Proc. Nati. Acad. Sci. USA
`Vol. 84, pp. 3439-3443, May 1987
`Medical Sciences
`
`Chimeric mouse-human IgGi antibody that can mediate lysis of
`cancer cells
`(immunoglobulin domain cDNA/DNA transfection/tumor
`cytotoxicity)
`ALVIN Y. LIU*t, RANDY R. ROBINSON*, KARL ERIK HELLSTROMt§, E. DAVID MURRAY, JR.*¶,
`C. PAUL CHANG*, AND INGEGERD HELLSTR6M*"
`*International Genetic Engineering, Inc., 1545 17th Street, Santa Monica, CA 90404; *Oncogen, 3005 First Avenue, Seattle, WA 98121; and Departments of
`§Pathology and ilMicrobiology, University of Washington Medical School, Seattle, WA 98195
`Communicated by Paul D. Boyer, January 28, 1987 (receivedfor review December 22, 1986)
`
`antigen/complement-dependent :cytolysis/antibody-dependent cellular
`
`A chimeric mouse-human antibody has been
`ABSTRACT
`created that recognizes an antigen found on the surface of cells
`from many carcinomas. Immunoglobulin constant (C) domains
`of the mouse monoclonal antibody L6, CA and CM, were
`substituted by the human C-y and CM by recombining cDNA
`modules encoding variable or C domains. The cDNA constructs
`were transfected into lymphoid cells for antibody production.
`The chimeric antibody and mouse L6 antibody bound to
`carcinoma cells with equal affinity and mediated complement-
`dependent cytolysis. In the presence ofhuman effector cells, the
`chimeric antibody gave antibody-dependent cellular cytotox-
`icity at 100 times lower concentration than that needed for the
`mouse L6 antibody. The chimeric antibody, but not the mouse
`L6 antibody, is effective against a melanoma line expressing
`small amounts of the L6 antigen. The rindings point to the
`usefulness of the chimeric antibody approach for obtaining
`agents with strong antitumor activity for possible therapeutic
`use in man.
`
`The presence of tumor-associated antigens at the cell surface
`is a characteristic of many cancers. Since these antigens are
`either absent or found in much lower amounts in normal cells,
`it should be possible to use antibodies for targeting oftumors.
`A sizeable collection of relatively tumor-specific monoclonal
`antibodies (mAb) of mouse origin is available (1). Some of
`these mAb possess tumoricidal activity in the presence of
`human effector cells [antibody-dependent cellular cytotox-
`icity (ADCC)] or serum [complement-dependent cytotoxicity
`(CDC)] (2, 3). It has been shown (4) that partial tumor
`regression can be achieved when mAb possessing such
`functional activity are given to patients. One complication
`preventing repeated use of mouse mAb in man is that they are
`immunogenic. Furthermore, mouse mAb may interact less
`efficiently with human effector cells to mediate tumor de-
`struction.
`A method made possible by recombinant DNA technology
`was chosen to generate chimeric mouse-human antibodies. It
`entails the replacement of the mouse constant (C) domain
`regions with the corresponding human equivalents (5-7). In
`principle, antibody molecules obtained by this approach
`should retain their specificity for antigen and thus their
`usefulness for targeting, be much less immunogenic to man,
`and perhaps have increased antitumor activity.
`The mouse mAb L6 [IgG2a(K)] binds to a carbohydrate
`antigen found at the surface of cells from human carcinomas
`ofthe lung, breast, colon, and ovary (8). L6 can mediate CDC
`with human complement or ADCC with human effector cells
`(2). mAb L6 may thus be of use for tumor targeting, either in
`its native form or after conjugation of anticancer agents.
`
`The publication costs of this article were defrayed in part by page charge
`payment. This article must therefore be hereby marked "advertisement"
`in accordance with 18 U.S.C. §1734 solely to indicate this fact.
`
`In this study we have generated a mouse-human chimeric
`L6 antibody in which the mouse constant domains Cy25 and
`C, are substituted by the human Cy1 and C,. First, the cDNAs
`encoding the immunoglobulin genes were isolated. Next,
`restriction enzyme recognition sites were created in the
`cDNA sequences at the V/C junction (where V stands for
`variable) (9) by in vitro mutagenesis using oligodeoxyribo-
`nucleotides (10). The chimeric cDNAs thus constructed were
`then introduced into lymphoid cells by DNA transfection.
`The chimeric antibody isolated from the transfectants was
`compared with the mouse L6 for effector functions.
