Structural studies of human gamma-G-myeloma proteins of different antigenic subgroups and genetic specificities

A comparison of the ability of the human IgG1 allotypes G1m3 and G1m1,17 to stimulate T-cell responses from allotype matched and mismatched donors

The immunogenicity of clinically administered antibodies has clinical implications for the patients receiving them, ranging from mild consequences, such as increased clearance of the drug from the circulation, to life-threatening effects. The emergence of methods to engineer variable regions resulting in the generation of humanised and fully human antibodies as therapeutics has reduced the potential for adverse immunogenicity. However, due to differences in sequence referred to as allotypic variation, antibody constant regions are not homogeneous within the human population, even within sub-classes of the same immunoglobulin isotype. For therapeutically administered antibodies, the potential exists for an immune response from the patient to the antibody if the allotype of patient and antibody do not match. Allotypic distribution in the human population varies within and across ethnic groups making the choice of allotype for a therapeutic antibody difficult. This study investigated the potential of human IgG1 allotypes to stimulate responses in human CD4(+) T cells from donors matched for homologous and heterologous IgG1 allotypes. Allotypic variants of the therapeutic monoclonal antibody trastuzumab were administered to genetically defined allotypic matched and mismatched donor T cells. No significant responses were observed in the mismatched T cells. To investigate the lack of T-cell responses in relation to mismatched allotypes, HLA-DR agretopes were identified via MHC associated peptide proteomics (MAPPs). As expected, many HLA-DR restricted peptides were presented. However, there were no peptides presented from the sequence regions containing the allotypic variations. Taken together, the results from the T-cell assay and MAPPs assay indicate that the allotypic differences in human IgG1 do not represent a significant risk for induction of immunogenicity.

Human Gm, Km, and Am allotypes and their molecular characterization: a remarkable demonstration of polymorphism

Human immunoglobulin allotypes are antigenic determinants (or "markers") determined serologically, classically by hemagglutination inhibition, on the human immunoglobulin (IG) heavy and light chains. The allotypes have been identified on the gamma1, gamma2, gamma3, and alpha2 heavy chains (they are designated as G1m, G2m, G3m, and A2m allotypes, respectively), and on the kappa light chain (Km allotypes). Gm-Am allotypes are inherited in fixed combinations, or Gm-Am haplotypes, owing to the linkage of the human IGHC genes (IGHG3, IGHG1, IGHA1, IGHG2, IGHG4, IGHE, and IGHA2 from 5' to 3' in the IGH locus on chromosome 14). Gm and Am allotypes have been one of the most powerful tools in population genetics and very instrumental in molecular characterization of the human IGHC genes (gene conversion, copy number variation, gene order). They represent a major system for understanding immunogenicity of the polymorphic IG chains, in relation with amino acid and conformational changes. The correlation between G3m allotypes and amino acid changes has been possible with the sequencing of many alleles of the IGHG3 gene, from individuals from different populations and with known allotypes. In this chapter, we integrate genetics and sequence data and provide an updated overview of the Gm-Am haplotypes and Km allotypes. We propose, for the first time, a complete elucidation of the G3m allotypes, illustrated by the "IMGT G3m allele butterfly" concept that allows a graphical representation of the G3m alleles (variants of a gene expressing a given set of allotypes). Knowledge of allotypes is important in antibody engineering and humanization of monoclonal antibodies to improve immunotherapy.

FURTHER STUDIES ON THE gammaG-HEAVY CHAIN GENE COMPLEXES, WITH PARTICULAR REFERENCE TO THE GENETIC MARKERS Gm(g) AND Gm(n)

