BIOLOGICAL PRODUCTS

Abstract

There is disclosed antibody molecules containing at least one CDR derived from a mouse monoclonal antibody having specificity for human CD22. There is also disclosed a CDR grafted antibody wherein at least one of the CDRs is a modified CDR. Further disclosed are DNA sequences encoding the chains of the antibody molecules, vectors, transformed host cells and uses of the antibody molecules in the treatment of diseases mediated by cells expressing CD22.

Claims

1. An antibody molecule that binds human CD22 comprising a heavy chain and a light chain, wherein each chain comprises three complementarity determining regions (CDRs), wherein CDR-H1 comprises the amino acid sequence of SEQ ID NO:1; CDR-H3 comprises the amino acid sequence of SEQ ID NO:3; CDR-L1 comprises the amino acid sequence of SEQ ID NO:4; CDR-L2 comprises the amino acid sequence of SEQ ID NO:5; and CDR-L3 comprises the amino acid sequence of SEQ ID NO:6.

2. The antibody molecule of claim 1, which is a CDR-grafted antibody molecule.

3. The antibody molecule of claim 2, wherein the variable domain comprises human acceptor framework regions and non-human donor CDRs.

4. A composition comprising the antibody molecule of claim 1.

5. The composition according to claim 4, comprising a pharmaceutically acceptable excipient, diluent, or carrier.

6. The composition according to claim 4, additionally comprising anti-T cell, anti-IFNγ, anti-LPS antibodies, or a non-antibody ingredient.

Description

[0088] The present invention is further described by way of illustration only in the following examples, which refer to the accompanying Figures, in which:

[0089] FIG. 1 shows the amino acid sequence of the CDRs of mouse monoclonal antibody 5/44 (SEQ ID NOs:1 to 6);

[0090] FIG. 2 shows the complete sequence of the light chain variable domain of mouse monoclonal antibody 5/44 (nucleotide sequence-SEQ ID NO:48; amino acid sequence-SEQ ID NO: 7); antisense nucleotide strand-SEQ ID NO:67;

[0091] FIG. 3 shows the complete sequence of the heavy chain variable domain of mouse monoclonal antibody 5/44 (nucleotide sequence—SEQ ID NO:49; amino acid sequence-SEQ ID NO:8); antisense nucleotide strand-SEQ ID NO:68;

[0092] FIG. 4 shows the strategy for removal of the glycosylation site and reactive lysine in CDR-H2 (SEQ ID NOs:9-12);

[0093] FIG. 5 shows the graft design for the 5/44 light chain sequence (V.sub.L-SEQ ID NO:7; DPK9-SEQ ID NO:17, SEQ ID NO:69, and SEQ ID NO:70, respectively; JK1-SEQ ID NO:18 gL1-SEQ ID NO:19; and gL2-SEQ ID NO:20);

[0094] FIG. 6 shows the graft design for the 5/44 heavy chain sequence (V.sub.H-SEQ ID NO:8, DP7-SEQ ID NO:24, gH5-SEQ ID NO:25, gH6-SEQ ID NO:26, gH7-SEQ ID NO:27, and JH4-SEQ ID NO:22);

[0095] FIGS. 7A-7B show the vectors pMRR14 and pMRR10.1;

[0096] FIG. 8 shows the Biacore assay results of the chimeric 5/44 mutants;

[0097] FIG. 9 shows the oligonucleotides for 5/44 gH1 (SEQ ID NOs:32-39, respectively) and gL1 (SEQ ID NOs:40-47, respectively) gene assemblies;

[0098] FIGS. 10A-10B show the intermediate vectors pCR2.1(544gH1) and pCR2.1(544gL1);

[0099] FIG. 11 shows the oligonucleotide cassettes used to make further grafts (gH4-SEQ ID NOs:52, 53, and 62, respectively, gH5—SEQ ID NOs:54, 55, and 63, respectively; gH6—SEQ ID NOs:56, 57, and 64, respectively; gH7—SEQ ID NOs: 58, 59, and 65, respectively; and gL2—SEQ ID NOs:60, 61, and 66, respectively;

[0100] FIGS. 12A-12B show the competition assay between fluorescently labelled mouse 5/44 antibody and grafted variants; and

[0101] FIG. 13 shows the full DNA and protein sequence of the grafted heavy and light chains—a) SEQ ID NO:30 (amino acid), SEQ ID NO:31 (nucleotide), and SEQ ID NO:63 (antisense nucleotide strand); b) SEQ ID NO: 28 (amino acid), SEQ ID NO:29 (nucleotide), and SEQ ID NO:74 (antisense strand).

