A directory of human germ-line V kappa segments reveals a strong bias in their usage

Towards the design of an antibody that recognises a given protein epitope

We have explored the possibility of designing repertoires of antibodies complementary to a given protein epitope, specifically the face of the ribonuclease inhibitor barstar that binds to the enzyme barnase. An antibody repertoire was created by mutation of ten residues in the hypervariable loops of a synthetic antibody fragment and displayed on filamentous bacteriophage. The positions of three of the ten residues of the antibody (VL 32, 50 and 94) were chosen to match a triangle of three negative charges on the face of barstar and mutated to favour residues of opposite charge or those with hydrogen-bonding potential. The other seven residues, chosen to allow for variation in the surface of interaction, were mutated at random. One of the antibody fragments isolated after selection of the repertoire (10(8) clones per library) was shown to bind to barstar with an affinity of 1.0x10(-7) M and the binding was competed by barnase. Furthermore, the binding of the antibody to barstar was highly sensitive to mutation of any of five residues of barstar known to contact barnase. This indicates that it may be possible, by a combination of design and selection, to build antibodies to a given epitope.

Immunoglobulin variable region sequences of human monoclonal anti-DNA antibodies

OBJECTIVE

Anti-DNA antibodies are believed to be important in the pathogenesis of systemic lupus erythematosus (SLE). Antibodies that bind specifically and with high affinity to dsDNA are most closely involved in tissue damage. Analysis of the sequences of the variable regions of human monoclonal anti-DNA antibodies is useful in defining the structural features that give rise to these binding properties. This article systematically reviews the evidence derived from such sequences.

METHOD

Previous reviews of this subject have been hampered by incomplete knowledge of the human immunoglobulin variable region repertoire. In this article, the original sequence data from reports of over 50 human monoclonal antibodies (mAb) are reinterpreted by alignment to the most similar alleles of the most similar germline genes. This allows accurate estimation of the site and nature of somatic mutations.

RESULTS

Human IgG monoclonal anti-DNA antibodies generally carry more mutations than IgM. In many cases these have been selected by an antigen-driven process. In many of the more specific, higher affinity dsDNA binders, there is an accumulation of basic residues in the complementarity determining regions. However, many exceptions to this rule exist, particularly among IgM mAb.

CONCLUSIONS

Unlike murine anti-DNA antibodies, these human mAb show little evidence for preferential use of particular V(H), V(K) and V(lambda) genes or families to encode antibodies of this specificity.

Primary sequence and molecular model of the variable region of a mouse monoclonal anti-Der p 1 antibody showing a similar epitope specificity as human IgE

BACKGROUND

Der p 1, a major mite allergen, elicits IgE antibody responses in 80% of patients suffering from dust mite allergy. Given the potent IgE eliciting properties of Der p 1, there is considerable interest in studying the molecular architecture of the variable (Fv) region of IgE antibodies specific for this allergen.

OBJECTIVES

IgE is present in human serum at extremely low concentrations, and as such it is practically impossible to purify sufficient quantities for structural studies. We have therefore sought to sequence and model a representative murine monoclonal (MoAb) anti-Der p 1 antibody, as a surrogate human IgE.

METHODS

The cDNA coding for the Fv region of an anti-Der p 1 MoAb (2C7), that mimics the binding of human IgE to Der p 1, was amplified by PCR, cloned and sequenced. The predicted amino acid sequences were then compared with a directory of human germline V-gene segments. Modelling of the Fv region of MoAb 2C7 was carried out using the extensive database of existing immunoglobulin structures in the Brookhaven PDB.

RESULTS

The MoAb 2C7 heavy chain showed greater than 70% homology with three members of the VH3 family, DP-35, DP-53 and DP-54. Similarly, the light chain showed greater than 70% homology with 11 VK sequences, including the VKII sequences DPK18, DPK19 and DPK28. A molecular model of the Fv region of MoAb 2C7 was generated and can be accessed from the EMBL databank.

CONCLUSIONS

Antibodies similar to MoAb 2C7 could be generated as part of the human repertoire. The availability of 3-dimensional model of MoAb 2C7, as a surrogate human IgE antibody, combined with further data on its epitope specificity, will facilitate studies into IgE antibody responses to Der p 1.

Clonal Analysis of a Human Antibody Response. III. Nucleotide Sequences of Monoclonal IgM, IgG, and IgA to Rabies Virus Reveal Restricted Vκ Gene Utilization, Junctional VκJκ and VλJλ Diversity, and Somatic Hypermutation

In previous work, we generated four IgM, five IgG1, and one IgA1 mAbs to rabies virus using B cells from four subjects vaccinated with inactivated rabies virus, a thymus-dependent (TD) mosaic Ag, and sequenced the mAb VHDJH genes. Here, we have cloned the VκJκ and VλJλ genes to complete the primary structure of the Ag-binding site of these mAbs. While the anti-rabies virus mAb selection of Vλ genes (2e.2.2 twice, DPL11, and DPL23) reflected the representation of the Vλ genes in the human haploid genome (stochastic utilization), that of Vκ genes (O2/O12 twice, O8/O18, A3/A19, A27, and L2) did not (p = 0.0018) (nonstochastic utilization). Furthermore, the selection of both Vκ and Vλ genes by the anti-rabies virus mAbs vastly overlapped with that of 557 assorted VκJκ rearrangements, that of 253 VλJλ rearrangements in λ-type gammopathies, and that of other Abs to thymus-dependent Ags, including 23 anti-HIV mAbs and 51 rheumatoid factors, but differed from that of 43 Abs to Haemophilus influenzae type b polysaccharide, a prototypic thymus-independent (TI) Ag. The anti-rabies virus mAb VκJκ and VλJλ segments displayed variable numbers of somatic mutations, which, in mAb58 and the virus-neutralizing mAb57, entailed a significant concentration of amino acid replacements in the complementarity-determining regions (p = 0.0028 and p = 0.0023, respectively), suggesting a selection by Ag. This Ag-dependent somatic selection process was superimposed on a somatic diversification process that occurred at the stage of B cell receptor for Ag rearrangement, and that entailed V gene 3′ truncation and N nucleotide additions to yield heterogeneous CDR3s.

Molecular characterization of five neutralizing anti-HIV type 1 antibodies: identification of nonconventional D segments in the human monoclonal antibodies 2G12 and 2F5

We have stabilized a panel of 33 hybridomas producing human monoclonal antibodies (MAbs) against HIV-1 gp160 and p24. Five of these antibodies were able to neutralize different HIV-1 isolates, and two of them (2F5 and 2G12) revealed remarkable potential to neutralize primary virus isolates of different clades in several in vitro tests. To determine whether a structural basis for neutralization could be identified, we analyzed the antibodies at the molecular level. This study reports the primary nucleotide and deduced amino acid sequences of the rearranged heavy and light chain V segments (VH, Vkappa) of the neutralizing MAbs (1B1, 1F7, 2F5, 2G12, and 3D5) and the nonneutralizing anti-gp41 MAb 3D6. Aligning the V segments with the nearest related germline genes illustrated the occurrence of somatic mutations. The neutralizing MAbs show mutational rates comparable to those of antibodies that appear in patients in whom the immune system is under constant antigenic pressure over a long period of time. In contrast, 3D6, which recognizes the immunodominant region on gp41, displays homologies as high as 97 and 98% compared with its VH and Vkappa germline genes. The diversity segments [D(H)] of 1B1, 1F7, 3D5, and 3D6 were assigned to single D(H) segments on the chromosomal D(H) locus. 2F5 presents a D(H) segment 52 nucleotides in length, which could be explained by fusion of two segments on the immunoglobulin heavy chain locus that have not yet been described as rearranged regions. 2G12 D(H) shows best homologies to a D(H) segment between D3-22 and D4-23. This D(H) segment could be the reason for the rare occurrence of antibodies competing with 2G12. Since this nearest related chromosomal region on the D(H) locus does not display recombination signals at the flanking regions, this segment is normally not taken into consideration as a site for immunoglobulin heavy chain rearrangement.

