.) 1991 Oxford University Press
`
`Nucleic Acids Research, Vol. 19, No. 9 2471
`
`Polymerase chain reaction facilitates the cloning, CDR-
`grafting, and rapid expression of a murine monoclonal
`antibody directed against the CD18 component of
`leukocyte integrins
`
`Bruce L.Daugherty*, Julie A.DeMartino, Ming-Fan Law, Douglas W.Kawka1, Irwin l.Singer1 and
`George E.Mark
`Departments of Cellular and Molecular Biology and 1Biochemical and Molecular Pathology,
`Merck Sharp and Dohme Research Laboratories, Rahway, NJ 07065, USA
`
`Received November 12, 1990; Revised and Accepted March 25, 1991
`
`ABSTRACT
`Two novel approaches of recombinant PCR technology
`were employed to graft the complementarity
`determining regions from a murine monoclonal
`antibody (mAb) onto human antibody frameworks. One
`approach relied on the availability of cloned human
`variable region templates, whereas the other strategy
`was dependent only on human variable region protein
`sequence data. The transient expression of
`recombinant humanized antibody was driven by the
`adenovirus major late promoter and was detected 48
`hrs post-transfection into non-lymphoid mammalian
`cells. The application of these new approaches enables
`the expression of a recombinant humanized antibody
`just 6 weeks after initiating the cDNA cloning of the
`murine mAb.
`
`INTRODUCTION
`The immunogenicity of murine-derived monoclonal antibodies
`(mAb) precludes the therapeutic use of these antibodies (1) in
`chronic or recurrent human diseases, such as inflammation or
`cancer. To minimize the unwanted human anti-mouse immune
`response, chimeric Abs were engineered which combined light
`and heavy chain variable (VL, VH) regions of murine origin and
`constant (C) regions from human sequences (2,3). Subsequently,
`Winter and colleagues (4) postulated that a 'humanized' version
`of a murine-derived mAb might reduce further the foreign
`character of the therapeutic Ab so as to be less immunogenic.
`A humanized Ab is one in which only the antigen-recognition
`sites or CDRs (complementarity-determining regions whose
`sequences are hypervariable, and thus antigen-specific, relative
`to the rest of the V regions) are of non-human origin, whereas
`all framework (FR) and C regions are products of human genes.
`In the construction of humanized Abs, several groups (4-7) have
`employed strategies requiring annealing, extension and ligation
`of many long synthetic oligos to graft human immunoglobulin
`
`* To whom correspondence should be addressed
`
`(Ig) VL, VH containing rodent CDRs. In addition, once a
`suitable humanized Ab is identified and expressed, considerable
`time is required for drug selection and expansion of transfected
`clones prior to analysis of recombinant Ig.
`We have devised a novel, rapid and effective means of
`substituting murine CDRs for their human counterparts through
`using overlapping polymerase chain reaction (PCR) fragments
`(8-13). Variations of this method are applicable to CDR-grafting
`regardless of the availability of human Ig V region clones. The
`grafted V regions are interchangeable in our expression vectors
`in cassette-like fashion. In addition, we have developed a transient
`expression assay in non-lymphoid mammalian cells which allows
`detection of recombinant human Abs (by ELISA) within 48 hours
`post transfection. Our system enables rapid expression for testing
`the folding, pairing assembly, secretion and resultant activity of
`intact recombinant humanized Abs.
`
`MATERIALS & METHODS
`Oligodeoxynucleotide synthesis
`Oligos were synthesized on an Applied Biosystems 381A DNA
`synthesizer or Milligen Cyclone, removed from the resin by
`treatment with concentrated NH40H followed by desalting on
`a Pharmacia NAP-S column (for oligos <40 bases in length)
`with H20 elution or by the use of an OPC column (Applied
`Biosystems) eluting with 20% acetonitrile (for oligos >40 bases
`in length).
