- Email: [email protected]
human (IgG4) antibody
Molecular Immunology, Vol. 30, No. 1, pp. 105-108, 1993 Printed in Great Britain.
SHORT
A SINGLE
0161-5890/93 $5.00 + 0.00 0 1992 Pergamon Press Ltd
COMMUNICATION
AMINO ACID SUBSTITUTION ABOLISHES THE HETEROGENEITY CHIMERIC MOUSE/HUMAN (IgG4) ANTIBODY
OF
S. ANGAL, D. J. KING, M. W. BODMER. A. TURNER, A. D. G. LAWSON, G. ROBERTS, B. PEDLEY’ and J. R. ADAIR* Celltech
Ltd.. 216 Bath Rd.. Slough,
Berks.. SLl 4EN U.K. and lDepartment
of Clinical
Oncology,
Royal
Free Hospital.
London
U.K.
*Author for correspondence
(First received 28 July 1992; accepted in revised form 8 September 1992)
Abstract - Human immunoglobulin G4 (IgG4) exists in two molecular forms due to the heterogeneity of the inter-heavy chain disulphide bridges in the hinge region in a proportion of secreted human IgG4. This heterogeneity is only revealed under denaturing, non-reducing conditions in which an HL “half antibody” is detected, a phenomenon not seen in other human IgG isotypes. In native conditions noncovalent
interactions
hold
the antibody
together
as the H2I2
tetramer.
Analysis
of the hinge
sequences of human IgG heavy chains
suggested that the presence of serine at residue 241 might be the cause of this heterogeneity. We therefore changed the wine at 241 to proline (found at that position in lgG1 and IgG2) in a mousekuman chimeric heavy chain. This single residue substitution leads to the production of a homogeneous antibody. Further, the variant IgG4 has significantly extended serum half-life and shows an improved tissue distribution compared to the original chimeric lgG4.
to mouse or rat monoclonai antibodies and to chime& antibodies of oU1crhuman isotypes in a number of studies (Bruggemann er cl!., 1988; Bindon et al., 1988; Gamed et a/.,1989; Hutzcll et of., 1991; Marchitto er al., 1989; Michaclscn et al., 1991; Priius er al., lY% Riechmann er al., 1988; Tan et al., 1990; Sbcarman et (11..1YYO), and 13 ‘1 labehed cB723(IgG4) (cB72.3(74)) is being evaluated i11clinical studies (Baker er al., 1991; Khaxaeli et al.. 1991, Meredith et al., 1992).
INTRODUCTION IgG4 differs from the other hutnnn IgG isotypes in thal upon SDSPAGE under non-reducing conditiotis it shows lwo protein species, the major species being tctrameric IgG (H2L2) of M, approx. 150 kDa and the
second,
miuor species, having a M, of approx. 80
kDa. Charactcrizatio11 of Uiis 80 kDa material indicates that it is a
“half i11unuooglobulin”containing si11gleheavy and light chains of the cxpcc~cd molecular weight (Kiug er al., lY92). The fact that this molccute is only seen under denaturing, non-reducing. co11ditious suggests that its presence reflects hetcrogcncity of disulphidc btmd formatio11 betwee the two heavy chains in the hinge region. III native solutions the fraction of the antibody which is not fully disulphide bridged is held together as a tetramer by noncovalcnt iutcractions and cannot be scparatcd from the major spccics by gel filtration. However the 80 kDa species is not an artifact
of electrophoresis
since high resolution
TABLE 1 HINGE AMINO ACID SEQUENCES IgGl IgG2 IgG3
ion-exchange
IgG4
chromatography can effect its partial rcsolulion from the 150 kDa spccics (S.A., unpublished observations). These observations have txm IIOIIXI by ourselves (Colcher n al., 1989; King ef al., 1992). a11d omcrs (Bruggemann er al., 1988) where human lgG4 antibod1cs, rodent/human chimeric (cIgG) antibodies of IgG4 isotypc, aud rcco1nbinant hybrid antibodies with IgG4 hinge scqucuccs have bceu examined (Morrison et ul., 1988; Tan et al., 1YY)o; CanIicld and Morrison, 1991). It is bclicvcd that the HL spccics is a major intermcdiatc in the a\scmbly of lgG4. The assembly of lgG4 has been shown to be a much slower process than that of IgGl (Dorrington. 1978) and may rcflcct a reduced capability for inter-heavy chain disulphide britfgc ftumation. An u11derstanding of the consequences of this physical hcterugcneily of IgG4 is important as it may affect the dccisiou to USCU1cIgG4 isutypc for the constructio11of cIgG and II~II~~I~IX~~ antibodies for Uierapeutic applications. Thcrc ;1rcpotc11tiaJadvantages iu the use of IgG4 when chimeric autibodic~ arc being used as delivery agents. for example in ratlrl,ir111nul1~,Uher;lpy (RAIT) of tumours, or as a carrier of drugs LO tumo11r~;I&lgG4 biuds less well as a mo11omerto the high afBnity I? rcccptor (FcRI) loused OII monocytcs compared to IgGl and is less ahlc to recruit cffcctor cells into ADCC (Canfield and Monixott, IYYI), ml IS uuablc to aclivatc the co1nplcmcnt cascade bccausc of poor binding of Clq (Bruggcmann et a/., 1988, Gatred PI ul.. IYXY). Thcrc arc no know11 allulypic variants of IgG4 (Gor11ut11 :uid Clark, IYYO).clgG4 antibodies bavc been compared
EAPKSC--DKTHT-____--_______-ER-KCCVE______-_--____-----__
241 cppcp cppt-.p
EL-KTPLGDTTHT(C~RC~EPKSCDT~~P)~CPRCP ES-KYGPP----______-______---cpgcp
Nmnbcri11gaccording to Kabat et al., (1987).
