The effect of IgG aggregate size and concentration on reactivity in immune complex assays

Insoluble immune complexes are most effective at triggering IL-10 production in human monocytes and synergize with TLR ligands and C5a

PURPOSE

In systemic lupus erythematosus (SLE), a disease of immune complex (IC) deposition, interleukin-10 (IL-10) is thought to promote B-lymphocyte hyperactivity and autoantibody production. Both ICs and Toll-like receptor (TLR) ligands have been shown to stimulate the production of IL-10 by human monocytes. Using an in vitro model, we studied how IC solubility, complement activation products, and TLR ligands could affect IL-10 production by human monocytes stimulated with ICs.

METHODS

Human monocytes were stimulated with soluble or insoluble heat-aggregated human IgG with or without TLR ligands or C5a. Cytokine levels in cell culture supernatants were measured by ELISA. To study cytokine signaling, cell lysates were analyzed by Western blot for total or tyrosine-phosphorylated STAT3.

RESULTS

Insoluble ICs were most effective at stimulating production of IL-10, and costimulation LPS enhanced synthesis of IL-10. In addition, stimulation with insoluble ICs together with C5a enhanced the production of IL-10 by 2-4 fold in either the presence or absence of TLR ligands. Increased STAT3 phosphorylation correlated temporally with enhanced IL-10 production and was reduced by an IL-10 receptor blocking antibody, suggesting that IL-10 was responsible for observed STAT3 phosphorylation.

CONCLUSIONS

Because the immune deposits of SLE are, by definition, insoluble; and because IL-10 is thought to be important for B-cell hyperactivity and autoantibody production, these observations provide a critical link, bridging current views of B-cell hyperactivity with the early concept that SLE may arise from defective clearance of immune complexes.

C1q and systemic lupus erythematosus

In this chapter we review the association between SLE and C1q. In the first part of the chapter we discuss the clinical associations of C1q deficiency, and tabulate the available information in the literature relating to C1q deficiency and autoimmune disease. Other clinical associations of C1q deficiency are then considered, and we mention briefly the association between other genetically determined complement deficiencies and lupus. In the review we explore the relationship between C1q consumption and lupus and we discuss the occurrence of low molecular weight (7S) C1q in lupus, which raises the possibility that increased C1q turnover in the disease may result in unbalanced chain synthesis of the molecule. Anti-C1q antibodies are also strongly associated with severe SLE affecting the kidney, and with hypocomplementaemic urticarial vasculitis, and these associations are also examined. We address the question of how C1q deficiency may cause SLE, discussing the possibility that this may be due to abnormalities of immune complex processing, which have been well characterised in a umber of different human models. There is clear evidence that immune complex processing is abnormal in patients with hypocomplementaemia, and this is compatible with the hypothesis that ineffective immune complex clearance could cause tissue injury, and this may in turn stimulate an autoantibody response. We have also considered the possibility that C1q-C1q receptor interactions are critical in the regulation of apoptosis, and we explore the hypothesis that dysregulation of apoptosis could explain important features in the development of autoimmune disease associated with C1q deficiency. An abnormally high rate of apoptosis, or defective clearance of apoptotic cells, could promote the accumulation of abnormal cellular products that might drive an autoimmune response. Anti-C1q antibodies have been described in a number of murine models of lupus, and these are also briefly discussed. We focus on the recently developed C1q "knockout" mice, which have been developed in our laboratory. Amongst the C1q deficient mice of a mixed genetic background high titres of antinuclear antibodies were detected in approximately half the animals, and around 25% of the mice, aged eight months had evidence of a glomerulonephritis with immune deposits. Large numbers of apoptotic bodies were also present in diseased glomeruli, and this supports the hypothesis that C1q may have a critical role to play in the physiological clearance of apoptotic cells.

Anti-endothelial cell antibodies in inflammatory bowel disease

Antibodies to endothelial cells may be involved in the pathogenesis of vasculitic disorders. In view of recent evidence implicating intestinal vascular injury in the pathogenesis of inflammatory bowel disease, we have sought anti-endothelial cell antibodies in affected patients, examined their relationship to vascular injury, and tested their ability to mediate endothelial cell cytotoxicity in vitro. Anti-endothelial cell antibody levels were elevated in ulcerative colitis (P < 0.0001) and Crohn's disease (P < 0.05) compared with healthy controls. In ulcerative colitis, anti-endothelial cell antibody levels were related to disease activity and correlated with circulating levels of von Willebrand factor (r = 0.58, P < 0.01), a marker of vascular injury. Anti-endothelial cell antibodies, however, were not directly cytotoxic to endothelial cells in vitro. These data indicate, for the first time, an association between anti-endothelial cell antibody levels and vascular injury in vivo and suggest that they may be important in the pathogenesis of inflammatory bowel disease, particularly ulcerative colitis.