`cDNA cloning
`Two micrograms of total cellular RNA isolated (14) from 108
`murine 1B4 (IgG2Ax) myeloma cells (15) were reverse-
`transcribed for 30 min at 42°C using 200 units of Moloney MuLV
`reverse transcriptase (RT;BRL) and 10 pmol of the C region
`complementary strand primers representing either VL (5'-
`TCTCAAGCTTTGGTGGCAAGAT(AG)GATACAGTTGG-
`TGCAGC-3') or VH (5'-TCTCAAGCTTACCGATGG(AG)G-
`
`Genzyme Ex. 1037, pg 911
`
`

`
`2472 Nucleic Acids Research, Vol. 19, No. 9
`
`CTGTTGTTTTGGC-3') chain in 20 ,1l of a buffer containing
`50 mM Tris-HCl, pH 8.3, 75 mM KCl, 3 mM MgC12, 10 mM
`DTT, and 20 units of RNAsin (Pharmacia). The RT was heat-
`inactivated at 95'C for 5 min and to each reaction was added
`50 pmol of each of the paired primers (C region complementary-
`strand primers listed above plus FRI primers for the kappa chain:
`[5 '-TCTCGGATCCGA(CT)AT(CT)GTG(AC)T(CG)ACCC-
`AG-3']; or the IgG2A heavy chain: [5'-TTCTGGATCCG(CG-
`)AGGT(CGT)AAGCTGGTG(CG)AGTC(AT)GG-3')], 2.5 units
`of Taq DNA polymerase (Perkin Elmer Cetus) in 100 1l of PCR
`buffer (10 mM Tris-HCl, pH 8.3, 50 mM KCI, 1.5 mM
`MgC12, 0.01 % (w/v) gelatin, 200P4M each dNTP). PCR was
`performed for 45 cycles (2', 94°C; 2', 55°C; 2', 72°C) followed
`by gel purification. The resulting 400 bp PCR product (see Figure
`1) was phosphorylated, subcloned into the blunt-ended (SmaI cut)
`vector pSP72 (Promega) and sequenced using Sequenase®
`(United States Biochemical Corp.)
`CDR grafting
`CDR grafting utilizing PCR recombination was accomplished by
`synthesizing eight oligos (4 primer pairs RI-R6, SI, S2) [See
`Figure 2. The appropriate primer pair (50 pmol of each) was
`combined with 10 ng of plasmid DNA representing the human
`FR (obtained from Winter and colleagues)(5), and 2.5 units of
`Taq DNA polymerase in 100 11 of PCR buffer. These primers
`directed PCR for 25 cycles (cycle periods, as above). The
`products of the four reactions were purified by agarose gel
`electrophoresis and electroelution, and 10 ng of each of the
`purified products were combined with terminal primers Al and
`A2 (50 pmol of each) and 2.5 units of Taq DNA polymerase
`in 100 /.l of PCR buffer and amplified for 25 cycles. CDR
`grafting via long oligos and PCR recombination (see Figure 3)
`was accomplished by adding 1 pmol of each of 4 long l00mers
`(G1-G4) together with 50 pmol terminal primers A3 and A4 and
`2.5 units Taq DNA polymerase in 100 1l of PCR buffer and
`amplified for 25 cycles. The combined 400 bp fragment was
`purified by agarose gel electrophoresis and electroelution. In
`parallel, two DNA fragments representing amino-terminal
`sequences encoding a 5' untranslated sequence and the signal
`peptide and carboxy-terminal sequences encoding FR 4, splice
`donor, and a 5'-intronic sequence between V and C regions were
`used for PCR with primer pairs S1, G5 and S2, G6 (G5 and
`G6 contain an 18 base overlapping sequence between GI and
`G4 respectively). PCR was performed with 50 pmol each primer,
`10 ng plasmid DNA containing both the signal and intronic
`sequences, 2.5 units of Taq polymerase in 100 Al of PCR buffer
`for 25 cycles and the product purified as above. These two DNA
`fragments (10 ng of each) were combined with 10 ng of the
`amplified 1B4 CDR-grafted VH region, 50 pmol terminal
`primers Al and A2, and 2.5 units of Taq polymerase in 100 ,l
`PCR buffer and amplified for 25 cycles.
`Heavy chain expression vector construction
`The vectors used for the expression of both H- and L-chain genes
`were derived from the pD5 eukaryotic expression vector (16)
`which contains the origin of adenovirus replication (ori), the SV40
`enhancer, the adenovirus major late promoter (MLP), the
`adenovirus 2 tripartite leader, a 5' splice donor from the
`adenovirus third leader and a 3' splice acceptor derived from
`an Ig locus, a multiple cloning site (containing the sites BamHI,
`SpeL, Sacd, NheI, Clal, NdeI, and NotI), and the SV40 late
`polyadenylation signal.
`
`The adenovirus ori was removed by EcoRI and KpnI digestion
`and replaced by two fragments representing the NeoR selectable
`marker gene (17) derived from plasmid pCMVIE-AK1-DHFR
`(18) as an EcoRI/BamHI 1.8 Kbp fragment and the Ig H chain
`enhancer sequence (19) following digestion with BglII and KpnI.