TablC 1 shows Ihe sequences of the human gamma chain hinge regions as Ueduccd frum the gent sequences. The core hinge regions of the IgGs (CPXCP) are similar. The IgG3 hinge reflects exon duplication and is considered to adopt a different conformation to that of U1cother IgGs. WC postulated that the presence of the serine at position 241 rather than the proline seen in IgGl and IgG2 was inducing an alternative structure in the hinge and so preventing correct disulphidc bridge formation in a proportion of tbc IgG4 protein. Wb show in this paper that a serinc to prohnc substitution at 241 (S241P)of cB72.3(@) (Whittle et al., 1987) does lcad to the production of a homogeneous form of the antibody. WC then show ma1 U1cpropcrtics of this antibody are similar to but not identical witb (hat of tbc original cIgG4. M ATEHIALS AND METHODS Ohgonucleotide directed mutagenesis (Kmmcr er al., 1984) of the codon for setinc 241 (TCA) to proline (CCA) of the IgG4 constant rcgio11bcqucncc (amino acid numbering as in Kabat ef al., 1987) was carried out a11dthe mutant transformants propagated following Utc protocols provided by Anglian Biotechnology Ltd. 105
of the the dctcsucdbinge sequcrtce was done using the S~~CW$CjtrU&Ko~ ~~~~~~~~~~. Tra&&ion by cukaryotie expression vectors of the Chinese
Co~13~U~j
tiamsIcr ovary (CIiO) ccl1 lint, CHO-ctl8, which sccrctcs tbc ~872.3 Iigbt chain (Colcber ef ul., 1Y8Y)was done by the calcium pbosphatc precipitation procedure (Bcbbington, 1991). Chimeric antibodies were purified frwroiii CHQ cc11 culture su~)a~~~ by affinity eli~~inatogmpl~yusing Protein-A Scpharosc (Colchcr el ul., IYXY).Chin&c antibody was BnaIIy dialyzed into an approprialc buffer for intcndcd USCc.g. phosphate buffcrcd sahnc (PBS) for animd studies. aud conccnuatcd by uhraIiltration. Purity and assembly oCthe chin&c antibody was tcstcd by reducing and non-reducing SDS-PAGE fLacmmIi, 19X), Relative afFiirity fur antigen was assessed in an ELLSA format c~n~titioli assay (Lyons et ui., iYY0) using cB72.3(~~~ radish pcruxidasc (HRPO) cor~uga~ as a sIandard in the presence of antigen (bovine 5ubma~iI1,~ mucin (Sigma) prcscnted on a solid phase. T Antibodies were fabclted wiUI I’? using tbc chior~inc lcchniquc (Hunter and Greenwood, 1YG2)and with i 1IIu via tbc macrocyclc (a funcIionaIiscd dcrivalivc of 1,4,7triazaeyclononanc uiacctic acid) wbicb bad been altachcdraudomly to 2-iminothiolanc treated antibodies (Pbipps ei at., unpublisbcd). Groups of four fcmaIc nude mice bearin SU~U~COUS 2-3 week old xcnografts of the human coloreclal carcinoma cell Iinc tS174T on tbc Bank wcrc injected into tbe tail vein with (740-925 kBeyucrels of I*51 iabeficd antibody. Groups of animais wcrc sacrificed at 24,48 and I68 h for IIICc&lee&n of tissue samples whid wcrc weigh&, dissofvcrl in 7M KOH ad cwntd iti a LKB Model I270 gamma comdc~: Tissue uptake was caIeufatuI as me mean pcrccntagc i!tjcctccJtkrx per gram 0r lissuc(%i.dJg). YN3
KESUCI’S The cB723(y4) hc;rvy chain gcnc (WhittIc et al,, 1987) was used as the target scqucttcc for oligolluclcotide-dirccleJ Inutag~Rcsis (Krama ef of., 1984) of the codon for scrinc 241 (TCA) to prolinc (CCA). A putative mutalil clone, mjA155, was s~qucw.xd 10 confinn the prcscncc of Uic muuuion. Tbc muIan gcnc was tcnncd IgG4P and cB72.3 antibody conuuuing its I.ransIalion producl was Icnncd c3?23@4Pf. The IgG4P gene was ~~~j~~~~~ iatu llhc &&I attd Bcil silts of p~~liCMVg~~ (S~cpttcctsand Cockctt, 1989) 10 give pMB~#l.
This plasmid dirfcrs from the cxprcssion vector pJAY6 which contains the ulun~iried cB72.3(@) gcnc (WhitlIe ef al., 1987) only at Ihe single S24 1P ntrtino acid subslilution. Ccl1 fines wcrc
dcvclopcd which secreted rccom~na~~& cB72.3(@P) by rctransfection of the cbimcric B72.3 Iight &in sccrc&g CHO c& liilc, et-18 (Colchcr eraf., 1989)). Chimwic antibody was purified from the smbk ccl1 hncs by Protein A puritication. Yield of antibody was in the range 20 f3 mg/L. Figure 1 shows tbc results of a non-reducing SDS-PAGE am&is of eB723(1yIP) compared lo the inurinc B72.3 and ~372.3 (yl, y2, y3 and y4) antibodies (WhittIe er ai., 1987; Phipps er al., unpublished). The cB72.3(y4) antibody in lane 6 clearIy shows tbc 80 kDa material whcrcas in the cB72.3(y4P) antibody shown in fanc 7 tbc 80 kDa material is absent. The 80 kDa band is not seen in the otbcr ~372.3 antibodies &mcs 3-S) nor the murinc B72.3 (Ianc 2.X Tbc cB72.3@4P) antibody ciutcs from a ~u~n1 GF-250 HPLC column in Phosphate buffer (0.2M) at me same point as tbc o&r IgGs, with an cfution time of 8.5 mins. On HPLC the cBX?.3(y4) antibody also shows a single peak of fufly assembled IgG, the 80 IcDa material not being seen under native conditions (data not shown). I1 is rcasoncd mat non covalent interactions arc sufficient to bold Ihc molcculc togctbcr as a teuamcr. The antibodies were shown to be functional by competition for antigen against cB72.3(y4) antibody in an ELISA format assay. Each of tbc cbiicric antibodies and murinc 872.3 antibody were cquahy able to compctc for antigen suggesting that Ihe avidity of rhc antibodies was idcnlicsl ( Figure 2). Each ehicric antibody was mdi~I~iied with i 251or 1f lln aad injcetcd inlo the tail vein of nude mice bearing tbc human coIorecIaI carcinoma ceil Iinc LSI74T. This c&I fine, when grown as a s&id tumour xcnograft in nude mice, cxprcsses the antigen TAG-72 rccogniLcd by B72.3. Radio-labeilcd mminc B72.3 and cB72.3 autibodics can IocaIisc to such grafted tumours after injection. This animal m&l system has been dcscribcd in dctaii by Colcber et al., (1984). Figure 3A shows tbc bi~is~bu~on of tbc t 251lab&cd chiicric antibodies at 2448 and 168 hours post-injection. The cB72.3(yy1P) antibody appears to be rcmincd in mc circulation for longer and Urc Icvcl of antibody in tissues and on the tumour is higher than with the cIgG4. This is more pronounced at 168 hours where f0.fM3.25 %i.dfg is Iocatcd on tbc tumonr for the cB72.3cyulP) antibody compared to 2.36kLfl %i.dJg for the cB72.3(y4). Figure 33 shows the bio~s~bution of 111Xn1abcBcd cbiicric an~bodics a& 24, 48 and 168 h. As for the I25I Iabeilcd components tbc l 1IIn fahcUcdcB72.3Q4P) results in higher
94kDa
67kDa- m
.i
10
1
~T~ODY
567 12 34 FlGUHE 1 Non-reducing SDS-PAGE analysis of chimeric B72.3 antibodies. Lane 1, mokxuhu weight markers; Lane 2 rnuriue B72.3; Lane 3, cB72.3QI); Lane 4, cB72.3fy2); Lane 5, cB72.3fy3); Lane 6, cB72.3(@); Lane 7, eB72.3~y4P)
loo
1000
(u&nL)
FIGURE 2 Competition betweencB72.3(y4)-HRPO
and mB72.3,
cB723@4) or cB?2_3~~4P~ for binding to antigen, 0, mB72.3; B, ~372.3~~); A, cB72.3~~4P)
IgG4 antibody
168h
24h
FIGURE
107
3
Biodistribution
of cB72.3fy4)
and cB72.3fy4P)
in nude mice with Ls 174T xenografts.
