Polyreactive antigen-binding B cells in the peripheral circulation are IgD+ and B7−

Polyreactive antibodies are naturally occurring antibodies, primarily of the IgM isotype, that are capable of reacting with a wide variety of different self and non-self antigens. Previously, we reported that a B cell capable of making polyreactive antibody has Ig receptors on its surface that can bind different antigens. The present investigation was initiated to characterize these polyreactive antigen-binding B cells further. A panel of fluorescein isothiocyanate-labeled antigens (insulin, IgG Fc fragment or β-galactosidase) served as probes to select polyreactive antigen-binding B cells by cell sorting. Our experiment revealed that these polyreactive antigen-binding B cells were mainly of the IgD isotype. They expressed high levels of CD40 and major histocompatibility complex class II molecules, but little or no B7-1, B7-2, or Fas. In contrast to the binding of antigens to monoreactive receptors (usually high affinity), the binding of antigens to polyreactive receptors (usually moderate or low affinity) did not up-regulate the expression of B7-1 or B7-2. Antigens that bound to polyreactive receptors, however, were internalized and degraded, although not as efficiently as antigens that bound to monoreactive receptors. Despite the ability of these B7⁻ cells to process antigens, they were not able to activate T cells in a mixed leukocyte reaction. It is concluded that polyreactive antigen-binding B cells have properties that are consistent with the ability to induce immunological tolerance.     

Polyreactivity as an acquired artefact, rather than a physiologic property, of antibodies: evidence that monoreactive antibodies may gain the ability to bind to multiple antigens after exposure to low pH

Evidence is presented that monoreactive antibodies exposed to low pH may acquire the ability to bind to multiple antigens. M11, a murine, monoclonal, IgM(K) anti-goat IgG (GIgG) was purified from a hybridoma supernatant by elution at low pH from an anti-mu-Sepharose 4B affinity column. By measuring the specific antiGIgG activities and the affinity constants for the interactions of M11, pre- and post-affinity-purification, with GIgG, M11 was shown to be monoreactive before purification. Quite unexpectedly, however, the affinity-purified M11 reacted extensively with size-fractionated liver proteins when tested in an immunoblot, clearly indicating that it was polyreactive. It was concluded that the exposure to low pH had altered the M11 binding-site so as to allow it to bind to many different proteins. This phenomena provides an alternative basis for interpreting the polyreactivity observed following affinity-purification.     

Somatic mutations can lead to a loss of superantigenic and polyreactive binding

Although antibodies have been assumed to bind a specific antigen, evidence exists showing that a single antibody can bind to multiple unrelated antigens. We previously studied a human monoclonal antibody expressing a mutated form of the V(H)3-73 gene and displaying anti-tubulin activity in a patient suffering from an immunocytic lymphoma. Despite its expression of a V(H)3 family member, this immunoglobulin failed to react with protein A (SpA), suggesting that somatic mutations could account for its change in specificity. To examine this possibility, we produced recombinant Ig expressing germ-line (IgM kappa-Germ) or the mutated form (IgM kappa-PER) of the V(H)3-73 fragment. Comparison of the respective affinities of the two Ig demonstrated that IgM kappa-Germ restores its SpA-binding capacity, and shows a moderate decrease in its affinity for tubulin. Interestingly, IgM kappa-Germ displayed polyreactive specificity for different autoantigens, which contrasted to the monospecific binding of IgM kappa-PER to tubulin. These results suggest that the monoreactive IgM kappa-PER antibody may be derived from a natural polyreactive antibody through somatic mutation. In addition, both temperature modification and mild denaturation succeeded in recovering the polyreactivity of IgM kappa-PER, which favors the view that conformational modifications of the tertiary structure of antibodies may play a key role in the genesis of polyreactivity.     

