Humanization of mouse anti-human IL-2 receptor antibody B-B10

Mouse monoclonal anti–human IL–2 receptor antibody(BB10) inhibits EL–2–dependent human T–cellproliferation. It has been used in clinical trials in the transplantationfield and promising results are being accumulated. Mouse B–B10antibody was humanized by grafting all CDRs and some frameworkamino add residues onto human antibodies, KAS for V_H and PAYfor V_x. Nine humanized B–BlOs with differently graftedframework residues were constructed and assessed for their biologicalactivities. One of these humanized B–B10, M5, showed nearlythe same activity as the mouse B–B10. The 49th residueof V_x was demonstrated to play a crucial role in the antigen–antibodyinteraction by 3–D structure analysis with a computermodeling system.     

Humanization of a mouse monoclonal antibody by CDR-grafting: the importance of framework residues on loop conformation

A mouse monoclonal antibody (mAb 425) with therapeutic potentialwas ‘humanized’ in two ways. Firstly the mouse variableregions from mAb 425 were spliced onto human constant regionsto create a chimeric 425 antibody. Secondly, the mouse complementarity–determiningregions (CDRs) from mAb 425 were grafted into human variableregions, which were then joined to human constant regions, tocreate a reshaped human 425 antibody. Using a molecular modelof the mouse mAb 425 variable regions, framework residues (FRs)that might be critical for antigen-binding were identified.To test the importance of these residues, nine versions of thereshaped human 425 heavy chain variable (VH) regions and twoversions of the reshaped human 425 light chain variable (VJregions were designed and constructed. The recombinant DNAscoding for the chimeric and reshaped human light and heavy chainswere coexpressed transiently in COS cells. In antigen-bindingassays and competition-binding assays, the reshaped human antibodieswere compared with mouse 425 antibody and to chimeric 425 antibody.The different versions of 425–reshaped human antibodyshowed a wide range of avidities for antigen, indicating thatsubstitutions at certain positions in the human FRs significantlyinfluenced binding to antigen. Why certain individual FR residuesinfluence antigen-binding is discussed. One version of reshapedhuman 425 antibody bound to antigen with an avidity approachingthat of the mouse 425 antibody.     

Humanization of the anti-CEA T84.66 antibody based on crystal structure data

Chimeric T84.66 (cT84.66) is a monoclonal antibody (mAb) of high specificity and affinity for the tumor-associated carcinoembryonic antigen (CEA). Radiolabeled cT84.66 has demonstrated utility in the clinic as a reagent for the radioimmunoscintigraphy and radioimmunotherapy of CEA-positive colorectal and breast malignancies. To extend the therapeutic efficacy of T84.66, humanization by complementary determining region (CDR) grafting was employed. CDR grafting is a well-established technique, though often a series of framework back-mutations is required to restore high affinity. Recently, the crystal structure of the T84.66 diabody (scFv dimer) derived from the murine T84.66 mAb was determined, facilitating the humanization process by the availability of crystal structure data for both the graft donor and graft acceptor. A search of the Protein Data Bank revealed close structural similarity (r.m.s.d. of 1.07 A) between the Fv of T84.66 and the Fv of 4D5v8, a humanized anti-p185HER2 antibody marketed as Herceptin (Trastuzumab). This resulted in two humanized versions of the T84.66 M5A and M5B mAbs that differed only in the number of murine residues present in the C-terminal half of CDR-H2. Biochemical analysis and animal biodistribution studies were conducted to evaluate the humanized mAbs. The M5A, M5B and cT84.66 mAbs showed sub-nanomolar affinity for CEA and as radiolabeled mAbs exhibited specific tumor localization in tumor bearing mice. The T84.66 M5A mAb was selected for clinical development due to a slightly higher tumor uptake and a larger content of human residues, and was renamed hT84.66. A limited-scale production and animal imaging study have demonstrated hT84.66's ability to support clinical trials. Planned clinical trials will determine the effective utilization of this structure-based approach in the development of a promising new therapeutic.     