Distribution of cB72.3($)(solid bars) amI cB72Xy4P)(hatcbul bars) Iabcllcd with 1251(upper set of Iiguros) or ’ ’ ‘In (lowerSCL of ligurcs) in tissues at 24,48 and 168h post-injection.
lumour Icvcls. The major diffcrcncc bclwccn the two antibodies is UIC
lcvcl of antibody rctaincd by the kidney. At all of the time poinls UIC kidney lcvcls arc higher for cB72.3(+) such that at 168 II the level is 11.62fl.15 %i.d./g cornparcd to 3.41rt0.56 %i.d./g for cB723fy4P). This results in a sjgtlificantly improved turnourlkidney ratio for the cB72.3(y4P). Tlms a single amino acid nltcration in the hinge region appears to IN! suIIicicr~1 to afb.xl the biological half-life unnour loading and kidney rctcntion of a chimcric antibody . DISCUSSION Tbc lgG4 hinge Iins SC.& at position 241 instead of the prolinc found al that position in IgGl and lgG2. All published scqucnccs of IgG4 deduced from DNA scquc~~cespredict a scrinc at 241 (Ellison er of., 1981; Takahashi el uf., tY82; Ellison and Hood, IYX2) and all mycioma proteins cxamincd to date bavc, in our hands, shown a band at around 80 kDa on non-reducing SDSPAGE (Ring ef al, 1992).Thcrc is howcvcr an carticr report of the parlial amino acid X~UCIIW ol Ihc heavy chain COIIS~~II~region of a human IgG4 inycloma protein which predicts a prolinc at position 241 (Pink ef al., IY67, lY70). The scqucnce of the rcbvant pcptidc from the hinge region was dcduccd from Utc amino acid composition of Utc isolated pcptidc and by comparison with the scqucncc of the IgGl hinge (Frangionc and Milstcin, lY67). The .amino acid composition of UIC IgG4 hinge obtained by Pink ef al., (IY70) is however also corlsiskxl with UIC gcnctic dam if a u~tttsp~)sitiuti of Prolinc 241 au0 Scrinc 245 of the scqucncc dcduccd by Pink ef al., (lY70) is niadc. Thcrcforc it would stem Iikcly UIiIt in fact there is only one scqucncc for UIC igG4 hinge region and which conlahs a scrinc al 241. TIE first indication of an alteration in UK propcrtics of t.hc cB72.3(y4P) was lhc climinalion of the 80 kDa “halfitttrrtur~r~glubulilt”iiorn~alIy seen OIItiott-rcducirig SDS-PAGE. It .sccms likely that prolinc 241 is rcquircd to in&m an uptimal structure within the IIiIIgc hr UIC iutcr-chaiu bonding of the cyslcinc sulphydryls in IgGl and lyG2. The prcscncc of scrinc at 24 1 in IgG4 may allow murc main chain flcxibilily chaII proline mill inay pcmIi1 other suuclurcs to form, for example iutrachain disulplii~ bonds, or Urcoxidation of cystcincs which wiI1 prevent Utc formation of the correct disulphidc bonds. Of more practical
importance arc the observations that the hinge modification in the cB72.3(y4P) protcin affects the in vivo biodistribution of the antibody comparcd to cB72.3(T4) (Figure 3A). The biodiitribution profile of the 1251-IabcllcdcB72.3(flP) more closely rcscmbics that seen previously for cB72.3(71) than that of cB72.3(*) (Hutzcll ef al., 1991). However of particular importance is the difcrcncc in the kidney uptake of 11‘In- cB72.3(y4P) compared to 11‘III- cB72.3(19) (Figure 3B). a difference not seen when the ~tti~i~ arc lab&cd with *251. 11 is suggested that the nondisulphidc bridged form of the cB72.3(-@) is more easily broken down and tbc fragments arc clcarcd via the kidney. Due to dchalogcnation this kidney uptake is not noticcablc when the antibodies are 1251-labeIlcd but is noticcablc when the antibodies arc Iabcilcd with a metallic isotope. This diffcrcncc has hnpli~tions for the use of the natural IgG4 isotypc in clinical studies when iabcllcd with a mctaIIic isotope as in vivo USC would bc limited by Utc kidney toxicily. It is probable that the obscrvcd biological propcrtics of native human IgG4 and the recombinant chimcric IgG4 antibodies arc a consqucncc of the prcscncc of a mixture of bridged and n~b~dg~ material. We have shown hcrc that fully cross-Iinkcd IgG4 has propcrtics mat differ from the natural mixture. Recently Gilics and WcsoIowski, (1991) have examined the propcrtics of a cIgC1 antibody in which the hinge cystcincs have been mutaicd to scrincs. While forming HzL2 h solution, the antibody was found to bc pmdo~~Uy of the HL form on non-reducing SDS-PAGE gels. This mutant antibody had sIighUy reduced avidity for antigen, was unable to effect ADCC and showed rcduccd ability to interact with complcmcnt to cffcct CDCcomparcd to the normal chimcric IgG I. It is clear thercfore that the correct formaton of disulphidc bridges in the hinge can have ~por~t conscqucnccs on Utc biological propertics of the JgGs. WCarc currently comparing cB72.3(flP) and cB72.3(74) in smdics to dctcrminc whether the hinge alteration has affected the Fc functions of the antibody (Clq binding, interactions with FcRl and FcRII) to further chatactcrlzc this antibody and to dcterminc whcthcr this altered constant region would bc a nscful alternative to the naturally occuring isotypcs of IgG in clinical situations.