The recently described Gm (g) and Gm (n) genetic markers of the γG3- and γG2-subgroups of γ-globulin were characterized in detail primarily through studies of myeloma proteins, their polypeptide chains and fragments. Antisera derived from rabbits, non-human primates and rheumatoid arthritis patients gave identical results. This contrasted with the Gm (b) system where the rabbit antisera react with a different genetic determinant (b⁰) than the sera from rheumatoid arthritis patients (b). The Gm (g) and Gm (n) antigens were detected both by precipitin analysis and by hemagglutination inhibition. The Gm (g) antigen was not associated with any of the other genetic antigens of the γG3-proteins which all belonged in the Gm (b) class. The genes for the latter were always allelic to the gene coding for Gm (g), with that for Gm (b⁰) constantly present when that for Gm (g) was absent. The Gm (g) and Gm (n) markers were of particular value in tracing the various gene complexes made up of the closely linked subgroup genes. Further support was gained for the concept that the different gene complexes of various population groups arose primarily through crossing-over. The Gm^g and Gm^b genes for the γG3-subgroup were extremely closely linked to those for the γG1-subgroup. However the Gm (n) marker indicated that the γG2-subgroup genes were probably further separated on the chromosome. Additional evidence was obtained for the γG2-γG3-γG1-order of the subgroup cistrons. Among the wide range of gene complexes a new type (γG2,—,γ/G1) was described. This complex appeared to have a deletion of the γG3-cistron. Lower levels of γG3-globulin were found in the sera of the individuals with this gene in the heterozygous state. The possibility that this unusual complex arose through an unequal nonhomologous crossing-over is discussed.

The C-terminal sequences of the heavy chains of human immunoglobulin G myeloma proteins of differing isotopes and allotypes

The sequences of the C-terminal octadecapeptides obtained by cyanogen bromide cleavage of the gamma-chains of myeloma proteins of the four subclasses, and a urinary heavy-chain-disease protein, have been determined. Although the sequences were markedly homologous, unique replacements were identified that distinguished between the gamma(2b), gamma(2c) and gamma(2d) subclasses. The data are in accord with the postulated existence of four genetic loci or cistrons, these having arisen by the process of gene duplication.

Core hinge of human immunoglobulin G3 as a system of four independent co-operative blocks

On the heat absorption curves of human immunoglobulin G3 (hIgG3) Kuc melting the scanning calorimetry method reveals a high-temperature (high-T(m)) peak of high intensity that is absent at the curves of other hIgG subclasses and IgG of other species. An analogous peak is observed also at the curves of melting of hIgG3 fragments containing the hinge segments. The high-T(m) peak is accompanied by characteristic changes in circular dichroism (CD) spectra at 220-230 nm. This allows relating the peak to the melting of a poly-L-proline conformation of an extremely long hIgG3 core hinge. The comparison of deltaH(cal) and deltaH(eff) testifies that the core hinge can be considered as a system of four independent co-operative blocks connected by flexible sites. These sites may provide additional flexibility to the hIgG3 molecule and also permit a transition of the rod-like shape of the hinge to compact globule-like conformation.

The specificity of synovial IgM rheumatoid factors (RF) for genetically engineered IgG antibodies is not affected by the method used to immortalize RF-producing B cells

Previously, we have shown that some rheumatoid factors (RFs) produced by Epstein-Barr virus (EBV)-transformed B cells from patients with rheumatoid arthritis (EBV-RA-RF) appear to be disease-specific autoantibodies that bind differently to defined epitopes on genetically engineered IgG antibodies, compared with RFs expressed by patients with Waldenstrom's macroglobulinaemia (Wmac-RFs) and healthy immunized donors (HID-RFs). To exclude the possibility that EBV transformation is responsible for these differences, we have now studied 15 other monoclonal IgM RFs from patients with RA that were produced by heterohybridoma-B-cell fusion (HRA-RFs). These HRA-RFs show the same gross specificity profiles for IgG as do their EBV-RA-RF counterparts. However, when the specificities of the HRA-RF and EBV-RA-RF panels were combined and compared with those RFs from patients with Wmac or HID, significant differences in binding specificity were again observed. Hybrid IgG3/4 antibodies made by exon shuffles between the IgG3 and IgG4 wild-type genes, and families of IgG variant antibodies made by site-directed mutagenesis, were used to map the fine specificity of HRA-RFs. The fine specificity of HRA-RFs were also similar to those of EBV-RA-RFs. These studies demonstrate that the method used for immortalizing IgM, RF-producing B cells from RA patients does not influence the specificity of the RFs obtained. Furthermore, some RFs expressed in RA have distinct and unique specificities, and may therefore represent disease-specific autoantibodies.