`The Ig H chain enhancer fragment was obtained as a 300 bp
`amplified fragment using PCR applied to human genomic DNA
`as template, the 2 primers: (5'-TTTTAGATCTGTCGACAG-
`ATGGCCGATCAGAACCAG-3' and 5 '-TTGGTCGACGGT-
`ACCAATACATTTTAGAAGTCGAT-3') and standard
`procedures described above. The resultant expression vector was
`found to lack a small portion of the TK promoter responsible
`for the transcription of the NeoR gene. The promoter was
`reconstituted by insertion into the EcoRI site of a 140 bp PCR-
`amplified fragment derived from pCMVIE-AK1-DHFR using the
`following primers: 5'-TATAGAATTCGGTACCCTTCATCC-
`CCGTGGCCCG-3' and 5 '-TGCGTGTTCGAATTCGCC-3'.
`This resulted in the Neo-selectable vector and was designated
`pD5/IgH-En/Neo. The 1B4 CDR-grafted VH region PCR
`product was digested with BglIH and BamHI, purified, cloned
`into pD5/IgH-En/Neo (BamHI cut) and subjected to DNA
`sequence analysis to determine orientation and verify the sequence
`of the reconstructed V region. The appending of the human IgG4
`(hIgG4) C region to the variable region was performed as follows.
`Plasmid pAT84 (20) containing the human genomic IgG4 C
`region as a 6.7 Kbp HindIll fragment was used as template DNA
`to amplify a 1.8 Kbp IgG4 fragment containing only the exons
`encoding CHI, CH2 and CH3 and the introns separating them.
`PCR was performed with the following primers: 5'-ATTTGG-
`ATCCTCTAGACATCGCGGATAGACAAGAAC-3' and
`5 '-AATAATGCGGCCGCATCGATGAGCTCAAGTATGTA-
`GACGGGGTACG-3'. Following digestion with BamHI and NotI
`(contained on the 5' and 3'-primers, respectively), the IgG4
`fragment was cloned into the BamHI and NotI sites of pD5/IgH-
`En/Neo containing the grafted VH region. The resulting plasmid
`designated pD5/IgH-En/Neo/VH/Human Cy4 (Figure 4) encodes
`an entire Ig H chain.
`Light chain expression vector construction
`The 1B4 CDR-grafted VL region PCR product was digested
`with HindIII and XbaI, purified from an agarose gel and
`subcloned into these same sites of vector pSP72 (Promega) which
`contained the human kappa CL region (21), obtained as follows.
`DNA (lg) purified from a human B cell line (GM01018A;
`NIGMS Human Genetic Mutant Cell Repository, Institute for
`Medical Research, Camden, NJ 08103) was used as a template
`with primers (5'-TCTCGGATCCTCTAGAAGAATGGCTG-
`CAAAGAGC-3' and 5'-TCTCGCTAGCGGATCCTTGCAG-
`AGGATGATAGGG-3') to PCR amplify a 920 bp fragment
`containing the splice acceptor for the human kappa CL domain,
`the exon and a portion of its 3'-untranslated region. The PCR
`product was purified by agarose gel electrophoresis, digested with
`BamH1, and cloned into BamH 1-digested pSP72. The
`pSP72-based intermediate vector containing both the 1B4 grafted
`VL region and the human CL region was digested with Spel and
`Clal resulting in a 1.5 Kbp fragment (containing the CDR-grafted
`variable and human constant region of the kappa L chain) and
`purified by agarose gel electrophoresis.
`The Neo-selectable vector (pD5/IgH-En/Neo) for H chain
`expression was converted into one expressing the hygromycin
`B (Hyg B) selectable marker for L chain expression as follows:
`The NeoR cassette was removed by digestion first with EcoRI
`
`Genzyme Ex. 1037, pg 912
`
`

`
`followed by DNA polymerase-directed fill in of the 5' overhang,
`then subsequent SalI digestion. The Hyg B expression cassette
`(1.9 Kbp BamHl fragment) containing the TK promoter and TK
`polyadenylation signal flanking the Hyg B gene from plasmid
`pLG90 (22) was removed from the plasmid pAL2 (Albert Lenny,
`personal communication) by BamHI digestion and subcloned into
`the BamHI site of the vector pSP72. The Hyg B cassette was
`removed from pSP72/Hyg B by digestion with SmaI and SalI
`and subcloned into the expression vector linearized as described
`above. This resulted in the Hyg B-selectable vector and was
`designated pDS/IgH-En/Hyg B. The vector was digested with
`Spel and Clal, purified and ligated to the 1.5 Kbp SpeI/ClaI L
`chain fragment purified from above. The resulting plasmid
`designated pD5/IgH-En/Hyg B/Vx/Human Cx (Figure 4)
`encoded an entire Ig kappa L chain.