Design and use of a phage display library. Human antibodies with subnanomolar affinity against a marker of angiogenesis eluted from a two-dimensional gel

We report the construction and the use of a phage display human antibody library (>3 x 10(8) clones) based on principles of protein design. A large repertoire of functional antibodies with similar properties was produced by appending short variable complementarity-determining region 3 (CDR3) onto the two antibody germ line segments most frequently found in human antibodies. With this strategy we concentrated sequence diversity in regions of the antibody structure that are centrally located in the antigen binding site, while leaving residues in more peripheral positions available for further mutagenesis aimed at improving the affinity of the selected antibodies. In addition, the library was tested by selecting antibodies against six biologically relevant antigens. Using only 0.3 microg of antigen eluted from a two-dimensional gel spot, we isolated binders specific for the ED-B domain of fibronectin, a marker of angiogenesis. These antibodies recognize the native antigen with affinities in the 10(7)-10(8) M-1 range, and perform well in immunosorbent assays, in two-dimensional Western blotting and in immunohistochemistry. The affinity of one anti-ED-B antibody was improved by 27-fold by combinatorially mutating six strategically selected residues in the heavy chain variable domain. A further 28-fold affinity improvement could be achieved by mutating residues 32 and 50 of the light chain. The resulting antibody, L19, bound to the ED-B domain of fibronectin with very high affinity (Kd = 54 pM), as determined by real-time interaction analysis with surface plasmon resonance detection, band shift analysis, and by competition experiments with electrochemiluminescent detection.

Novel diagnostic and therapeutic strategies with genetically engineered human antibodies

Human antibodies can be generated by recombinant technology. The human immunoglobulin repertoire can be tapped in an effective manner by the so called V gene phage display technology. Using this technique the genes encoding the variable domains of an antibody of interest can be captured from the B cell repertoire. With these V genes tailor-made immune recognition molecules can be obtained by a 'design and build' strategy. These novel developments will lead to the introduction of a manifold of antibody based therapies into the clinic in coming decades. Recombinant antibody technology also provides unique opportunities to study the molecular structure of antibody variable domains against blood cell antigens. Such studies might possibly lead to the development of new therapies for antibody mediated blood cell destruction.

Molecular Analysis of Polyreactive Monoclonal Antibodies from Rheumatic Carditis: Human Anti-N-Acetylglucosamine/Anti-Myosin Antibody V Region Genes

Anti-myosin Abs are associated with inflammatory heart diseases such as rheumatic carditis and myocarditis. In this study, human cross-reactive anti-streptococcal/anti-myosin mAbs 1.C8, 1.H9, 5.G3, and 3.B6, produced from peripheral blood lymphocytes of patients with rheumatic carditis, and mAb 10.2.5, produced from a tonsil, were characterized, and the nucleotide sequences of their VH and VL genes were analyzed. Human mAbs 1.C8, 1.H9, 10.2.5, and 3.B6 reacted with human cardiac myosin while mAb 5.G3 did not. The mAbs were strongly reactive with N-acetyl-β-d-glucosamine, the dominant epitope of the group A streptococcal carbohydrate. mAb 1.H9 was moderately cytotoxic to rat heart cells in vitro in the presence of complement. The anti-myosin mAbs from rheumatic carditis were found to react with specific peptides from the light meromyosin region of the human cardiac myosin molecule. Anti-streptococcal/anti-myosin mAbs from normal individuals reacted with distinctly different light meromyosin peptides. The mAbs were encoded by VH3 gene segments V3-8, V3-23, and V3-30 and by the VH4 gene segment V4-59. The variable region genes encoding the anti-streptococcal/anti-myosin repertoire were heterogeneous and exhibited little evidence of Ag-driven somatic mutation.

Regulation of B cell development by variable gene complexity in mice reconstituted with human immunoglobulin yeast artificial chromosomes

The relationship between variable (V) gene complexity and the efficiency of B cell development was studied in strains of mice deficient in mouse antibody production and engineered with yeast artificial chromosomes (YACs) containing different sized fragments of the human heavy (H) chain and kappa light (L) chain loci. Each of the two H and the two kappa chain fragments encompasses, in germline configuration, the same core variable and constant regions but contains different numbers of unique VH (5 versus 66) or Vkappa genes (3 versus 32). Although each of these YACs was able to substitute for its respective inactivated murine counterpart to induce B cell development and to support production of human immunoglobulins (Igs), major differences in the efficiency of B cell development were detected. Whereas the YACs with great V gene complexity restored efficient development throughout all the different recombination and expression stages, the YACs with limited V gene repertoire exhibited inefficient differentiation with significant blocks at critical stages of B cell development in the bone marrow and peripheral lymphoid tissues. Our analysis identified four key checkpoints regulated by VH and Vkappa gene complexity: (a) production of functional mu chains at the transition from the pre B-I to the pre B-II stage; (b) productive VkappaJkappa recombination at the small pre B-II stage; (c) formation of surface Ig molecules through pairing of mu chains with L chains; and (d) maturation of B cells. These findings demonstrate that V gene complexity is essential not only for production of a diverse repertoire of antigen-specific antibodies but also for efficient development of the B cell lineage.

Somatic mutations in the Ig variable region genes and expression of novel Cmu-germline transcripts in a B-lymphoma cell line ("Farage") not producing Ig polypeptide chains

Non-Hodgkin's B-lymphomas (B-NHL) are a very heterogeneous group of B-cell neoplasias originating from the germinal centers of lymphatic follicles. Thus, they represent a suitable experimental model to study the molecular basis of certain key events which take place in the lymphatic follicles, including somatic hypermutation and heavy chain isotypic switch. An unusual B-NHL cell line ("Farage") not producing Ig polypeptide chains was previously shown to rearrange its IgH and Igkappa genes and transcribe seemingly normal size mu and kappa mRNAs. In an attempt to characterize the phenotype of Farage cells better and to elucidate the molecular basis of the failure of Farage cells to synthesize Ig chains, we sequenced its VH and Vkappa rearranged gene segments by PCR and RT-PCR. It was found that both V genes are somatically, heavily mutated compared to their germline counterparts. In addition, this rearranged VDJ gene of the heavy chain is not transcribed. Instead, the Farage cells express a low level of a new family of germline transcripts starting with a VH like sequence, continuing with a small segment of the 3'VH germline flanking region, and ending within the Cmu region. These transcripts lack D and J segments and do not contain the open reading frame of the full-length Cmu protein. Thus, Farage cells fail to produce mu heavy chains due to silencing of the expression of the conventional VDJCmu transcript and expression of unusual Cmu-germline transcripts. In contrast to the IgH genes, the rearranged VJ gene of Farage is transcribed and gives rise to a full-size kappa-mRNA. This transcript, however, is not translated to a full-length kappa-chain, as it contains a stop codon in its coding region. All the above show that Farage cells are unable to produce Ig polypeptide chains, due to somatic mutations altering the kappa-chain gene, and mutations and/or regulatory events that shutoff the transcription of the IgH gene. The heavily mutated Vkappa and Vkappa genes found, support the conclusion that the Farage cell line originated either from germinal center cells or from the mantle zone of the lymphoid follicle.