DETAILED DESCRIPTION OF THE INVENTION

Example 1: Generation of Candidate Antibodies

[0102] A panel of antibodies against CD22 were selected from hybridomas using the following selection criteria: binding to Daudi cells, internalisation on Daudi cells, binding to peripheral blood mononuclear cells (PBMC), internalisation on PBMC, affinity (greater than 10.sup.−9M), mouse γ1 and production rate. 5/44 was selected as the preferred antibody.

Example 2: Gene Cloning and Expression of a Chimeric 5/44 Antibody Molecule

[0103] Preparation of 5/44 Hybridoma Cells and RNA Preparation Therefrom

[0104] Hybridoma 5/44 was generated by conventional hybridoma technology following immunisation of mice with human CD22 protein. RNA was prepared from 5/44 hybridoma cells using a RNEasy kit (Qiagen, Crawley, UK; Catalogue No. 74106). The RNA obtained was reverse transcribed to cDNA, as described below.

[0105] Distribution of CD22 on NHL Tumours

[0106] An immunohistochemistry study was undertaken to examine the incidence and distribution of staining using the 5/44 anti-CD22 monoclonal antibodies. Control anti-CD20 and anti-CD79a antibodies were included in the study to confirm B cell areas of tumours.

[0107] A total of 50 tumours were studied and these were categorised as follows by using the Working Formulation and REAL classification systems:

[0108] 7 B lymphoblastic leukaemia/lymphoma (High/l)

[0109] 4 B-CLL/small lymphocytic lymphoma (Low/A)

[0110] 3 lymphoplasmacytoid/Immunocytoma (Low/A)

[0111] 1 Mantle cell (Int/F)

[0112] 14 Follicle center lymphoma (Low to Int/D)

[0113] 13 Diffuse large cell lymphoma (Int to High/G,H)

[0114] 6 Unclassifiable (K)

[0115] 2 T cell lymphomas

[0116] 40 B cell lymphomas were positive for CD22 antigen with the 5/44 antibody at 0.1 μg/ml and a further 6 became positive when the concentration was increased to 0.5 μg/ml. For the remaining 2 B cell tumours that were negative at 0.1 μg/ml, there was insufficient tissue remaining to test at the higher concentration. However, parallel testing with another Celltech anti-CD22 antibody 6/13, which gave stronger staining than 5/44, resulted in all 48 B cell lymphomas staining positive for CD22.

[0117] Thus, it is possible to conclude that the CD22 antigen is widely expressed on B cell lymphomas and thus provides a suitable target for immunotherapy in NHL.

[0118] PCR Cloning of 5/44 V.sub.H and V.sub.L

[0119] cDNA sequences coding for the variable domains of 5/44 heavy and light chains were synthesised using reverse transcriptase to produce single stranded cDNA copies of the mRNA present in the total RNA. This was then used as the template for amplification of the murine V-region sequences using specific oligonucleotide primers by the Polymerase Chain Reaction (PCR).

[0120] a) cDNA Synthesis

[0121] cDNA was synthesised in a 20 μl reaction volume containing the following reagents: 50 mM Tris-HCl pH 8.3, 75 mM KCl, 10 mM dithiothreitol, 3 mM MgCl.sub.2, 0.5 mM each deoxyribonucleoside triphosphate, 20 units RNAsin, 75 ng random hexanucleotide primer, 2 μg 5/44 RNA and 200 units Moloney Murine Leukemia Virus reverse transcriptase. After incubation at 42° C. for 60 minutes, the reaction was terminated by heating at 95° C. for 5 minutes.