`Expression and assay of recombinant antibody
`Expression plasmids encoding the 1B4 CDR-grafted H and L
`chain were purified through CsCl gradients. The plasmids were
`co-transfected (10,tg of each) into human 293 cells and monkey
`COS-7 and CV1P cells (2.5 x 106 cells in 100 mm plates) by
`standard calcium phosphate precipitation methods (23,24). The
`culture supernatants were assayed 48 hrs post-transfection by a
`sandwich ELISA described as follows: Immulon-2 (Dynatech
`Labs.) 96-well plates were coated overnight with a 5 14g/mL
`solution of mouse anti-human Cx mAb (cat. # MC009, The
`Binding Site, Inc., San Diego, CA) in 0.1 M NaHCO3 buffer
`(pH 8.2) at 4°C, and blocked with 1 % bovine serum albumin
`(BSA) in 0.1M NaHCO3 for lh at 25°C. After this and all
`subsequent steps, washing was performed with phosphate
`buffered saline (PBS). The wells then were incubated with culture
`supernatants containing recombinant humanized 1B4
`IgG4-antibody, or with predetermined quantities of hIgG4
`purified by protein A Sepharose (Pharmacia Fine Chemicals)
`chromatography from hIgG4 myeloma serum (cat. # BP026,
`The Binding Site, Inc.) All samples were diluted in PBS
`containing 0.05% (w/v) Tween-20. 100-FL aliquots were
`incubated for lh at 37°C in triplicate, and standard calibration
`curves were constructed using hIgG4 concentrations of 10-100
`
`Nucleic Acids Research, Vol. 19, No. 9 2473
`
`ng/mL. Bound and fully assembled hIgG4 (either native or
`humanized IgG) were detected with 100-/AL aliquots of a 1:500
`dilution of mouse anti-hIgG4 Fc mAb conjugated to alkaline
`phosphatase (cat #05-3822, Zymed Laboratories, Inc., South
`San Francisco, CA) in 0.1 M Tris, 0.15 M NaCl ph 7.5
`containing 1 % (w/v) BSA. After incubation for lh at 37°C and
`subsequent washing, the quantities of bound conjugate were
`detected by incubating the wells with a lmg/mL solution of p-
`nitrophenyl phosphate in 0. IM 2,2' amino-methyl-propanediol
`buffer, pH 10.3, for 30 min at 25°C and measuring the
`adsorbance at 405 nm with a UV Max plate reader (Molecular
`Devices).
`
`RESULTS
`cDNA cloning of an anti-CD18 mAb
`The VH and VL regions of murine 1B4 anti-CD18 mAb (15)
`were cloned using PCR and was similar to the approach taken
`by other groups (26,27).
`Analysis of the murine 1B4
`immunoglobulin with an isotyping ELISA kit (ScreenType,
`Boehringer Mannheim, Indianapolis, IN) indicated it is comprised
`of an IgG2A H chain and a kappa L chain (15). The 5' primers
`for PCR were chosen from conserved N-terminal FRI kappa and
`IgG2A V region sequences (28). The H chain 5' primer was
`24-fold degenerate and the L chain 5' primer 16-fold (Figure
`1). Both 3' PCR primers were chosen from conserved regions
`of either CHI or CL, each being 2-fold degenerate. Agarose gel
`analysis (Figure 1) of the PCR products of the VH and VL
`regions clearly indicates fragments migrating at approximately
`400 bp as was predicted from the chosen primers. These PCR
`products were subcloned and sequenced and the CDR and FR
`regions were delineated based on the
`locations of the
`hypervariable CDR domains in available sequenced V genes (28).
`Grafting of murine anti-CD18 mAb CDRs onto human V
`region FRs
`Once the murine FR and CDRs had been identified, two general
`approaches were devised to graft these CDRs onto human V
`region FRs. The first approach uses a rearranged genomic DNA
`
`PCR /
`
`NH2
`
`VL
`
`CDR1
`
`M 1 2 3 4 5 6
`
`COOH
`
`Figure 1. Schematic of the coding sequence for an antibody molecule depicting the position where PCR primers were chosen to clone the anti-CD18 VH and VL
`regions. Arrow indicates size of fragments on the agarose gel depicted on the right after RNA/PCR with primers shown (see text). CDR: complementarity-determining
`region; FR: Framework region; V: variable region; C: constant region; L: light chain; H: heavy chain; D: diversity; J: joining; NH2: amino termiinus; COOH:
`carboxyl terminus; -S-S-: disulfide bridge; lanes 1-4, VH RT/PCR products; lanes 5-6, VL RT/PCR products; M: X HindIH, pBR322 MspI.