A phage display approach for rapid antibody humanization: designed combinatorial V gene libraries

The development of a new strategy for antibody humanization is described. This strategy incorporates key recognition sequences from the parental rodent antibody into a phage display-based selection strategy. The original sequences of the third complementarity-determining regions (CDRs) of heavy and light chains, HCDR3 and LCDR3, were maintained and all other sequences were replaced by human sequences selected from phage-displayed antibody libraries. This approach was applied to the humanization of mouse mAb LM609 that is directed to human integrin alphav beta3 and has potential applicability in cancer therapy as an antiangiogenic agent. We demonstrate this approach (i) provides a rapid route for antibody humanization constraining the content of original mouse sequences in the final antibodies to the most hypervariable of the CDRs; (ii) generates several humanized versions with different sequences at the same time; (iii) results in affinities as high as or higher than the affinity of the original antibody; and (iv) retains the antigen and epitope specificity of the original antibody. The production of multiple humanized variants may present advantages in the selection of antibodies that are more readily expressed on a large scale and could be important in therapeutic regimens that call for long-term treatment with antibodies in which antiidiotypic responses might be avoided by administration of alternative antibodies.

Characterization of anti-mouse Fc gamma RII single-chain Fv fragments derived from human phage display libraries

BACKGROUND

Few antibodies are available to study the function of the Fc gamma RII murine immunoglobulin receptor. Human phage display libraries represent a potential source of single-chain Fv (sFv) to facilitate the study of the Fc gamma RII murine immunoglobulin receptor.

OBJECTIVES

To isolate human sFv specific for mouse Fc gamma RII.

STUDY DESIGN

Two human phage display libraries were selected for reactivity to mouse Fc gamma RII. Those human anti-mouse Fc gamma RII sFv that were derived from the libraries were characterized with respect to kinetics, cellular binding, epitope specificity and amino acid sequence.

RESULTS

Nine anti-mouse Fc gamma RII sFv molecules were isolated from two human phage display libraries (Marks et al., J Mol Biol 1991;222:581-597; Sheets et al., Proc Natl Acad Sci USA, in press). Surface plasmon resonance (SPR) analysis revealed that the human anti-mouse Fc gamma RII sFv had off-rates ranging from 10(-2) to 10(-3) s-1, with KD values calculated to range between 10(-7) and 10(-9) M. The binding of the FITC-labeled human anti-mouse Fc gamma RII sFv to mouse peritoneal neutrophils was not detected by flow cytometry, due to the rapid off-rates of these monomeric proteins. However, when the human anti-mouse Fc gamma RII sFv were coated on yellow-green latex particles, all of the human sFv were found to specifically bind to mouse peritoneal neutrophils. Deglycosylation of mouse Fc gamma RII did not diminish the binding of these sFv, suggesting that the sFv molecules recognize a polypeptide epitope on murine Fc gamma RII. In contrast, denaturation of mouse Fc gamma RII dramatically reduced the binding of the human sFv, suggesting that the epitopes are conformational. Sequence analysis of the human anti-mouse Fc gamma RII sFv revealed a high degree of structural similarity among the nine sFv. The DP73 VH gene segment was utilized by four of the nine sFv, while seven of the nine sFv contained the DPL16 V lambda gene segment. The sequence similarities between these sFv suggested that several of the human sFv may recognize a common epitope on mouse Fc gamma RII. Epitope mapping studies demonstrated that eight of the nine human anti-mouse Fc gamma RII sFv recognized overlapping epitopes. All of these human anti-mouse Fc gamma RII sFv competed with the 2.4G2 rat monoclonal anti-mouse Fc gamma RII/III antibody for binding with mouse Fc gamma RII, suggesting that the targeted epitopes reside in or near the Fc binding pocket of mouse Fc gamma RII.

CONCLUSIONS

The availability of novel sFv recognizing mouse Fc gamma RII will facilitate the study of receptor triggering events. Such sFv may prove useful to engage murine Fc gamma RII for targeted cytotoxicity or immunization strategies.

Efficient construction of a large nonimmune phage antibody library: the production of high-affinity human single-chain antibodies to protein antigens

A large library of phage-displayed human single-chain Fv antibodies (scFv), containing 6.7 x 10(9) members, was generated by improving the steps of library construction. Fourteen different protein antigens were used to affinity select antibodies from this library. A panel of specific antibodies was isolated with each antigen, and each panel contained an average of 8.7 different scFv. Measurements of antibody-antigen interactions revealed several affinities below 1 nM, comparable to affinities observed during the secondary murine immune response. In particular, four different scFv recognizing the ErbB2 protein had affinities ranging from 220 pM to 4 nM. Antibodies derived from the library proved to be useful reagents for immunoassays. For example, antibodies generated to the Chlamydia trachomatis elementary bodies stained Chlamydia-infected cells, but not uninfected cells. These results demonstrate that phage antibody libraries are ideally suited for the rapid production of panels of high-affinity mAbs to a wide variety of protein antigens. Such libraries should prove especially useful for generating reagents to study the function of gene products identified by genome projects.

Ligand recognition by alpha beta T cell receptors

While still incomplete, the first data concerning the biochemistry of T cell receptor-ligand interactions in cell-free systems seem to have considerable predictive value regarding whether a T cell response is strong or weak or suppressive. This data will help considerably in elucidating the mechanisms behind T cell responsiveness. Also of great interest are the first structures of T cell receptor molecules and, particularly, TCR-ligand complexes. These appear to confirm earlier suggestions of a fixed orientation for TCR engagement with peptide/MHC and should form the basis for understanding higher oligomers, evidence for which has also just emerged. We conclude with an analysis of the highly diverse CDR3 loops found in all antigen receptor molecules and suggest that such regions form the core of both TCR and antibody specificity.

Development of more efficacious antibodies for medical therapy and diagnosis

Two procedures for improving the efficacy of medically important antibodies are described. The first procedure is designed to reduce the immunogenicity of nonhuman antibodies to the barest minimum--the "humanization" is accomplished by transplanting only the specificity-determining residues of the nonhuman antibody onto a human antibody template. The second procedure is designed to permit the easy production of multispecific/multivalent antibodies via heterodimer formation of electrostatically complementary Fc regions.

Sequence of the spacer in the recombination signal sequence affects V(D)J rearrangement frequency and correlates with nonrandom Vkappa usage in vivo

Functional variable (V), diversity (D), and joining (J) gene segments contribute unequally to the primary repertoire. One factor contributing to this nonrandom usage is the relative frequency with which the different gene segments rearrange. Variation from the consensus sequence in the heptamer and nonamer of the recombination signal sequence (RSS) is therefore considered a major factor affecting the relative representation of gene segments in the primary repertoire. In this study, we show that the sequence of the spacer is also a determinant factor contributing to the frequency of rearrangement. Moreover, the effect of the spacer on recombination rates of various human Vkappa gene segments in vitro correlates with their frequency of rearrangement in vivo in pre-B cells and with their representation in the peripheral repertoire.

Preparation of phage antibodies to the ED-A domain of human fibronectin

The fibronectin (FN) isoform containing the alternative spliced ED-A domain is much more expressed in fetal, tumoral, and regenerating tissues than in normal adult tissues. The ED-A containing FN is up-regulated by numerous cytokines, such as TGF-beta, and, although in normal adult liver the ED-A domain is undetectable, in regenerating rat liver the expression of ED-A is increased and mediates the conversion of fat storing cells to myofibroblasts. Here we describe the selection from a phage display library and the characterization of human antibody fragments directed against the ED-A sequence of FN. As they can be easily radiolabeled with 32P, these antibodies are very highly sensitive reagents for the determination of ED-A levels in tissues and biological fluids; in fact, use of these scFv induced a more than 10-fold increase in sensitivity with respect to the murine monoclonal IST-9. The possibility of preparing a range of human engineered antibodies should facilitate the development of antibody reagents with suitable pharmacokinetics, valency, functional affinity, and effector functions and that could be useful for clinical purposes.