Molecular modelling and site-directed mutagenesis on a bovine anti-testosterone monoclonal antibody

A three-dimensional (3D) molecular model of the antigen-combiningsite of a bovine anti-testosterone monoclonal antibody has beenconstructed. In the model, the CDRs, and a single heavy chainframework region residue (Trp47), associate to form a hydrophobiccavity large enough to accommodate a single molecule of testosterone.Tyr97 of CDR-H3 lies at the bottom of the cavity with its hydroxylgroup exposed to solvent. Using the model and data from bindingstudies, we predicted that the cavity forms the antibody's paratopeand on binding testosterone a hydrogen bond is formed betweenTyr97 of CDR-H3 and the hydroxyl group on the D-ring of testosterone.This prediction has subsequently been tested by site-directedmutagenesis. An antibody with phenylalanine in place of tyrosineat position 97 in CDR-H3 has its affinity reduced by {smalltilde}800 fold. The reduction in binding energy associated withthe reduced affinity has been calculated to be 3.9 kcal/molwhich is within the range (0.5–4.0 kcal/mol) expectedfor the loss of a single hydrogen bond. The model has been usedto suggest ways of increasing the antibody's affinity for testosterone.     

Humanization of 60.3, an anti-CD18 antibody; importance of the L2 loop

Monoclonal antibody 60.3 binds to the CD18 component of theß2 integrin family of adhesion molecules. 60.3 haspotential clinical application in blocking the neutrophilmediatedorgan damage which occurs following a myocardial infarct orhemorrhagic shock. Analysis of the nucleotide and deduced aminoacid sequences of murine 60.3 shows that the light chain containsno amino acid substitutions relative to the closest germlinesequence, while the heavy chain is heavily substituted. We reporthere the humanization of 60.3. The humanized antibody bindsto CD18-bearing cells with {small tilde}4-fold less affinitythan the murine or chimeric antibody. We have shown that modificationof amino acid L50 in the L2 loop of the humanized antibody resultsin loss of binding, while modification of a structural determinant(H71) for the H2 loop has no effect.     

'Knobs-into-holes' engineering of antibody CH3 domains for heavy chain heterodimerization

‘Knobs-into-holes’ was originally proposed by Crickin 1952 as a model for the packing of amino acid side chainsbetween adjacent -helices. ‘Knobs-into-holes’ isdemonstrated here as a novel and effective design strategy forengineering antibody heavy chain homodimers for heterodimerization.In this approach a ‘knob’ variant was first obtainedby replacement of a small amino acid with a larger one in theC_H3 domain of a CD4-IgG immuno-adhesin: T366Y. The knob wasdesigned to insert into a ‘hole’ in the C_H3 domainof a humanized anti-CD3 antibody created by judicious replacementof a large residue with a smaller one: Y407T. The anti-CD3/CD4-IgGhybrid represents up to 92% of the protein A purified proteinpool following co-expression of these two different heavy chainstogether with the anti-CD3 light chain. In contrast, only upto 57% of the anti-CD3/CD4-IgG hybrid is recovered followingco-expression in which heavy chains contained wild-type C_H3domains. Thus knobs-into-holes engineering facilitates the constructionof an antibody/ immunoadhesin hybrid and likely other Fc-containingbifunctional therapeutics including bispecific immuno-adhesinsand bispecific antibodies.     

Molecular modeling of antibody-antigen complexes between the Brucella abortus O-chain polysaccharide and a specific monoclonal antibody

A molecular model of the binding site of an anti-carbohydrateantibody (YsT9.1) has been developed using computer-assistedmodeling techniques and molecular dynamics calculations. Sequencehomologies among YsT9.1 and the Fv regions of McPC603, J539and human Bence-Jones protein REI, all of which have solvedcrystal structures, provided the basis for the modeling. Thegroove-type combining site model had a topography which wascomplementary to low energy confonners of the polysaccharide,a Brucella O-antigen, and the site could be almost completelyfilled by a pentasaccharide epitope in either of two dockingmodes. Putative interactions between this epitope and the antibodyare consistent with the known structural requirements for bindingand lead to the design of oligosaccharide inhibitors that probethe veracity of the modeled docked complex. Ultimately boththe Fv model and the docked complex will be compared with independentcrystal structures of YsT9.1 Fab with and without pentasaccharideinhibitor, currently at the stage of refinement.     

Multimerization behaviour of single chain Fv variants for the tumour-binding antibody B72.3

A systematic study has been performed on the relationship betweenlinker length, relative orientation of variable domains, multimerizationbehaviour and antigen binding activity for single chain Fvs(scFvs) of the tumour-binding antibody B72.3. Thirteen scFvvariants with linkers comprising up to six repeats of the motifGly-Gly-Gry-Gly-Ser were studied. All these scFvs showed a tendencyto form dimers or higher molecular weight species, and thistendency decreased with increasing linker length. The dimersand higher molecular weight forms may arise from head to tailintermolecular association of V_H and V_L domains. For each linkerlength, scFvs with the organization VL-linker-VH showed greaterbinding activity than those with the organization VH-linker-VL.In fact, for the latter organization only the variant with a30 amino acid linker showed good binding activity, suggestingthat (0 for B72.3 the C-tenninus of VH or the N-tenninus ofVL makes a structural contribution to antigen binding, and (ii)shorter linkers interfere with this contribution. Antigen bindingstudies on scFvs should be interpreted with caution becauseof their tendency to multlmerize. Such multimerization can beminimized by using linkers longer than those in common use     