Localization of circulating immune complexes from patients with rheumatoid arthritis in murine spleen germinal centres

In previous studies we have demonstrated high levels of rheumatoid factor (RF) and large-size (greater than 22S) circulating immune complexes (CIC) in the serum of rheumatoid arthritis (RA) patients with extra-articular disease. These findings were paralleled by a concurrent increase in the level of RF-associated cross-reactive idiotypes (CRI) and an apparent diversification of the RF repertoire detected in the serum of the same patients. In the present study we examine the ability of CICs to activate the complement system in vivo, and its possible influence on expanding the RF repertoire in RA patients with extra-articular disease. Activation of complement by CICs is the key for germinal centre localization and long-term retention of such complexes on the surface of follicular dendritic cells (FDC), and so provides a source for the selection of cells with high affinity receptors for IgG and leads to the establishment of immunological memory. CICs containing different immunoglobulin isotypes and from different patients localized in mouse spleen germinal centres. However, intense localization was mainly seen for IgG-containing complexes from the serum of patients with large-size (greater than 22S) IgG-IgM RF complexes. The ability of these complexes to localize in mouse spleen germinal centres was related to activation of the complement system via the classical pathway in the patients' sera. Localization of IgG complexes was significantly (P less than 0.05) higher in sera from RA patients with extra-articular disease than those with articular disease alone. This study demonstrates the ability of large-size (greater than 22S) IgG-IgM RF complexes to activate complement, and suggests a possible role for such complexes in modulating the immune response to IgG in RA patients with extra-articular disease.

Characterisation of the size and composition of circulating immune complexes in patients with rheumatoid arthritis

The size and composition of circulating immune complexes in the sera of patients with rheumatoid arthritis (RA) were studied in relation to different manifestations of the disease. Circulating immune complexes from the sera of 94 patients (50 with extra-articular disease) and 10 matched controls were fractionated by sucrose density gradient ultracentrifugation. The composition, immunoglobulin and rheumatoid factor (RF) concentrations within each of the fractions were determined by a sensitive enzyme linked immunosorbent assay (ELISA). Intermediate size (14S-21S) IgG complexes containing RF activity and 22S IgG-IgM RF complexes were found in the sera of 40 patients with RA, while intermediate size complexes of self associated IgG RF and larger size complexes (greater than 22S) of IgG RF and IgM RF were associated with extra-articular features of RA (50% of extra-articular disease). Complexes containing IgA were found in the sera of many patients with RA, and dimeric IgA RF mainly in patients with extra-articular disease. These results support the view that whereas small size circulating immune complexes are of no primary pathogenic importance in synovitis, large size (greater than 22S) circulating immune complexes may play a role in extra-articular disease in RA. Current understanding of the formation of large complexes provides a biological explanation for their occurrence and effects.

Spectroscopic studies on IgG aggregate formation

Spectroscopic techniques have been used to examine the effects of IgG of heating, irradiation by ultraviolet light and exposure to glutaraldehyde. Relatively few changes were observed in treated IgG which remained unaggregated but several significant size-dependent changes were observed in aggregated IgG. These results suggest that IgG aggregate formation by cross-linking with glutaraldehyde involves the least perturbation of the basis IgG structure, whereas aggregate formation by heating involves the most, with ultraviolet irradiation occupying an intermediate position.