[0122] b) PCR

[0123] Aliquots of the cDNA were subjected to PCR using combinations of primers specific for the heavy and light chains. Degenerate primer pools designed to anneal with the conserved sequences of the signal peptide were used as forward primers. These sequences all contain, in order, a restriction site (V.sub.L SfuI; V.sub.H HindIII) starting 7 nucleotides from their 5′ ends, the sequence GCCGCCACC (SEQ ID NO:50), to allow optimal translation of the resulting mRNAs, an initiation codon and 20-30 nucleotides based on the leader peptide sequences of known mouse antibodies (Kabat et al., Sequences of proteins of immunological interest, 5.sup.th Edition, 1991, U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health).

[0124] The 3′ primers are designed to span the framework 4 J-C junction of the antibody and contain a restriction site for the enzyme BsiWI to facilitate cloning of the V.sub.L PCR fragment. The heavy chain 3′ primers are a mixture designed to span the J-C junction of the antibody. The 3′ primer includes an ApaI restriction site to facilitate cloning. The 3′ region of the primers contains a mixed sequence based on those found in known mouse antibodies (Kabat et al., 1991, supra).

[0125] The combinations of primers described above enable the PCR products for V.sub.H and V1 to be cloned directly into an appropriate expression vector (see below) to produce chimeric (mouse-human) heavy and light chains and for these genes to be expressed in mammalian cells to produce chimeric antibodies of the desired isotype.

[0126] Incubations (100 μl) for the PCR were set up as follows. Each reaction contained 10 mM Tris-HCl pH 8.3, 1.5 mM MgCl2, 50 mM KCl, 0.01% w/v gelatin, 0.25 mM each deoxyribonucleoside triphosphate, 10 pmoles 5′ primer mix, 10 pmoles 3′ primer, 1 cDNA and 1 unit Taq polymerase. Reactions were incubated at 95° C. for 5 minutes and then cycled through 94° C. for 1 minute, 55° C. for 1 minute and 72° C. for 1 minute. After 30 cycles, aliquots of each reaction were analysed by electrophoresis on an agarose gel.

[0127] For the heavy chain V-region, an amplified DNA product was only obtained when a primer pool annealing within the start of framework I replaced the signal peptide primer pool. The fragments were cloned into DNA sequencing vectors. The DNA sequence was determined and translated to give a deduced amino acid sequence. This deduced sequence was verified by reference to the N-terminal protein sequence determined experimentally. FIGS. 2 and 3 shows the DNA/protein sequence of the mature light and heavy chain V-regions of mouse monoclonal 5/44 respectively.

[0128] c) Molecular Cloning of the PCR Fragments

[0129] The murine v-region sequences were then cloned into the expression vectors pMRR10.1 and pMRR14 (FIG. 7). These are vectors for the expression of light and heavy chain respectively containing DNA encoding constant regions of human kappa light chain and human gamma-4 heavy chain. The V.sub.L region was sub-cloned into the expression vector by restriction digest and ligation from the sequencing vector, using SfuI and BsiWI restriction sites, creating plasmid pMRR10(544cL). The heavy chain DNA was amplified by PCR using a 5′ primer to introduce a signal peptide, since this was not obtained in the cloning strategy—a mouse heavy chain antibody leader from a different in-house hybridoma (termed 162) was employed. The 5′ primer had the following sequence:

TABLE-US-00001 (SEQ ID NO: 51) .sup.5′GCGCGCAAGCTTGCCGCCACCATGGACTTCGGATTCTCTCTCGTGTT CCTGGCACTCATTCTCAAGGGAGTGCAGTGTGAGGTGCAGCTCGTCGA GTCTGG.sup.3′.