A~KNOWL~~~~MENT~ V& thank J. BODEN and R. BODEN I’or help with the animal studies ad K. MILLAR and B. BOYCE for Urn 9N3 macrocycle REFERENCES Baker. T S.. Begent. R. H. J.. Dewji, M. R., Conlan, J., Secher, D. S.. (lQ91). Characterization of the antibody response in patients undergoing radioimmuntherapy with chimcric B72.3. Anribody Iwtuuwconj. Rodiophorru., 4 , 799-809. Hcbhington. nonlymphoid
C. R., (1991). Expression of’ antibody mammalian cells. Mcfhods. 2, 136-14.5
genes
in
Hruggcmann, M.. Williams, G. T., Bindon. C.I., Clark, M. R., #lker, M. R., Jcfferis. R.. Waldmann, H.. Neuberger, M. S., (1988). Comparison of the effector functions of human immunoglobulins using a matched set of chimeric antibodies. 1. tip. Ned., 166. 1351-1361 Bindon, C. I., Hale, G.. Bntggemann. M., Waldmann, H., (1988). Human monoclonal IgG isotypcs differ in complement activating function at the level of C4 as well as CIq. J. Exp. Med.. 168, 127142 Canficld. S., Morrison S. L.. (1991). The binding affinity of human IgG for its high affinity Fc receptor is determined by multiple amino acids in the CH2 domain and is mediated by the hinge region. J. Exp. Med., 173. 1483-1491 Colchet, D., Keenan, A. M.. Larson. S. M., Schlom, J., (1984). Prolonged binding of a radiolabeled monoclonal antibody (372.3) used for the in sirrt radio~i~unode~ction of human colon carcinoma xenografts. Cancer Res., 4 4, 5744.5751 C&her. D., Milenic. D., Roselli, M.. Raubitshek, A., Yarranton, G.. King. D.. Adair. J.. Whittle. N.. Rodmer. M.. Schlom, J., (IQ89). Characrcrization and biodistribution of recombinant and recombinant/chimcric constructs of the monoclonal antibody B72.3 Cancer Res.. 4 9, 173X-1745
Khzaeli, M. B., Salch, M. N., Liu, T. P., Meredith, R. E. Wheeler, R. H.. Baker. T. S., King, D., Secher. D., Allen, L., Rogers. K., Colchcr, D., Scblom, J., Shuchat. T, D., LoBugIio. A. F.. (1991). Pharmacokinetics and immune response of 13lI-chime& mouse/human 13’72.3 (human gamma 4) monoclonal antibody in man. Cancer Rcs.. 5 1. 5461-5466 King. D. 1.. Adair, J. R.. Angal. S.. Low. D. C.. Proudfoot. K. A.. Lloyd. J. C.. Bodmcr, M.. Yarranton. G, T., (1992). Expression, purification and characterisation of a mouse:human chimeric antibody and cbimcric Fab’ fragment. Biocltern. J.. 2 8 1, 3 I7-323
Kramer. W., Drulsa, V.. Jansen, H.-W., Kramer, B., Pflugfclder, M.. Fritz. H. 1.. (1984). ‘I%e gapped duplex approach to oligonucicotidcdirected mutation construction. Nucf. Acids Rex., 1 2, 9441-9456 Lyons. A., King. D. J., Owens, R. J.. Yarranlon, G. T., Millican. A., Whittle, N. R., Adair, J. R., (1990). Site-specific attachment to recombinant antibodies via introduced surface cystcine r&dues. Prof. En&, 3, 703-708
Marchitto, K. S.. Kindsvogel, W. R., Bcaumier. P. L., Fine, S. K.. Gilbert, T., Levine, S. D.. Woodhouse. C. S., Morgan, A. C. Jr.. (1989). Characterization of a human-mouse chimcric antibody reactive with a human melanoma associated antigen. ful,,trtniry to cwlcer, 2 1. 101-105 Meredith, R. I?, Khazaeli. M. B., Plott. W. E., Saleh, M. N., Liu. T, Allen. L. F.. Russell, C. D., Orr. R. A., Colchcr, D., Schlom. J., Shochat. D., Wheeler, R. H., LoBuglio. A. F.. (1992). Phase I trial of Iodine-131-chimeric B72.3 (Human IgG4) in metastatic cclorectal cancer. J. Nucl. Med., 3 3, 23-29 Michaclsen, T. E., Garred, P., Aase, A., (1991). Human IgG subclass pattern of inducing complet~cnt-mediated cytolysis dcpcnds on antigen concentration and to a lesser extent on epitope patchiness, antibody affinity and complement concentration. Eur: J. It~rmunol., 2 1, 11-16
1)orrinpton.K. I.. (1978). The structural basis for the functional diversity of immunoglobulin G. Con. 1. Biochew, 5 6. 1087-1101
Morrison. S. L.. Canfield, S., Porter. S.. Tan, 1.. K.. l?~o, M. Wims. L. A.. (1988). Production and characterization or gcnctically cnginecrcd antibody molecules. Clin. C/wn.. 3 4. 1668- 1675
J., Uuxbaum, J.. Hood, L.. (1981). Nucleotide sequence of a human immunoglobulin Cy4 gene. DNA, 1, 11-18
Pink, J. R .L., Milstcin. C., (1967). Disulphidc bridges of a human immunoglobulin G protein. Nanwe, 2 16, 941-942
Ellisoa. J. W., Berson. B. 1.. Hood, L. E., (1982). The nudeotide
Pink, J. R .L., Buttuy, S. H., DcVries. G. M.. Milstein, C., (1970). Human immunoglobulin subclass. Partiai amino acid sequence of the constant region of a y4 chain. Uioslretn 1.. I 17,33-47
Ellisou.
sequence of a human i~unoglobuIi[l 1 0, 407 I-4079.
Cyl gene. Nucf. Acids Res.,
Ellison. J. W., Hood. I_ E.. (1982). Linkage and squence homology of two i~unoglobulin y heavy chain constant region genes. Proc. Nail. Acod. Sci. USA. 7 9. 1984-1988
Primus, F. J., Pendurthi, T. K., Hutzcll. P.. Kashmiri, S., Slavin, D. C., Callahan, R., Schlom, J., (f990). Chimcric B72.3 inouse~uman (IgGl) antibody directs the lysis of tumour cells by lymphokineactivated killer cells. Cancer hnunol. Irrrntunorher., 3 1. 349-357
Frangione, 8.. Milstein. C., (1967). Disulphidc bridges immunoglobulin G-l heavy chain. Nafurc. 2 16, 939-941
Riechmann, I,.. Clark. M., Waldmann. H.. Winter. C., (1988). Reshaping human antibodies for therapy. Nowe, 3 3 2, 323-327
of
Garrcd. P.. Michaelsen. T E.. Aase. A., (1989). The IgG subclass pattern of complement activation depends on epitope density and antibody and complement concenuation. &and. 1. In~munof., 3 0, 319.382 Gillies. S. D.. Wesolowski. 1. S., (1991). Antigen binding and biological activities of engineered mutant chimeric Abs with human tumor specificities. Hum. Anribod. Hybridornas, 1. 47-54 German. S. D., Clark, M. R., (19901. HuI~aIlisatiotl of lnono~lonal antibodies ior therapy. Scirrninnrs in ~~~~~;funo~ogy, 2, 457-466 Hunter, W. M.. Greenwood. F. C., (1962). Preparation ol’ iodine-13t labelled human growth hormone of high specific activity. Nature, 1 9 4. 495.496 Hut& P.. Kashmiri. S.. Colcher, D.. Primus. E 1.. Horan Hand, P., Rosclli. M., Finch, M.. Yarranton, G., Bodmer. M.. WhitUc. N.. King. D.. Loullis, C. C., McCoy. D. W., Callahan, R.. Schlom, J., (1991). Generation and characterization of a recombinant/chimeric W72.3 (human ~1). CancerRes., 5 I, 181-189 Kabat, E. A., Wu. T. T., Reid-Miller, M., Perry, H. M., Gottesman, KS., (1987). Sequences of pro&ins of imrnunologicai inrcresr. 4th edition. Washington DC: United States Department of He&b and Human Services
Shcarman, C. W, Kanzy, E. J., Lawric, D. K.. Li, Y.-W.. Thammana, I?. Moore. G. P., Kurrle. R.. (1990). Construction, expression, and biological activity of murinc/human chimcric antibodies with specilicity for the human a/J3 T ccl1 receptor. J. I~wutnol.. I4 6, 928-935 Stephens, P. E., Cockett, M. I., (1989). The construction of a highly efficient versatile set of mammalian expression vectors. Nuci. Acids Rer., 1 7, 7110 ‘Ikn. L. K.. Shapes, R. J., Oi, V. T., Morrison. S. L., (1990). Influence of the binge region on complcmcnt activation, Clq binding, and segmental flexibility in chime& human immunoglobulins. Proc. Nail. Acad. Sci. USA, 8 7, 162-166 lkkahashi, N., Ueda, S., Obata. M.. Nikaido. T., Nakai, S.. Honjo, T. (1982). Structure of human immunoglobulin gamma genes. Implications for evolulion of a gent family. Cell. 2 9. 671-679 Wbittlc, N., Adair, J., Lloyd, C., Jenkins, L., Dcvinc, J.. Scblom, J.. Raubitshek, A., Colcher, D., Bodmcr. M.. (1987). Expression in COS -11~ of a mousci%uman chimaoric B72.3 anribody. Pror. &g.. I, 499-50s
SHORT
A SINGLE
0161-5890/93 $5.00 + 0.00 0 1992 Pergamon Press Ltd
COMMUNICATION
AMINO ACID SUBSTITUTION ABOLISHES THE HETEROGENEITY CHIMERIC MOUSE/HUMAN (IgG4) ANTIBODY
OF
S. ANGAL, D. J. KING, M. W. BODMER. A. TURNER, A. D. G. LAWSON, G. ROBERTS, B. PEDLEY’ and J. R. ADAIR* Celltech
Ltd.. 216 Bath Rd.. Slough,
Berks.. SLl 4EN U.K. and lDepartment
of Clinical
Oncology,
Royal
Free Hospital.