Antibodies to the collagen-like region of C1q in sera of patients with autoimmune rheumatic diseases

Antibodies to the collagen-like region of C1q have recently been observed in sera of patients with systemic lupus erythematosus (SLE). In this study, we documented that these antibodies were present in 47.3% of SLE patient sera, whereas they were uncommon in sera from patients with rheumatoid arthritis (2.8%) and Sjögren's syndrome (12.8%), as well as in normal sera (6.4%). Markedly elevated antibody levels (greater than 4 SD above the normal mean) were observed almost exclusively in sera of patients with SLE. Levels of antibodies to the collagen-like region correlated highly with levels of solid-phase C1q-binding IgG when analyzed by the C1q solid-phase assay for immune complexes (r = 0.87). We previously found that, after sucrose density gradient ultracentrifugation, a predominance of the solid-phase C1q-binding IgG in SLE sera sediments as monomeric IgG. These findings, together with the present data, indicate that reactivity of SLE patients' sera in the C1q solid-phase assay reflects primarily the presence of antibodies to the collagen-like region, and not the presence of immune complexes.

Binding of monomeric and aggregated immunoglobulin to enzymes. A source of artefact in antibody assays

Pepsinogen has previously been shown to bind non-specifically to immune complexes and aggregated immunoglobulins. We demonstrate here using a solid-phase immunoassay that immunoglobulins aggregated by heat or glutaraldehyde bind non-specifically to several different enzymes. Some of these, including pepsinogen (marketed as pepsin), hyaluronidase and trypsin, are used in the breakdown of tissues or biochemical preparations during the preparation of antigens. Contamination of impure antigens by enzyme is likely to lead to products which bind non-specifically to immune complexes. This can cause misidentification of complexes as antibodies. We recommend that all tests for specific antibody involving the use of antigens prepared by these or other enzymes should include a control with aggregated immunoglobulin substituted for the test serum.

An analysis of the fluid phase C1q binding assay. The effect of endogenous C1q on the precipitation and detection of an immune complex model

We examined the effect of endogenous C1q on the sensitivity of the fluid-phase C1q binding assay (C1qBA) in detecting an immune complex (IC) model, heat-aggregated IgG (HAIgG), at concentrations of 10-10,000 micrograms/ml sample. Results in normal human serum (NHS) or plasma (NHP) were compared with those in heat-inactivated NHS (NHS/56) in which most endogenous C1q was depleted by heat denaturation. Higher HAIgG concentrations were required in NHP and NHS to produce the same 125I-C1q precipitation seen in NHS/56. This decreased sensitivity varied from 70% at low HAIgG concentrations to 0% at high concentrations, as predicted for a large pool of endogenous C1q, in equilibrium with 125I-C1q, but in excess of that which could bind to all but the highest concentrations of IC model. In serum depleted of functional C1q on an immunoadsorbant of HAIgG, the precipitation of radiolabeled HAIgG under C1qBA conditions was concentration dependent and generated a saturation curve, showing that only a fraction of IC are usually precipitated in this assay. HAIgG precipitation was enhanced 1.4-fold in NHS/56 (8 micrograms C1q/ml) and three-fold in NHS (67 micrograms C1q/ml) suggesting that IC size is increased by endogenous C1q. In dual label experiments using 131I-HAIgG, the precipitation of 125I-C1q in NHS/56 was directly proportional to IC model precipitation, but markedly discordant in NHP, showing the measurement of IC in heat-inactivated sera superior to that in native serum. A comparison of the C1q:HAIgG ratio in PEG precipitates with that in samples, indicated that equilibrium was established between C1q and IC model. Thus the precipitation of 125I-C1q in the C1qBA represents (1) the fraction of total C1q bound to IC, and (2) the fraction of IC precipitated by PEG.

Prognostic and pathogenetic implications of immune complexes in human cancer

The dynamic relationships of circulating immune complexes (CIC) and associated immune reactants to stage of disease and prognosis in cancer patients have been evaluated for malignant melanoma. Patients were divided into group I patients, with no evidence of disease at sample time; group II patients, with minimal tumor burden; and group III patients, with advanced metastatic disease. By means of single- and serial-sample determinations, we have demonstrated correlation of CIC concentration, size, and composition to tumor burden, disease progression, and high risk of disease recurrence. In view of the phenotypic heterogeneity of antigen expression in malignant melanoma, we have extracted, purified, and utilized melanoma-associated antigens (MAA) from CIC, and human oncofetal antigens (HOFA) from fetal liver cells, and reacted them with patients' sera in an in vitro model that simulates tumor burden changes. The observed changes in CIC concentration and lattice size upon reaction with MAA and HOFA demonstrated a significant (p less than or equal to 0.001) relationship to patients' in vivo tumor burden, as well as to the origin of serum and antigen involved. Our results suggest a homeostasis of CIC size and concentration to associated antibodies and antigens operational in malignant melanoma patients' sera. In addition, we have demonstrated how the monitoring of more than one immune reactant, characterization of CIC size and composition, and evaluation of resulting reequilibrium caused by in vitro addition of antigen or antibody, could allow for an improved assessment of humoral immunity and its related pathogenetic effects in cancer patients.

