Carl G.M. Magnusson

Comparison of Surface Properties of Human IgA, IgE, IgG and IgM Antibodies with Identical and Different Specificities

In this paper, the authors report the use of liquid‐liquid partition chromatography (LLPC) in an aqueous polyethylene glycol (PEG)/dextran two‐phase system to compare the surface properties (partition properties) of human antibodies and fragments thereof. The surface properties of all the monoclonal antibodies of different classes and subclasses investigated were within the same broad range as that observed for the polyclonal antibodies and no relationship was found between the exposed surfaces of the immunoglobulins (Ig) and their heavy chain isotype. Moreover, Fc fragments from various IgG1, 2 and 4 myeloma proteins were found to exhibit similar surface properties. Employing chimeric antibodies with identical variable regions the authors found that intact IgG1, 2 and 4 displayed identical surface properties, while the corresponding IgA1, IgA2, IgG3, IgE and IgM antibodies differed both from each other and from the IgGs. The surface properties of chimeric IgG3 could be made similar to those of the IgG1, 2 and 4 chimers by partially reducing the length of the hinge section, but new differences in surface properties appeared when their hinges were of similar length. Thus, LLPC can be used to detect differences or similarities in the surface properties of the antigen‐binding regions as well as the Fc part in the various isotypes. This can shed light on biological activities such as antigen binding and effector function.

N-glycosylation influences epitope expression and receptor binding structures in human IgE

Although human IgE is relatively rich in carbohydrates, there are few studies concerning their structural and functional importance. The low serum concentration of IgE has limited carbohydrate characterisation to a few IgE myeloma proteins. Four to six of the seven potential N-glycosylation sites in the constant region of the epsilon chain seem occupied together with some residual microheterogeneity. We have used a panel of 28 anti-Cepsilon2, 7 anti-Cepsilon3 and 18 anti-Cepsilon4 domain-specific anti-IgE mAbs, and rFcepsilonRIalpha to examine the effect of N-glycosylation on epitope expression of human IgE. Myeloma proteins IgE(DES)-kappa, IgE(ND)-lambda and IgE(UD)-kappa as well as polyclonal IgE were deglycosylated with PNGF and/or sialidase and tested in different ELISA. In all ELISA approaches, the reactivity of most domain-specific anti-IgE mAbs was independent of the glycosylation state of IgE(DES), except for one-third of the anti-Cepsilon2 mAbs. These mAbs reacted better with deglycosylated IgE(DES) in the order of treatment PNGF/sialidase > PNGF > or = sialidase > buffer control. In sharp contrast, the reactivity of IgE(DES) with rFcepsilonRIalpha was not influenced by sialidase but markedly reduced following PNGF or PNGF/sialidase treatment. These findings were neither myeloma restricted nor caused by aggregation, since monomeric IgE demonstrated the same reactivity pattern. Thus. N-glycosylation seems to influence both structure and function of human IgE. The oligosaccharides modulate epitope expression, mainly in the Cepsilon2-domain, as revealed by a subset of mAbs. They also promote subtle changes in the Cepsilon3-domain, leading to a reduced FcepsilonRIalpha binding. These findings suggest physiological implications of carbohydrates in human IgE.

Human IgG is substrate for the thioredoxin system: Differential cleavage pattern of interchain disulfide bridges in IgG subclasses