Evidence That Amyloidogenic Light Chains Undergo Antigen-Driven Selection

AL amyloidosis is characterized by fibrillar tissue deposits (amyloid) composed of monoclonal light chains secreted by small numbers of indolent bone marrow plasma cells whose ontogenesis is unknown. To address this issue and to provide insights into the processes that accompanied pathogenic light chain formation, we isolated the complete variable (V) regions of 14 light (VL) and 3 heavy (VH) chains secreted by amyloid clones at diagnosis (10 Bence Jones and 4 with complete Igs, 9 λ and 5 κ) by using an inverse polymerase chain reaction-based approach free of primer-induced biases. Amyloid V regions were found to be highly mutated compared with the closest germline genes in the databases or those isolated from the patients' DNA, and mutations were not associated with intraclonal diversification. Apparently high usage of the λIII family germline gene V λIII.1 was observed (4 of 9 λ light chains). Analysis of the nature and distribution of somatic mutations in amyloid V regions showed that there was statistical evidence of antigen selection in 8 of 14 clones (7 in VL and 1 in VH). These results indicate that a substantial proportion of the amyloid clones developed from B cells selected for improved antigen binding properties and that pathogenic light chains show evidence of this selection.

Production and selection of antigen-specific fully human monoclonal antibodies from mice engineered with human Ig loci

The ability to produce highly specific fully human monoclonal antibodies to human antigens has potential significant applications to human therapy. This review describes the creation of novel mouse strains engineered to produce a diverse repertoire of fully human antibodies in the absence of mouse antibodies. These mouse strains have been generated by introducing megabase-sized human immunoglobulin loci, containing the majority of the human antibody gene repertoire, in nearly germline configuration, into mice deficient in mouse antibody production. The mice produce high levels of human IgMkappa and IgGkappa antibodies with a diverse adult-like repertoire. Upon immunization with multiple human antigens the mice generate high affinity, antigen-specific fully human monoclonal antibodies with neutralization activity. Comparison of these mice to other strains containing limited human antibody gene repertoire underscores the importance of the large number of variable genes for faithful reproduction of functional and diverse human antibody response in mice.

Characterization of human IgG1 monoclonal antibody against gangliosides expressed on tumor cells

A human IgG1.k monoclonal antibody (MAb) designated GMA1 was developed by fusing pooled lymph node lymphocytes from cancer patients with the human lymphoblastoid cell line, SHFP-1. The GMA1 MAb reacted with several melanoma and neuroblastoma cell lines. Normal tissue derived from human brain and tumor-cell lines derived from colon, ovary, and breast were not reactive. FACS analysis performed using live cells demonstrated that the antibody recognizes a cell-surface antigen. Enzyme immunoassay (EIA) and thin layer chromatography (TLC) immunostaining with purified gangliosides indicated that the antibody has specificity for the major tumor associated gangliosides GD3, GM3, and GD2. GMA1 heavy and light chain genes were isolated by RT-PCR and a recombinant derivative of this human antibody was expressed in Chinese hamster ovary (CHO) cells. High-level antibody synthesis and secretion was achieved using a vector designed to maximize expression. FACS analysis and TLC immunostaining indicated recombinant GMA1 reacted with human tumor cell lines and gangliosides GD3, GM3, GD2 in a manner similar to the antibody produced by the hybridoma cell line, demonstrating that the specificity of the antibody was not altered during molecular cloning.

The signature of somatic hypermutation appears to be written into the germline IgV segment repertoire

We present here a unifying hypothesis for the molecular mechanism of somatic hypermutation and somatic gene conversion in IgV genes involving reverse transcription using RNA templates from the V-gene loci to produce cDNA which undergoes homologous recombination with chromosomal V(D)J DNA. Experimental evidence produced over the last 20 years is essentially consistent with this hypothesis. We also review evidence suggesting that somatically generated IgV sequences from B lymphocytes have been fed back to germline DNA over evolutionary time.

Recombination signature of germline immunoglobulin variable genes

In human and mouse, the germline contains a tandem array of highly homologous variable (V) gene elements which encode part of the antigen-binding region of the antibody protein. During evolution this array apparently arose by gene duplication followed by diversification of duplicated genes via point mutation and recombination. Analysis of germline V gene sequences using a novel algorithm shows that major recombination sites coincide with the borders of the leader intron and the cap site, consistent with the hypothesis that over evolutionary time cDNA derived by reverse transcription of pre-mRNA in B lymphocytes has recombined with germline DNA.

Relationship between autoantibody epitopic recognition and immunoglobulin gene usage

An immunodominant region recognized by serum autoantibodies has been defined on the autoantigen thyroid peroxidase (TPO) using recombinant human TPO-specific Fab or a panel of mouse MoAbs. We have now analysed the epitopic relationships between the four recombinant Fab that identify the A and B domains of the TPO immunodominant region and (i) the mouse TPO MoAb as well as (ii) nine new TPO-specific Fab isolated independently. Competition between mouse MoAbs and recombinant Fab for binding to 125I-TPO revealed three patterns. First, for MoAbs 15, 59, 64 and 18, TPO binding was virtually abolished (approximately 90%) by Fab which define the A domain of TPO, with less inhibition by B domain Fab. Second, for MoAbs 2, 9 and 47, the Fab competed much less for TPO binding, and, when detectable, inhibition was predominantly with B domain Fab (65-20%). Third, for MoAbs 53, 30, 1, 24 and 40, none of the Fab competed effectively for 125I-TPO binding. Thus, the epitopes for MoAbs 18, 59, 64 and 15 correspond to those of the A domain defined by the human Fab, and the epitopes for MoAbs 2, 9 and 47 correspond to those of the B domain. In the second part of the study, competition studies demonstrated that the epitopes of nine new Fab corresponded to those of the four Fab that define the immunodominant region. For four new Fab, TPO binding was inhibited to a greater extent by B- than by A-domain Fab (65-95% versus <50%). In contrast, for five new Fab the A-domain Fab were more effective inhibitors (approximately 90%) than the B-domain Fab. In addition, consistent with previous observations, all five new Fab with 02/012 kappa L chains, but none of the new Fab with non-O2/O121 chains, interacted with A-domain epitopes. In conclusion, we have established the epitopic relationships between recombinant human Fab and mouse MoAbs that define the TPO immunodominant region on TPO. Further, analysis of recombinant TPO Fab isolated from patients on three continents strengthens the paradigm of a relationship between autoantibody epitopic recognition and immunoglobulin gene usage.

Lack of Intraclonal Diversification in Ig Heavy and Light Chain V Region Genes Expressed by CD5+IgM+ Chronic Lymphocytic Leukemia B Cells: A Multiple Time Point Analysis

To analyze the modalities of clonal expansion of chronic lymphocytic leukemia (CLL) cells, we sequenced at multiple time points the V(D)J genes expressed by CD5+IgM+CLL B cells in three patients. All three V(D)J gene sequences were found to be point mutated. The mutation frequency in the Ig VH (3.96 × 10−2 and 2.41 × 10−2 change/bp) and Vκ and Vλ (6.67 × 10−2 and 1.74 × 10−2 change/bp) genes of two CLLs (1.19 and 1.32, respectively) was similar, and higher than that in the corresponding gene segments of the third CLL (1.69; 3.4 × 10−3 and 6.67 × 10−3 change/bp). In all three CLLs, there was no preferential representation of nucleotide changes yielding amino acid replacement (R mutations), nor was there any preferential segregation of R mutations within the Ig V gene complementarity-determining regions. In all three CLLs, the somatic mutations were all identical in multiple Ig VHDJH transcripts at any given time point, and were all conserved at multiple time points throughout a 2-yr period. The lack of concentration of R mutations in the complementarity-determining regions and the lack of intraclonal heterogeneity suggest that Ag may no longer be able to play a significant role in the clonal expansion of these cells. This conclusion would be strengthened further by the germline configuration of the bcl-1 and bcl-2 proto-oncogenes that are translocated in neoplastic B cells that display significant traces of intraclonal diversification and Ag-dependent selection, such as B-prolymphocytic leukemia and low grade follicular non-Hodgkin lymphoma.