Investigation of the cooperative structure of Fc fragments from myeloma immunoglobulin G

The cooperative structure of Fc fragments prepared from myeloma human IgG1 was studied using scanning microcalorimetry and fluorescence at pH 4.2-8.0. It was shown that the first to be melted are CH2 domains whose interaction with each other is rather weak, while that with CH3 domains is strong. Then CH3 domains which form a single cooperative block are melted. The data for the structure of the Fc fragment in solution agree with the X-ray data according to which the interaction between CH2 domains is mediated by the carbohydrate moiety while the two CH3 domains are strongly associated. The presence of intensive CH2-CH3 interaction is a distinctive feature of the state of the Fc fragment in the given pH region as compared to that at pH <4.1 [Tischenko, V. M., et al. (1982) Eur. J. Biochem. 126, 517-521; Ryazantsev, S., et al. (1990) Eur. J. Biochem. 190, 393-399]. First, cis interactions greatly increase the free energy of the native structure stabilization in CH2 domains. Second, they decrease this energy for CH3 domains when compared to the state of the latter at pH 3.8 or within the Fc' fragment (the dimer of CH3 domains). The temperature and enthalpy of melting of CH2 domains coincide in all the samples studied despite heterogeneity of the carbohydrate moiety. Thus, it may be postulated that the conservative part of CH2 domains makes a cardinal contribution to the interaction of these domains with the carbohydrate moiety.

Mapping IgG Epitopes Bound by Rheumatoid Factors from Immunized Controls Identifies Disease-Specific Rheumatoid Factors Produced by Patients with Rheumatoid Arthritis

We have mapped the specificity of 28 monoclonal IgM rheumatoid factors (RFs) produced by heterohybridomas derived from five healthy blood donors immunized with mismatched human red blood cells (HID). The HID-RFs did not differ in their binding specificity for IgG epitopes from RFs that we previously analyzed from patients with Waldenström’s macroglobulinemia. However, IgM RFs produced by HID differed in their specificity for IgG compared with RFs expressed by patients with rheumatoid arthritis (RA-RFs). Only 1 of 28 HID-RFs bound all IgG subclasses (pan binding pattern) compared with 7 of 19 RA-RFs (p = 0.006). Three HID-RFs bound IgG3 compared with 9 RA-RFs (p = 0.007). Fine specificity differences were also identified between HID- and RA-RFs. Therefore, some RA-RFs show novel specificities for IgG not found among RFs from HID or individuals with Waldenström’s macroglobulinemia who do not have joint disease. These Abs with unique specificities may represent disease-specific autoantibodies in patients with RA. Nine of the HID-RFs from the same individual were clonally related, and several contained somatic mutations. Even when the clonally related HID-RFs were considered as one RF for comparison, the reactivity of the HID-RFs differed significantly from RA-RFs in their inability to recognize all IgG subclasses (p = 0.044) and recognize IgG3 (p = 0.041). Interestingly, among the clonally related RFs, considerable differences in the specificity for IgG were also observed, with the RF containing the most somatic mutations in VH and VL showing the most distinctive specificity changes. Therefore, these studies also demonstrate a correlation between somatic mutation and binding specificity.

pH6 antigen (PsaA protein) of Yersinia pestis, a novel bacterial Fc-receptor

It was found that recombinant pH6 antigen (rPsaA protein) forming virulence-associated fimbriae on the surface of Yersinia pestis at pH 6.7 in host macrophage phagolysosomes or extracellularly in abscesses such as buboes, is a novel bacterial Fc-receptor. rPsaA protein displays reactivity with human IgG1, IgG2 and IgG3 subclasses but does not react with rabbit, mouse and sheep IgG.