[0130] The reverse primer was identical to that used in the original V.sub.H gene cloning. The resultant PCR product was digested with enzymes HindIII and ApaI, was sub-cloned, and its DNA sequence was confirmed, creating plasmid pMRR14(544cH). Transient co-transfection of both expression vectors into CHO cells generated chimeric c5/44 antibody. This was achieved using the Lipofectamine reagent according to the manufacturer's protocols (InVitrogen: Life Technology, Groningen, The Netherlands. Catalogue no. 11668-027).

[0131] Removal of Glycosylation Site and Reactive Lysine

[0132] A potential N-linked glycosylation site sequence was observed in CDR-H2, having the amino acid sequence N-Y-T (FIG. 3). SDS-PAGE, Western blotting and carbohydrate staining of gels of 5/44 and its fragments (including Fab) indicated that this site was indeed glycosylated (not shown). In addition, a lysine residue was observed at an exposed position within CDR-H2, which had the potential to reduce the binding affinity of the antibody by providing an additional site for conjugation with an agent with which the antibody may be conjugated.

[0133] A PCR strategy was used to introduce amino acid substitutions into the CDR-H2 sequence in an attempt to remove the glycosylation site and/or the reactive lysine, as shown in FIG. 4. Forward primers encoding the mutations N55Q, T57A or T57V were used to remove the glycosylation site (FIG. 4) and a fourth forward primer containing the substitution K60R, was generated to remove the reactive lysine residue (FIG. 4). A framework 4 reverse primer was used in each of these PCR amplifications. The PCR products were digested with the enzymes XbaI and ApaI and were inserted into pMRR14(544cH) (also cleaved with XbaI and ApaI) to generate expression plasmids encoding these mutants. The N55Q, T57A and T57V mutations ablate the glycosylation site by changing the amino acid sequence away from the consensus N-X-T/S whilst the K60R mutation replaces the potentially reactive lysine with the similarly positively charged residue arginine. The resultant cH variant plasmids were co-transfected with the cL plasmid to generate expressed chimeric antibody variants.

[0134] Evaluation of Activities of Chimeric Genes

[0135] The activities of the chimeric genes were evaluated following transient transfection into CHO cells.

[0136] c) Determination of Affinity Constants by BiaCore Analysis.

[0137] The affinities of chimeric 5/44 or its variants, which have had their glycosylation site or their reactive lysine removed, were investigated using BIA technology for binding to CD22-mFc constructs. The results are shown in FIG. 8. All binding measurements were performed in the BIAcore™ 2000 instrument (Pharmacia Biosensor AB, Uppsala, Sweden). The assay was performed by capture of CD22mFc via the immobilised anti-mouse Fc. The antibody was in the soluble phase. Samples, standard, and controls (50 ul) were injected over immobilised anti-mouse Fc followed by antibody in the soluble phase. After each cycle the surface was regenerated with 50 ul of 40 mM HCl at 30 ul/min. The kinetic analysis was performed using the BIAevaluation 3.1 software (Pharmacia).

[0138] Removal of the glycosylation site in construct T57A resulted in a slightly faster on-rate and a significantly slower off-rate compared to the chimeric 5/44, giving an affinity improvement of approximately 5-fold. The N55Q mutation had no effect on affinity. This result was unexpected as it suggests that the removal of the carbohydrate itself apparently has no effect on binding (as with the N55Q change). The improved affinity was observed only with the T57A change. One possible explanation is that, regardless of the presence of carbohydrate, the threonine at position 57 exerts a negative effect on binding that is removed on conversion of threonine to alanine. The hypothesis that the small size of alanine is important, and that the negative effect of threonine is related to its size, is supported from the result obtained using the T57V mutation: that replacement with valine at position 57 is not beneficial (results not shown).

[0139] Removal of the lysine by the K60R mutation had a neutral effect on affinity, i.e. the introduction of arginine removes a potential reactive site without compromising affinity.

[0140] The mutations for removal of the glycosylation site and for removal of the reactive lysine were therefore both included in the humanisation design.