Lupus-specific antibodies reveal an altered pattern of somatic mutation

The F4 idiotype is a heavy chain determinant expressed almost exclusively on IgG immunoglobulins and is highly associated with specificity for double-stranded DNA. Since high-titered F4 expression is present predominantly in sera of patients with systemic lupus erythematosus (SLE), we thought F4+ IgG antibodies might constitute a useful subset of immunoglobulins in which to investigate lupus-specific alterations in variable (V) region gene expression or in the process of somatic mutation. This molecular analysis of F4+ B cell lines generated from lupus patients demonstrates that despite the strong association of F4 reactivity with specificity for native DNA, there is no apparent VH gene restriction. Furthermore, VH gene segments encoding these antibodies are also used in protective immune responses. An examination of the process of somatic mutation in F4+ antibodies showed no abnormality in frequency of somatic mutation nor in the distribution of mutations in complementarity-determining regions or framework regions. However, there was a decrease in targeting of mutations to putative mutational hot spots. This subtle difference in mutations present in these antibodies may reflect an intrinsic defect in mutational machinery or, more likely, altered state of B cell activation that affects the mutational process and perhaps also negative selection.

Molecular single-cell analysis reveals that CD5-positive peripheral blood B cells in healthy humans are characterized by rearranged Vkappa genes lacking somatic mutation

B cells expressing the CD5 cell surface antigen are involved in certain B cell malignancies and autoimmune diseases. From studies in the mouse, it emerged that CD5+ B cells represent a separate lineage of B lymphocytes that, in contrast to conventional (CD5-) B cells, are not driven into T cell-dependent immune responses in which rearranged variable (V) region genes are diversified by somatic hypermutation. Against this background it came as a surprise that human disease-involved CD5-positive autoreactive B cells as well as B cell chronic lymphocytic leukemias can harbor somatically mutated V region genes. Recent V gene analyses on CD5+ B cells in healthy adults did not give rise to a clear picture about the fraction of somatically mutated among all CD5+ B cells. In this work we used a molecular single-cell analysis to determine reliably the frequency of mutated CD5+ B cells in healthy humans: single, kappa light chain-expressing CD5+ peripheral blood B cells were isolated by flow cytometry, and rearranged Vkappa genes were amplified by PCR. From one donor, CD5+CD19+ B cells were analyzed. Since CD5+ B cells were found among IgM+IgD+ and IgM+IgD- cells (but almost not among class-switched cells) from two other donors, individual cells corresponding to these IgM-expressing subsets were investigated separately. The sequence analysis of rearranged Vkappa genes revealed that most if not all CD5+ B cells in healthy humans carry unmutated V region genes. From one of the donors, a novel polymorphic Jkappa2 gene segment was identified. To explain the discrepancy between the frequent occurrence of disease-associated somatically mutated CD5+ B cells and the low incidence or absence of somatic mutation in normal CD5+ B cells, we speculate that CD5+ B cells usually do not participate in germinal center reactions, but if they occasionally do so, they may be at an increased risk to become involved in autoimmune diseases or B cell malignancies.

The epitopic "fingerprint" of thyroid peroxidase-specific Fab isolated from a patient's thyroid gland by the combinatorial library approach resembles that of autoantibodies in the donor's serum

A new thyroid peroxidase (TPO)-specific Fab (KM1) was obtained from an immunoglobulin gene combinatorial library of patient KM containing L chain genes amplified with a single "promiscuous" V kappa oligonucleotide primer. The KM1 L chain is encoded by a mutated B3 gene (V kappa IV family). Another mutated B3 L chain had been identified previously in a TPO-specific Fab (WR1.223) isolated from a different patient (WR). In contrast to patient KM, the WR L chains were amplified with a panel of V kappa family-specific primers. Both KM1 and WR1.223 bind TPO with high affinity (approximately 1 x 10(-9) M) and interact with an epitope in the B domain of the TPO immunodominant region. TPO-specific Fab previously isolated from a WR combinatorial library constructed with the promiscuous V kappa primer recognised the TPO A domain and none used a B3-like L chain. Remarkably, for both patients, Fab isolated from L chains generated with the promiscuous V kappa primer had epitopic profiles similar to autoantibodies in the donor's serum (KM-B domain; WR-A domain). Our data indicate that the promiscuous primer preferentially amplifies the dominant L chain present in vivo. However, to obtain a relatively rare Fab (such as the B domain Fab from WR), family-specific kappa primers are required. These findings provide insight into the relationship between TPO autoantibody gene usage, epitopic recognition, and the effectiveness of the combinatorial library approach.

Does base composition help predispose the complementarity-determining regions of antibodies to hypermutation?

A survey of the base usage in genes coding for human antibodies reveals that more (A+T) and less (C+G) are found in the segments coding for the complementarity-determining regions, while the opposite is true for the segments coding for framework and constant regions. The possibility that this bias in base usage may contribute to hypermutation is explored.

Diffuse large cell lymphomas are derived from mature B cells carrying V region genes with a high load of somatic mutation and evidence of selection for antibody expression

In the Revised European American Lymphoma (REAL) classification, several subtypes of high-grade lymphomas were combined in the entity diffuse large cell lymphoma (DLL). In the present study, a total of 19 cases of DLL (10 cases of centroblastic lymphoma, 5 cases of mediastinal B cell lymphoma, 2 cases of immunoblastic lymphoma, 1 case of T cell-rich B lymphoma and one case of large cell anaplastic lymphoma) were analyzed for somatically mutated immunoglobulin V region genes. Somatic mutations are acquired in the course of the germinal center (GC) reaction and are thus found in GC B cells and their descendants, i.e. memory B cells. The V gene sequences revealed that the tumor cells of all five subtypes of DLL harbored mutated V region genes and are thus derived from antigen-experienced (post) GC B cells. This indicates that from the point of view of the stage of development of the tumor precursor, the combination of those five subtypes to one entity, i.e. DLL, seems reasonable. In some cases, an unusually high frequency of somatic mutations was detected. This may indicate that DLL are derived from GC B cells, which, due to transforming events, stayed in the GC for prolonged periods of time, thereby accumulating a high load of somatic mutation. An analysis of the mutation pattern suggests that the tumor clone or its precursor were selected for antibody expression while acquiring somatic mutations. The latter observation discriminates DLL from classical Hodgkin's disease, where we recently also observed a high load of somatic mutation within rearranged V region genes, but a frequent occurrence of crippling mutations.

Population studies of the human V kappa A18 gene polymorphism in Caucasians, blacks and Eskimos. New functional alleles and evidence for evolutionary selection of a more restricted antibody repertoire

Immunoglobulin gene polymorphisms are interesting because they reflect differences in the available antibody repertoire which may affect the susceptibility to specific infections. Until recently, the human V kappa gene, A18, was known as a nonfunctional gene only. In this study, we cloned and sequenced four apparently functional alleles and determined the gene frequencies in three well-defined populations: Danish Caucasians, eastern Greenland Eskimos and Mozambican blacks. The A18b allele that was recently described in Native American Navajos by Atkinson et al. was found in all three populations with gene frequencies of 8%, 45% and 23% in Caucasians, Eskimos and blacks, respectively. Conversely, the frequencies of the nonfunctional A18a allele were 92%, 55% and 57%. Further, three new A18 alleles, c, d, and e were found exclusively in blacks, among whom they had an total frequency of 19%. These data indicate that both the A18a and A18b alleles originated before the diversification of Africans and non-Africans 90,000 years ago, whereas the A18c, A18d and A18e alleles may have a more recent origin. The functionality of the A18b allele was documented by the demonstration of properly rearranged and somatically hypermutated A18b messenger RNA present in the blood lymphocytes of individuals carrying this allele. The expression clearly exceeded that of a known functional V gene, A2, indicating that functional A18 alleles contribute significantly to the available antibody repertoire. In this context, it is surprising that the functional A18b allele apparently has been negatively selected in the Caucasian population, among whom 85% completely lack a functional gene.