Behaviour of two IgG subclasses in transport of immunoglobulin across the human placenta

The human IgG subclasses are a family of highly related yet distinct molecules. Each of these four subclasses performs a discrete function within the human immune system. Previous studies have shown that one of these molecules, hIgG2, may be discriminated against in transport across the human placenta. We have aimed to elucidate the mechanism of this discrimination in order to gain a more comprehensive understanding of the process of transport of immunoglobulin across the human placenta. We have used a combination of immunocyctochemical localisation and biochemical analysis to detail the behaviour of hIgG2. Confocal laser scanning microscopy was used to compare the localisation of hIgG1 (chosen as representative of the efficiently transported subclasses) and hIgG2 in term and first trimester chorionic villi. Complementary evidence was provided from immunoblot analysis of isolated placental coated vesicles. The data presented here suggest that the hIgG2 is transported into the syncytiotrophoblast and appears to accumulate in the stroma of the villi. This leads us to the hypothesis that the fetal capillary endothelium is the cellular impediment to the transport of hIgG2 into the fetal circulation.

A key amino acid determining G3m(b) allotypic markers

A key amino acid substitution specific for allotypic Gm b markers, b0, b3, b5, were determined through sequence analyses of the pFc' fragments of IgG1 (Su) and IgG3 (Ba[Gm(g)], Bu[Gm(b1b3)], and Kam[Gm(b3st)]) myeloma proteins. The results indicate that serine at position 384 is responsible for the specificities. It is considered from crystallographic data of IgG-Fc [Deisenhofer et al. (1981) Biochemistry 20:2361] that two residues, the serine and isoleucine specific for IgG3 subclass at position 422, cause the structural change responsible for b markers. The two residues are close to each other in the CH3 domain. The allocations of the epitopes are estimated to be on two bends (residue no. 382-392, 411-424) between the beta-strands, whose amino acid residues are present in wide contact area [Novotny et al. (1987) Immunol. Today 8:26).

Structural study of the carbohydrate moieties of two human immunoglobulin subclasses (IgG2 and IgG4)

Asparagine-linked sugar chains were quantitatively released as oligosaccharides from human IgG2 and IgG4 myeloma proteins by hydrazinolysis followed by N-acetylation and NaB3H4 reduction. Each oligosaccharide was isolated by serial lectin column chromatography. Study of their structures by sequential exoglycosidase digestion and methylation analysis, revealed that all of them were of the bi-antennary complex-type containing Man alpha 1-6(+/- GlcNAc beta 1-4)(Man alpha 1-3)Man beta 1-4GlcNAc beta 1-4(+/- Fuc alpha 1-6)GlcNAc as core structures, and GlcNAc beta 1-, Gal beta 1-4GlcNAc beta 1- and Sia alpha 2-6Gal beta 1- in their outer chain moieties. However, the molar ratio of each oligosaccharide was different in each IgG sample, indicating that clonal variation is included in the sugar chain moieties of IgG molecules. One of the IgG2 contained four asparagine-linked sugar chains in one molecule, two on the Fc fragment and the remainder on the Fab fragment. The sugar chains in the Fc fragment contained much less galactose as compared with the Fab fragment.

Variations in the secretion of monoclonal antibodies by human-human hybridomas

A series of human-human hybridomas derived from a single fusion of UC 729-6 with lymph node lymphocytes were examined for the type and nature of macromolecules synthesized and secreted. One hybrid, VLN3G2, secreted fourfold higher IgG than that present in the cytoplasm over 4 days of growth, while the IgM distribution was opposite to that of IgG. VLN5C7, contrary to VLN3G2, contained several-fold more cytoplasmic IgG as well as IgM than the amounts secreted over the same period of time. Of the secreted IgG and IgM by both of these hybridomas, only the IgG showed immunoreactivity against target A431 cell surface antigen(s). Another hybridoma, termed VLN1H12, secreted immunoreactive IgM against target A431 cells, but no detectable IgG. Cytoplasmic proteins prepared by repeated freeze-thaw of the hybridoma cells, membrane proteins obtained by NP-40 extraction of the cell membrane, and secreted proteins present in the supernates of the various hybridomas were assayed by an enzyme-linked immunoassay (ELISA), to understand the discrepancy observed in the immunoglobulins of the cellular and extracellular compartments. The parental UC 729-6 cell line used in these cell fusions produced only trace amounts of immunologically inactive IgM and no detectable IgG. Molecular sieving column chromatography of these hybridoma supernates suggested the presence of intact IgG and IgM molecules and the absence of free heavy chains or hybrid antibodies containing both mu and gamma heavy chains. Intrinsic labeling of VLN3G2 hybridoma cells with 35S-methionine demonstrated the presence of not only a nonimmunoglobulin protein but also a small molecular weight protein-A-binding polypeptide in the culture supernatant. 35S-methionine-incorporated IgG and IgM antibodies, isolated from spent media, cytoplasm, and cell membranes of VLN3G2, also showed binding to protein-A-bearing bacteria. In conclusion, the differences observed in the amounts of secreted MAbs by the human-human hybridomas were not due to the decreased synthesis of these molecules.