Monoreactive and Polyreactive Rheumatoid Factors Produced by in Vitro Epstein-Barr Virus-Transformed Peripheral Blood and Synovial B Lymphocytes from Rheumatoid Arthritis Patients

The CD5 membrane molecule, initially identified as an exclusive T-cell marker, also defines a phenotypically and functionally distinct B-lymphocyte population. In normal individuals, CD5+ B cells are committed to secrete 'natural' polyreactive (auto)antibodies, that is antibodies, mainly IgM, endowed with multiple antigen-binding capabilities, including rheumatoid factor (RF) activity. At variance with this, in rheumatoid arthritis (RA) as well as in other autoimmune conditions, monoreactive autoantibodies binding with high affinity and specificity to a given self antigen have been isolated and the cells from which they originate differently related to the CD5+ B-lymphocyte subset. Here, we studied the proportions of CD5+ B cells and the characteristics, in terms of polyreactivity and monoreactivity, of RF produced by B lymphocytes in RA patients with classified disease activity. Our results suggest that patients with a more severe disease activity have higher proportions of CD5+ B cells and higher frequencies of B lymphocytes committed to secrete RF, with the characteristics of polyreactive antibodies. On the other hand, we did not find a significant difference between the proportions of peripheral B cells producing monoreactive RF in patients with high- versus patients with low-activity RA. However, in two highly active RA patients, we found that synovial fluid, compared with peripheral blood, was significantly enriched for (IgG and IgA) monoreactive RF-producing B cells.     

The use of in vitro immunisation, as an adjunct to monoclonal antibody production, may result in the production of hybridomas secreting polyreactive antibodies.

The possibility that immortalisation of in vitro immunised splenocytes may result in hybridomas secreting polyreactive antibodies was investigated. A panel of nine murine hybridomas, secreting IgM(kappa) anti-goat immunoglobulin G (anti-GIgG), was produced by immortalising splenocytes that had been immunised in vitro with GIgG. The ability of the corresponding monoclonal antibodies (Mabs) to bind multiple antigens was investigated using two techniques. First, the affinity constants characterising the interactions of each of the nine Mabs with each of a panel of six antigens were determined. Second, the specific anti-GIgG activities of each hybridoma supernatant and its corresponding affinity-purified IgM fraction were determined and compared. In total, these experiments indicated that eight of the nine hybridomas were polyreactive.     

Complete sequence of the genes encoding the VH and VL regions of low- and high-affinity monoclonal IgM and IgA1 rheumatoid factors produced by CD5+ B cells from a rheumatoid arthritis patient.

We have characterized the VH and VL genes of three low-affinity polyreactive and two high-affinity monoreactive IgM and IgA1 rheumatoid factor (RF) mAb generated using circulating CD5+ B cells from a single rheumatoid arthritis patient. We found that four and one RF mAb utilized genes of the VHIV and VHIII families, respectively. The VHIV gene usage by these RF mAb differs from the preferential VHIII, VHI, and, to a lesser extent, VHII gene usage by the IgM with RF activity found in patients with mixed cryoglobulinemia, Waldenstrom's macroglobulinemia, and other monoclonal gammopathies. In addition, in contrast to the preponderant kappa L chain usage by the RF in these patients, a lambda L chain was utilized by all RF mAb from our rheumatoid arthritis patient. Two RF mAbs utilized V lambda I, two V lambda IV, and one V lambda III L chains. The VH genes of the two low-affinity polyreactive IgM RF mAb were in germline configuration. When compared with the deduced amino acid sequence of the putatively corresponding genomic segment, the VH gene of the high-affinity monoreactive IgM RF mAb displayed five amino acid differences, all of which are in the complementarity determining regions (CDR), possibly the result of a process of somatic point mutation and clonal selection driven by Ag. The unavailability of the corresponding genomic VH segment sequences made it impossible to infer whether the VH genes utilized by the two IgA1 RF were in a germline or somatically mutated configuration. Sequencing of the genes encoding the H chain CDR3 (D segments) revealed that all three low-affinity polyreactive RF mAb displayed a much longer D segment (36-45 bases) than their high-affinity monoreactive counterparts (15-24 bases), raising the possibility that a long D segment may be one of the factors involved in antibody polyreactivity.     