Ongoing Ig Gene Hypermutation in Salivary Gland Mucosa-Associated Lymphoid Tissue-Type Lymphomas

Salivary gland mucosa-associated lymphoid tissue (MALT) type lymphomas are typically indolent B-cell neoplasms that are often associated with Sjogren's syndrome. To better define the cell of origin and evaluate whether antigen receptor stimulation may be playing a role in tumor growth, the Ig heavy and light chain variable genes (VH and VL) expressed by five salivary gland MALT lymphomas were cloned and sequenced. Comparison to known germline sequences indicated that three of the lymphoma VH genes were derived from 51p1, a member of the VH1 family, while the other two used different VH gene segments from the VH3 family, 22-2B and HG19. All five of the VL genes belonged to the VkIII family, with three derived from Humkv325 and the other two from the Vg and Humkv328 genes. Numerous point mutations relative to the proposed germline genes were present in all of the lymphoma VH and VL genes. In addition, the VH and VL genes from each lymphoma showed intraclonal sequence heterogeneity indicative of ongoing somatic hypermutation. Because the process of Ig gene hypermutation is thought to occur at the germinal center stage of B-cell development, these findings suggest the MALT lymphoma cell of origin may be a germinal center B cell. Selection against mutations that result in replacement of amino acids suggested that Ig stimulation may be important for lymphoma growth. The possibility that antigen receptor stimulation may be involved in the growth of salivary gland MALT lymphomas is further suggested by the noted restricted use of VH and VL gene segments.

Immunoglobulin lambda light chain orphons on human chromosome 8q11.2

We have identified two V lambda genes outside the major lambda locus on chromosome 22q11.2, and shown that they reside on chromosome 8q11.2. One gene (Orphée1), hybridizing strongly to the V lambda probes, was sequenced and found to belong to the V lambda 8 family; the other gene (Orphée2) only hybridized weakly. Orphée1 was present in all individuals tested (140) from three different populations, and was also found in gorillas. We envisage that these genes were generated by duplication and translocation of the V lambda 8a gene (and a V lambda pseudogene) from the major locus, and that this event occurred before the evolutionary divergence of humans and gorillas. As there is no other evidence for V lambda genes outside the major locus, it appears that the human lambda locus has undergone considerably less evolutionary shuffling than either the human light chain kappa locus or the heavy chain locus.

The creation of diversity in the human immunoglobulin V(lambda) repertoire

Sequence diversity in the human antibody repertoire is generated in two steps: by the combinatorial assembly of V gene segments and by somatic hypermutation. Here, we have characterised these processes for the lambda (lambda) light chain using a library of 7600 lambda cDNA clones from peripheral blood lymphocytes. By hybridisation and sequencing we found that most lambda chains are derived from the cluster of V(lambda) segments closest to the J(lambda)-C(lambda) pairs and that there is considerable variation in the use of individual V(lambda) segments (ranging from 0.02% to 27%): three of the 30 functional V(lambda) segments encode half the expressed V(lambda) repertoire. As a result of these biases, sequence diversity in the primary repertoire is focused at the centre of the antigen binding site. By contrast, somatic hypermutation spreads diversity to the periphery. Comparison with the human kappa (kappa) light chain indicates that both kappa and lambda use the same strategy for searching sequence space and have almost identical patterns of diversity in the mature antibody repertoire.

Molecular mechanisms and selective influences that shape the kappa gene repertoire of IgM+ B cells

To analyze the human kappa chain repertoire and the influences that shape it, a single cell PCR technique was used that amplified Vkappa Jkappa rearrangements from genomic DNA of individual human B cells. More than 350 productive and 250 nonproductive Vkappa Jkappa rearrangements were sequenced. Nearly every functional Vkappa gene segment was used in rearrangements, although six Vkappa gene segments, A27, L2, L6, L12a, A17, and O12/O2 were used preferentially. Of these, A27, L2, L6, and L12a showed evidence of positive selection based on the variable region and not CDR3, whereas A17 was overrepresented because of a rearrangement bias based on molecular mechanisms. Utilization of Jkappa segments was also nonrandom, with Jkappa1 and Jkappa2 being overrepresented and Jkappa3 and Jkappa5 underrepresented in the nonproductive repertoire, implying a molecular basis for the bias. In B cells with two Vkappa Jkappa rearrangements, marked differences were noted in the Vkappa segments used for the initial and subsequent rearrangements, whereas Jkappa segments were used comparably. Junctional diversity was generated by n-nucleotide addition in 60% and by exonuclease trimming in 75% of the Vkappa Jkappa rearrangements analyzed. Despite this large degree of diversity, a strict CDR3 length was maintained in both productive and nonproductive rearrangements. More than 23% of the productive rearrangements, but only 7% of the nonproductive rearrangements contained somatic hypermutations. Mutations were significantly more frequent in Vkappa sequences derived from CD5- as compared with CD5+ B cells. These results document that the gene segment utilization within the Vkappa repertoire is biased by both intrinsic molecular processes as well as selection after light chain expression. Moreover, IgM+ memory cells with highly mutated kappa genes reside within the CD5- but not the CD5+ B cell compartment.

Comparison of strategies for the construction of libraries of artificial antibodies

Construction of libraries of artificial antibodies has been reported by several groups of investigators. Various forms of antibody fused to surface protein, cpII, are expressed on the surface of filamentous phage. Since phages that encode desired antibodies can be easily grown in Escherichia coli after selection with target antigens, the phage-display antibody system appears to be very useful for various biological purposes. In this brief review, recent progress in research into the production of artificial antibodies is summarized and the strategies used for the construction of libraries of artificial antibodies are compared.

Cell display library for gene cloning of variable regions of human antibodies to hepatitis B surface antigen

A novel cell display system was developed for cloning the variable region (V) genes of antigen-specific human antibodies. The system is based on an antibody library displayed on the surface of COS cells, using a plasmid vector designed to direct expression of membrane-bound antibodies. COS cells expressing antigen-specific antibodies were separated using a flow cytometer for their binding to a fluorescent dye-labeled antigen. To test the performance of this system. We cloned V genes of 4 antibodies directed against hepatitis B surface antigen (HBsAg) from a library prepared from peripheral blood lymphocytes of a vaccinated donor. These membrane-bound anti-HBsAg antibodies were easily converted to soluble forms, all of which showed a size similar to human serum IgG in SDS-PAGE and the same specific binding to HBsAg as membrane-bound forms in ELISA. All VH and VK gene segments of the 4 clones isolated in this study belonged to VHIII and VKI subgroups, respectively. These findings demonstrate the potential and selection capabilities of our cell display system for cloning the V genes of antigen-specific human antibodies.

An integrated vector system for the eukaryotic expression of antibodies or their fragments after selection from phage display libraries

Phage display is now an established method to select antibody fragments specific for a wide range of diverse antigens. In particular, isolation of human monoclonal antibodies has become a reality and for most purposes bacterial expression of the selected recombinant antibody fragments is sufficient. However, there are some cases where the expression of complete human immunoglobulin in mammalian cells is, if not essential, at least desirable. For this reason we have designed and constructed a set of mammalian expression vectors which permit facile and rapid cloning of antibody genes for both transient and stable expression in mammalian cells. Immunoglobulin genes may be cloned into these expression vectors as V regions or as Fabs for expression as either complete antibodies or as Fab fragments, using restriction sites which are rare in human V genes. All the important elements in the vectors--promoter, leader sequence, constant domains and selectable markers--are flanked by unique restriction sites, allowing simple substitution of elements. The vectors have been evaluated using the variable regions from the neutralizing anti-nerve growth factor (NGF) antibody, alphaD11, and the V regions from 2E10, a scFv selected from a scFv phagemid library.