Deleted IgG1 and IgG2 H chains in a patient with an IgG subclass imbalance

The serum of a female patient studied over 7 months initially showed, in addition to normal residual IgG, two abnormal IgG proteins and in the last 3 months of the disease showed only one abnormal IgG protein. Gm typing and serological subclass determinations revealed an imbalance of allelic forms within the IgG1 subclass during the disease. The IgG2 level remained markedly elevated throughout the study. The two abnormal IgG-related proteins were devoid of light chains. The abnormal gamma cathodic immunoglobulin and the abnormal beta 2-immunoglobulin were Fc-like and covalently disulphide linked, with molecular weights of 60,000 daltons (N terminal, SER) and 72,000 daltons (N terminal, GLY) respectively. The first belonged to the IgG1 subclass and the second to the IgG2 subclass. No abnormal proteolytic activity was noted and plasma cells reacted with anti-gamma-chain antisera only. We hypothesize that the molecular defect leading to the deleted chains was an early event, preceding the differentiation into plasma cells which produced the two IgG1 and IgG2 deleted H chains.

Gene deletion and gene duplication within the cluster of human heavy-chain genes. Selective absence of IgG sub-classes

Individuals with selective absence of IgG1 and IgG2 were discovered by testing for allotypes and isotypes of the respective sub-classes. These individuals were homozygous for sub-class deleted Gm-Am haplotypes, as shown by allotype studies in two families (Gm--;..;g;A2m1/Gm--;n;b;A2m1 and Gm--;n;b;A2m1/Gm--;..;b;A2m1) and by a population study of New Guineans (Gm fa;--;b;A2m2). The individuals with IgG1 sub-class deficiency showed elevation of IgG2, IgG4 and in particular of IgG3. Gene deletion can result from unequal crossing over which renders a complementary chromosome with a duplication of a sub-class gene. In one family, duplication of gamma 3 genes was observed to have happened in one of a twin pair. Quanitation of sub-classes in families with gamma 1- and with gamma 3-duplicated haplotypes did not show increased levels of the gene involved.

The first example of an isoallotype of human IgG located on the Fd fragment

An isoallotype of IgG is described that is an isotype on heavy chains of IgG3 and IgG4 proteins and an allotype for IgG1 proteins, in which it is associated with Glm(f). The isoallotypic determinant is located on Fd. Binding of light chains (independent of their type) is required for its antigenic expression, since isolated heavy chains are not reactive in inhibition tests. It is suggested that the coding for the amino acid sequence related to this isoallotype was present on the ancestral gene before the duplication of subclasses. The presence in many non-human primate species proves that it is strongly conserved in evolution.

The first isoallotype of human IgA proteins. An antigenic determinant occurring as allotype in the IgA2 subclass and as isotype in the IgA1 subclass

An antigenic determinant that occurs on all IgA proteins except on IgA2 proteins of the A2m2 allotype is described. The determinant is an isoallotype, being an isotype of IgA1 and at the same time an allotype of IgA2 proteins. It is proposed that the symbol nA2m2 be used for the new determinant, where n stands for isoallotype (formerly nonmarker) and A2m2 refers to the allotype to which it is antithetically related.

Antigenic heterogeneity among monoclonal IgM: observations on guinea-pig sera

Antisera against purified monoclonal IgM from patients with Waldenström's macroglobulinemia were raised in guinea-pigs. Absorption with normal human serum and heterologous macroglobulinemic serum resulted in an antiserum to a subgroup present in 5 of 54 sera.