`
`Genzyme Ex. 1037, pg 913
`
`

`
`2474 Nucleic Acids Research, Vol. 19, No. 9
`
`U~~~~~~~~~~~~~~~~~~~~~~~~~~-eU
`
`Figure 2. Strategy for CDR-grafting utilizing PCR recombination. Primer pairs SI-Ri, R2-R3, R4-R5, and R6-S2 generate 4 PCR products in round 1 as shown
`in the figure and lanes 1-4 on the agarose gel depicted on the right. The 4 fragments are combined and PCR amplified with external amplifiers Al and A2 in
`round 2 (see text). The arrow pointing to the band in lane 5 depicts a murine CDR-grafted/human L chain FR of 600 bp. RE: restriction endonuclease sites; M:
`X HindIII, (f X174 HaeIII.
`
`.~~ -
`
`m
`
`Figure 3. Strategy for CDR-grafting utilizing long oligos and PCR recombination. In round 1 long oligos GI, G2, G3, G4 and short termninal primers A3 and A4
`generate the PCR product as shown in the figure and lane 1 on the agarose gel depicted on the right. In parallel, signal peptide .(primers Si1 and G5) and intron
`(primers G6 and S2) PCR products are generated as shown (lanes 2 and 3, respectively). The 3 fragments are combined and PCR amplified with external amplifiers
`Al and A2 in round 2 (see text). The arrow pointing to the band in lane 4 depicts a CDR-graftedlheavy chain FR of 850 bp.
`
`clone (containing intact VDJ for H chain, VJ for L chain, each
`flanked by signal peptide and intron, and V-C intron sequences)
`of a human V region as template for PCR. Both FRs contained
`a leader sequence and 3' intronic sequences. The grafting of
`murine CDRs onto a human V region FR or so-called 'CDR
`grafting via PCR Recombination' is shown in Figure 2. Eight
`representing the primers
`oligo primers were synthesized,
`necessary to generate four DNA fragments by PCR amplification.
`Each internal primer (Rl-R6) contains a sequence that anneals
`to the beginning or end of each human FR plus some or all of
`the murine 1B4 CDR sequence to be grafted. The murine CDR
`sequences do not anneal to the template and effectively hang off
`the human FRs as shown. The dotted lines indicate an 18 base
`complementary sequence between primers. Four DNA fragments
`generated in the first round of PCR by the appropriate primer
`pairs of sizes 314, 97, 134, and 113 bp (Figure 2, Lanes 1-4)
`were recombined and amplified in a second round of PCR using
`external amplifiers Al and A2. These primers anneal to a
`complementary sequence of 18 bases (chosen at random) present
`on primers Si and S2, their subsequent extensions will exclude
`amplification of the original V region cDNA template in favor
`of the CDR grafted recombinant, resulting in a PCR-amplified
`600-bp DNA fragment. As depicted in figure 2 (Lane 5), a DNA
`fragment encoding a fully grafted murine CDR/human L chain
`FR was generated, demonstrating the success of this novel
`technique.
`
`Often the protein sequence of antibodies is available (ie, from
`protein databases) when DNA clones encoding human FRs are
`not. Therefore, a second approach was devised which uses long
`synthetic oligos in conjunction with PCR recombination to
`produce a fully grafted (humanized) human FR/murine CDR V
`region. Codons for murine 1B4 H chain CDRs and the human
`H chain V region FR were built into the design. The 4 long oligos
`(100 bases in length) each contained complementary ends of 18
`bases (Figure 3) which, during PCR using external amplifiers
`A3 and A4, result in the formation and subsequent amplification
`of the combined sequences of a 400-bp DNA fragment (Figure
`3, Lane 1). This grafted sequence is combined with PCR products
`encoding the signal peptide and splice donor (225 bp, Figure 3,
`Lane 2) and intronic sequences (230 bp, Figure 3, Lane 3) in
`a second round of PCR recombination using external amplifiers
`Al and A2, resulting in a PCR-amplified 850-bp DNA fragment
`(Figure 3, Lane 4). Although three other minor PCR-amplified
`DNA fragments were generated in this round, they were not as
`highly resolved by agarose gel eletrophoresis and were not of
`the predicted size. The major DNA fragment encodes a fully
`grafted murine CDR/human H chain FR thereby definitively
`demonstrating
`a second successful
`application of PCR
`recombination in CDR grafting.
`The fully grafted 1B4 CDR/human H- and L-chain FR-
`encoding DNAs generated by the two approaches were subcloned
`and sequenced to determine the accuracy of PCR recombination.
