146x; correlation coefficient r@= 0.95.

4000 80100dcr•aa•chimiric

MAb

pircantB/

Theca@stedregression sWSig@

linehasthecharactedstlcs of

@40

0

surgery.

—.

Overall,asimilarserumdisappearance khietlc of bothantibodyformsisapparentwltha mar

ginallyshorterhat-lifeforthemouseMAb.

The alpha and beta half-lives have been calculated for the chimeric MAb from a group of 12 patients, 9 of whom hadbeen injectedwith the chimeric MAb alone because no surgerywas scheduled. In these patients with a follow-up time of 2 to 6 days (mean 3.3 days) (Fig. 4), the median alpha and beta half-lives were 7.2 hr (range 1.4 to 18.4 hr) and 91 hr (range 30 to 292 hr), respectively.

did not

Immunoreactive Fraction and Immune Complex.. The immunoreactive fraction of both MAbs was mea and beta half-lives. It appears, however, from the individ sured in buffer immediately after radiolabeling and com ual data shown in Figure 2A that there is only a minor pared to binding in serum of patients collected 1 to 4 hr difference in the circulation kinetics between the two anti after injection (Table 2). For half of the patients, the im bodies. Thus, a high correlation (correlation coefficient r@ munoreactive fraction ofradiolabeled chimeric and original = 0.95) was observed for the two radiolabeled antibody mouse MAb in serum was similar to that observed before fonns in the serum samples as shown in a double plot injection (less than 15%decrease in serum as compared to analysis (Fig. 2B). preinjection value). The remaining patients showed a more In two patients (Patients 2 and 4), the serum half-lives of important decrease of binding activity for the two antibod both the chimeric MAb and of the original mouse MAb were very short (5 and 7.1 hr) and only between 6 to 13% of radioactivity remained in circulation after one day for both antibody forms. In one of these two patients (Patient 300Ca@Wb CI IIj@@*30 I 4) the accelerated half-lifewas most likely due to bindingof the antibodies to CEA, since no HAMA and no anti-idio types were detected in his serum. Indeed, he had a high I amount of serum CPA (342 ng/ml) and 47% of the chimeric MAlIand 43%of the mouse MAt, were in aggregatedform 1000 early after injection, as determined by size chromatogra phy on Sephadex 0200 (Fig. 3C). For the second patient 500 (Patient 2), no obvious reason for the short half-life of the 0 two antibody forms was found. CEA in serum was rela 20 30 40 50 60 70 3070tubs 40 50 60 40 50 60 70 numbar tively low (24 ng/ml), and both antibody forms appeared to circulate at a high percent in monomeric form: the serum FIGURE3. Sephadex G 200radlochromatographs ofthreerep collected4 hr after injectioncontainedonlyabout4%of resentative patientswithlow,meciumandhighamountsof serum aggregates as determined by Sephadex G200 radiochro CEA Patient8 (a) hed 2.8 ng/n@of CEA and 2% of Immunecom matography. No indication of increased dehalogenation in serum was found since only low amounts of free iodine CEAand 14%(chimerlcMAb,4 and 8% (mouseMAb,x) of permit us to obtain a clear result concerning serum alpha

C

@,1500

@

U

ic.i: —..— chi..•.@

—forboth MM@s were found. Patient II(b)tied 26.7 n@of

(