Properties of a single-chain antibody containing different linker peptides

Single-chain antibodies were constructed using six differentlinker peptides to join the V_H and V_L domains of an anti-2-phenyloxazolone(Ox) antibody. Four of the linker peptides originated from theinterdomain linker region of the fungal cellulase CBHI and consistedof 28, 11, six and two amino acid residues. The two other linkerpeptides used were the (GGGGS)₃ linker with 15 amino acid residuesand a modified IgG2b hinge peptide with 22 residues. Proteolyticstability and Ox binding properties of the six different scFvderivatives produced in Escherichia coli were investigated andcompared with those of the corresponding Fv fragment containingno joining peptide between the V domains. The hapten bindingproperties of different antibody fragments were studied by ELISAand BIAcore^TM. The interdomain linker peptide improved the haptenbinding properties of the antibody fragment when compared withFv fragment, but slightly increased its susceptibility to proteases.Single-chain antibodies with short CBHI linkers of 11, six andtwo residues had a tendency to form multimers which led to ahigher apparent affinity. The fragments with linkers longerthan 11 residues remained monomeric.     

Expression in COS cells of a mouse--human chimaeric B72.3 antibody

B72.3 is a mouse hybndoma cell-line secreting an IgG1 antibodywhich recognises an epitope on a tumour-associated antigen,TAG-72. This high molecular weight mucin-like molecule is foundon a variety of human neoplasms, including colon, breast andovarian carcinomas. Chimaeric Immunoglobulin genes with theB72.3 specificity have been constructed by joining the mousevariable regions from cDNA clones to human genomic constantregions using recombinant DNA techniques. The chimaeric heavyand light chain Immunoglobulin genes were placed under the controlof a strong viral promoter, and co-transfected into COS-1 cells.SDS–PAGE analysis of the ³⁵S-labelled products demonstratedthat the transiently expressed antibodies were correctly synthesisedand assembled. The specific binding characteristics of the parentB72.3 antibody were retained by the chimaeric antibody in anantigen-based ELISA. This system gave sufficiently high transientexpression of the chlmaerlc antibody molecules to allow rapidphysical and Immunological characterisation of the engineeredgene products.     

Linker modification introduces useful molecular instability in a single chain antibody

SCA 4-4-20/212, is a recombinant single chain antibody directedagainst fluorescein (Fl) composed of the variable light (V_L)and variable heavy (V_H) domains of the monoclonal antibody 4-4-20,tethered by a 14 amino acid linker. Binding of SCA 4-4-20/212to Fl quenches its fluorescence, thus enabling the distinctionbetween bound and free Fl. This was used to follow antibodydenaturation which followed a two-step process: rapid selectedand restricted denaturation followed by slow and progressivedenaturation. This two-phase phenomenon might reflect selectivesusceptibility of the CDR loops to denaturation. Furthermore,a new SCA, SCA 4-4-20/9, was constructed by site-directed mutagenesisof SCA 4-4-20/212 using PCR methodology. SCA 4-4-20/9 was similarto SCA 4-4-20/212, but for a nine residue linker. The two SCAswere compared for Fl binding, heat stability, the effect ofdenaturing agents and susceptibility to proteolysis. The modificationof the linker caused a general conformational rearrangementin the SCA molecule, rendering it more sensitive to denaturationand proteolysis. This molecular instability may find utilityin the application of SCAs in analytical systems or as the recognitioncomponent in biosensors.     

