Antibodies to human myeloperoxidase in glomerular immune deposits of systemic lupus erythematosus.

Antibodies to human myeloperoxidase and cathepsin G have been detected in the serum of some patients with systemic lupus erythematosus. Therefore, the presence of antibodies to human myeloperoxidase and cathepsin G was examined in glomerular immune deposits. Glomerular basement membrane fragments were prepared from renal tissues obtained at autopsy from 19 patients with systemic lupus erythematosus. IgG was extracted from the glomerular basement membrane fragments and tested with sensitive immunoassays for antibodies to myeloperoxidase and cathepsin G. Antibodies to cathepsin G were not detected in the extracts but antibodies to human myeloperoxidase were found in extracts of one specimen. In the extract with 6M guanidine hydrochloride these antibodies were enriched 103-fold, compared to the initial supernatant of glomeruli, which served as a serum surrogate. The recovered antibodies to myeloperoxidase accounted for 12% of the recovered IgG. These findings add autoantibodies to human myeloperoxidase to the list of antibodies that have been shown to be present in glomerular immune deposits of patients with lupus glomerulonephritis.

Diagnostic value of antibodies to filaggrin in rheumatoid arthritis.

OBJECTIVE: To determine the prevalence of antibodies to filaggrin in a cross sectional sample of patients with rheumatoid arthritis (RA). METHODS: Filaggrin from human skin was either extracted with 0.05% Nonidet P-40 and then partially purified by precipitating in ethanol and resuspending in water (Nonidet preparation) or extracted with 9 M urea and then purified by sequential fractionation on a DEAE Sephadex column and on a strong cation exchange column (purified preparation). Antibodies to filaggrin were detected using immunoblotting techniques with sera diluted 1:50. Antikeratin antibodies (AKA) were detected using indirect immunofluorescence microscopy on sections of rat esophagus. RESULTS: Antibodies to filaggrin were detected in 5 of 30 sera of patients with RA using filaggrin from the Nonidet preparation and 6 of 49 sera using filaggrin from the purified preparation. AKA were detected in 13 of 40 sera. A positive correlation existed between the presence of AKA and the presence of antibodies to filaggrin using the purified preparation (p=0.017). CONCLUSION: These data indicate that the reactivity of RA sera with filaggrin is not identical to the presence of AKA and is variable depending upon the preparation of filaggrin used. The diagnostic value of antibodies to filaggrin remains to be proven.

Deposition of antibodies to the collagen-like region of C1q in renal glomeruli of patients with proliferative lupus glomerulonephritis.

OBJECTIVE: To determine if antibodies to the collagen-like region of C1q (C1q-CLR) are present in the glomerular immune deposits of patients with systemic lupus erythematosus (SLE). METHODS: Kidney tissues were obtained at autopsy, glomeruli were isolated, and glomerular basement membrane fragments were prepared. Antibodies were extracted with low pH or with DNase. RESULTS: The concentrations of antibodies to C1q-CLR recovered from the glomeruli were > or =50-fold higher per unit of IgG than that found in the serum or in the serum and interstitial fluid entrained in glomeruli. Antibodies to C1q-CLR were recovered from glomeruli of 4 of 5 patients with proliferative glomerulonephritis at autopsy. CONCLUSION: This is the first demonstration that antibodies to C1q-CLR are deposited and concentrated in the renal glomeruli of patients with SLE. These antibodies, thus, have the potential of contributing to the pathogenesis of lupus glomerulonephritis.

In patients with rheumatoid arthritis IgG binding to denatured collagen type II is in part mediated by IgG-fibronectin complexes.

The presence of autoantibodies to native and denatured collagen type II has been reported in some patients with rheumatoid arthritis (RA). This study examined the molecular nature of IgG binding to native and denatured collagen type II. The binding of IgG to native and denatured collagen was detected with an ELISA. Serum proteins were separated by size with gel filtration. Gel-filtered fractions were further characterized by binding to staphylococcal protein A. Among plasmas or serums from 60 patients with RA, 12 patients had IgG binding to native collagen type II and 12 had IgG binding to denatured collagen type II. Decomplementing normal serums by heating to 56 degrees C for 30 min markedly increased IgG binding to denatured collagen type II, but not to native collagen. This binding was shown to result from a complex formed between IgG and fibronectin. Nine unheated RA serums were separated by gel filtration. The IgG binding to native collagen was limited to monomeric IgG, but in these serums 65.6 +/- 22.0% of the IgG binding to denatured collagen type II was in the excluded protein fractions. The binding of IgG to denatured collagen type II in the excluded fractions was inhibited by fibronectin and did not bind to staphylococcal protein A. In patients with RA, the IgG binding to native collagen type II represents true autoantibodies. In contrast, in these patients the majority of IgG binding to denatured collagen type II is caused by macromolecular complexes formed between fibronectin and IgG or between fibronectin and IgG-containing immune complexes.