Example 2: CDR-Grafting of 5/44

[0141] The molecular cloning of genes for the variable regions of the heavy and light chains of the 5/44 antibody and their use to produce chimeric (mouse/human) 5/44 antibodies has been described above. The nucleotide and amino acid sequences of the mouse 5/44 V.sub.L and V.sub.H domains are shown in FIGS. 2 and 3 (SEQ ID NOs:7 and 8), respectively. This example describes the CDR-grafting of the 5/44 antibody onto human frameworks to reduce potential immunogenicity in humans, according to the method of Adair et al., (WO91/09967).

[0142] CDR-Grafting of 5/44 Light Chain

[0143] Protein sequence alignment with consensus sequences from human sub-group I kappa light chain V region indicated 64% sequence identity. Consequently, for constructing the CDR-grafted light chain, the acceptor framework regions chosen corresponded to those of the human VK sub-group I germline 012,DPK9 sequence. The framework 4 acceptor sequence was derived from the human J-region germline sequence JK1.

[0144] A comparison of the amino acid sequences of the framework regions of murine 5/44 and the acceptor sequence is given in FIG. 5 and shows that there are 27 differences between the donor and acceptor chains. At each position, an analysis was made of the potential of the murine residue to contribute to antigen binding, either directly or indirectly, through effects on packing or at the V.sub.H/V.sub.L interface. If a murine residue was considered important and sufficiently different from the human residue in terms of size, polarity or charge, then that murine residue was retained. Based on this analysis, two versions of the CDR-grafted light chain, having the sequences given in SEQ ID NO:19 and SEQ ID NO:20 (FIG. 5), were constructed.

[0145] CDR-Grafting of 5/44 Heavy Chain

[0146] CDR-grafting of 5/44 heavy chain was accomplished using the same strategy as described for the light chain. The V-domain of 5/44 heavy chain was found to be homologous to human heavy chains belonging to sub-group I (70% sequence identity) and therefore the sequence of the human sub-group I germline framework VH1-3,DP7 was used as an acceptor framework. The framework 4 acceptor sequences were derived from human J-region germline sequence JH4.

[0147] A comparison of 5/44 heavy chain with the framework regions is shown in FIG. 6 where it can be seen that the 5/44 heavy chain differs from the acceptor sequence at 22 positions. Analysis of the contribution that any of these might make to antigen binding led to 5 versions of the CDR-grafted heavy chains being constructed, having the sequences given in SEQ ID NO:23, SEQ ID NO:24, SEQ ID NO:25, SEQ ID NO:26 and SEQ ID NO:27 (FIG. 6).

[0148] Construction of Genes for Grafted Sequences.

[0149] Genes were designed to encode the grafted sequences gH1 and gL1, and a series of overlapping oligonucleotides were designed and constructed (FIG. 9). A PCR assembly technique was employed to construct the CDR-grafted V-region genes. Reaction volumes of 100 ul were set up containing 10 mM Tris-HCl pH8.3, 1.5 mM MgCl2, 50 mM KCl, 0.001% gelatin, 0.25 mM each deoxyribonucleoside triphosphate, 1 pmole each of the ‘internal’ primers (T1, T2, T3, B1, B2, B3), 10 pmole each of the ‘external’ primers (F1, R1), and 1 unit of Taq polymerase (AmpliTaq, Applied BioSystems, catalogue no. N808-0171). PCR cycle parameters were 94° C. for 1 minute, 55° C. for 1 minute and 72° C. for 1 minute, for 30 cycles. The reaction products were then run on a 1.5% agarose gel, excised and recovered using QIAGEN® spin columns (QIAquick® gel extraction kit, cat no. 28706). The DNA was eluted in a volume of 30 μl. Aliquots (1 μl) of the gH1 and gL1 DNA were then cloned into the InVitrogen TOPO® TA cloning vector pCR2.1 TOPO® (catalogue no. K4500-01) according to the manufacturer's instructions. This non-expression vector served as a cloning intermediate to facilitate sequencing of a large number of clones. DNA sequencing using vector-specific primers was used to identify correct clones containing gH1 and gL1, creating plasmids pCR2.1 (544gH1) and pCR2.1(544gL1) (FIG. 10).