Antigen-binding B cells and polyreactive antibodies

The present experiments were initiated to see if cells capable of binding antigens could make polyreactive antibodies. Fluorescein isothiocyanate-labeled self and non-self antigens were incubated with B cells from normal individuals. Antigenbinding cells were separated from non-antigen-binding cells by flow cytometry, immortalized with Epstein-Barr virus and analyzed at the clonal level for their capacity to make polyreactive antibodies. Four to six times more cells making polyreactive antibodies were found in the B cell subset that bound antigens than in the B cell subset that did not bind antigens. The majority of the polyreactive antibodies were of the immunoglobulin (Ig)M isotype. Immunoflow cytometry revealed that cell lines making polyreactive antibodies bound a variety of antigens (e.g., insulin, IgGFc and β-galactosidase), whereas cell lines making monoreactive antibodies bound only a single antigen. The binding of antigens to B cell lines that made polyreactive antibodies could be inhibited (range, 28%–57%) by both homogeneous and heterogeneous antigens. Both CD5⁺ and CD5^? antigen-binding B cells made polyreactive antibodies, but the frequency was slightly higher in the CD5⁺ antigen-binding (85%) as compared to the CD5^? antigen-binding (50%) population. Comparison of CD5⁺ B cells that bound antigens with CD5⁺ B cells that did not bind antigens showed that approximately 86% of the former, but only 15% of the latter, made polyreactive antibodies. It is concluded that cells capable of binding a variety of different antigens can make polyreactive antibodies and that antigen binding is a good marker for identifying polyreactive antibody-producing cells.     

Binding of polyreactive antibodies to self versus foreign antigens

Aside from their ability to bind to multiple antigens, the classic hallmark of polyreactive antibodies is their autoreactivity. Because of their ability to bind a number of common autoantigens, it has long been speculated that polyreactive antibodies are involved in the clearance of self-antigens. However, it has been demonstrated more recently that polyreactive antibodies are also capable of binding to some foreign and synthetic antigens. Although data regarding the relative reactivity of polyreactive antibodies with self versus foreign antigens is lacking, it is generally thought that both activities may play an important biological role. In this study, the relative reactivity of polyclonal human polyreactive IgM with human proteins and tissue extracts versus foreign (xenogeneic) proteins and tissue extracts was probed. The binding of affinity purified anti-ssDNA IgM from adult human serum and the binding of polyreactive IgM in human cord serum and in human adult serum were evaluated. Using competitive and direct binding assays, human polyreactive IgM were found to be generally more reactive with foreign (xenogeneic) proteins than with self or allogeneic proteins. These data shed light on the fundamental nature of polyreactive antibodies, and may provide additional insight into their putative biological roles.     

Induced antigen-binding polyreactivity in human serum IgA

Previous studies have shown that polyreactive antibodies play an important role in the frontline defense against the dissemination of pathogens in the pre-immune host. Interestingly, antigen-binding polyreactivity can not only be inherent, but also acquired post-translationally. The ability of individual monoclonal IgG and IgE antibodies to acquire polyreactivity following contact with various agents that destabilize protein structure (urea, low pH) or have a pro-oxidative potential (heme, ferrous ions) has been studied in detail. However, to the best of our knowledge this property of human IgA has previously been described only cursorily. In the present study pooled human serum IgA and two human monoclonal IgA antibodies were exposed to buffers with acidic pH, to free heme or to ferrous ions, and the antigen-binding behavior of the native and modified IgA to viral and bacterial antigens were compared using immunoblot and ELISA. We observed a dose-dependent increase in reactivity to several bacterial extracts and to pure viral antigens. This newly described property of IgA may have therapeutic potential as has already been shown for pooled IgG with induced polyreactivity.     

Structural properties of the glycoplasmanylinositol anchor phospholipid of the complement membrane attack complex inhibitor CD59.

Peripheral blood B cells from patients with systemic autoimmune disease and healthy volunteers were immortalized using EBV and the frequencies of B cell precursors that produced immunoglobulin class-specific antibodies against anti-nRNP, a specific marker for mixed connective tissue disease, were assessed using limiting dilution analysis. The frequencies of EBV-induced B cell precursors that produced IgG anti-nRNP were correlated closely with the serum titres of the corresponding autoantibodies, which indicates that B cell precursors that produced potentially pathogenic autoantibodies could be immortalized from the peripheral blood of the patients by EBV. In contrast, the frequency of EBV-induced B cell precursors that produced IgM anti-nRNP in patients with systemic autoimmune disease was comparable to that in healthy volunteers and greater than those that produced IgG and IgA anti-nRNP. Moreover, many of the clones that produced IgM antibodies against nRNP reacted with other autoantigens, such as double-stranded DNA, single-stranded DNA and rabbit IgG. These polyreactive IgM antibodies are believed to belong to the 'natural antibodies', to be coded by the germline immunoglobulin V genes, and to react with evolutionarily conserved structural cellular components, including nRNP. Our finding that nRNP is one of the target antigens for this polyreactive autoantibody may lead to the elucidation of the origin of the pathogenic IgG and IgA anti-nRNP antibodies found in sera from patients with systemic autoimmune diseases.     