`
`Genzyme Ex. 1037, pg 914
`
`

`
`EcoRI EcoRI
`
`Bho,HI
`
`pD5/IgH-Enhaner/Neo/
`CDR-grafted 1 B4-VH/
`Human C74 (8743 bp)
`
`i
`
`anH
`CDR-rafted I U-Vd
`I IHI
`xb.l
`X
`Human Cc (7650 bp)
`S~~~~~~~~~~~~~~~~W
`
`ia41
`
`Spel~~~~~~~~~X
`a
`
`Figure 4. Vectors used for L and H chain Ig gene expression. IgH-En:
`Immunoglobulin heavy chain enhancer; SV40: SV40 origin and enhancer;
`MLP+L: Adenovirus major late promoter and tripartite leader; p(A): SV40 late
`polyadenylation signal; ori: pBR322 origin of replication; Amp: ampicillin
`resistance gene; tk: herpes simplex virus thymidine kinase promoter; Neo:
`neomycin (G418) resistance gene; Hyg B: hygromycin B resistance gene: lB4-VH:
`CDR-grafted/H chain FR; 1B4-Vx: CDRy-grafted/L chain FR; Human C-y4:
`human gamma-4 C region; Human Cx: human kappa C region.
`
`Table 1. Expression level of recombinant antibodies
`
`Cell Line
`
`293
`COS-7
`CV1-P
`
`N=2
`
`ng/mL
`
`385 ±438
`82
`1.2
`50±3.9
`
`The Taq polymerase was found to be error prone during PCR
`recombination. Indeed, many clones lacked a completely intact
`open reading frame (ORF). For expediting the selection process
`of clones with intact ORFs, we screened clones by sequencing
`with a single dideoxy termination lane (A track) alone. In one
`case, we were able to quickly eliminate 18/31 prospective clones
`using this simple screening method. The lack of fidelity of the
`polymerase has been addressed in
`the context of our
`recombination strategy, and improvements on the method have
`been found which greatly diminish the number of undesirable
`clones (manuscript in preparation).
`Expression and activity of recombinant CDR-grafted Ig genes
`The vectors for IgG expression driven by the adenovirus MLP
`and tripartite leader are depicted in Figure 4. Both L and H chain
`expression vectors are identical, except that the L chain vector
`carries the Hyg BR gene and the H chain vector the NeoR gene.
`Transfections for transient expression were performed by co-
`transfection of equal amounts of each plasmid into three primate
`kidney cell lines; human 293 cells and monkey COS-7 and CV1-P
`cells. The culture supernatants were assayed 48 hours post-
`transfection by a trapping ELISA that specifically measures
`secreted human kappa L chain linked to a human IgG4 H chain.
`As shown in Table 1, recombinant IgG4x antibodies were
`synthesized by all three cell lines. Human 293 cells expressed
`the highest transient level while the 2 monkey cell lines expressed
`significantly lower levels and this was not suprising since human
`293 cells constitutively express the adenoviral EIA protein (29)
`which stimulates transcription directed from the adenovirus major
`late promoter (30).
`In preliminary binding experiments, medium from transiently
`transfected 293 cultures containing fully-grafted humanized 1B4
`
`Nucleic Acids Research, Vol. 19, No. 9 2475
`
`Table 2. Comparative effects of unlabeled fully-grafted humanized 1B4 versus
`murine 1B4 on the binding of 125I-murine 1B4 to stimulated human
`polymorphonuclear leukocytes (PMNs)
`
`Molarity IgG Added
`
`Quantity of 125I-Murine 1B4 Bound (CPM)
`Humanized 1B4
`Native Murine 1B4
`
`1 x 10- 12
`1x 10-0
`I x 10-9
`5 x 10-9
`
`1941
`1761
`858
`460
`
`2211
`1697
`783
`99
`
`Purified murine 1B4 was radiolabeled with 125I to a specific activity of 17 uCi/ug
`using chloramine-T. PMNs were isolated and activated with phorbol myristate
`acetate as previously described (25). Aliquots of PMNs were incubated in
`1.8 x 10-11 M 1251-murine 1B4 in the presence of various concentrations of either
`fully grafted humanized 1B4, or unlabeled murine 1B4, and the quantity of 1251_
`murine 1B4 bound to the cells determined as described (25). For humanized 1B4,
`various dilutions of conditioned medium from 293 cultures transfected 48hr
`previously with the plasmids encoding the fully grafted 1B4 were used to compete
`with 125I -1B4 binding to PMNs. The concentration of humanized 1B4 present
`in this medium was determined by ELISA as described in materials and methods.
`
`was observed to compete with the binding of 125I-murine 1B4
`to CD18-containing receptors on stimulated human polymorpho-
`nuclear leukocytes (Table 2). The level of competition was very
`similar to that observed when unlabeled native 1B4 was used as
`the competitor. Similar results were obtained with CV1-P
`transfected cultures, while media from mock transfected cells or
`purified human IgG4 were completely uninhibitory (data not
`shown).
