The influence of lipid composition and divalent cations on annexin V binding to phospholipid mixtures.

Annexin V is a 36-kDa protein which, it has been suggested, is a factor in protecting the vascular endothelium from attack by antibodies to other phospholipid-binding proteins. Competition between annexin V and beta2-glycoprotein I (beta2GPI) for phospholipid surfaces is complicated by empirical observations regarding alterations in binding to anionic phospholipid, primarily phosphatidylserine. In order to elucidate the effect of phospholipid composition and divalent cations (Ca(+2) and Mg(+2)) on annexin V binding to phospholipid, we used biotinylated annexin V and peroxidase-conjugated avidin D to probe the binding of annexin V to phospholipid-coated wells of polystyrene microtiter plates. Binding of annexin V to anionic phospholipid is Ca(+2)-dependent and, in its absence, annexin V was found to bind most avidly to 100% phosphatidylcholine in a saturable manner, followed by decreasing percentages of phosphatidylcholine. Ca(+2) was found to inhibit phosphatidylcholine binding and promote the binding of phospholipid mixtures containing phosphatidylserine. Phosphatidylserine (100%) did not bind annexin V as strongly as mixtures of 50% and 75% phosphatidylserine. The effect with Ca(+2) suggests saturation of Ca(+2)-binding sites on annexin V, reached under our experimental conditions at approximately 1 mM. Under the same conditions, Mg(+2) slightly enhanced the binding of all of the phospholipid compositions studied. Ca(+2)-dependent binding of annexin V was competitively inhibited by Mg(+2); 5 mM Mg(+2) reduced binding significantly (p < 0.0001 by ANOVA, p < 0.05 for post hoc test of 5 mM vs 0 mM). These data suggest that the translocation of membrane phospholipid under the dynamics of ion transport in vascular endothelium may alter annexin V binding.

Binding of complement component C1q to anti-beta2 glycoprotein I antibodies from patients with antiphospholipid syndrome.

OBJECTIVE: To determine if anti-beta2 GPI reactive with surface-bound beta2 GPI can bind C1q, i. e. to determine whether surface-bound beta2 GPI-anti-beta2 GPI immune complexes can initiate the classical pathway of complement activation. METHODS: Beta2 GPI was bound to chemically-activated microtiter plates which had previously been shown to promote anti-beta2 GPI reactivity with bound beta2 GPI. Wells with surface-bound beta2 GPI (capped with bovine serum albumin) were then reacted with complement-inactivated sera from antiphospholipid syndrome patients (APS) or with control sera. Following removal of unbound serum components, the wells were incubated with biotinylated C1q and probed with peroxidase-conjugated avidin D. Bound C1q was detected at 450 nm using tetramethyl benzidine/peroxidase as a substrate system and expressed as absorbance units (Abs). RESULTS: The identified 20 APS with elevated anti-beta2 GPI: 4 with IgG only, 4 with IgM only, 1 with IgA only, 1 with IgG and IgA, 6 with IgG and IgM and 4 with IgG, IgA and IgM. C1q binding from 20 healthy controls was 0.039 +/- 0.029 (SD). Of the APS, 17/20 (85%) had Abs >5 SD above controls. The 3 APS with C1q Abs within normal limits had, respectively, IgM only (1), IgA only (1), and both IgG and IgM (1). Statistical analyses (Kruskal-Wallis followed by Dunn's post test) suggest differences in IgG and IgG + IgM groups compared to con (Kruskal-Wallis: p = 0.0002; Dunn's: con vs. IgG, p < 0.05; con vs. IgG + IgM, p < 0.01). CONCLUSIONS: Anti-beta2 GPI from APS appear to have a variable degree of C1q affinity. Those patients with strong C1q binding responses are likely to have an inflammatory component to their disease processes.

Modification of beta 2glycoprotein I by glutardialdehyde. Conformational changes and aggregation accompany exposure of the cryptic autoepitope.