London
U.K.
*Author for correspondence
(First received 28 July 1992; accepted in revised form 8 September 1992)
Abstract - Human immunoglobulin G4 (IgG4) exists in two molecular forms due to the heterogeneity of the inter-heavy chain disulphide bridges in the hinge region in a proportion of secreted human IgG4. This heterogeneity is only revealed under denaturing, non-reducing conditions in which an HL “half antibody” is detected, a phenomenon not seen in other human IgG isotypes. In native conditions noncovalent
interactions
hold
the antibody
together
as the H2I2
tetramer.
Analysis
of the hinge
sequences of human IgG heavy chains
suggested that the presence of serine at residue 241 might be the cause of this heterogeneity. We therefore changed the wine at 241 to proline (found at that position in lgG1 and IgG2) in a mousekuman chimeric heavy chain. This single residue substitution leads to the production of a homogeneous antibody. Further, the variant IgG4 has significantly extended serum half-life and shows an improved tissue distribution compared to the original chimeric lgG4.
to mouse or rat monoclonai antibodies and to chime& antibodies of oU1crhuman isotypes in a number of studies (Bruggemann er cl!., 1988; Bindon et al., 1988; Gamed et a/.,1989; Hutzcll et of., 1991; Marchitto er al., 1989; Michaclscn et al., 1991; Priius er al., lY% Riechmann er al., 1988; Tan et al., 1990; Sbcarman et (11..1YYO), and 13 ‘1 labehed cB723(IgG4) (cB72.3(74)) is being evaluated i11clinical studies (Baker er al., 1991; Khaxaeli et al.. 1991, Meredith et al., 1992).
INTRODUCTION IgG4 differs from the other hutnnn IgG isotypes in thal upon SDSPAGE under non-reducing conditiotis it shows lwo protein species, the major species being tctrameric IgG (H2L2) of M, approx. 150 kDa and the
second,
miuor species, having a M, of approx. 80
kDa. Charactcrizatio11 of Uiis 80 kDa material indicates that it is a
“half i11unuooglobulin”containing si11gleheavy and light chains of the cxpcc~cd molecular weight (Kiug er al., lY92). The fact that this molccute is only seen under denaturing, non-reducing. co11ditious suggests that its presence reflects hetcrogcncity of disulphidc btmd formatio11 betwee the two heavy chains in the hinge region. III native solutions the fraction of the antibody which is not fully disulphide bridged is held together as a tetramer by noncovalcnt iutcractions and cannot be scparatcd from the major spccics by gel filtration. However the 80 kDa species is not an artifact
of electrophoresis
since high resolution
TABLE 1 HINGE AMINO ACID SEQUENCES IgGl IgG2 IgG3
ion-exchange
IgG4
chromatography can effect its partial rcsolulion from the 150 kDa spccics (S.A., unpublished observations). These observations have txm IIOIIXI by ourselves (Colcher n al., 1989; King ef al., 1992). a11d omcrs (Bruggemann er al., 1988) where human lgG4 antibod1cs, rodent/human chimeric (cIgG) antibodies of IgG4 isotypc, aud rcco1nbinant hybrid antibodies with IgG4 hinge scqucuccs have bceu examined (Morrison et ul., 1988; Tan et al., 1YY)o; CanIicld and Morrison, 1991). It is bclicvcd that the HL spccics is a major intermcdiatc in the a\scmbly of lgG4. The assembly of lgG4 has been shown to be a much slower process than that of IgGl (Dorrington. 1978) and may rcflcct a reduced capability for inter-heavy chain disulphide britfgc ftumation. An u11derstanding of the consequences of this physical hcterugcneily of IgG4 is important as it may affect the dccisiou to USCU1cIgG4 isutypc for the constructio11of cIgG and II~II~~I~IX~~ antibodies for Uierapeutic applications. Thcrc ;1rcpotc11tiaJadvantages iu the use of IgG4 when chimeric autibodic~ arc being used as delivery agents. for example in ratlrl,ir111nul1~,Uher;lpy (RAIT) of tumours, or as a carrier of drugs LO tumo11r~;I&lgG4 biuds less well as a mo11omerto the high afBnity I? rcccptor (FcRI) loused OII monocytcs compared to IgGl and is less ahlc to recruit cffcctor cells into ADCC (Canfield and Monixott, IYYI), ml IS uuablc to aclivatc the co1nplcmcnt cascade bccausc of poor binding of Clq (Bruggcmann et a/., 1988, Gatred PI ul.. IYXY). Thcrc arc no know11 allulypic variants of IgG4 (Gor11ut11 :uid Clark, IYYO).clgG4 antibodies bavc been compared
EAPKSC--DKTHT-____--_______-ER-KCCVE______-_--____-----__
241 cppcp cppt-.p
EL-KTPLGDTTHT(C~RC~EPKSCDT~~P)~CPRCP ES-KYGPP----______-______---cpgcp
Nmnbcri11gaccording to Kabat et al., (1987).