Molecular cloning of anti-SS-A/Ro 60-kDa peptide Fab fragments from infiltrating salivary gland lymphocytes of a patient with Sjögren's syndrome

Anti-SS-A/Ro antibodies are commonly found in systemic autoimmune diseases such as Sjögren's syndrome and systemic lupus erythematosus, and some of these antibodies appear to be responsible for certain pathological lesions including congenital heart block in neonatal lupus. In this study, we generated three human antibody Fab fragments that specifically bind to SS-A/Ro 60-kd peptide from salivary gland lymphocytes of a patient with Sjögren's syndrome by using a phage-display technique. Sequence analysis demonstrated that two of the three Fab clones (E-42 and E-60) used homologous heavy chains derived from the germline VH gene DP73 in combination with different light chains which were derived from germline V kappa gene L6 and V lambda gene DPL23. The third Fab clone (E-56) used another heavy chain derived from the germline VH gene DP31 in combination with the identical light chain as that of E-42. All three Fab clones revealed a high number of somatic mutations that likely occurred in the context of antigen selection. These findings suggest the restricted usage of VH and VL genes of anti-SS-A/Ro antibodies in salivary gland lymphocytes of the patient.

Dimerization of Fab fragments enables ready screening of phage antibodies that affect hepatocyte growth factor/scatter factor activity on target cells

A number of applications of antibodies in diagnosis and therapy require multivalent reagents either because of the polymeric nature of the antigens or because biological activity depends on an effect on the formation of homodimeric species. Here, we report a procedure for mass screening of phage-derived monomeric antibody fragments that depend on valency for activity. As a model system, a set of 13 phage-derived human Fab fragments were first selected against mouse and human recombinant hepatocyte growth factor/scatter factor (HGF/SF), a high molecular weight polypeptide growth factor related to the blood protease plasminogen and involved in development and cancer. These Fab fragments were subsequently screened for an effect on HGF/SF activity either as monomeric fragments or after dimerization with a monoclonal antibody (9E10) directed against a peptide tag on the fragments. Fab were identified that either inhibited or enhanced HGF/SF activity on target cell lines, but dimerization was required for this effect. The approach proposed should facilitate mass screening of phage-derived antibody fragments that depend on multiple valency for activity.

Evidence for a Large Compartment of IgM-Expressing Memory B Cells in Humans

The recent finding of somatically mutated μ heavy chain transcripts in human peripheral blood (PB) B lymphocytes suggests that T-dependent B-cell memory might not be restricted to class-switched cells. We provide here evidence that IgM-only PB B cells are likely to be the IgM-expressing counterpart of classical (IgM−IgD−) memory B cells in humans. As shown by molecular single cell analysis, most IgM-only cells carry mutated V region genes, like class-switched cells. Although both subsets represent populations of nonactivated, resting cells, they express higher levels of Ig mRNA than naive (IgM+IgD+) B cells. IgM-only and class-switched cells are CD38−CD77−, and mostly CD23−, thus neither resembling germinal center nor naive B cells. Because many IgM-expressing B cells located in secondary lymphoid tissues resemble IgM-only PB B cells in terms of cell phenotype, we propose that the human lymphoid system contains a large compartment of IgM-expressing memory cells. Moreover, these cells seem to represent the nonmalignant counterparts of IgM-expressing tumor cells in sporadic Burkitt's lymphoma, MALT lymphoma, monocytoid B-cell lymphoma, and diffuse large-cell lymphoma that were found to harbor somatically mutated V genes.

Sequence and structure specific antibodies from phage display libraries

A large combinatorial phage display library was panned against five nucleic acid antigens, calf thymus DNA, poly[d(GC)], poly[d(AT)], poly(dA) x poly(dT) and poly(rA) x poly(dT). After the third and fourth rounds of panning, many positive clones were selected against poly[d(GC)], poly(dA) x poly(dT) and poly(rA) x poly(dT). The specificity of these antibodies was tested by both direct and competitive solid phase radioimmune assays. All the clones derived from panning with poly[d(GC)] were non-specific and bound to all nucleic acids. The poly(rA) x poly(dT) derived clones were specific for single-stranded nucleic acids, with some sequence preferences, and the poly(dA) x poly(dT) derived clones showed considerable specificity for this antigen. The sequences of these phage-derived antibodies showed no similarities with DNA-binding antibodies from other sources. Even after six rounds of panning no positive clones were detected which bound to poly[d(AT)] and after seven rounds only two were derived from panning with calf thymus DNA. Therefore, sequence- and structure specific antibodies can be recovered from phage display libraries but not all sequences may be represented in the repertoire.

Functional transplant of megabase human immunoglobulin loci recapitulates human antibody response in mice

We constructed two megabase-sized YACs containing large contiguous fragments of the human heavy and kappa (kappa) light chain immunoglobulin (Ig) loci in nearly germline configuration, including approximately 66 VH and 32 V kappa genes. We introduced these YACs into Ig-inactivated mice and observed human antibody production which closely resembled that seen in humans in all respects, including gene rearrangement, assembly, and repertoire. Diverse Ig gene usage together with somatic hypermutation enables the mice to generate high affinity fully human antibodies to multiple antigens, including human proteins. Our results underscore the importance of the large Ig fragments with multiple V genes for restoration of a normal humoral immune response. These mice are likely to be a valuable tool for the generation of therapeutic antibodies.

Myeloma VL and VH Gene Sequences Reveal a Complementary Imprint of Antigen Selection in Tumor Cells

In multiple myeloma, sequence studies of VH genes used to encode clonal Ig in neoplastic plasma cells have shown a common pattern of extensive somatic hypermutation. A further consistent feature of these VH sequences is a complete lack of intraclonal variation. These findings indicate that the malignant cell arises at a mature, postfollicular stage of B-cell development. However, only a minority of cases have a distribution of somatic mutations in VH consistent with a prior role for antigen in selecting the B cell of origin. To complement these studies, and to take further the investigation of a role for antigen in the clonal history of myeloma, we have investigated tumor-derived VL sequences from bone marrows of 15 patients. All sequences (9Vκ and 6Vλ) were potentially functional and 5 of 15 had evidence for N-region additions. All had undergone extensive somatic hypermutation, and showed no intraclonal variation. In 4 of 15 cases, the distribution of mutations revealed a significant (P < .05) clustering of replacement mutations in the CDR sequences, indicating a role for VL in selection by antigen. Comparison with the VH sequences used by the same tumor cells showed that, if significant clustering was present, it was in either VH or VL, but not both. Altogether, 10 of 15 V-regions showed evidence for antigen selection, suggesting that the B cell of origin has behaved as a normal germinal center B cell. Deductions concerning a role for antigen selection may require both VH and VL sequences for validation.