`
`MATERIALS AND METHODS
`DNA Transfection ofMouse Sp2/0 Lymphoid Cells. Expres-
`sion plasmid pING2114 (50 pug), linearized at a unique site
`(Aat II) in the nonessential bacterial region (see Fig. 3A), was
`transfected into 107 mouse Sp2/0 cells (CRL 1581, ATCC) by
`electroporation (11, 12). Transformants were selected in
`Dulbecco's modified Eagle's medium (DMEM) supplement-
`ed with 10% (vol/vol) fetal bovine serum (HyClone, Logan,
`UT) and G418 at 0.8 mg/ml (GIBCO). The transfection
`frequency was between 10-5 and 10-4. Human antibody in
`the medium was detected by ELISA (13).
`Isolation of Chimeric Antibody. Antibody-producing cells
`were grown to a density of 106 cells per ml and then shifted
`to serum-free DMEM 24 hr before harvest. Antibody secret-
`ed by the cells was concentrated by ultrafiltration, then
`chromatographed on a DEAE-cellulose column equilibrated
`in 40 mM NaCl/10 mM sodium phosphate, pH 8.0. The
`antibody in the flow-through was further purified to apparent
`homogeneity on protein A-Sepharose (14). For production of
`ascites fluid, 106 cells were injected into pristane-primed
`BALB/c mice. The chimeric antibody was purified by anti-
`human IgG-Sepharose chromatography (14).
`Functional Tests of the Chimeric L6 Antibody. The follow-
`ing tests were included: (i) measurement of antibody binding
`to target cells, either positive or negative for reactivity with
`the mouse L6; (it) competitive inhibition of binding of L6 to
`these cells; (iii) assays for CDC and ADCC. The binding tests
`were performed using a Coulter model EPIC-C cell sorter (8).
`The assays for CDC and ADCC were carried out on 51Cr-
`labeled target cells (2, 3) that were exposed to antibodies and
`human serum or peripheral blood leukocytes over a 4-hr
`period.
`
`Abbreviations: V, variable; C, constant; J,joining; mAb, monoclonal
`antibody(ies); CDC, complement-dependent cytolysis; ADCC, an-
`tibody-dependent cellular cytotoxicity; SV40, simian virus 40; H,
`heavy.
`tTo whom reprint requests should be addressed.
`VPresent address: Department of Biochemistry, University of Cali-
`fornia, Riverside, CA 92521.
`
`3439
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1032 Page 1 of 5
`
`

`
`3440
`
`Medical Sciences: Liu et al.
`
`roc. Natl. Acad. Sci. USA 84 (1987)
`Pr
`
`A
`
`Lb VH
`
`met asp trp leu
`C23AGTTTGTCTTAAGGCACCACTGAGCCCAAGTCTTAGACATCATG GAT TGG CTG
`Aft II
`BAL31 del.
`GTCGACTCT
`JFR1
`leader peptide
`Sal
`trp asn leu leu phe leu met ala ala ala gin ser ala gin ala gin
`TGG AAC TTG CTA TTC CTG ATG GCA GCT GCC CAA AGT GCC CAA GCA CAG
`
`0 ~~ ~~~0 0
`ile gin leu val gin ser gly pro gTu leu lys lys pro gly gTu thr
`ATC CAG TTG GTG CAG TCT GGA CCT GAG CTG AAG AAG CCT GGA GAG ACA
`
`9
`
`0
`
`FRlICDR1
`val lys ile ser cys lys ala ser gly tyr thr phe thr asn tyr gly
`GTC AAG ATC TCC TGC AAG GCT TCT GGG TAT ACC TTC ACA AAC TAT GGA
`Bgl II
`FR 21
`CDR 1FR2
`met asn trp val lys gin ala pro gly lys gly leu lys trp met gly