Thioredoxin, a 12,000 mol. wt protein with two redox-active cysteine residues, together with thioredoxin reductase and NADPH, may reduce protein disulfides and thereby act as a molecular probe of their structure and reactivity. Interchain and intrachain disulfides are structural elements in all immunoglobulins and therefore potential substrates for the reduced thioredoxin, Trx(SH)2. It was investigated whether such disulfides are cleaved in human polyclonal IgG and IgG subclass myeloma proteins by both the human and the Escherichia coli thioredoxin systems. The reactions were monitored spectrophotometrically as oxidation of NADPH at 340 nm, and by following the kinetics of the cleavage patterns with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS PAGE). Human IgG was a substrate for both prokaryotic and eukaryotic Trx(SH)2, which directly reduced IgG disulfides in a time and dose-dependent manner. Stoichiometric analyses indicated near-complete reduction of mainly inter-heavy light chain and inter-heavy chain disulfides, and SDS PAGE corroborated that the buried intrachain disulfides were left intact. The kinetic studies showed that IgG1, IgG3 and IgG4 were readily reduced into heavy and light chains via the formation of half-molecules with slightly slower kinetics for IgG4. In sharp contrast, IgG2 was not cleaved at all, even with increased thioredoxin concentrations or reduction times. A small but significant NADPH consumption by IgG2 myeloma proteins suggested reduction of a labile interchain or surface-exposed mixed disulfide. Consistent results were obtained for several IgG myeloma proteins within each subclass. The structural and functional importance of interchain disulfides in immunoglobulins suggests physiological implications of the thioredoxin system.

Autoantibodies of the IgM Class against a Human Myeloma Protein IgE(DES). I. Occurrence

We report on the natural occurrence of human serum antibodies with specificity for a human monoclonal myeloma IgE(DES). These antibodies were of the IgM class, based on their susceptibility to reduction, sedimentation in sucrose gradients, gel filtration and inhibition of agglutination by anti-IgM antiserum. Autoantibody levels were studied in several groups of patients by particle-counting immunoassay using latex particles to which purified monoclonal IgE(DES) was coupled. Only sera of patients suffering from parasitosis had significantly higher levels (p less than 0.0005) than those of healthy blood donors. Cord sera had very low levels, followed by an age-dependent increase during early infancy. There was no relation (p greater than 0.10) between serum IgE and IgM antibody level. On the other hand, significant relations between IgM anti-IgE(DES) levels and serum IgM (p less than 0.0005), serum IgA (p less than 0.001) and serum IgG (p less than 0.05) were observed suggesting that high levels were caused by or related to polyclonal activation of the immune system.

Monoclonal antibodies against human IgE. Identification of a determinant restricted to IgE of the lambda light-chain type.

The specificity of five monoclonal anti-IgE antibodies (Mabs) was studied in direct latex agglutination and agglutination-inhibition experiments by particle-counting immunoassay. Twenty IgE myeloma proteins and several purified D epsilon O-, D epsilon 2-containing pepsin and papain fragments of IgE-DES(kappa) were used in the evaluation. The results demonstrate two Mabs with isotypic specificity for two distinct epitopes of the Fc epsilon-fragment within the D epsilon 1- and D epsilon 2-determinants. One Mab recognized only the immunizing IgE protein and was directed against determinants on the Fd epsilon-fragment probably related to the idiotype. Anti-Em(1) allotypic Mabs recognized all 20 IgE myeloma proteins including two of Japanese origin and the Em(1)-allotype was confined to D epsilon-determinants. Interestingly, one Mab (ALE) reacted with all 8 IgE myeloma proteins of the lambda light-chain type but none out of 12 bearing kappa chains. ALE seems therefore to recognize a new marker on IgE besides the known idiotypic, allotypic and isotypic ones. These results illustrate that a critical specificity control of Mabs is always warranted. Moreover, one should be aware of possible interference in IgE assays from the kind of determinants recognized by ALE whenever intact IgE myeloma proteins are used to raise polyclonal antisera, to get immunosorbent-purified anti-IgE antibodies or when used as tracers and standards.

Antigen-binding sites dominate the surface properties of IgG antibodies

A new technique, liquid-liquid partition chromatography in an aqueous polyethylene glycol-dextran two-phase system, was used to detect differences in surface properties of antibodies with different antigen-binding sites. Employing well-characterized monoclonal IgG antibodies and Fab and Fc fragments thereof as well as chimeric IgG antibodies we found a remarkable relationship between structure of the antibody combining site and chromatographic behaviour. The surface properties of the IgG antibodies were dominated by those of its antigen-binding regions. In addition, our results indicated that the constant parts of the IgGs form similar scaffoldings, on to which CDRs of variable shapes and sizes are interspaced and constitute the major dominant differences in exposed surface properties.