`
`DISCUSSION
`In this report, we have demonstrated novel adaptations of PCR
`to the successful cloning, humanization, and expression of a
`murine mAb directed against the CD18 component of leukocyte
`integrins (15). Notably, PCR has enabled us to expediently carry
`out the process of cDNA cloning of murine Ig V regions, CDR-
`grafting onto human FRs, and expression as measured by ELISA
`of a fully assembled recombinant human antibody in as little as
`six weeks. The novel technical approaches we have taken enable
`a considerable time and cost savings as compared to more
`traditional methods. In addition, the preliminary competitive
`binding experiments described here also suggest that the
`humanized CDR-grafted 1B4 molecule has an avidity for CD18
`at the surface of activated PMNs nearly comparable to that of
`the native murine 1B4 mAb. The choice of particular human VL
`and VH region FRs used in the humanization of the murine 1B4
`mAb play a critical role in the outcome of the humanization and
`will be reported elsewhere.
`It must be emphasized that the techniques we describe
`necessitated a substantial amount of DNA sequencing in order
`to distinguish those clones that contained an intact ORF from
`those containing insertions, deletions, or substitutions. Initially,
`the method of CDR-grafting using DNA templates for PCR was
`much more successful for producing clones that contained intact
`ORFs than the synthetic long oligo method. Indeed, the majority
`of the clones derived from the latter approach had significant
`deletions of greater than 6 bp, as well as numerous point
`mutations. Retrospectively, upon sequence analysis using a dyad
`symmetry computer program (Intelligenetics)
`ascertain
`to
`predicted regions of stable secondary structure, it was noted that
`the particular sequences subject to deletion gave hairpin loops
`with AG values of -13 to -26 Kcal/mol. Furthermore, the stem-
`
`Genzyme Ex. 1037, pg 915
`
`

`
`2476 Nucleic Acids Research, Vol. 19, No. 9
`
`loops predicted in the sequence by the symmetry program closely
`approximated the region of observed deletions in size and
`location. Subsequently, the codon choice was altered in this region
`of the long oligonucleotides. This resulted in a decrease in the
`predicted AG values and a reduction in the frequency of deletions
`during PCR recombination was observed in the product. In
`addition, a decrease in cycle number has also resulted in an
`increase in fidelity and a reduction in the number of deletions.
`Other parameters (i.e., cycle time, temperature, etc.) are being
`investigated actively. Although we do not know the upper limit
`of either the length or the number of fragments for which PCR
`recombination will be practical, we have been successful at
`combining up to 5 fragments and up to a total of approximately
`1700 bp. PCR recombination need not be limited to CDR
`grafting, but will lend itself to a variety of other applications such
`as domain shuffling, domain editing, construction of hybrid
`molecules, etc. Thus, it is a general method to insert specific
`new sequences into other molecules.
`We have generated expression vectors which enable us to
`exchange CDR-grafted FR regions as a single casette. This
`affords us the opportunity to combine a variety of human. FRs
`with identically grafted murine CDR regions, and thus to explore
`the effect of the FR regions on the avidities of a particular human
`recombinant antibody. In this way we can test the hypothesis that
`the recipient FRs for CDR-grafting are generic, ie, any FR region
`is capable of supporting all CDR loops. Likewise, we can easily,
`rapidly and systematically introduce amino acid changes in the
`CDR or FR regions as part of a structure-function analysis of
`antibody/antigen interaction.
`Our rapid transient expression system in monolayer culture has
`served a two-fold purpose. Firstly, due to the high frequency
`of deletions during PCR recombination, it provides a rapid screen
`that allows us to ascertain whether or not the CDR-grafted
`recombinant contains an intact ORF. Secondly, upon expression,
`the activity (avidity) of the recombinant antibody can be quickly
`determined and evaluated.
`It is hopeful that CDR-grafting of the murine V-regions would
`reduce the immunogenicity of the mAb in the context of the
`human immune system. Experiments performed in mice (31)
`suggest that this would indeed be the case, in that a significant
`antigenic response was produced against the foreign V-region
`in a chimeric antibody. Although it is beyond the scope of this
`paper, it will be necessary to produce sufficient quantities of the
`recombinant humanized anti-CD18 mAb for clinical evaluation
`in humans. The establishment of stable mammalian expressor
`cell lines is currently in progress and will subsequently be
`followed by scale-up production. Ideally, the reduction in
`immunogenicity would result in increased serum half-life of the
`therapeutic mAb requiring reduced dosage to achieve efficacy.
`Furthermore, it might enable repeated administration of the
`recombinant humanized mAb for longer time periods.