`ATG AAC TGG GTG AAG CAG GCT CCA GGA AAG GGT TTA AAG TGG ATG GGC
`Aha IlIl
`
`CDR 2
`trp ile asn thr tyr thr gly gin pro thr tyr ala asp asp ohe lys
`TGG ATA AAC ACC TAC ACT GGA CAG CCA ACA TAT GCT GAT GAC TTC AAG
`Nde
`
`CDR2 1 FR3
`gly ara phe ala phe ser leu glu thr ser ala tyr thr ala tyr leu
`GGA CGG TTT GCC TTC TCT TTG GAA ACC TCT GCC TAC ACT GCC TAT TTG
`
`gin ile asn asn leu lys asn glu asp met ala thr tyr Dhe cys ala
`CAG ATC AAC AAC CTC AAA AAT GAG GAC ATG GCT ACA TAT TTC TGT GCA
`
`FR31CDR3
`arg phe ser tyr gly asn ser
`AGA TTT AGC TAT GGT AAC TCA
`Dsp2.2
`JH2
`thr thr leu thr val ser ser
`ACC ACT CTC ACA GTC TCC TCA
`
`CDR 31 FR4
`arg tyr ser asp tyr trp gly qln gly
`CGT TAC TCT GAC TAC TGG GGC CAA GGC
`
`ala lys thr thr ala pro ser
`GCC AAA ACA ACA GCC CCA TCG
`
`GC
`
`AG-G
`
`MJH2Apal
`
`B
`
`L6 VK
`
`met asp phe gln
`CgCCCCAAGACAAAATG GAT TTT CAA
`-GTC
`5'Sa
`
`leader peptide
`val gin ile phe ser phe leu leu
`GTG CAG ATT TTC AGC TTC CTG CTA
`
`ile ser ala ser val ile met ser
`ATC AGT GCT TCA GTC ATA ATG TCC
`
`JFR1
`arg gly gln ile val leu ser qln
`AGA GGA CAA ATT GTT CTC TCC CAG
`
`ser pro ala ile leu ser ala ser
`TCT CCA GCA ATC CTG TCT GCA TCT
`
`pro gly glu lys val tfr leu thr
`CCA GGG GAG AAG GTC ACA TTG ACT
`
`FORl1CD§l
`o
`o
`o
`0
`0
`cys arg ala ser ser ser val ser
`TGC AGG GCC AGC TCA AGT GTA AGT
`
`0
`pro gly ser ser pro lys pro trp
`CCA GGA TCC TCC CCC AAA CCC TGG
`Bam Hi
`CDR 21FR3
`ser glu phe pro gly arg phe ser
`TCT GAG TTC CCT GGT CGC TTC AGT
`
`o CR11FRg
`0
`phe met asn trp tyr qln gin 1ys
`TTC ATG AAC TGG TAC CAG CAG AAG
`Kpn
`
`FR 21 CDR 2
`ile tyr ala thr ser asn leu ala
`ATT TAT GCC ACA TCC AAT TTG GCT
`
`gly glu tro ser gly thr ser tyr
`GGC GAG TGG TCT GGG ACC TCT TAC
`
`ser leu ala ile ser arg val glu
`TCT CTC GCA ATC AGC AGA GTG GAG
`
`ala glu aso ala ala thr tyr tyr
`GCT GAA GAT GCT GCC ACT TAT TAC
`
`F,?3gCD 3
`o
`0
`0
`0
`0
`cys gln gln trp asn ser asn pro
`TGC CAG CAG TGG AAT AGT AAC CCA
`
`J 5-
`CDR31FR4
`K
`0
`leu thr'phe gly ala gly thr lys
`CTC ACG TTC GGT GCT GGG ACC AAG
`
`leu glu leu lys
`CTG GAG CTG AAA
`-T- JKHindlll
`
`Nucleotide sequences and predicted
`FIG. 1.
`amino acid sequences of the L6 VH (A) and VK (B).
`The framework (FR) and complementarity deter-
`mining region (CDR) segments are indicated. The
`diversity (D) segment in VH is underlined. Circles
`above the amino acid residues indicate that these
`residues matched to those obtained from peptide
`sequencing. The VH sequence is present in plasmid
`pH3-6a. The C23 at the 5' end was removed by
`BAL-31 nuclease digestion. The resultant DNA
`sequence at the 5' end is shown below the first line.
`An Apa I site was introduced by the oligonucleo-
`tide primer MJH2ApaI. The VK sequence is pres-
`ent in plasmid pL3-12a. The C13 at the 5' end was
`removed by oligonucleotide-mediated mutagene-
`sis. A HindIII site was introduced by the oligonu-
`cleotide primer JKHindIII.
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1032 Page 2 of 5
`
`

`
`Medical Sciences: Liu et al.