Molecular defect in a gamma-2 heavy chain

The first gamma-2 (gamma2) heavy chain disease protein Gif has pyrroli-dinecarboxylic acid as its amino terminal residue, much of the Fd variable region, and an internal deletion of the heavy chain of about 100 residues corresponding to most of the Fd constant region. Normal sequence resumes with a glutamic acid residue at position 216 in the hinge region. This is the third gamma heavy chain disease protein where normal sequence resumes at the same position after the deletion.

Evidence for recent duplications among certain gamma globulin heavy chain genes

A survey of a large number of human sera with the heavy chain genetic markers of the gamma-globulin system has revealed an unusual gene complex which is inherited as a unit through two different families. The gene complex involves two pairs of gammaG1 genetic markers which ordinarily behave as homoalleles, Gm(z) and Gm(f) for the Fd part of gammaG1 molecules, and Gm(a) and non-a for the Fc part. Isolation of the gammaG1 fraction from the unusual sera demonstrated the presence of the important non-a antigen in the gammaG1 fraction. Through the use of immunoadsorbents it was shown that these antigens were not part of a single molecule but that separate molecules were involved. The accumulated evidence indicated that the appearance of such homoalleles on the same chromosome probably resulted from a recent gene duplication, giving rise to two gammaG1 cistrons on one chromosome.

Disulphide bridges of the heavy chain of human immunoglobulin G2

Amino acid sequences around the disulphide bridges of the heavy chain of an immunoglobulin of the gamma2 subclass have been studied. The protein was digested with pepsin and the digest fractionated by Sephadex. Screening of the eluate by one-dimensional electrophoresis of oxidized and unoxidized samples was used as an assay and pools of fractions were prepared. Identification by diagonal electrophoresis of several inter- and intra-chain disulphide bridges was done on the pooled fractions. The inter-heavy-chain bridged peptide included four cystine residues. Comparison with proteins of other human subclasses indicated that the intrachain bridges identified are the bridges of the invariable section of gamma2 heavy chains. The amino acid sequence of one cysteic acid peptide that may have been derived from the variable part of the molecule was determined. Partial reduction followed by carboxymethylation with radioactive iodoacetate of two proteins of the gamma2 class showed a number of labelled peptides that could be identified as being related to the inter-chain bonded cystine residues.

The molecular defect in a protein (CRA) found in gamma-1 heavy chain disease, and its genetic implications

A gamma1 protein, designated CRA, found in heavy chain disease contains three inter-heavy disulfide bridges instead of the two normally found in gamma1 immunoglobulin heavy chains. Almost the entire Fd fragment is missing. The NH(2) terminal region is heterogeneous and contains carbohydrate; after 11 residues that do not resemble any of the known heavy-chain variable-region subclasses, normal synthesis seems to be resumed at the same amino acid residue as in another heavy-chain disease protein (ZUC). This finding raises the possibility that glutamic acid at position 216 represents the beginning of the Fe region, synthesized under the direction of another gene.

Human immunoglobulin subclasses. Partial amino acid sequence of the constant region of a gamma 4 chain

The heavy chain of a human myeloma protein (Vin) belonging to the gamma4 subclass was subjected to tryptic digestion after reduction and carboxymethylation. Cyanogen bromide fragments were also prepared and all 19 tryptic peptides that account for one of them (the Fc-like fragment) were studied. Selected peptic peptides were isolated and provided evidence for the order of 15 of the tryptic peptides. In addition the sequence of two large peptic peptides derived from two sections of the molecule including all the interchain bridges is presented. Comparison with published data on other chains allows us to propose a sequence of gamma4 chains that extends from just before the presumed starting point of the invariable region (at about residue 113) to the C-terminal end of the chain (approx. residue 446), except for a section of about 50 residues. The results of the comparison suggest that the immunoglobulin subclasses have a recent independent evolutionary origin in different species. Implications for complement fixation and for the evolutionary origin of antibody diversity are also discussed.