`
`ACKNOWLEDGEMENTS
`The authors wish to gratefully acknowledge the expert technical
`assistance of Susan Zavodny. They also thank Dr. John A.
`Schmidt and Dr. Alan R. Williamson for valuable suggestions
`and discussions, Dr. Ronald W. Ellis for critical review of the
`manuscript and Susan Pols for manuscript preparation.
`
`REFERENCES
`1. Houghton, A.N., Mintzer, D., Cordon-Cardo, C., Welt, S., Fliegel, B.,
`Vadhan, S., Carswell, E., Melamed, M.R., Oettgen, H.F. and Old, L.J.
`(1985) Proc. Natl. Acad. Sci. USA 82: 1242-1246.
`2. Morrison, S.L., Johnson, M.J., Herzenberg, L.A. and Oi, V.T. (1984) Proc.
`Natl. Acad. Sci. USA 81: 6851-6855.
`3. LoBuglio, A.F., Wheeler, R.H., Trang, J., Haynes, A., Rogers, K., Harvey,
`E.B., Sun, L., Ghrayeb, J. and Khazaeli, M.B. (1989) Proc. Natl. Acad.
`Sci. USA 86: 4220-4224.
`4. Jones, P.T., Dear, P.H., Foote, J., Neuberger, M.S. and Winter, G. (1986)
`Nature 321: 522 - 525.
`5. Riechmann, L., Clark, M., Waldmann, H. and Winter, G. (1988) Nature
`332: 323-327.
`6. Verhoeyen, M., Milstein, C. and Winter, G. (1988) Science 239: 1534-1536.
`7. Queen, C., Schneider, W.P., Selick, H.E., Payne, P.W., Landolfi, N.F.,
`Duncan, J.F., Avdalovic, N.M., Levitt, M., Junghans, R.P. and Waldmann,
`T.A. (1989) Proc. Natl. Acad. Sci. USA 86: 10029-10033.
`8. Saiki, R.K., Scharf, S.J., Faloona, F., Mullis, K.B., Horn, G.T., Erlich,
`H.A. and Arnheim, N. (1985) Science 230: 1350-1354.
`9. Mullis, K.B., Faloona, F., Scharf, S.J., Saiki, R.K., Horn, G.T. and Erlich,
`H.A. (1986) Cold Spring Harbor Symp. Quant. BioL 51: 263-273.
`10. Higuchi, R., Krummel, B. and Saiki, R.K. (1988) Nucleic Acids Res. 16:
`7351-7367.
`11. Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K. and Pease, L.R. (1989)
`Gene 77: 51-59.
`12. Horton, R.M., Hunt, H.D., Ho, S.N., Pullen, J.K. and Pease, L.R. (1989)
`Gene 77: 61-68.
`13. Vallette, F., Mege, E., Reiss, A. and Adesnik, M. (1989) Nucleic Acids
`Res. 17: 723-733.
`14. Chirgwin, J.M., Przybyla, A.E., MacDonald, R.J. and Rutter, W.J. (1979)
`Biochemistry 18: 5294-5299.
`15. Wright, S.D., Rao, P.E., Van Voorhis, W.C., Craigmyle, L.S., Ilda, K.,
`Talle, M.A., Westberg, E.F., Goldstein, G. and Silverstein, S.C. (1983)
`Proc. Natl. Acad. Sci., USA 80: 5699-5703.
`16. Berkner, K.L. and Sharp, P.A. (1985) Nucleic Acids Res. 13: 841-857.
`17. Rothstein, S.J. and Reznikoff, W.S. (1981) Cell 23: 191-199.
`18. Silberklang, M., Kopchick, J., Munshi, S. Lenny, A., Livelli, T. and Ellis,
`R.W. (1987) In Spier, R.E. and Griffiths, B., (ed.) Modern Approaches
`to Animal Cell Technology, Butterworth, U.K., pp. 199-214 .
`19. Gillies, S.D., Morrison, S.L., Oi, V.T. and Tonegawa, S. (1983) Cell 33:
`717-728
`20. Flanagan, J.G. and Rabbitts, T.H. (1982) Nature 30: 709-713.
`21. Hieter, P.A., Max, E.E., Seidman, J.G., Maizel, J.V. and Leder, P. (1980)
`Cell 22: 197-207.
`22. Gritz, L. and Davies, J. (1983) Gene 25: 179-188.
`23. Graham, F.L. and van der Eb, A.J. (1973) Virology, 52: 456-467.
`24. Wigler, M., Silverstein, S., Lee, L.S., Pellicer, A., Cheng, Y.C. and Axel,
`R. (1977) Cell 11: 223 -232.
`25. Singer, I.I., Sc