`
`Proc. Natl. Acad. Sci. USA 84 (1987)
`
`3441
`
`RESULTS
`Isolation of Mouse cDNA. A cDNA library was generated
`from the L6 hybridoma cells by priming poly(A)+ RNA with
`oligo(dT) as described (9, 15). The probes used to screen the
`library were a JK5 oligonucleotide, d(GGTCCCAGCAC-
`CGAACG), for the light chain and a JH2 oligonucleotide,
`d(TGGCTGAGGAGACTGTGAGAG) for the heavy chain
`(where J stands for joining and H stands for heavy). Two
`methods (16, 17) were used to determine that the L6 K mRNA
`contains JKS sequences and that the L6 y2a mRNA contains
`JH2 sequences.
`Preparation of Mouse V-Region cDNA Modules. Restriction
`enzyme sites were engineered into the immunoglobulin
`cDNA around the V/C border for recombining mouse V
`regions to human C modules. The oligonucleotide MJH2ApaI
`[d(ATGGGCCCTTTGTGCTGGCTGAGGAGACTGT)
`(with the restriction enzyme site underlined)] was used for
`mutagenesis of the VH cDNA; and the oligonucleotide
`JKHindIII [d(CTCAAGCTTGGTCCC)] for that of the VK
`cDNA. Restriction sites on the 5' side ofthe ATG codon were
`also created. The oligonucleotide d(GAAAATCCATTT-
`TGTCGACGGG) was used to generate a Sal I site eight
`residues on the 5' side ofthe VK ATG codon. By cleaving with
`Sal I the oligo[d(GC)] segment on the 5' side of the cDNA
`insert was removed. To remove the oligo[d(GC)] segment on
`the 5' side of the VH cDNA, the nuclease BAL-31 (18) was
`used. The digested products were inserted into the vector
`M13mp19 (19) in such a way that the M13 Sal I site became
`a convenient site on the 5' side of the ATG codon. The DNA
`sequences of VH and VK are shown in Fig. 1.
`Human C-Region cDNA Modules. Human C-region cDNAs
`were isolated from libraries generated from the mRNA from
`two human lymphoblastoid cell lines, GM1500 and GM2146
`(Human Genetic Mutant Cell Repository). The human C,,1
`module has been described (9). The cDNA clone pGMH6
`contains an Apa I site 16 nucleotide residues on the 3' side of
`the V/C border (Fig. 2).
`The human CK module is a composite of two K cDNAs
`isolated from the GM1500 and GM2146 libraries. In plasmid
`pGML60, the 3'-untranslated region was derived from the K
`mRNA of GM2146 while the coding region was from that of
`GM1500. The JKHindIII oligonucleotide described above
`was used to engineer a HindIII recognition site at a position
`of the human JK segment analogous to that in the mouse JK
`segment.
`Chimeric L6 Heavy- and Light-Chain Expression Plasmids.
`The cDNA constructs were inserted into the vector se-
`quences of plasmid pING2012E (9). Directionality of inser-
`tion was achieved by using a Sal I-BamHI bracket.
`pING2012E contains regulatory sequences derived from
`plasmid pLl (20) that furnished the early promoter and splice
`donor-acceptor of simian virus 40 (SV40); and from plasmid
`pSV2neo (21) that furnished the transcription termination/
`polyadenylylation signals of SV40. We added the mouse
`immunoglobulin heavy-chain gene transcription enhancer,
`
`pGMH6 human Cyl module
`
`placed upstream of the SV40 promoter (9). The selectable
`marker is the TnS neo gene that confers resistance to the drug
`G418.
`The heavy-chain plasmid pING2111 was constructed by
`first joining the mouse VH cDNA module in a Sal I-Apa I
`DNA fragment with the human C,,1 cDNA module in an Apa
`I-BamHI DNA fragment. The ligated fragments were then
`inserted into pING2012E cleaved by Sal I-BamHI. The
`light-chain plasmid pING2119 was constructed byjoining the
`mouse VK cDNA module in a Sal 1-HindIlI DNA fragment
`with the human CK cDNA module in a HindIII-BamHI DNA
`fragment. The same vector fragment was used (Fig. 3A). In
`both plasmids the cDNA gene is placed 11 nucleotide
`residues downstream of the SV40 19S 3'-splice acceptor (9).
`The cDNA ends in a segment approximately A70G20, where
`it is joined to the SV40 transcription-termination/polyaden-
`ylylation sequences. Fig. 3B shows the incident nucleotide
`sequence changes made at the V/Cjunction as a result of the
`gene construction.