Human monoclonal antibodies demonstrate polyreactivity for histones and the cytoskeleton.

Systemic lupus erythematosus (SLE) and other autoimmune diseases are characterized by immune responses to intracellular, highly conserved antigens such as DNA and histone. In this study, peripheral blood lymphocytes (PBL) from a patient with histone autoantibodies were used to prepare IgM human-human hybridoma cell lines. Indirect immunofluorescence (IIF) was used to identify monoclonal antibodies that bound to cytoskeletal and other cytoplasmic constituents. These supernatants did not bind double-stranded or single-stranded DNA. However, immunoblotting revealed that 7/20 hybridomas selected for their binding to cytoskeletal components produced antibodies that also bound mammalian and avian histones. When peptide fragments of histone were used in immunoblotting experiments, it was found that the monoclonal antibodies bound to the carboxyl terminus of H1, a region previously shown to bind autoantibodies from sera of patients with SLE and drug-induced lupus (DIL). When the amino acid sequences of histones and cytoskeletal components were compared using the Swiss-Prot protein data bank, it was confirmed that there are eight regions of similarity. While the significance of polyreactive human monoclonal antibodies to cytoskeletal components and histones is not understood at present, it is possible that the human histone antibodies represent polyreactive antibodies that arise through the mechanism of molecular mimicry.     

Molecular determinants of polyreactive antibody binding: HCDR3 and cyclic peptides

Human monoclonal antibody 63 (mAb63) is an IgM/lambda polyreactive antibody that binds to multiple self and non-self antigens. The molecular basis of polyreactivity is still unclear. The present study was initiated to prepare a recombinant Fab of mAb63 and use it to study the determinants involved in polyreactivity. The baculovirus system was employed to express large amounts of mAb63 Fab in Sf9 cells. Our experiments showed that infected Sf9 cells secreted a soluble 50-kD Fab heterodimer that bound to multiple self and non-self antigens. The antigen-binding activity of mAb63 Fab was inhibited by both homologous and heterologous antigens. To study in more detail the molecular determinants involved in polyreactivity, the heavy chain complementarity-determining region 3 (HCDR3), which is known to play a key role in the binding of monoreactive antibodies to antigens, was subjected to site-directed mutagenesis. A single substitution, alanine for arginine, at position 100A resulted in complete loss of antigen-binding activity. The 19 amino acids comprising the HCDR3 of mAb63 were then synthesized and a cyclic peptide prepared. The cyclic peptide showed the same antigen-binding pattern as the parental mAb63 and the recombinant mAb63 Fab. A five amino acid motif (RFLEW), present in the HCDR3 of mAb63, was found by searching the GenBank in three of 50 other human polyreactive antibodies, but in none of nearly 2500 human antibodies thought to be monoreactive. It is concluded that HCDR3 plays a major role in polyreactivity and that in some cases cyclic peptides comprising the HCDR3, by themselves, may be polyreactive.     

Natural autoantibodies: from 'horror autotoxicus' to 'gnothi seauton'.

The immune system of normal unimmunized animals is characterized by the presence of B cells synthesizing and secreting mainly polyreactive, but also monoreactive, IgM and IgG natural antibodies that can react with a variety of self constituents. These antibodies, like the autoantibodies appearing in several immunopathological states, use the same genetic elements as the antibodies directed against environmental antigens, and seem to be encoded by unmutated germ-line genes. Accumulating evidence indicates that these natural auto-antibodies exert various biological roles, both related and unrelated to the immune system. In this article, Stratis Avrameas proposes that natural auto-antibodies, by interacting with the large number of self constituents present in an organism, establish an extensive dynamic network that contributes to the general homeostasis of the organism.     

Specific and natural antibody production during Salmonella typhimurium infection in genetically susceptible and resistant mice.