Are Increased Levels of von Willebrand Factor in Chronic Coronary Heart Disease Caused by Decrease in von Willebrand Factor Cleaving Protease Activity?: A Study by an Immunoassay with Antibody Against Intact Bond 842Tyr–843Met of the von Willebrand Factor Protein

Low levels of von Willebrand factor (VWF) in von Willebrands disease type 2A (VWD 2A) result from increased cleavage of the bond 842Tyr-843Met in the VWF protein by VWF cleaving protease. On the other hand, decreased levels of this protease result in unusually large VWF in thrombotic thrombcytopenic purpura with thrombotic complications. In the present study, we designed an enzyme-liked immunosorbent assay of VWF cleaving protease activity to be used to assess whether the high levels of VWF in coronary heart disease (CHD) relate to a deficiency of this protease. Plasma samples with added Pefabloc and CaCl(2) were incubated with purified VWF coated on a microtiter plate. The remaining undigested multimers were quantified by an antibody directed against the intact 842Tyr-843Met bond of the VWF protein. Phosphate-buffered saline (PBS), instead of plasma, was used to obtain the initial level of coated undigested VWF. The reduction in absorbance at 492 nm between PBS and the unknown sample was taken as a measure of the protease activity. The assay was applied to plasma samples from 21 senior women with chronic CHD (cases) and 34 age-matched controls, as well as to samples from three patients with VWD 2A. The protease activity was similar in the two women groups (P>.05), although the VWF antigen levels were higher in the cases (P<.01). The VWD 2A patients had similar plasma levels of the protease to that in normal pooled plasma (NPP). In the senior controls, the protease activity correlated with the subject age (rs=-.61, P<.01, n=34). In conclusion, the developed method is specific for evaluating the protease function on VWF cleavage. The moderate increase of VWF antigen in chronic CHD may not depend on the protease activity. The age influence on the protease levels supports earlier findings of higher VWF levels in healthy older subjects. A high sensitivity of the mutated protein of VWF for the protease effect rather than increases in activity or quantity of the enzyme is probably involved in the pathogenesis of VWD 2A.

Major differences in specificity among naturally occurring human IgG-subclass anti-IgE autoantibodies.

BACKGROUND The specificity of naturally occurring IgG-subclass anti-IgE autoantibodies (a-IgE Ab) was studied by ELISA inhibition assay with a highly purified IgE-DES myeloma protein as a solid-phase antigen. Because the IgG isotypes differ in effector functions, detailed specificity studies of IgG subclass anti-IgE antibodies might clarify their putative role. METHODS Selected sera containing a high concentration of a single IgG a-IgE Ab subclass were allowed to react with purified immunoglobulins of all five classes including five different IgE myeloma proteins and papain-derived Fab epsilon and Fc epsilon fragments of IgE-DES. RESULTS The inhibition results indicated that IgG1, IgG2, and IgG3 a-IgE Ab reacted in a low-affinity reaction with IgE myeloma protein-restricted determinants because only IgE-DES was inhibitory. Moreover, the reacting epitopes were heat resistant (2 hours, 56 degrees C) and localized in the Fab epsilon-DES fragment. In sharp contrast, IgG4 a-IgE Ab reacted with high affinity to all five IgE myeloma proteins involving heat-susceptible epitopes confined to the Fc epsilon fragment. CONCLUSIONS It seems that a majority of IgG a-IgE Ab have a specificity for nonisotypic myeloma-restricted determinants whereas IgG4 a-IgE Ab are mainly isotype specific. These findings ought to be taken into account in the consideration of physiologic implications of IgG a-IgE Ab and in the interpretation of previous investigations in which intact IgE myeloma proteins have been used to detect IgG a-IgE Ab.

Characterisation of recombinant human IgE-Fc fragments expressed in baculovirus-infected insect cells.