`A two-gene plasmid, pING2114, was constructed from
`pING2111 and pING2119 in which the light- and heavy-chain
`gene transcription units are in tandem (Fig. 3A). By using this
`plasmid, we introduced an equal ratio of heavy- and light-
`chain genes into recipient cells. Unexpectedly, we observed
`that there was a consistently higher expression of heavy than
`of light chain in all transfected cell lines examined (data not
`shown). The two transcription units differ in that the light-
`chain gene is about 700 base pairs shorter than the heavy-
`chain gene, and the C, gene segment has a higher A+T
`content. This imbalance was reduced by introducing more
`light-chain gene copies carried on a second plasmid with a
`different selectable marker [pING2121a, an Eco-gpt (22)
`version of pING2119].
`Two initial Sp2/0 transformants, D7 and 3E3, obtained by
`transfection with pING2114 were cultured for the isolation of
`chimeric antibody. D7 secretes 10% ofthe antibody produced
`by 3E3-K (17 ,ug/liter) and y (77 ,ug/liter) chains for D7
`compared to K (100 ug/liter) and y (700 ,g/liter) chains for
`3E3.
`Binding Characteristics of Chimeric Versus Mouse L6 An-
`tibody. Table 1 shows that the chimeric L6 antibody binds to
`cells from a human colon carcinoma (line C-3347) that
`express 5 x 105 molecules per cell of the antigen defined by
`the mouse L6 mAb (8). In a competition assay, 50% inhibition
`of binding was achieved by the same amount of the chimeric
`and mouse L6 (Fig. 4). Cells from a T-cell line, HSB-2, did
`not bind either mouse L6 or the chimeric antibody. Data on
`the melanoma line M-2669, clone 13 (3), are also included in
`Table 1, since this line, which expresses a low level of the
`L6-defined antigen, was used for the functional studies (see
`below).
`Chimeric L6 Antibody Mediates CDC and ADCC. Fig. 5
`shows that both the chimeric and mouse L6 antibodies lysed
`tumor cells in the presence of human complement. The
`experiment further showed that the chimeric L6 gave higher
`CDC at all dilutions of the complement.
`
`JH
`CH
`G GTC ACC GTC TCT TCAJGCC TCC ACC AAG GGC CCA
`Bst E 11
`Apa
`
`TCG G
`
`pGML60 human CK module
`
`-IGAT CAT CTC CCT CTC ACT TTC GGC GGA GGG ACC AAG
`Sau 3A
`
`Hind
`
`GTG GAG ATG AAA
`
`C-T
`III
`
`FIG. 2. Human C-domain cDNA
`gene modules. The relevant sequences
`at the V/C junction of human Cy, and
`C, cDNA clones are shown. The Cyl
`clone contains a BstEII and an Apa I
`site that can be used to recombine with
`different V-domain cDNA genes. The
`C,, clone does not contain a convenient
`recombination site; and an oligonucle-
`otide containing a HindIII site was used
`to introduce this site into the cDNA.
`The C-domain cDNA modules are
`pGMH6 and pGML60.
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1032 Page 3 of 5
`
`

`
`3442
`
`Medical Sciences: Liu et al.
`
`Proc. Natl. Acad. Sci. USA 84 (1987)
`
`Binding of chimeric L6 and mouse L6 antibodies to cell
`Table 1.
`lines used as targets for functional assays
`Antibody
`concentration,
`,g/ml
`
`Binding rati*
`GAM
`GAH
`
`Antibody
`Human colon carcinoma line C-3347
`Mouse L6
`
`Chimeric L6 (ascites)
`
`Chimeric L6 (cell culture)
`
`30
`10
`3
`30
`10
`3
`30
`10
`3
`
`38
`49
`40
`2
`2
`1
`1
`1
`1
`
`7
`3
`1
`NT
`NT
`NT
`
`4
`4
`3
`108
`84
`42
`105
`86
`44
`
`NT
`NT
`NT
`4
`2
`1
`
`Human melanoma line M-2669 (clone 13)
`Mouse L6
`30
`10
`3
`30
`10
`3
`
`Chimeric L6 (cell culture)
`
`Human T-cell line HSB-2
`Mouse L6
`1
`10
`1
`Chimeric L6 (ascites)
`1
`1
`10
`Chimeric L6 (cell culture)
`1
`10
`1
`*The binding ratio is the number of times a test sample is brighter
`than a control sample when treated with GAM (fluorescein isothio-
`cyanate-conjugated goat anti-mouse immunoglobulin) or with GAH
`(fluorescein isothiocyanate-conjugated goat anti-human immuno-
`globulin). For example, a ratio of 1 means that the test sample is as
`bright as the control; a ratio of 2 means that the test sample is twice
`as bright as the control. NT, not tested.