T cell receptor biochemistry, repertoire selection and general features of TCR and Ig structure

T cell recognition is a central event in the development of most immune responses, whether appropriate or inappropriate (i.e. autoimmune). We are interested in reducing T cell recognition to its most elemental components and relating this to biological outcome. In a model system involving a cytochrome c-specific I-Ek restricted T cell receptor (TCR) derived from the 2B4 hybridoma, we have studied the interaction of soluble TCR and soluble peptide-MHC complexes using surface plasmon resonance. We find a striking continuum in which biological activity correlates best with the dissociation rate of the TCR from the peptide-MHC complex. In particular, we have found that weak agonists have significantly faster off-rates than strong agonists and that antagonists have even faster off-rates. This suggests that the stability of TCR binding to a given ligand is critically important with respect to whether the T cell is stimulated, inhibited or remains indifferent. It also suggests that the phenomenon of peptide antagonists might be explained purely by kinetic models and that conformation, either inter- or intramolecular, may not be a factor. We have also studied TCR repertoire selection during the establishment of a cytochrome c response, initially using an anti-TCR antibody strategy, but more recently using peptide-MHC tetramers as antigen-specific staining reagents. These tetramers work well with either class I or class II MHC-specific TCRs and have many possible applications. Lastly, we have also tried to correlate the structural and genetic features of TCRs with their function. Recent data on TCR structure as well as previous findings with antibodies suggest that both molecules are highly dependent on CDR3 length and sequence variation to form specific contacts with antigens. This suggests a general "logic' behind TCR and Ig genetics as it relates to structure and function that helps to explain certain anomalous findings and makes a number of clear predictions.

Human B cell biology

Human B lymphocytes share one major distinctive feature with B cells of other higher animals, namely the ability to generate and secrete immunoglobulins. These highly specialized proteins are capable of tremendous diversity, and thereby account for much of our immune protection against invading organisms. Despite the great potential diversity possible in the specificities of immunoglobulin molecules, however, the binding of antibody to antigen initiates a limited spectrum of biologically important effector functions, such as complement activation and/or adherence of the immune complex to receptors on leukocytes. A variety of mechanisms have been elucidated that account for this, not all of which are shared by the different types of animals capable of making these proteins. The purpose of this chapter is to review the genetic, developmental, and physiologic mechanisms critical for human B cell expression of immunoglobulin.

Myeloma VL and VH Gene Sequences Reveal a Complementary Imprint of Antigen Selection in Tumor Cells

In multiple myeloma, sequence studies of VH genes used to encode clonal Ig in neoplastic plasma cells have shown a common pattern of extensive somatic hypermutation. A further consistent feature of these VH sequences is a complete lack of intraclonal variation. These findings indicate that the malignant cell arises at a mature, postfollicular stage of B-cell development. However, only a minority of cases have a distribution of somatic mutations in VH consistent with a prior role for antigen in selecting the B cell of origin. To complement these studies, and to take further the investigation of a role for antigen in the clonal history of myeloma, we have investigated tumor-derived VL sequences from bone marrows of 15 patients. All sequences (9Vκ and 6Vλ) were potentially functional and 5 of 15 had evidence for N-region additions. All had undergone extensive somatic hypermutation, and showed no intraclonal variation. In 4 of 15 cases, the distribution of mutations revealed a significant (P < .05) clustering of replacement mutations in the CDR sequences, indicating a role for VL in selection by antigen. Comparison with the VH sequences used by the same tumor cells showed that, if significant clustering was present, it was in either VH or VL, but not both. Altogether, 10 of 15 V-regions showed evidence for antigen selection, suggesting that the B cell of origin has behaved as a normal germinal center B cell. Deductions concerning a role for antigen selection may require both VH and VL sequences for validation.

Affinity maturation of recombinant antibodies using E. coli mutator cells

BACKGROUND

Phage libraries can display repertoires of antibodies which are greater in number than the mammalian immune response. However, the selected antibodies often have low binding affinity to their target antigen or hapten (KD below 10(-6) M), which is characteristic of the primary immune repertoire. There is a need for procedures to mimic somatic hypermutation through antigen driven affinity maturation, thereby increasing the affinity of selected immunoglobulins.

OBJECTIVE

To investigate the effectiveness of mutation and affinity selection of recombinant antibody genes with mutator E. coli cells, incorporating phage-display strategies.

STUDY DESIGN

Unique human scFvs were selected from a naive Fd-phage library. These genes were mutated by propagation in mutD5 mutator E. coli cells (mutD5-FIT) which were competent for Fd (M13) based phagemid transfections and generated point mutations (transversions and transitions) in the scFv genes. Individual phage-displayed scFvs were affinity selected from the mutation library and were assayed as soluble scFvs by ELISA and BIAcore for binding to antigen.

RESULTS

The in vivo mutation of phage-displayed scFvs in E. coli mutD5-FIT, combined with affinity selection against antigen, produced scFv molecules with improved binding activity. The point mutations which resulted in single amino acid substitutions frequently produced ten fold increases in apparent binding affinity. Structural comparisons revealed that these point mutations were in framework regions (adjacent to the CDRs) and within the CDRs. In one case the apparent affinity of an anti-glycophorin scFv after mutation in the VL framework region close to CDR3 increased by 10(3). However, this increase in apparent affinity was accompanied by an increased propensity to dimerise and form aggregates.

CONCLUSIONS

A strategy for the rapid affinity maturation of scFv and Fab antibody fragments has been developed which utilises mutator strains of E. coli and incorporates phage display of antibody repertoires (libraries).

Characteristics of human antibody repertoires following active immune responses in vivo

Possibilities to develop human monoclonal antibody specificities have recently been much facilitated by improvements of human hybridoma technology but even more so by the emerging phage-display technique. However, until recently very little has been known about the characteristics at the molecular level of the induced, T cell-dependent human antibody response, frequently targeted by these techniques. Rather, the major part of available sequence information has been related to tumor-derived or autoreactive antibodies. We have now investigated high affinity, monospecific, human antibody repertoires as developed by hybridoma technology. The VH region gene usage among such in vivo-induced repertoires is in only some respects similar to that found in the total B cell population. A limited number of heavy-chain variable segment loci account for the majority of all induced antibodies. A particular VH gene locus (4-34) frequently employed by peripheral B cells and associated with autoreactive antibodies was rarely used by the induced repertoire. Furthermore, in particular antigen systems, V region usage differs from the total available repertoire, and heavy-chain CDR3 is generally longer among antibodies induced against foreign protein antigens than in the average B cell population. Light-chain gene usage is often restricted to just a few dominant genes frequently found among B cells in general. In comparison, variable regions derived by phage-display technology in some antigen systems display even longer heavy-chain CDR3 than hybridoma-derived antibodies. This technique also appears to select a different set of germline genes preferentially (both with respect to VH and JH) as compared to hybridoma technology. In summary, the T cell-dependent antibody response against foreign antigens appears to differ from the average circulating B cell in several ways, and thus does not seem to represent a random selection of the available repertoire.

Characterization of human variable domain antibody fragments against the U1 RNA-associated A protein, selected from a synthetic and patient-derived combinatorial V gene library

This is the first study describing recombinant human antibody fragments directed to the U1 RNA-associated A protein (U1A). Three anti-U1A antibody fragments (Fab) were isolated from a semi-synthetic human Fab library and one anti-U1A single-chain variable fragment (scFv) was isolated from a library which was derived from the IgG-positive splenic lymphocytes of an autoimmune patient. Competition studies with autoantibodies against the U1 small nuclear ribonucleoprotein (snRNP) particle from patients with systemic lupus erythematosus (SLE) and SLE-overlap syndromes revealed that U1A binding of these antibody fragments can be inhibited by about 40% of the patient sera. All antibody fragments recognized the native U1 snRNP in immunoprecipitation assays. Two of three Fab clones as well as the scFv clone derived from the repertoire of an autoimmune patient use the same heavy chain germ-line gene DP-65. Epitope mapping revealed that these three clones appear to recognize an identical epitope domain present on the C-terminal RNP motif of the U1A protein. The DP-65 heavy chain gene is used in less than 1% of the B cells in healthy individuals, while three out of four anti-U1A antibody fragments use this gene. This points to a restricted VH gene usage in the case of U1A, suggesting that the DP-65 heavy chain has a natural shape complementarity to the U1A protein.