TablC 1 shows Ihe sequences of the human gamma chain hinge regions as Ueduccd frum the gent sequences. The core hinge regions of the IgGs (CPXCP) are similar. The IgG3 hinge reflects exon duplication and is considered to adopt a different conformation to that of U1cother IgGs. WC postulated that the presence of the serine at position 241 rather than the proline seen in IgGl and IgG2 was inducing an alternative structure in the hinge and so preventing correct disulphidc bridge formation in a proportion of tbc IgG4 protein. Wb show in this paper that a serinc to prohnc substitution at 241 (S241P)of cB72.3(@) (Whittle et al., 1987) does lcad to the production of a homogeneous form of the antibody. WC then show ma1 U1cpropcrtics of this antibody are similar to but not identical witb (hat of tbc original cIgG4. M ATEHIALS AND METHODS Ohgonucleotide directed mutagenesis (Kmmcr er al., 1984) of the codon for setinc 241 (TCA) to proline (CCA) of the IgG4 constant rcgio11bcqucncc (amino acid numbering as in Kabat ef al., 1987) was carried out a11dthe mutant transformants propagated following Utc protocols provided by Anglian Biotechnology Ltd. 105
of the the dctcsucdbinge sequcrtce was done using the S~~CW$CjtrU&Ko~ ~~~~~~~~~~. Tra&&ion by cukaryotie expression vectors of the Chinese
Co~13~U~j
tiamsIcr ovary (CIiO) ccl1 lint, CHO-ctl8, which sccrctcs tbc ~872.3 Iigbt chain (Colcber ef ul., 1Y8Y)was done by the calcium pbosphatc precipitation procedure (Bcbbington, 1991). Chimeric antibodies were purified frwroiii CHQ cc11 culture su~)a~~~ by affinity eli~~inatogmpl~yusing Protein-A Scpharosc (Colchcr el ul., IYXY).Chin&c antibody was BnaIIy dialyzed into an approprialc buffer for intcndcd USCc.g. phosphate buffcrcd sahnc (PBS) for animd studies. aud conccnuatcd by uhraIiltration. Purity and assembly oCthe chin&c antibody was tcstcd by reducing and non-reducing SDS-PAGE fLacmmIi, 19X), Relative afFiirity fur antigen was assessed in an ELLSA format c~n~titioli assay (Lyons et ui., iYY0) using cB72.3(~~~ radish pcruxidasc (HRPO) cor~uga~ as a sIandard in the presence of antigen (bovine 5ubma~iI1,~ mucin (Sigma) prcscnted on a solid phase. T Antibodies were fabclted wiUI I’? using tbc chior~inc lcchniquc (Hunter and Greenwood, 1YG2)and with i 1IIu via tbc macrocyclc (a funcIionaIiscd dcrivalivc of 1,4,7triazaeyclononanc uiacctic acid) wbicb bad been altachcdraudomly to 2-iminothiolanc treated antibodies (Pbipps ei at., unpublisbcd). Groups of four fcmaIc nude mice bearin SU~U~COUS 2-3 week old xcnografts of the human coloreclal carcinoma cell Iinc tS174T on tbc Bank wcrc injected into tbe tail vein with (740-925 kBeyucrels of I*51 iabeficd antibody. Groups of animais wcrc sacrificed at 24,48 and I68 h for IIICc&lee&n of tissue samples whid wcrc weigh&, dissofvcrl in 7M KOH ad cwntd iti a LKB Model I270 gamma comdc~: Tissue uptake was caIeufatuI as me mean pcrccntagc i!tjcctccJtkrx per gram 0r lissuc(%i.dJg). YN3
KESUCI’S The cB723(y4) hc;rvy chain gcnc (WhittIc et al,, 1987) was used as the target scqucttcc for oligolluclcotide-dirccleJ Inutag~Rcsis (Krama ef of., 1984) of the codon for scrinc 241 (TCA) to prolinc (CCA). A putative mutalil clone, mjA155, was s~qucw.xd 10 confinn the prcscncc of Uic muuuion. Tbc muIan gcnc was tcnncd IgG4P and cB72.3 antibody conuuuing its I.ransIalion producl was Icnncd c3?23@4Pf. The IgG4P gene was ~~~j~~~~~ iatu llhc &&I attd Bcil silts of p~~liCMVg~~ (S~cpttcctsand Cockctt, 1989) 10 give pMB~#l.
This plasmid dirfcrs from the cxprcssion vector pJAY6 which contains the ulun~iried cB72.3(@) gcnc (WhitlIe ef al., 1987) only at Ihe single S24 1P ntrtino acid subslilution. Ccl1 fines wcrc
dcvclopcd which secreted rccom~na~~& cB72.3(@P) by rctransfection of the cbimcric B72.3 Iight &in sccrc&g CHO c& liilc, et-18 (Colchcr eraf., 1989)). Chimwic antibody was purified from the smbk ccl1 hncs by Protein A puritication. Yield of antibody was in the range 20 f3 mg/L. Figure 1 shows tbc results of a non-reducing SDS-PAGE am&is of eB723(1yIP) compared lo the inurinc B72.3 and ~372.3 (yl, y2, y3 and y4) antibodies (WhittIe er ai., 1987; Phipps er al., unpublished). The cB72.3(y4) antibody in lane 6 clearIy shows tbc 80 kDa material whcrcas in the cB72.3(y4P) antibody shown in fanc 7 tbc 80 kDa material is absent. The 80 kDa band is not seen in the otbcr ~372.3 antibodies &mcs 3-S) nor the murinc B72.3 (Ianc 2.X Tbc cB72.3@4P) antibody ciutcs from a ~u~n1 GF-250 HPLC column in Phosphate buffer (0.2M) at me same point as tbc o&r IgGs, with an cfution time of 8.5 mins. On HPLC the cBX?.3(y4) antibody also shows a single peak of fufly assembled IgG, the 80 IcDa material not being seen under native conditions (data not shown). I1 is rcasoncd mat non covalent interactions arc sufficient to bold Ihc molcculc togctbcr as a teuamcr. The antibodies were shown to be functional by competition for antigen against cB72.3(y4) antibody in an ELISA format assay. Each of tbc cbiicric antibodies and murinc 872.3 antibody were cquahy able to compctc for antigen suggesting that Ihe avidity of rhc antibodies was idcnlicsl ( Figure 2). Each ehicric antibody was mdi~I~iied with i 251or 1f lln aad injcetcd inlo the tail vein of nude mice bearing tbc human coIorecIaI carcinoma ceil Iinc LSI74T. This c&I fine, when grown as a s&id tumour xcnograft in nude mice, cxprcsses the antigen TAG-72 rccogniLcd by B72.3. Radio-labeilcd mminc B72.3 and cB72.3 autibodics can IocaIisc to such grafted tumours after injection. This animal m&l system has been dcscribcd in dctaii by Colcber et al., (1984). Figure 3A shows tbc bi~is~bu~on of tbc t 251lab&cd chiicric antibodies at 2448 and 168 hours post-injection. The cB72.3(yy1P) antibody appears to be rcmincd in mc circulation for longer and Urc Icvcl of antibody in tissues and on the tumour is higher than with the cIgG4. This is more pronounced at 168 hours where f0.fM3.25 %i.dfg is Iocatcd on tbc tumonr for the cB72.3cyulP) antibody compared to 2.36kLfl %i.dJg for the cB72.3(y4). Figure 33 shows the bio~s~bution of 111Xn1abcBcd cbiicric an~bodics a& 24, 48 and 168 h. As for the I25I Iabeilcd components tbc l 1IIn fahcUcdcB72.3Q4P) results in higher
94kDa
67kDa- m
.i
10
1
~T~ODY
567 12 34 FlGUHE 1 Non-reducing SDS-PAGE analysis of chimeric B72.3 antibodies. Lane 1, mokxuhu weight markers; Lane 2 rnuriue B72.3; Lane 3, cB72.3QI); Lane 4, cB72.3fy2); Lane 5, cB72.3fy3); Lane 6, cB72.3(@); Lane 7, eB72.3~y4P)
loo
1000
(u&nL)
FIGURE 2 Competition betweencB72.3(y4)-HRPO
and mB72.3,
cB723@4) or cB?2_3~~4P~ for binding to antigen, 0, mB72.3; B, ~372.3~~); A, cB72.3~~4P)
IgG4 antibody
168h
24h
FIGURE
107
3
Biodistribution
of cB72.3fy4)
and cB72.3fy4P)
in nude mice with Ls 174T xenografts.