IgE mediates its effector functions through the Fc region and it has been demonstrated that structures in the Cvarepsilon3-domain are crucial for FcvarepsilonR-binding. In order to further study structures of importance for the function of IgE, such as the carbohydrates, fragments with unmodified amino acid sequence were blunt-end cloned and expressed in baculovirus-infected Sf9 cells. Two fragments of human IgE, one encompassing the entire Fc-region (rCvarepsilon2-4) and a smaller one comprising the second and third domain (rCvarepsilon2-3), were produced and characterised with respect to epitope expression, glycosylation and FcvarepsilonR-binding. N-terminal analysis showed the expected VCSRDF-sequence of the Cvarepsilon2-domain, confirming correct cleavage of the secretion signal. Immunoblotting and gel permeation chromatography demonstrated that rCvarepsilon2-4 mainly formed a dimer, whereas rCvarepsilon2-3 also existed as monomers and oligomers. Endoglycosidase-treatment revealed that both fragments were N-glycosylated. In inhibition ELISA, rCvarepsilon2-4 and myeloma protein IgE(DES) reacted in a near equimolar way with monoclonal antibodies against the Cvarepsilon2-, Cvarepsilon3- and Cvarepsilon4-domains, whereas rCvarepsilon2-3 only reacted with anti-Cvarepsilon2 mAbs. Moreover, in FACS analysis rCvarepsilon2-4 interacted with two cell-lines constitutively expressing FcvarepsilonRI or FcvarepsilonRII, whereas rCvarepsilon2-3 lacked reactivity. A substantial reduction in the ability of rCvarepsilon2-4, following endoglycosidase treatment, to react with recombinant alpha-chain of the high affinity receptor for IgE in sandwich ELISA, indicated a role of N-linked oligosaccharides in stabilising receptor binding structures. Taken together, our results show that rCvarepsilon2-4, but not rCvarepsilon2-3, will be useful in studies of structure-function relationships of IgE, including the role of N-glycosylation, since it demonstrated appropriate epitope expression, conformation and ability to bind Fcvarepsilon-receptors.

Disproportional distribution of isotype and non-isotype-specific IgG subclass anti-IgE autoantibodies in human cord serum

The levels of naturally occurring IgG and IgG subclass anti-IgE autoantibodies (a-E Ab) were studied in 71 randomly collected cord sera with ELISA using solid-phase IgE-DES myeloma protein. IgG a-E Ab were present in all cord sera, and the range was 300-fold (1.8-540 arbitrary units/ML; median = 11.8). However, this activity is the sum of two major types of a-E Ab that we refer to as isotype-specific (IS) and non-isotype-specific (NIS) because they react with E-chain-specific and myeloma-restricted epitopes, respectively. These two types of a-E Ab were distinguished in IgG subclass a-E Ab inhibition ELISA for all samples using unheated IgE-PS and heated IgE-DES as inhibitors. IS and NIS a-E Ab were found among all four IgG subclasses though with different prevalences. No significant influence of gender was observed. Comparisons within each subclass indicate that NIS a-E Ab were more common than IS a-E Ab for the IgG1 (p < 0.00005), IgG2 (p = 0.07) and IgG3 (p < 0.005) subclasses while the reverse was true for IgG4 (p < 0.00005). In fact, only a minor part of the IgG1 (7%), IgG2 (13%) and IgG3 (9%) a-E Ab activity towards IgE-DES was IS while for IgG4 it was a major part (82%). Attempts to quantify IS and NIS IgG subclass a-E Ab using chimeric IgG subclass antihapten Ab suggested that the pool of IgG a-E Ab against IgE-DES was dominated by NIS a-E Ab particularly of the IgG1 subclass. The 75 percentiles for NIS IgG1, IgG2, IgG3 and IgG4 a-E Ab were 24, < 2, 2.1 and 0.27 ng/ml, respectively, whereas the corresponding figures for IS a-E Ab were 7, < 2, < 2 and 0.90 ng/ml. The findings raise questions on the definition of a-E Ab and suggest that caution should be exercised in the interpretation of any a-E Ab results that are detected with IgE myeloma proteins. The potentially more interesting IS a-E Ab may be overshadowed by the bulk of ubiquitous NIS a-E Ab and background. Consequently, discriminatory assays are necessary if the physiological implication of naturally occurring IS IgG subclass a-E Ab is to be elucidated and these considerations are not limited to studies of cord serum.