Autoantibodies from patients with antiphospholipid syndrome (APS) recognize an epitope on beta 2glycoprotein I (beta 2 GPI) only when native beta 2 GPI is adsorbed on surfaces composed of anionic phospholipids or oxidized polystyrene. beta 2 GPI was modified with the crosslinking agent, glutardialdehyde (GDA), which induced exposure of the anti-beta 2 GPI epitope at GDA:beta 2 GPI mol ratios in the range of 500-2000. A second crosslinking agent, dimethyl-suberimidate (DMS), did not expose the epitope, which may be a consequence of its having less tendency than GDA to form intermolecular links. SDS-PAGE experiments demonstrate that GDA does promote extensive intermolecular crosslinking of beta 2 GPI, and DMS does not. Formaldehyde also reacts with the lysine residues of beta 2 GPI, but does not expose the epitope. The circular dichroism spectra of native and modified beta 2 GPI confirm that GDA induces changes in conformation that are qualitatively different from those caused by formaldehyde. These data provide evidence that binding of lysine residues is not a sufficient condition for exposure of the autoepitope, and also support the likelihood that anti-beta 2 GPI antibodies bind only to aggregates of the protein. Thus, by synthesizing an active holoantigen of beta 2 GPI, conditions were defined that are necessary for binding of human autoantibodies.

Standardized measurement of major immunoglobulin class (IgG, IgA, and IgM) antibodies to beta2glycoprotein I in patients with antiphospholipid syndrome.

Patients with the antiphospholipid syndrome (APS) have autoantibodies directed against epitopes on beta2 glycoprotein I (beta2GPI). We describe herein the performance characteristics of standardized enzyme-linked immunosorbent assays (ELISAs) for anti-beta2GPI of the three major immunoglobulin classes: IgG, IgA, and IgM. All three assays generated highly linear standard curves (5 points, r > or = 0.993 for each); precision was excellent both intra-assay and run-to-run, with coefficients of variation (CV) ranging from 2.3% to 6.6%. Values for IgG anti-beta2GPI correlated strongly with those obtained by an earlier method (r = 0.80, P< 0.0001). A study group consisting of 203 healthy subjects was used to generate percentile-based reference intervals for all three classes of anti-beta2GPI. APS subjects' anti-beta2GPI were found to differ significantly (P <0.0001 for each) from those of the control population. All three assays correlated well with beta2GPI-dependent anticardiolipin antibody (aCL) measurements; IgG: r = 0.94 (P <0.0001), for IgA: r = 0.82 (P<0.001) and for IgM: r = 0.84 (P<0.0001). We suggest that these ELISAs may provide valuable standardized measurements of IgG, IgA, and IgM anti-P2GPI.

A stable, multi-subunit complex of beta2glycoprotein I.

Beta2glycoprotein I (beta2GPI) is a 54-kDa plasma protein which is recognized as an autoantigen for antibodies from patients with antiphospholipid syndrome (APS). SDS-PAGE (under reducing conditions) of beta2GPI from three sources indicates that the 54-kDa beta2GPI band is accompanied by a band corresponding to an 8-kDa protein. In the absence of detergent and reducing agents (native PAGE), beta2GPI demonstrated a large complex (molecular mass approximately 320 kDa) which is dissociable by boiling in 6-8 M urea, yielding several lower molecular mass bands, one of which corresponds to the 8-kDa protein observed in SDS-PAGE. Sera from five healthy adults demonstrated native beta2GPI migration equivalent to the commercially purified protein. Atomic force microscopy (AFM) images of native beta2GPI show aggregrates of particles each having a diameter of 30-35 nm. This is consistent with a globular unit the size of which would be substantially larger than that expected for a 54-kDa protein. These experiments suggest that the 54-kDa beta2GPI monomer subunits exist as a multimeric complex with the 8-kDa protein.