Distribution of cB72.3($)(solid bars) amI cB72Xy4P)(hatcbul bars) Iabcllcd with 1251(upper set of Iiguros) or ’ ’ ‘In (lowerSCL of ligurcs) in tissues at 24,48 and 168h post-injection.
lumour Icvcls. The major diffcrcncc bclwccn the two antibodies is UIC
lcvcl of antibody rctaincd by the kidney. At all of the time poinls UIC kidney lcvcls arc higher for cB72.3(+) such that at 168 II the level is 11.62fl.15 %i.d./g cornparcd to 3.41rt0.56 %i.d./g for cB723fy4P). This results in a sjgtlificantly improved turnourlkidney ratio for the cB72.3(y4P). Tlms a single amino acid nltcration in the hinge region appears to IN! suIIicicr~1 to afb.xl the biological half-life unnour loading and kidney rctcntion of a chimcric antibody . DISCUSSION Tbc lgG4 hinge Iins SC.& at position 241 instead of the prolinc found al that position in IgGl and lgG2. All published scqucnccs of IgG4 deduced from DNA scquc~~cespredict a scrinc at 241 (Ellison er of., 1981; Takahashi el uf., tY82; Ellison and Hood, IYX2) and all mycioma proteins cxamincd to date bavc, in our hands, shown a band at around 80 kDa on non-reducing SDSPAGE (Ring ef al, 1992).Thcrc is howcvcr an carticr report of the parlial amino acid X~UCIIW ol Ihc heavy chain COIIS~~II~region of a human IgG4 inycloma protein which predicts a prolinc at position 241 (Pink ef al., IY67, lY70). The scqucnce of the rcbvant pcptidc from the hinge region was dcduccd from Utc amino acid composition of Utc isolated pcptidc and by comparison with the scqucncc of the IgGl hinge (Frangionc and Milstcin, lY67). The .amino acid composition of UIC IgG4 hinge obtained by Pink ef al., (IY70) is however also corlsiskxl with UIC gcnctic dam if a u~tttsp~)sitiuti of Prolinc 241 au0 Scrinc 245 of the scqucncc dcduccd by Pink ef al., (lY70) is niadc. Thcrcforc it would stem Iikcly UIiIt in fact there is only one scqucncc for UIC igG4 hinge region and which conlahs a scrinc al 241. TIE first indication of an alteration in UK propcrtics of t.hc cB72.3(y4P) was lhc climinalion of the 80 kDa “halfitttrrtur~r~glubulilt”iiorn~alIy seen OIItiott-rcducirig SDS-PAGE. It .sccms likely that prolinc 241 is rcquircd to in&m an uptimal structure within the IIiIIgc hr UIC iutcr-chaiu bonding of the cyslcinc sulphydryls in IgGl and lyG2. The prcscncc of scrinc at 24 1 in IgG4 may allow murc main chain flcxibilily chaII proline mill inay pcmIi1 other suuclurcs to form, for example iutrachain disulplii~ bonds, or Urcoxidation of cystcincs which wiI1 prevent Utc formation of the correct disulphidc bonds. Of more practical
importance arc the observations that the hinge modification in the cB72.3(y4P) protcin affects the in vivo biodistribution of the antibody comparcd to cB72.3(T4) (Figure 3A). The biodiitribution profile of the 1251-IabcllcdcB72.3(flP) more closely rcscmbics that seen previously for cB72.3(71) than that of cB72.3(*) (Hutzcll ef al., 1991). However of particular importance is the difcrcncc in the kidney uptake of 11‘In- cB72.3(y4P) compared to 11‘III- cB72.3(19) (Figure 3B). a difference not seen when the ~tti~i~ arc lab&cd with *251. 11 is suggested that the nondisulphidc bridged form of the cB72.3(-@) is more easily broken down and tbc fragments arc clcarcd via the kidney. Due to dchalogcnation this kidney uptake is not noticcablc when the antibodies are 1251-labeIlcd but is noticcablc when the antibodies arc Iabcilcd with a metallic isotope. This diffcrcncc has hnpli~tions for the use of the natural IgG4 isotypc in clinical studies when iabcllcd with a mctaIIic isotope as in vivo USC would bc limited by Utc kidney toxicily. It is probable that the obscrvcd biological propcrtics of native human IgG4 and the recombinant chimcric IgG4 antibodies arc a consqucncc of the prcscncc of a mixture of bridged and n~b~dg~ material. We have shown hcrc that fully cross-Iinkcd IgG4 has propcrtics mat differ from the natural mixture. Recently Gilics and WcsoIowski, (1991) have examined the propcrtics of a cIgC1 antibody in which the hinge cystcincs have been mutaicd to scrincs. While forming HzL2 h solution, the antibody was found to bc pmdo~~Uy of the HL form on non-reducing SDS-PAGE gels. This mutant antibody had sIighUy reduced avidity for antigen, was unable to effect ADCC and showed rcduccd ability to interact with complcmcnt to cffcct CDCcomparcd to the normal chimcric IgG I. It is clear thercfore that the correct formaton of disulphidc bridges in the hinge can have ~por~t conscqucnccs on Utc biological propertics of the JgGs. WCarc currently comparing cB72.3(flP) and cB72.3(74) in smdics to dctcrminc whether the hinge alteration has affected the Fc functions of the antibody (Clq binding, interactions with FcRl and FcRII) to further chatactcrlzc this antibody and to dcterminc whcthcr this altered constant region would bc a nscful alternative to the naturally occuring isotypcs of IgG in clinical situations.
A~KNOWL~~~~MENT~ V& thank J. BODEN and R. BODEN I’or help with the animal studies ad K. MILLAR and B. BOYCE for Urn 9N3 macrocycle REFERENCES Baker. T S.. Begent. R. H. J.. Dewji, M. R., Conlan, J., Secher, D. S.. (lQ91). Characterization of the antibody response in patients undergoing radioimmuntherapy with chimcric B72.3. Anribody Iwtuuwconj. Rodiophorru., 4 , 799-809. Hcbhington. nonlymphoid
C. R., (1991). Expression of’ antibody mammalian cells. Mcfhods. 2, 136-14.5
genes
in
Hruggcmann, M.. Williams, G. T., Bindon. C.I., Clark, M. R., #lker, M. R., Jcfferis. R.. Waldmann, H.. Neuberger, M. S., (1988). Comparison of the effector functions of human immunoglobulins using a matched set of chimeric antibodies. 1. tip. Ned., 166. 1351-1361 Bindon, C. I., Hale, G.. Bntggemann. M., Waldmann, H., (1988). Human monoclonal IgG isotypcs differ in complement activating function at the level of C4 as well as CIq. J. Exp. Med.. 168, 127142 Canficld. S., Morrison S. L.. (1991). The binding affinity of human IgG for its high affinity Fc receptor is determined by multiple amino acids in the CH2 domain and is mediated by the hinge region. J. Exp. Med., 173. 1483-1491 Colchet, D., Keenan, A. M.. Larson. S. M., Schlom, J., (1984). Prolonged binding of a radiolabeled monoclonal antibody (372.3) used for the in sirrt radio~i~unode~ction of human colon carcinoma xenografts. Cancer Res., 4 4, 5744.5751 C&her. D., Milenic. D., Roselli, M.. Raubitshek, A., Yarranton, G.. King. D.. Adair. J.. Whittle. N.. Rodmer. M.. Schlom, J., (IQ89). Characrcrization and biodistribution of recombinant and recombinant/chimcric constructs of the monoclonal antibody B72.3 Cancer Res.. 4 9, 173X-1745