Sera of patients with rheumatoid arthritis contain antibodies to recombinant human T-lymphotrophic virus type I/II envelope glycoprotein p21.

A possible retroviral etiology for rheumatoid arthritis (RA) has been raised by results of recent studies. Therefore, we examined sera of patients with RA, including those with coexisting Felty's syndrome or leukemia of large granular lymphocytes, for the presence of antibodies to retroviral proteins of human T-lymphotrophic virus type I and type II (HTLV-I/II). Reactivity to recombinant HTLV-I envelope protein rgp21 alone was the primary pattern observed. Twenty-five percent of RA sera, 28% of Felty's syndrome sera, and 30% of large granular lymphocyte leukemia/RA sera reacted with rgp21, each significantly more than the 8% of normal sera (P less than 0.01). Removing rheumatoid factor did not abolish reactivity with rgp21 in any of six RA sera tested. Immunoreactivity to the authentic viral protein was confirmed by using purified rgp21 that was cleaved by CNBr to remove the bacterial fusion peptide, or by blocking sera with a synthetic peptide corresponding to the fusion peptide. Only one serum, from a patient with RA, showed definite evidence for prior infection with prototypic HTLV-II. These data indicate that 25% of RA sera have IgG antibodies to recombinant HTLV-I envelope protein rgp21, which is highly homologous to envelope protein gp21 of HTLV-II. These findings provide potentially novel clues regarding the pathogenesis of RA.

Blood clearance kinetics and liver uptake of mononucleosomes in mice.

Nucleosomes generated by apoptosis have become of considerable interest in relation to pathogenesis of systemic lupus erythematosus in mice and humans. Therefore, the fate of circulating mononucleosomes was examined in normal C57Bl/6J mice. The mononucleosomes were prepared from chicken erythrocytes and radiolabeled on the histone component. The removal of nucleosomes from circulation at doses less than 11 micrograms of injected mononucleosomes was rapid, but with increasing doses of injected nucleosomes, the slopes of the removal curves decreased. Liver was the major organ for removal of circulating nucleosomes, accounting for 71.0 to 84.7% of nucleosomes removed from circulation at 10 min. After i.v. injection of nucleosomes, 0.52 +/- 0.15% localized in kidneys. With prior i.v. injection of histones, the glomerular localization of mononucleosomes increased threefold. The clearance of mononucleosomes was decreased sixfold by concurrent injection of ssDNA. These studies show that in mice, circulating mononucleosomes are handled similar to DNA, and they do not avidly localize in glomeruli unless histones have already bound to renal glomeruli.

Presence of covalent bonds between immune deposits and other macromolecules in murine renal glomeruli.

Microscopic studies have suggested that immune deposits persist for a long time in subepithelial areas of renal glomeruli in human diseases and in experimental models. This study showed quantitatively that in an experimental murine model radiolabelled antibodies in subepithelial immune deposits remained constant for 9 months at about 3 micrograms/mouse. The extraction of these antibodies was incomplete with 1% SDS at 8 months after immune deposit formation. Furthermore, in the extracted material some antibodies were covalently bound to other molecules. This study demonstrates that covalent bond formation contributes to the persistence of immune deposits in renal glomeruli.

C1q-binding immunoglobulin G in MRL/l mice consists of immune complexes containing antibodies to DNA.

Previous studies have shown that the majority of C1q-binding IgG in patients with systemic lupus erythematosus (SLE) is composed of autoantibodies to the collagen-like region of C1q. Mice of the MRL/l strain are considered as a murine model of human SLE and possess autoantibodies to nuclear antigens as well as IgM and IgG rheumatoid factors (RF). This study was undertaken to characterize the C1q-binding IgG in MRL/l mice. In contrast to human SLE, C1q-binding IgG in MRL/l mice showed immunochemical characteristics of immune complexes rather than those of autoantibodies to C1q. Namely, C1q-binding IgG in MRL/l mice was large-sized upon HPLC gel filtration and abolished by digestion with pepsin or by high salt concentration, and bound to the globular region of C1q. The C1q-binding activity in MRL/l mice was absorbed by double-stranded DNA- or single-stranded DNA-cellulose. The medium-sized immune complexes containing RF have been well documented in MRL/l mice. In this study, however, mouse IgG-Sepharose failed to absorb fully C1q-binding IgG. We conclude that the majority of C1q-binding IgG in MRL/l mice consists of large-sized immune complexes containing antibodies to DNA.

Deep penetration of antibodies into the articular cartilage of patients with rheumatoid arthritis.