The IgG detected in the C1q solid-phase immune-complex assay is not always of immune-complex nature

The properties of the solid-phase C1q immune-complex assay as well as the nature of the IgG detected by this assay in patients' sera were investigated. Aggregated IgG was used as a model for immune complexes. Aggregated IgG bound to solid-phase C1q was detected by 125I-anti-IgG. Fluid-phase C1q (either in normal human serum or purified) neither inhibited the binding of aggregated IgG to solid-phase C1q nor dissociated bound aggregated IgG from the solid-phase C1q. Therefore, we concluded that the solid-phase C1q has a higher affinity for aggregated IgG than the fluid-phase C1q, probably because of the polymerization of the solid-phase C1q. To get more insight into the nature of the IgG detected by the C1q solid-phase assay in patients' sera, we investigated whether C4 and/or C3 were present on it. With the use of 125I-anti-C4 and 125I-anti-C3 instead of 125I-anti-IgG, C4 and C3, respectively, were easily detected on the aggregated IgG that had bound to the solid-phase C1q. The lower limit of detection of these assays was 30 micrograms aggregated IgG/ml of normal human serum. Sera of patients suffering from rheumatoid arthritis and systemic lupus erythematosus were tested with these assays and, despite positive results with 125I-anti-IgG, no positive results were obtained with either 125I-anti-C4 or 125I-anti-C3. So, on the IgG detected by the C1q solid-phase assay in patients' sera, neither C4 nor C3 are present. Furthermore, in five of the six sera tested, this IgG sedimented as monomeric IgG. Therefore, it seems unjustified to refer to this IgG as circulating immune complexes.

The production of small IgG aggregates by glutaraldehyde cross-linking

Reaction conditions have been determined for the production of soluble IgG polymers in the size range 10 S to 30 S by covalent cross-linking with glutaraldehyde. This size range is comparable with that of the immune complexes which are frequently found in the circulation of patients with certain autoimmune diseases such as rheumatoid arthritis and systemic lupus erythematosus. The yield of IgG aggregates in this size range is far greater than has been reported for cross-linking by other bifunctional reagents or for aggregation by heating. Glutaraldehyde cross-linked IgG polymers are stable and biologically reactive. They can also be labelled with fluorescein and freeze-dried with minimal loss of integrity or reactivity.

Soluble IgG aggregates produced by heating remain stable on freeze-drying

IgG aggregates produced by heating gamma globulin solutions were freeze-dried, kept at 4 degrees C and reconstituted up to 4 months later. By comparison with frozen (-20 degrees C) preparations, only minimal changes in biological reactivity and in physical integrity occurred during this period. These results demonstrate that freeze-dried preparations of heat-aggregated IgG are potentially useful as a reference reagent for the comparative evaluation and standardisation of immune complex assays.

Immunoglobulin inclusions in rheumatoid arthritis polymorphonuclear cells: lack of correlation with circulating immune complexes

A discriminating direct immunofluorescent test has been used to identify immunoglobulin inclusions in polymorphonuclear leucocytes (PMNs) isolated from the blood of patients with rheumatoid arthritis. These inclusions are thought to represent phagocytosed immune complexes, since normal PMNs incubated in RA sera known to contain raised levels of immune complexes developed similar immunoglobulin inclusions. Inclusions did not develop in normal PMNs incubated in normal serum. No correlation was found between the percentage of either RA blood PMNs with immunoglobulin inclusions or normal PMNs developing inclusions after incubation in RA sera, and levels of immune complexes in the corresponding sera. Using heat-aggregated IgG as a laboratory model of immune complexes, a simple relationship has been demonstrated between the uptake of IgG aggregates by normal PMNs and the concentrations of IgG aggregates in the test solutions over a concentration range of 12.5-200 micrograms . ml-1. These results indicate that the C1q- PEG test gives no measure of the actual amounts of immune complexes available in serum for phagocytosis.