Phospholipid-independent binding of beta 2glycoprotein I by IgA from patients with antiphospholipid syndrome.

beta 2glycoprotein I (beta 2GPI) is a phospholipid-binding protein of the coagulation system. In patients with the antiphospholipid syndrome (APS), antibodies to beta 2GPI contribute to the population of "antiphospholipid antibodies" measured in the anticardiolipin antibody (aCL) assay. In fact, both IgG and IgM antibodies from patients with APS bind beta 2GPI in the absence of anionic phospholipids if the antigen is bound to a suitable surface, i.e., one which exposes the epitope. The binding of IgA was studied from patients with APS, using an enzyme-linked immunosorbent assay (ELISA), and significantly higher binding of IgA was observed from 39 patients compared to a control group of 50 healthy individuals (p < 0.0001). Moreover, 15 out of 39 APS subjects (38 percent) exhibited binding greater than 5 standard deviations (SD) above the mean of the control group. All 39 APS patients had elevated IgG anti-beta 2GPI; however, depletion of IgG from two APS sera diminished, rather than enhanced, binding of IgA. Pre-incubation with purified IgG from a subject with APS led to inhibition of IgA binding at inhibitor levels > 125 micrograms IgG/well. These data demonstrate that patients with APS have IgA anti-beta 2GPI autoantibodies and that the epitope(s) which are recognized by these antibodies can be presented in the absence of cardiolipin or other anionic phospholipids.

The association of antibodies to beta 2 glycoprotein I with the antiphospholipid syndrome: a meta-analysis.

The precise target of marker autoantibodies in the antiphospholipid syndrome (APS) has been controversial. Recently, the theory that surface-bound beta 2 glycoprotein I (beta 2 GPI) presents a normally encrypted autoepitope and that antibodies to beta 2 GPI (anti-beta 2 GPI) are associated with APS, has been prominent. The literature has been searched from 1990 to 1995 (inclusive) to find studies in which: (1) anti-beta 2 GPI antibodies were measured and, (2) adequate clinical data describing the patients were included. These conditions were met in four studies, from August 1992 to December 1995, in which patient samples ranged from 15 to 39 cases, the total of the four studies being 90 cases. Applying the diagnostic criteria of > or = 2 clinical manifestations of APS, 65 cases had APS while 25 did not. For the group with APS, 58/65 (89 percent) were anti-beta 2 GPI (+); among those who did not meet the criteria, 11/25 (44 percent) were anti-beta 2 GPI (+) (p < 0.0001). The presence of either the primary (1 degree) or secondary (2 degrees) form of APS made no difference in association with anti-beta 2 GPI; 13/16 (81 percent) patients with 1 degree APS and 45/49 (92 percent) with 2 degrees APS had anti-beta 2 GPI. The presence of lupus anticoagulant (LAC) did not predict APS: 25/58 (38 percent) APS (+) cases were LAC (+); 11 APS (-) were all LC (+) (p < 0.0001). IgG anticardiolipin also showed a closer association with APS absence (25/25 cases; 100 percent) than with APS presence (40/65; 62 percent, p < 0.0001). These data support the contention that anti-beta 2 GPI antibodies are closely associated with APS and that their measurement may facilitate the diagnosis.

Antibodies to beta 2 glycoprotein I: standardized immunoassay and a reference interval for healthy controls.

Patients with antiphospholipid syndrome (APS) demonstrate an antibody reactivity with beta 2 glycoprotein I (beta 2 GPI) independent of the anionic phospholipids that previously had been believed to be the relevant autoantigens associated with this syndrome. Immunoassays for IgG anti-beta 2 GPI have, however, suffered from a lack of standardization. In this article, we describe an assay based on reference standards that we developed recently. The assay exhibits excellent linearity, with regard to both application of the standards and dilution of out-of-range specimens. Precision was assessed both within the between run and was judged to be satisfactory. A reference interval was developed on the basis of a control group consisting of 111 men and 93 women, yielding a range of 0-19 standard IgG anti-beta 2 GPI units (SGU) for the ninety-fifth percentile of values. These data suggest that this standardized anti-beta 2 GPI assay may be useful in the clinical diagnostic laboratory.