`
`(data not shown). Efficacy of the chimeric L6 was further
`demonstrated by its ability to lyse M-2669 melanoma cells
`(35% at 10 Ag/ml, 27% at 0.1 /ig/ml); the mouse L6 had no
`effect on these cells (9% at 10 ,ug/ml, as compared to 10%
`lysis with lymphocytes alone, cf. Table 1).
`
`DISCUSSION
`The mouse mAb L6 recognizes a carbohydrate antigen
`present in abundance in a variety of carcinomas. Normal
`tissues express only trace amounts of the antigen (8). Blased
`on this specificity there is justification in considering L6 for
`cancer treatment with the mAb used either alone (2) or as a
`carrier of anticancer agents. However, the immunogenicity
`of mouse L6 mAb in man is a disadvantage for its sustained
`use in patients, and its functional activity (ADCC and CDC)
`may be insufficient to effect optimal tumor destruction at the
`1oor
`-
`>, 8
`x 6
`0
`
`C-3347 Cells
`
`I-
`
`0 i
`
`0
`
`f-Chimeric L6
`
`C)
`
`e 2
`
`C ~~~~~~~~os L6
`Undiluted
`1:2
`1:8
`1:4
`Dilution of Humon Serum
`Titration of human serum (as a source of complement) in
`FIG. 5.
`the presence of chimeric or mouse L6 at 2.5 jug/ml.
`
`s L
`
`/ (XbaIr/SmaVI)
`
`pING2114
`11.5kb
`
`V"
`4
`
`Aof1i
`
`pING2119
`7.6kb
`
`B
`
`VHCy1 joint
`
`mohu
`.
`*-
`CTC ACA GTC TCC TCA GCC AGC ACA AAG GGC CCA T-
`Apa
`
`VKCK joint
`
`mo hu
`ACC AhAG CTO GAG OTG AAA CGA ACT
`Hind III
`
`FIG.3.
`(A) Expression plasmids. pING2111 is the heavy-chain
`and pING2119 is the light-chain expression plasmid. They were used
`to construct pING2114, a two-gene plasmid. Solid circles, mouse
`heavy-chain immunoglobulin gene enhancer; small arrows, SV40
`early promoter; diamonds, bidirectional SV40 transcription
`termination/polyadenylylation signals. (B) Nucleotide changes made
`in the V/C junction. Dotted residues in the VHCY1 junction were
`introduced by oligonucleotide-mediated mutagenesis. They are silent
`changes. Circled residues in the VKCK junction are residues contrib-
`uted by the human cDNA module to the mouse VK gene.
`
`Fig. 6 shows ADCC tests with cells from two cell lines. At
`a ratio of 100:1 human peripheral blood leukocytes to the
`colon carcinoma line C-3347 cells, the chimeric L6 killed a
`greater fraction of the target cells (a maximum of 98% versus
`63%) and gave 50% ADCC at 100 times lower concentration
`than the mouse L6 (0.01 ,ug/inl versus 1 ug/ml, Fig. 6A).
`Significant ADCC of C-3347 cells (24% as compared to 3%
`lysis with lymphocytes alone) was observed down to a 3:1
`ratio of effector cells to target cells when the chimeric L6 (at
`2.5 gg/ml) was used (Fig. 6B). Cell killing was specific
`because ADCC was not observed with the following three
`cell lines lacking detectable L6 antigens: B-cell lines DHL-10
`(Fig. 6C) and T51 (data not shown) and the T-cell line HSB-2
`
`50% Inhibition:
`Mouse L6= 1.9j.g/ml
`Chimeric L6 = 1.9j g/ml
`
`F[
`
`Chimeric L6
`
`100
`
`80s
`6O*D, 60
`
`0.
`
`c __.
`
`_ C
`
`40H
`201
`0 b,-;.-
`0.1
`
`0
`
`_ .
`
`~~Mouse L6
`
`I
`
`I
`
`100
`
`/ /
`---or
`10
`Inhibiting Antibody (pg/ml)
`Comparison between chimeric and standard mouse L6 in
`FIG. 4.
`antibody inhibition assays, performed by fluorescence-activated cell
`sorting. C-3347 cells were incubated with the blocking antibodies
`before fluorescein isothiocyanate-conjugated mouse L6 (3 ,g/ml)
`was added.
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1032 Page 4 of 5
`
`

`
`Medical Sciences: Liu et al.