Khzaeli, M. B., Salch, M. N., Liu, T. P., Meredith, R. E. Wheeler, R. H.. Baker. T. S., King, D., Secher. D., Allen, L., Rogers. K., Colchcr, D., Scblom, J., Shuchat. T, D., LoBugIio. A. F.. (1991). Pharmacokinetics and immune response of 13lI-chime& mouse/human 13’72.3 (human gamma 4) monoclonal antibody in man. Cancer Rcs.. 5 1. 5461-5466 King. D. 1.. Adair, J. R.. Angal. S.. Low. D. C.. Proudfoot. K. A.. Lloyd. J. C.. Bodmcr, M.. Yarranton. G, T., (1992). Expression, purification and characterisation of a mouse:human chimeric antibody and cbimcric Fab’ fragment. Biocltern. J.. 2 8 1, 3 I7-323
Kramer. W., Drulsa, V.. Jansen, H.-W., Kramer, B., Pflugfclder, M.. Fritz. H. 1.. (1984). ‘I%e gapped duplex approach to oligonucicotidcdirected mutation construction. Nucf. Acids Rex., 1 2, 9441-9456 Lyons. A., King. D. J., Owens, R. J.. Yarranlon, G. T., Millican. A., Whittle, N. R., Adair, J. R., (1990). Site-specific attachment to recombinant antibodies via introduced surface cystcine r&dues. Prof. En&, 3, 703-708
Marchitto, K. S.. Kindsvogel, W. R., Bcaumier. P. L., Fine, S. K.. Gilbert, T., Levine, S. D.. Woodhouse. C. S., Morgan, A. C. Jr.. (1989). Characterization of a human-mouse chimcric antibody reactive with a human melanoma associated antigen. ful,,trtniry to cwlcer, 2 1. 101-105 Meredith, R. I?, Khazaeli. M. B., Plott. W. E., Saleh, M. N., Liu. T, Allen. L. F.. Russell, C. D., Orr. R. A., Colchcr, D., Schlom. J., Shochat. D., Wheeler, R. H., LoBuglio. A. F.. (1992). Phase I trial of Iodine-131-chimeric B72.3 (Human IgG4) in metastatic cclorectal cancer. J. Nucl. Med., 3 3, 23-29 Michaclsen, T. E., Garred, P., Aase, A., (1991). Human IgG subclass pattern of inducing complet~cnt-mediated cytolysis dcpcnds on antigen concentration and to a lesser extent on epitope patchiness, antibody affinity and complement concentration. Eur: J. It~rmunol., 2 1, 11-16
1)orrinpton.K. I.. (1978). The structural basis for the functional diversity of immunoglobulin G. Con. 1. Biochew, 5 6. 1087-1101
Morrison. S. L.. Canfield, S., Porter. S.. Tan, 1.. K.. l?~o, M. Wims. L. A.. (1988). Production and characterization or gcnctically cnginecrcd antibody molecules. Clin. C/wn.. 3 4. 1668- 1675
J., Uuxbaum, J.. Hood, L.. (1981). Nucleotide sequence of a human immunoglobulin Cy4 gene. DNA, 1, 11-18
Pink, J. R .L., Milstcin. C., (1967). Disulphidc bridges of a human immunoglobulin G protein. Nanwe, 2 16, 941-942
Ellisoa. J. W., Berson. B. 1.. Hood, L. E., (1982). The nudeotide
Pink, J. R .L., Buttuy, S. H., DcVries. G. M.. Milstein, C., (1970). Human immunoglobulin subclass. Partiai amino acid sequence of the constant region of a y4 chain. Uioslretn 1.. I 17,33-47
Ellisou.
sequence of a human i~unoglobuIi[l 1 0, 407 I-4079.
Cyl gene. Nucf. Acids Res.,
Ellison. J. W., Hood. I_ E.. (1982). Linkage and squence homology of two i~unoglobulin y heavy chain constant region genes. Proc. Nail. Acod. Sci. USA. 7 9. 1984-1988
Primus, F. J., Pendurthi, T. K., Hutzcll. P.. Kashmiri, S., Slavin, D. C., Callahan, R., Schlom, J., (f990). Chimcric B72.3 inouse~uman (IgGl) antibody directs the lysis of tumour cells by lymphokineactivated killer cells. Cancer hnunol. Irrrntunorher., 3 1. 349-357
Frangione, 8.. Milstein. C., (1967). Disulphidc bridges immunoglobulin G-l heavy chain. Nafurc. 2 16, 939-941
Riechmann, I,.. Clark. M., Waldmann. H.. Winter. C., (1988). Reshaping human antibodies for therapy. Nowe, 3 3 2, 323-327
of
Garrcd. P.. Michaelsen. T E.. Aase. A., (1989). The IgG subclass pattern of complement activation depends on epitope density and antibody and complement concenuation. &and. 1. In~munof., 3 0, 319.382 Gillies. S. D.. Wesolowski. 1. S., (1991). Antigen binding and biological activities of engineered mutant chimeric Abs with human tumor specificities. Hum. Anribod. Hybridornas, 1. 47-54 German. S. D., Clark, M. R., (19901. HuI~aIlisatiotl of lnono~lonal antibodies ior therapy. Scirrninnrs in ~~~~~;funo~ogy, 2, 457-466 Hunter, W. M.. Greenwood. F. C., (1962). Preparation ol’ iodine-13t labelled human growth hormone of high specific activity. Nature, 1 9 4. 495.496 Hut& P.. Kashmiri. S.. Colcher, D.. Primus. E 1.. Horan Hand, P., Rosclli. M., Finch, M.. Yarranton, G., Bodmer. M.. WhitUc. N.. King. D.. Loullis, C. C., McCoy. D. W., Callahan, R.. Schlom, J., (1991). Generation and characterization of a recombinant/chimeric W72.3 (human ~1). CancerRes., 5 I, 181-189 Kabat, E. A., Wu. T. T., Reid-Miller, M., Perry, H. M., Gottesman, KS., (1987). Sequences of pro&ins of imrnunologicai inrcresr. 4th edition. Washington DC: United States Department of He&b and Human Services
Shcarman, C. W, Kanzy, E. J., Lawric, D. K.. Li, Y.-W.. Thammana, I?. Moore. G. P., Kurrle. R.. (1990). Construction, expression, and biological activity of murinc/human chimcric antibodies with specilicity for the human a/J3 T ccl1 receptor. J. I~wutnol.. I4 6, 928-935 Stephens, P. E., Cockett, M. I., (1989). The construction of a highly efficient versatile set of mammalian expression vectors. Nuci. Acids Rer., 1 7, 7110 ‘Ikn. L. K.. Shapes, R. J., Oi, V. T., Morrison. S. L., (1990). Influence of the binge region on complcmcnt activation, Clq binding, and segmental flexibility in chime& human immunoglobulins. Proc. Nail. Acad. Sci. USA, 8 7, 162-166 lkkahashi, N., Ueda, S., Obata. M.. Nikaido. T., Nakai, S.. Honjo, T. (1982). Structure of human immunoglobulin gamma genes. Implications for evolulion of a gent family. Cell. 2 9. 671-679 Wbittlc, N., Adair, J., Lloyd, C., Jenkins, L., Dcvinc, J.. Scblom, J.. Raubitshek, A., Colcher, D., Bodmcr. M.. (1987). Expression in COS -11~ of a mousci%uman chimaoric B72.3 anribody. Pror. &g.. I, 499-50s