Binding of beta 2 glycoprotein I to activated polystyrene and its recognition by human IgG autoantibodies.

Beta 2 glycoprotein I (apolipoprotein H, beta 2GPI) is involved in the formation of epitope(s) recognized by clinically relevant autoantibodies from patients with antiphospholipid syndrome. We studied the binding of beta 2GPI to chemically activated polystyrene in a microtitre plate format. Adsorption isotherms (at 37 degrees C) were generated for beta 2GPI on activated polystyrene and on unactivated polystyrene, with both human serum antibodies and rabbit polyclonal IgG antibodies as probes, and horseradish peroxidase (HRP)-tagged anti-IgG to detect binding. Additionally, beta 2GPI was biotinylated and isotherms were developed by using HRP-streptavidin as the recognition sequence. Human serum autoantibodies, which did not precipitate beta 2GPI in solution, yielded a characteristic chemisorption isotherm on activated polystyrene but did not recognize beta 2GPI bound to untreated polystyrene. The rabbit IgG, which did precipitate beta 2GPI in solution, detected beta 2GPI bound to both activated polystyrene and, to a lesser extent, to untreated polystyrene. The binding of beta 2GPI to untreated polystyrene was confirmed by the use of biotinylated beta 2GPI. To assess the prevalence of IgG anti-beta 2GPI autoantibodies, we surveyed 113 sera submitted to our laboratory for anticardiolipin antibody (aCL) testing. Only nine (8%) had anti-beta 2GPI activity greater than two standard deviations above the mean for those sera in which aCL activity was within normal limits. We conclude that epitope presentation of beta 2GPI for human autoantibody binding is dependent on surface properties of the polystyrene, and that beta 2GPI autoantibodies are found only in a subpopulation of sera positive for aCL.

Analytical and clinical relationships between human IgG autoantibodies to beta 2 glycoprotein I and anticardiolipin antibodies.

OBJECTIVE: To examine IgG anti-beta 2 glycoprotein I (anti-beta 2 GPI) binding in 82 sera referred for anticardiolipin antibody (aCL) testing and to develop preliminary clinical correlations with antiphospholipid syndrome (APS). METHODS: Immunoassay of IgG cofactor dependent aCL and IgG anti-beta 2 GPI antibodies and retrospective chart review. RESULTS: Forty-four sera exhibited normal (< or = 22 GPL units) aCL activity, 18 had moderate binding activity (23-45 GPL units), and 20 had high (> or = 46 GPL units) binding activity to cardiolipin. Among these groups, 6 of the 20 sera in the high GPL group had elevated anti-beta 2 GPI. This correlated strongly with 2 or more clinical manifestations of APS. CONCLUSION: Anti-beta 2 GPI activity may be a more valuable indicator of APS than aCL.

Biphasic response of complement to heparin: fluid-phase generation of neoantigens in human serum and in a reconstituted alternative pathway amplification cycle.

We describe a study of the effects of heparin on complement activation through the use of assays for fragment C4d, fragment Bb, and the S-C5b-9 complex (S-MAC). In sera from healthy volunteers, virtually no change was observed in C4d either as a function of time or of heparin concentration, whereas changes in Bb and S-MAC were biphasic. This observation was explored in greater detail in the heparin concentration range 0.001-5.0 u/ml (5 x 10(-3) to 25 micrograms/ml). For both Bb and S-MAC, a significant (P < 0.05) increase in production was noted in the heparin concentration range, 0.01-0.5 u/ml (5 x 10(-2) to 2.5 micrograms/ml). At higher heparin concentrations, Bb and S-MAC production decreased markedly (P < 0.05). We reconstituted the alternative pathway amplification cycle (C3, factor B, and factor D) and studied Bb generation. With reactants at concentrations one tenth those of normal serum, we observed a maximum generation of 13.2 micrograms/ml Bb. Control and heparin at 5 x 10(-4) micrograms/ml generated Bb concentrations of 6.8 and 6.1 micrograms/ml, respectively, for a 2-min incubation; at 5 x 10(-3) micrograms/ml heparin, Bb was increased to 9.8 micrograms/ml. Using isoelectric focusing to study anionic pI shifts in heparin-bound factors B and D, it was found that factor B bound heparin only at the highest heparin concentration studied, i.e., 50 micrograms/ml; factor D, however, bound heparin at a much lower concentration (0.05 micrograms/ml). We conclude that, at low concentrations, heparin activates complement due to potentiation of the alternative pathway amplification cycle in the fluid phase.