Genetically susceptible (C57BL/6) and resistant (CBA) mice were infected with an avirulent strain of Salmonella typhimurium and studied over a 35-day period for the production of antibodies directed against bacterial antigens including lipopolysaccharide (LPS) (specific antibodies) and antibodies directed against self antigens [natural antibodies (NAb)]. Antibodies directed against LPS and self antigens were detected by enzyme immunoassay (EIA) and those directed against other bacterial antigens by immunoblotting. We found that serum natural antibody titres in C57BL/6 and CBA mice were similar and correlated with the bacterial load in the spleen and liver. In C57BL/6 mice, anti-LPS antibodies remained polyreactive and of the IgM isotype. In contrast, CBA mice, after an early increase in polyreactive IgM anti-LPS antibodies, mounted a specific anti-LPS IgG antibody response. The immunoblotting results demonstrated that the IgM polyreactive antibodies in the resistant and susceptible mice recognized bacterial antigens of different molecular weights and that CBA, but not C57BL/6 mice, were able to produce IgG antibodies recognizing bacterial components. Our results suggest that the synthesis of antibodies directed against bacterial antigens and natural antibodies follow, at least partially, distinct pathways, but they do not allow us to determine whether these two antibody populations are produced by the same or distinct B-cell subpopulations.     

IgA and IgG Subclass Deficiency in a Poor Population in a Developing Country

The levels of IgG, IgG subclasses, IgM and IgA were determined in serum from 17 patients with IgA deficiency and severe or frequent infections, allergy and/or autoimmunity (median age 7 years, range 2–19), 11 healthy IgA-deficient adults and 35 controls (median age 7 years, range 2–19). In serum from all groups IgG, IgM and IgA antibodies were determined against β-lactoglobulin, E. coli O antigens and poliovirus type 1 antigen. In saliva of 15 IgA-deficient patients and 12 of the controls IgG, IgM and secretory component-carrying antibodies against E. coli O antigens and poliovirus type I were determined. The majority of the studied individuals lived under poor socio-economic conditions in Brazil, with consequent heavy microbial exposure. One IgA-deficient patient with rheumatoid arthritis also had IgG2 deficiency but no infectious problems. Four out of the 35 controls without any obvious infectious problems were found with IgA or IgG subclass deficiency. One of the 11 healthy IgA-deficient adults was low in the IgG2 subclass, one in IgGl and one in IgG3. Those with symptomatic IgA deficiency had significantly higher serum IgG than the controls, especially in the age group 6–11 years. This latter group also had significantly increased serum IgG 1 and IgG2 levels when compared with the age-matched controls. Salivary IgM antibodies to E. coli and poliovirus antigens were significantly higher among the symptomatic IgA-deficient individuals than among the controls. It is not clear at present whether these increased Ig levels are secondary to frequent infections and/or part of mechanisms that may compensate for the IgA deficiency.     

Human Polyreactive IgM Monoclonal Antibodies with Blocking Activity Against Self-Reactive IgG

Natural IgM antibodies have been found to be involved in the control of IgG reactivity in normal serum. The authors investigated the blocking activity of four human IgM monoclonal antibodies (BY-2, BY-7, BY-10 and IRM-7) derived from B-cells from blood samples of three renal dialysis patients, which had shown multispecific properties similar to those observed for natural polyreactive autoantibodies. To achieve this, competitive inhibition assays were performed with these MoAbs on the binding of IgG purified from a healthy control, three patients with SLE, and two patients with autoimmune thyroiditis, to histone, dsDNA, RNP and thyroglobulin. MoAbs inhibited binding of self-reactive IgG to histone and dsDNA, but not to thyroglobulin or RNP, of natural and active or inactive phase disease-associated autoreactive IgG. The inhibitory effect of the MoAbs was mediated by V-region dependent interactions with autoreactive IgG, as shown by the ability of these MoAbs to block the binding of F(ab')₂ fragments of autoreactive IgG to antigens (histone and dsDNA). The blocking of autoantibody activity was dose-dependent with maximal inhibition occurring at a specific molar ratio between the patient's IgG and a given MoAb. In contrast, MoAbs did not inhibit binding of IgG alloantibodies present in the sera of four polytransfused renal dialysis patients to target antigens on the surface of different cells. These results support the concept of a functional idiotypic network regulating autoimmune responses, and suggest that the IgM MoAbs under study may be natural polyreactive antibodies belonging to the physiological network of autoantibodies with highly connected V-regions, capable of binding and functionally neutralizing V-regions of natural and pathologic autoantibodies.     