`
`Proc. Natl. Acad. Sci. USA 84 (1987)
`
`3443
`
`the functional attributes of chimeric L6, should make it a
`strong candidate for therapeutic trials. Some ofthe antibodies
`induced in man to mouse mAb were directed to idiotypic
`determinants (26, 27). It remains to be seen whether the
`immunogenicity of those determinants ofthe chimeric L6 will
`be different from that of the mouse L6.
`The advantage of the cDNA approach lies in the ease with
`which immunoglobulin gene cDNAs can be isolated. The
`technology used for the present work should make it possible to
`convert many other mouse mAb to chimeric antibodies with
`improved antitumor activity via ADCC and CDC mechanisms.
`The chimeric antibodies will augment the relatively few human
`mAb currently used in the treatment of cancer (28).
`We thank Cathy Shapiro, Phil Mack, Phil Mixter, Pam Smith,
`Susan Azemove, Grethe Lovold, and Pat McGowan for excellent
`technical assistance. We also thank Randy Wall for discussion, and
`Randy Wall, Carol Hersh, Arup Sen, Gary Wilcox, Perry Fell, Jeff
`Ledbetter, Peter Linsley, and Erik Milner for useful comments on
`the manuscript. The work was supported by INGENE and
`ONCOGEN.
`Hellstrom, K. E. & Hellstrom, I. (1985) in Monoclonal Antibodies
`1.
`for Cancer Detection and Therapy, eds. Baldwin, R. S. & Byers,
`V. S. (Academic, New York), pp. 17-51.
`Hellstrom, I., Beaumier, P. L. & Hellstrom, K. E. (1986) Proc.
`Nati. Acad. Sci. USA 83, 7059-7063.
`Hellstrom, I., Brankovan, V. & Hellstrom, K. E. (1985) Proc.
`Nati. Acad. Sci. USA 82, 1499-1502.
`D., Cordon-Cardo, C., Welt, S.,
`Houghton, A. N., Mintzer,
`Fliegel, B., Vadhan, S., Carswell, E., Melamed, M. R., Oettgen,
`(1985) Proc. Natl. Acad. Sci. USA 82,
`H. F. & Old, L. J.
`1242-1246.
`Morrison, S. L., Johnson, M. J., Herzenberg, L. A. & Oi, V. T.
`(1984) Proc. Natl. Acad. Sci. USA 81, 6851-6855.
`Boulianne, G. L., Hozumi, N. & Shulman, M. J. (1984) Nature
`(London) 312, 643-646.
`Neuberger, M. S., Williams, G. T., Mitchell, E. B., Jouhal, S. S.,
`Flanagan, J. G. & Rabbitts, T. H. (1985) Nature (London) 314,
`268-270.
`Hellstr6m, I., Horn, D., Linsley, P., Brown, J. P., Brankovan, V.
`& Hellstrom, K. E. (1986) Cancer Res. 46, 3917-3923.
`Liu, A. Y., Mack, P. W., Champion, C. I. & Robinson, R. R.,
`Gene, in press.
`Zoller, M. J. & Smith, M. (1982) Nucleic Acids Res. 10, 6487-6500.
`Potter, H., Weir, L. & Leder, P. (1984) Proc. NatI. Acad. Sci. USA
`81, 7161-7165.
`Toneguzzo, F., Hayday, A. C. & Keating, A. (1986) Mol. Cell.
`Biol. 6, 703-706.
`Mixter, P. F., Wu, S. V., Studnicka, G. M. & Robinson, R. R.
`(1986) J. Immunol. Methods 91, 195-203.
`Johnstone, A. & Thorpe, R. (1982) Immunochemistry in Practice
`(Blackwell Scientific, Oxford), pp. 27-76.
`Gubler, U. & Hoffman, B. J. (1983) Gene 25, 263-269.
`Biol. Chem. 257,
`White, B. A. & Bancroft,
`F. C. (1982) J.
`8569-8572.
`Nobrega, F. G., Dieckmann, C. L. & Tzagoloff, A. (1983) Anal.
`Biochem. 131, 141-145.
`Legerski, R. J., Hodnett, J. L. & Gray, H. B., Jr. (1978) Nucleic
`Acids Res. 5, 1445-1463.
`Messing, J. (1983) Methods Enzymol. 101, 20-78.
`Okayama, H. & Berg, P. (1983) Mol. Cell. Biol. 3, 280-289.