Effects of humanization by variable domain resurfacing on the antiviral activity of a single-chain antibody against respiratory syncytial virus

HNK20 is a mouse monoclonal IgA that binds to the F glycoproteinof respiratory syncytial virus (RSV) and neutralizes the virus,both in vitro and in vivo. The single-chain antibody fragment(scFv) derived from HNK20 is equally active and has allowedus to assess rapidly the effect of mutations on affinity andantiviral activity. Humanization by variable domain resurfacingrequires that surface residues not normally found in a humanFv be mutated to the expected human amino acid, thereby eliminatingpotentially immunogenic sites. We describe the constructionand characterization of two humanized scFvs, hu7 and hu10, bearing7 and 10 mutations, respectively. Both molecules show unalteredbinding affinities to the RSV antigen (purified F protein) asdetermined by ELISA and surface plasmon resonance measurementsof binding kinetics (K_a 1x10⁹ M^–1). A competition ELISAusing captured whole virus confirmed that the binding affinitiesof the parental scFv and also of hu7 and hu10 scFvs were identical.However, when compared with the original scFv, hu10 scFv wasshown to have significantly decreased antiviral activity bothin vitro and in a mouse model. Our observations suggest thatbinding of the scFv to the viral antigen is not sufficient forneutralization. We speculate that neutralization may involvethe inhibition or induction of conformational changes in thebound antigen, thereby interfering with the F protein-mediatedfusion of virus and cell membranes in the initial steps of infection.     

Efficient generation of a reshaped human mAb specific for the {alpha} toxin of Clostridium perfringens

We have used the technique of antibody reshaping to producea humanized antibody specific for the a toxin of Clostridiumperfringens. The starting antibody was from a mouse hybridomafrom which variable (V) region nucleo-tide sequences were determined.The complementarity-determining regions (CDRs) from these Vregions were then inserted into human heavy and light chainV region genes with human constant region gene fragments subsequentlyadded. The insertion of CDRs alone into human frameworks didnot produce a functional reshaped antibody and modificationsto the V region framework were required. With minor frameworkmodifications, the affinity of the original murine mAb was restoredand even exceeded. Where affinity was increased, an alteredbinding profile to overlapping peptides was observed. Computermodelling of the reshaped heavy chain V regions suggested thatamino acids adjacent to CDRs can either contribute to, or distort,CDR loop conformation and must be adjusted to achieve high bindingaffinity.     

Toxicity-based selection of Escherichia coli mutants for functional recombinant protein production: application to an antibody fragment

We propose a novel approach to the selection of Escherichia coli bacterial strains improved for the production of recombinant functional proteins. This approach is based on aggregation-induced toxicity of recombinant proteins. We show that selection of clones displaying a reduced toxicity is an efficient means of isolating bacteria producing recombinant protein with reduced aggregation in favour of correct folding. For an efficient selection, we found that time of toxicity induction must be precisely determined and recombinant protein must be expressed as a fusion with a protein whose activity is easily detectable on plates, thus allowing elimination of non-productive mutants. Choosing the expression to the periplasmic space of an scFv fragment fused to the N-terminus of alkaline phosphatase as a model, we selected chromosomal mutations that reduce aggregation-induced toxicity and showed that they concomitantly improve production of a functional recombinant hybrid. The effects of the mutations isolated could then be cumulated with those of other strategies used for recombinant scFv production. Thus, we could ensure a 6- to 16-fold increase in production of a functional scFv-PhoA hybrid. This is the first report demonstrating the possibility of directly selecting on agar plates E.coli strains improved for functional recombinant protein production from a large bacterial mutant library.     

Prediction of the antigenic sites of the cystic fibrosis transmembrane conductance regulator protein by molecular modelling

Antibodies are powerful tools for studying the in situ localizationand physiology of proteins. The prediction of epitopes by molecularmodelling has been used successfully for the papilloma virus,and valuable antibodies have been raised [Muller et aL (1990)J. Gen. Virol, 71, 2709–2717]. We have improved the modellingapproach to allow us to predict epitopes from the primary sequencesof the cystic fibrosis transmembrane conductance regulator.The procedure involves searching for fragments of primary sequenceslikely to make amphipathic secondary structures, which are hydrophilicenough to be at the surface of the folded protein and thus accessibleto antibodies. Amphipathic helices were predicted using themethods of Berzofsky, Eisenberg and Jahnig. Their hydrophobichydrophilicinterface was calculated and drawn, and used to predict theorientation of the helices at the surface of the native protein.Amino acids involved in turns were selected using the algorithmof Eisenberg. Tertiary structures were calculated using ‘FOLDING’,a software developed by R.Brasseur for the prediction of smallprotein structures [Brasseur (1995) J. MoL Graphics, in press].We selected sequences that folded as turns with at least fiveprotruding polar residues. One important property of antibodiesis selectivity. To optimize the selectivity of the raised antibodies,each sequence was screened for similarity (FASTA) to the proteinsequences from several databanks. Ubiquitous sequences werediscarded. This approach led to the identification of 13 potentialepitopes in the cystic fibrosis transmembrane conductance regulator:seven helices and six loops.     