Immunoglobulin A-mediated hepatobiliary transport constitutes a natural pathway for disposing of bacterial antigens.

The possibility that hepatobiliary transport of immunoglobulin A (IgA) immune complexes might eliminate bacterial antigens was investigated in mice with pneumococcal type III capsular polysaccharide and C carbohydrate and corresponding monoclonal antibodies. Although all isotypes of antibody caused uptake by the liver, only IgA, but not IgG or IgM, antibodies transported these substances into bile where they were detected in the form of immune complexes and as free antigens. Small doses (10 micrograms or less) of passively administered IgA antibody were sufficient to induce measurable transport of capsular polysaccharide into bile. Transport of C carbohydrate was significantly correlated with the level of naturally occurring IgA antibodies specific for the phosphocholine determinant, but not with IgM or IgG antibodies. These results suggest that the continual process of hepatic uptake of circulating polymeric IgA may function to eliminate bacterial macromolecular products that are not readily susceptible to other mechanisms of disposal.     

Antibody responses to Toxoplasma gondii in sera, intestinal secretions, and milk from orally infected mice and characterization of target antigens.

Toxoplasma gondii-specific antibody responses in serum, intestinal secretions, and milk were identified with an enzyme-linked immunosorbent assay following a single oral infection of mice with strain 76K cysts of T. gondii. Immunoglobulin A (IgA) production began during week 2 of infection in serum and milk and during week 3 of infection in intestinal secretions and persisted in all three throughout the experiment (17 weeks). IgG but not IgM antibodies were detected in intestinal secretions later in the infection. Serum and milk IgG and IgM production began at the same time after infection as did the IgA response. In Western blotting (immunoblotting), intestinal IgA antibodies were shown to react with antigens comigrating with the T. gondii proteins p22, p23, p30, and p43, the 28-kilodalton antigen, and the 55- and 60-kilodalton rhoptry proteins, as recognized by specific monoclonal antibodies. Milk IgA antibodies reacted with antigens comigrating with p30 and p43. Most of the antigens recognized by IgA antibodies were also detected by IgG antibodies. IgA antibodies from all three biological samples detected the same major T. gondii antigens; thus, there was apparently no specific antibody production unique to one locality.     

Comparison of IgM, IgG and IgA responses to M.leprae specific antigens in leprosy

Antibodies of IgM, IgG and IgA classes against M.leprae specific antigens (PGL-I, ND-O-BSA, and NT-O-BSA) were determined in the sera of 80 leprosy patients (28 untreated, 34 treated lepromatous and 18 tuberculoid), 25 tuberculosis patients and 33 normal individuals of Northern Thailand. No strong distinction in reactivity could be found between the three antigens. The IgM antibody assay yielded more positive results than assays for IgG and IgA. It was found that the positivity rates of IgM antibodies to all three antigens were highest in untreated lepromatous leprosy (82%). In tuberculoid leprosy, the positivity rates of IgM, IgG and IgA to the antigens were more variable, ranging from 22 to 50 percent. Patients with tuberculosis and normal individuals did not produce IgM antibodies against the antigens. The results suggested that the determination of IgM against the three antigens is a more sensitive and specific test for active leprosy than those of IgG and IgA. The relationship between the duration of treatment and IgM antibody levels in lepromatous leprosy (LL) was studied. Untreated LL patients had significantly higher IgM and IgA antibody levels than treated patients. There was no difference in IgG antibody levels between the two groups, and the levels of both groups were higher than normal controls. Serial determination of IgM antibodies in 7 LL patients revealed that treatment was strongly associated with progressive decrease in IgM antibody levels against all three antigens.     

Polyreactivity of human monoclonal antibodies: Human IgM anti-rhesus monoclonal antibodies are frequently polyreactive

The aim of this study was to characterize human anti-Rhesus monoclonal antibodies cross-reacting with tissue antigens. Of the 155 monoclonal alloantibodies tested, 49 also reacted with intracellular antigens, as demonstrated by immunofluorescence assay on cryostat sections of animal and human tissues. This cross-reactivity was mainly a property of monoclonal alloantibodies belonging to the IgM isotype (among the 49 cross-reacting Mabs, 37 were IgM). The results confirm that during an immune response against a foreign antigen (alloantigen), B cells that produce polyreactive antibodies are not excluded from the pool of responding cells.