This study was conducted to determine the presence of immunoglobulin G (IgG) and albumin in deep layers of articular cartilage from patients with rheumatoid arthritis or osteoarthritis and from normal organ donors. Cartilage plugs were cut into 20-microns slices with a microtome and ten consecutive slices were pooled, dividing the specimen into 200 microns sections starting from the articular surface. Each pool was extracted overnight thrice with neutral buffer, thrice with 6 M guanidine hydrochloride, and then degraded with bacterial collagenase. IgG and albumin were quantified in each extract. From the surface and deep layers significantly more IgG and albumin were extracted from rheumatoid than from normal specimens, both with neutral buffer and with guanidine. In neutral buffer extracts the molar ratios of IgG to albumin were comparable from normal and rheumatoid specimens, with a molar excess of albumin. In contrast, the molar ratios of IgG to albumin in guanidine extracts from rheumatoid cartilages were significantly higher than in normal cartilages, and the IgG was in molar excess of albumin only in rheumatoid extracts. These results show for the first time that IgG has penetrated deep into the cartilage in rheumatoid arthritis and may contribute to the degradation of cartilage by inflammation.

The structural basis of germline-encoded VH3 immunoglobulin binding to staphylococcal protein A.

The ability of human VH3 immunoglobulins (Ig) to bind to staphylococcal protein A (SPA) via their Fab region is analogous to the binding of bacterial superantigens to T cell receptors. The present report establishes the structural basis for the interaction of SPA and VH3 Ig. We have studied a panel of 27 human monoclonal IgM that were derived from fetal B lymphocytes. As such, these IgM were expected to be encoded by unmutated germline genes. Binding to SPA in ELISA occurred with 15 of 15 VH3 IgM, but none of 12 IgM from the VH1, VH4, VH5, or VH6 families. The VH sequences of the 27 IgM were derived from 20 distinct VH elements, including 11 from the VH3 family. Use of D, JH, and CL genes was similar among VH3 and non-VH3 IgM. A comparison of the corresponding VH protein sequences, and those of previously studied IgM, identified a probable site for SPA binding that includes VH3 residues in framework region 3 (FR3), and perhaps FR1 and 3' complementary determining region 2. The results thus demonstrate that among human IgM, specificity for SPA is encoded by at least 11 different VH3 germline genes. Furthermore, like the T cell superantigens, SPA likely binds to residues in the VH framework region, outside the classical antigen-binding site of the hypervariable loops.

Antigens of varying size persist longer in subepithelial than in subendothelial immune deposits in murine glomeruli.

The kinetics of removal of immune deposits from the subendothelial and subepithelial areas of glomeruli were analyzed in mice. Radiolabeled, cationized Ag of different molecular size, including human serum albumin, bovine thyroglobulin, and human IgM, were used to form the immune deposits in mouse glomeruli with specific, purified rabbit antibodies to these proteins. The disappearance curves of the radiolabeled Ag from glomeruli consisted of two exponential components. The immune deposits and their location in glomeruli were identified by immunofluorescence and electron microscopy. The t1/2 of disappearance of immune deposits were assigned to subendothelial or subepithelial deposits on the basis of ultrastructural observations. The t1/2 of subendothelial immune deposits ranged from 0.6 to 1.9 days with the three different Ag-antibody systems. In contrast, the t1/2 of the subepithelial immune deposits ranged from 9.32 to 231 days. The cationized human serum albumin in subepithelial areas had the longest t1/2, and this was not altered by the endogenous immune response to the injected materials, as determined in studies with nude mice. The results constitute formal documentation of the prolonged t1/2 of an exogenous Ag in glomerular immune deposits. The described approach can serve to examine variables that alter this prolonged presence of subepithelial immune deposits in glomeruli.

Immunoglobulin G and serum albumin isolated from the articular cartilage of patients with rheumatoid arthritis or osteoarthritis contain covalent heteropolymers with proteoglycans.

The present study was undertaken to identify the cartilage matrix molecules that are bound with intermolecular disulfide bonds to IgG and serum albumin molecules recovered from the articular cartilage of patients with rheumatoid arthritis (RA) or osteoarthritis (OA). The cartilage specimens were extracted sequentially with three changes of neutral buffer, three changes of 6 M guanidine hydrochloride and then partially degraded with bacterial collagenase. The extracted IgG and albumin, along with matrix molecules bound to these proteins, were isolated with affinity chromatography using antibodies to IgG or human serum albumin conjugated to agarose beads. The isolated materials were characterized with sodium dodecyl sulfate polyacrylamide gel electrophoresis and transfer blotting, using specific antibodies to IgG, albumin, and proteoglycans. In the isolated materials, heteropolymers with IgG or albumin were identified. These polymers contained keratan sulfate and less frequently chondroitin-4-sulfate and chondroitin-6-sulfate. These findings identified the keratan sulfate rich proteoglycans, prevalent at the surface of joint cartilage, as the most common cartilage matrix molecules that are covalently bound to IgG or to serum albumin by disulfide bonds in the articular cartilage of patients with RA or OA.