Expression of an exogenous peptide epitope genetically engineered in the variable domain of an immunoglobulin: implications for antibody and peptide folding

Immunoglobulins bind antigens and express individual antigenicspecificities mainly through residues located in hypervariableloops of their N-terminal domains. Hyper-variable loops arekept in place by a molecular scaffold organized in a sandwich-likestructure with two ß-sheets stabilized by a disulfidebridge (the immunoglobulin fold). This structural feature, togetherwith the possibility of obtaining high level expression, extracellularsecretion, easy purification and stability of the protein product,render immunoglobulin an ideal ‘molecular vehicle’for the expression of exogenous peptides. Here we report onthe engineering of an immunoglobulin expressing an exogenousepitope, the repetitive tetrapeptide Asn-Ala-Asn-Pro (NANP)₃.By recombinant DNA techniques, we inserted three copies of thetetrapeptide (NANP)₃ in the third hypervariable loop (D region)of an immunoglobulin heavy chain variable domain. We show thatthe engineered antibody was properly assembled and secreted.A panel of polyclonal and monoclonal antibodies, including anti-syntheticpeptides and anti-(NANP)_n antibodies, were used to study themolecular configuration of the engineered domain's surface.The results indicate that (i) the exogenous sequence did notappreciably alter the overall fold of the variable domain; and(ii) the inserted epitope folded with a configuration immuno-logicallysimilar to the one assumed in the native protein, suggestingthat short- and medium- rather than long-range interactionsstabilized the structure of the (NANP)₃ peptide in the foldedprotein. We propose this system for the expression of peptidicsequences, and their structural and functional analysis.     

Engineering of protein epitopes: a single deletion in a snake toxin generates full binding capacity to a previously unrecognized antibody

Structural features associated with the ability of a monoclonalantibody (mAb) to discriminate between protein variants areidentified and engineered. The variants are the curaremimetictoxin from Naja nigricollis and erabutoxin a or b from Laticaudasemifasciata which differ from each other by 16 substitutionsand one insertion. The neutralizing mAb M1 recognizes with highaffinity a topographical epitope on the surface of toxin , butfails to recognize the erabutoxins although they possess mostof the residues forming the presumed epitope. Examinations ofthe toxin and erabutoxin 3-D structures and molecular dynamicssimulations reveal several differences between the variants.In particular, the region involving the ß-turn 17–24is organized differently. Analysis of the differences foundin this region suggests that the insertion (or deletion) atposition 18 of the variant amino add sequences is particularlyimportant in determining the differential cross-reactivity.To test this proposal, residue 18 was deleted in one erabutoxinusing sitedirected mutagenesis, and the biological propertiesof the resulting mutant were examined. We found that full antigenicitywas restored in the previously unrecognized variant. The implicationsof this finding are discussed.     

Association and dissociation kinetics of bobwhite quail lysozyme with monoclonal antibody HyHEL-5

The anti-hen egg lysozyme monoclonal antibody HyHEL-5 and itscomplexes with various species-variant and mutant lysozymeshave been the subject of considerable experimental and theoreticalinvestigation. The affinity of HyHEL-5 for bobwhite quail lysozyme(BWQL) is over 1000-fold lower than its affinity for the originalantigen, hen egg lysozyme (HEL). This difference is believedto arise almost entirely from the replacement in BWQL of thestructural and energetic epitope residue Arg68 by lysine. Inthis study, the association and dissociation kinetics of BWQLwith HyHEL-5 were investigated under a variety of conditionsand compared with previous results for HEL. HyHEL-5–BWQLassociation follows a bimolecular mechanism and the dissociationof the antibody–antigen complex is a first-order process.Changes in ionic strength (from 27 to 500 mM) and pH (from 6.0to 10.0) produced about a 2-fold change in the association anddissociation rates. The effect of viscosity modifiers on theassociation reaction was also studied. The large differencein the HEL and BWQL affinities for HyHEL-5 is essentially dueto differences in the dissociation rate constant.     

An active single-chain antibody containing a cellulase linker domain is secreted by Escherichia coli

Single-chain antibodies consist of the variable, antigen-bindingdomains of antibodies joined to a continuous polypeptide bygenetically engineered peptide linkers. We have used the flexibleinterdomain linker region of a fungal cellulase to link togetherthe variable domains of an anti-2-phenyloxazolone IgGl and showhere that the resulting single-chain antibody is efficientlysecreted and released to the culture medium of Escherichia coli.The yield of affinity-purified single-chain antibody is 1 -2mg/1 of culture medium and its affinity and stability are comparableto those of the corresponding native IgG.