[0150] An oligonucleotide cassette replacement method was used to create the humanised grafts gH4,5,6 and 7, and gL2. FIG. 11 shows the design of the oligonucleotide cassettes. To construct each variant, the vector (pCR2.1(544gH1) or pCR2.1(544gL1)) was cut with the restriction enzymes shown (XmaI/SacII for the heavy chain, XmaI/BstEII for the light chain). The large vector fragment was gel purified from agarose and was used in ligation with the oligonucleotide cassette. These cassettes are composed of 2 complementary oligonucleotides (shown in FIG. 11), mixed at a concentration of 0.5 pmoles/μ1 in a volume of 200 μl 12.5 mM Tris-HCl pH 7.5, 2.5 mM MgCl.sub.2, 25 mM NaCl, 0.25 mM dithioerythritol. Annealing was achieved by heating to 95° C. for 3 minutes in a waterbath (volume 500 ml) then allowing the reaction to slow-cool to room temperature. The annealed oligonucleotide cassette was then diluted ten-fold in water before ligation into the appropriately cut vector. DNA sequencing was used to confirm the correct sequence, creating plasmids pCR2.1 (5/44-gH4-7) and pCR2.1(5/44-gL2). The verified grafted sequences were then sub-cloned into the expression vectors pMRR14 (heavy chain) and pMR10.1 (light chain).

[0151] CD22 Binding Activity of CDR-Grafted Sequences

[0152] The vectors encoding grafted variants were co-transfected into CHO cells in a variety of combinations, together with the original chimeric antibody chains. Binding activity was compared in a competition assay, competing the binding of the original mouse 5/44 antibody for binding to Ramos cells (obtained from ATCC, a Burkitt's lymphoma lymphoblast human cell line expressing surface CD22). This assay was considered the best way to compare grafts in their ability to bind to cell surface CD22. The results are shown in FIG. 8. As can be seen, there is very little difference between any of the grafts, all performing more effectively than the chimeric at competing against the murine parent. The introduction of the 3 additional human residues at the end of CDR H2 (gH6 and gH7) does not appear to have affected binding.

[0153] The graft combination with the least number of murine residues was selected, gL1gH7. The light chain graft gL1 has 6 donor residues. Residues V2, V4, L37 and Q45 are potentially important packing residues. Residue H38 is at the V.sub.H/V.sub.L interface. Residue D60 is a surface residue close to the CDR-L2 and may directly contribute to antigen binding. Of these residues, V2, L37, Q45 and D60 are found in germline sequences of human kappa genes from other sub-groups. The heavy chain graft gH7 has 4 donor framework residues (Residue R28 is considered to be part of CDR-H1 under the structural definition used in CDR-grafting (se Adair et al (1991 WO91/09967)). Residues E1 and A71 are surface residues close to the CDR's. Residue I48 is a potential packing residue. Residue T93 is present at the V.sub.H/V.sub.L interface. Of these residues, E1 and A71 are found in other germline genes of human sub-group I. Residue I48 is found in human germline sub-group 4, and T73 is found in human germline sub-group 3.

[0154] The full DNA and protein sequence of both the light chain and heavy chain, including approximate position of introns within the constant region genes provided by the vectors, are shown in FIG. 13 and are given in SEQ ID NO:29 and SEQ ID NO:28 respectively for the light chain and SEQ ID NO:31 and SEQ ID NO:30 respectively for the heavy chain.

[0155] DNA encoding these light and heavy chain genes was excised from these vectors. Heavy chain DNA was digested at the 5′ HindIII site, then was treated with the Klenow fragment of E. coli DNA polymerase I to create a 5′ blunt end. Cleavage at the 3′ EcoRI site resulted in the heavy chain fragment which was purified from agarose gels. In the same way, a light chain fragment was produced, blunted at the 5′ SfuI site and with a 3′ EcoRI site. Both fragments were cloned into DHFR based expression vectors and used to generate stable cell lines in CHO cells.

[0156] All references and patents cited herein are hereby incorporated by reference in their entireties.