IgG is bound by antigen-antibody bonds and some IgG and albumin are bound by intermolecular disulfide bonds to cartilage in rheumatoid arthritis and osteoarthritis.

To elucidate the mechanisms for the presence of immunoglobulins and human serum albumin (HSA) in articular cartilage from patients with rheumatoid arthritis (RA) or osteoarthritis (OA), the recovery of these molecules was determined in several elution steps. These steps included serial elutions with a neutral buffer to extract entrapped molecules, elution with 6 M guanidine hydrochloride to extract molecules bound by noncovalent interactions, and digestion of cartilage with bacterial collagenase to release molecules covalently bound to cartilage matrix proteins. Significantly more IgG than HSA was recovered with 6 M guanidine after serial elutions with neutral buffer from the cartilages of patients with both RA and OA, consistent with the binding of IgG by antigen-antibody bonds. Degradation of cartilage with collagenase released additional IgG and HSA. Analysis of the IgG and HSA, recovered with guanidine or with collagenase, using SDS-PAGE and transfer blotting, indicated for the first time the presence of disulfide bonds between these molecules and cartilage matrix molecules.

Rheumatoid factors in the pathogenesis of rheumatoid arthritis.

Rheumatoid factors (RF) participate in the pathogenesis of rheumatoid arthritis by formation of immune complexes. IgM-RF form complement activating immune complexes with IgG-containing antigen-antibody complexes. IgG-RF form unique immune complexes without the presence of separate antigen molecules. The specificities of RF in rheumatoid arthritis have subtle differences from RF formed by B cell neoplasms. Immune deposits in rheumatoid articular cartilage have high potential for generating inflammation and contain RF. The immune deposits in articular cartilage need to be characterized further, and the mechanisms that initiate and perpetuate RF production in patients with rheumatoid arthritis should be elucidated.

Autoantibodies to the collagen-like region of C1Q deposit in glomeruli via C1Q in immune deposits.

The autoantibodies to the collagen-like region of C1q (CLR), purified from two patients with systemic lupus erythematosus, deposited in mouse glomeruli when human C1q was present in antigen-antibody complexes in glomeruli. The immune deposits with C1q in mouse glomeruli were achieved by the administration of cationized immune complexes containing human C1q. The presented data suggest that the autoantibodies to CLR could enhance the pathogenic role of immune complexes deposited in glomeruli by binding to C1q in immune deposits. These findings may explain the association of autoantibodies to CLR with proliferative lupus nephritis.

New antigenic determinants revealed on human IgG by binding to immunoblotting membranes [corrected].

During immunoblotting for detection of human IgG, rabbit antibodies to goat IgG reacted with human IgG, Fc fragments and gamma chains when these proteins were bound to nitrocellulose or Immobilon-P membranes. This reactivity could not be adsorbed with human IgG-agarose beads or blocked with fluid phase human IgG. It was readily abrogated with human IgG adsorbed to one of these membranes, indicating that new antigenic determinants were exposed and accounted for the cross-reactivity.

Human IgA and IgG F(ab')2 that bind to staphylococcal protein A belong to the VHIII subgroup.

Staphylococcal protein A (SPA) is a bacterial membrane protein that possesses, in addition to its Fc gamma-binding activity, a distinct specificity for the Fab region of some IgM, IgA, IgG, and IgE. The Fab site that binds to SPA has been localized to the V region of the Ig H chain. In a previous study of human monoclonal and polyclonal IgM, we demonstrated that binding to SPA was highly restricted to molecules of the VHIII subgroup, and that nearly all VHIII IgM were able to bind SPA. The present study examines the VH composition of SPA-binding and SPA-nonbinding fractions of purified human polyclonal IgA, and IgG F(ab')2 fragments. We found that 22% of the IgA and 15% of the IgG F(ab')2 bound to SPA-agarose. Analysis with VH subgroup-specific antisera indicated that the SPA-binding fraction of IgA was dominated by the VHIII subgroup, and the SPA-binding fraction of IgG F(ab')2 contained only VHIII molecules. Furthermore, substantial portions of the total VHIII protein in IgA and in IgG F(ab')2 bound to SPA. We conclude that Fab binding to SPA is both restricted to and highly prevalent among human VHIII molecules, regardless of Ig class. These results suggest that protein A is an Ig superantigen.