`Southern, P. J. & Berg, P. (1982) J. Mol. Appl. Genet. 1, 327-341.
`Mulligan, R. C. & Berg, P. (1981) Proc. NatI. Acad. Sci. USA 78,
`2072-2076.
`Novotny, J., Handschumacher, M. & Haber, E. (1986) J. Mol.
`Biol. 189, 715-721.
`Sun, L. K., Curtis, P., Rakowicz-Szulczynska, E., Ghrayeb, J.,
`Morrison, S. L., Chang, N. & Koprowski, H. (1986) Hybridoma 5
`Suppl. 1, S17-S20.
`Sahagan, B.G., Dorai, H., Saltzgaber-Muller, J., Toneguzzo, F.,
`Guindon, C. A., Lilly, S. P., McDonald, K. W., Morrissey, D. V.,
`Stone, B. A., Davis,G. L., McIntosh, P. K. & Moore,G. P. (1986)
`J. Immunol. 137, 1066-1074.
`Goodman,G. E., Beautimier, P. L., Hellstrom, I., Fernyhough, B.
`& Hellstrdm, K. E. (1985) J. Clin. Oncol. 3, 340-352.
`Koprowski, H., Herlyn, D., Lubeck, M., DeFreitas, E. & Sears,
`H. F. (1984) Proc. Nat-.
`Acad. Sci. USA81,216-219.
`Irie, R. F. & Morton, D. L. (1986) Proc. Natl. Acad. Sci USA 83,
`8694-8698.
`
`19.
`20.
`21.
`22.
`
`23.
`
`24.
`
`25.
`
`26.
`
`27.
`
`28.
`
`2.
`
`3.
`
`4.
`
`5.
`
`6.
`
`7.
`
`8.
`
`9.
`
`10.
`11.
`
`12.
`
`13.
`
`14.
`
`15.
`16.
`
`17.
`
`18.
`
`Antibody (jig/ml)
`
`100 -
`
`.80 _
`
`60 -
`
`C-3347 Cells
`
`0
`
`ahimeric L6
`
`/
`
`Mouse L6
`
`20
`
`20
`
`40
`
`40
`
`60
`
`60
`
`8
`80
`
`100
`
`Effector Cells per Target Cell
`
`DHL-10 Cells
`E/T =100/1
`
`F5S
`
`L6 Chimeric LG
`a
`
`l.
`
`,40
`
`6
`
`20 S
`
`o0
`
`0
`
`C
`
`80
`
`)F
`
`40
`
`'F
`20
`
`I
`01
`I
`
`00-
`
`0>
`
`U %
`
`A
`
`CD)
`
`00
`
`.0_
`x
`
`C)
`
`B
`
`u 0
`
`0Il
`
`Q02
`
`0.2
`
`2
`
`Antibody (jpg/ml)
`(A) 'Titration of chimeric and mouse L6 antibodies in
`FIG. 6.
`ADCC assays with human peripheral blood leukocytes. E/T, ef-
`fector-target cell ratio. Two preparations of chimeric L6 were used.
`(B) Titration of human peripheral blood leukocyte effector cells
`mediating ADCC in the presence of antibodies (2.5,ug/ml). (C)
`Titration of L6 (chimeric and mouse) in ADCC assays on the DHL-10
`T-cell line. iF5
`is a mouse mAb that recognizes the DHL-10 cells.
`
`concentrations that can be attained in tumors after intrave-
`nous administration.
`It is estimated that a major immunogenic site resides in the
`CH2 region (23) of the IgG molecule. As one approach to
`decrease this problem, one could generate chimeric mouse-
`human antibodies thereby replacing the immunogenic C do-
`mains of the mouse immunoglobulins with those of human
`immunoglobulins. We did this using cDNA rather than cloned
`genomic DNA (24, 25). We show here that this is a useful
`approach for producing chimeric antibody. In cell line 3E3 and
`its subclones close to 1pug/ml of IgG1 protein was detected.
`The chimeric antibody was found to bind to tumor cells as
`well as the mouse L6 antibody. The chimeric antibody was
`much more efficient than L6 in ADCC assays, killing a
`greater fraction of target cells at a concentration lower by a
`factor of 100. Furthermore, the chimeric L6 killed cells from
`a melanoma line that was refractory to ADCC by the mouse
`L6. In patients one may speculate that the chimeric L6 would
`remain longer in the circulation. This, in combination with
`
`PETITIONER'S EXHIBITS
`
`Exhibit 1032 Page 5 of 5