Spectrotype distributions of circulating IgG from patients with systemic lupus erythematosus.

Pathogenic autoantibodies from patients with systemic lupus erythematosus (SLE) may represent a relatively cationic fraction of IgG. We compared the spectrotype distributions of affinity-purified IgG from the sera of 10 SLE patients and 10 age- and sex-matched control subjects. Purified IgG was subjected to isoelectric focusing between pI 3 and 9. No significant difference was observed for pI 6.0-6.5 and 7.5-8.0. However, control subjects had a higher percent of total IgG at 6.5-7.0 (15.4 +/- 4.0 vs 11.1 +/- 2.0, P = 0.008) and at 7.0-7.5 (22.4 +/- 4.8 vs 18.2 +/- 3.8, P = 0.04) whereas SLE patients had a higher percent of total IgG at 8.0-8.5 (24.3 +/- 3.0 vs 20.5 +/- 4.0, P = 0.03) and at 8.5-9.0 (21.9 +/- 5.9 vs 15.1 +/- 3.7, P = 0.006). Spectrotype distributions of circulating IgG from SLE patients are skewed toward higher pI, providing further evidence of proliferation of B-cell clones that express more cationic IgG in patients with SLE. Longitudinal studies of serum IgG from several patients for > 1 year reveal distinct changes in both cationic and anionic clonotypes, suggesting clonal expansion of antiidiotypes to autoantibodies.

In vitro formation of complement activation products by lipopolysaccharide chemotypes of Salmonella minnesota.

We have applied immunoassays for complement activation products C4d, fragment Bb and the protein S-C5b-9 neoantigen (S-MAC) to assess activation of classical, alternative and terminal pathways, respectively, by lipopolysaccharides (LPS) from the smooth strain (SS) of Salmonella minnesota and the shallow rough (core) mutants R60, R345, R5 and R7. Incubations of sera (n = 6) with LPS generated small and insignificant quantities of Bb and S-MAC in the case of Rb, Rc and Rd chemotypes and slightly greater quantities with Ra. SS-LPS brought about significant (p = 0.01) increases in the formation of both Bb and S-MAC. No significant changes were observed in the concentration of C4d. Polymyxin B enhanced Bb and S-MAC production by SS-LPS, optimally at the lowest concentration of polymyxin B studied, 10 ng/ml. These data confirm and extend observations about complement activation by LPS and suggest that immunoassay may be useful in studying mechanisms of complement activation.

The development of a membrane-based screening method to detect antibodies to intermediate filament proteins.

Autoantibodies directed against the intermediate filament proteins (IF) arise in a variety of disease states. The authors have investigated the binding of the IF to solid membrane supports in a dot blot format in an attempt to develop a simple procedure to detect antibodies (ab) to IF. Commercially obtained, purified IF were utilized. These were: vimentin (VIM), cytokeratin 8 (CYK), glial fibrillary acidic protein (GFA), desmin (DES), and the neurofilament triplet proteins (68, 160, and 200 KDa, respectively designated LMW, MMW, and HMW). Murine monoclonal antibody (mAb) probes were used to detect the presence and immunoreactivity of IF. The mAb were visualized with HRP-anti-mouse conjugates using alpha-chloronaphthol/H 2O 2 as substrate. The membranes studied were nitrocellulose (NC), and two of modified nylon. Nitrocellulose provided the most reproducible binding; no advantage was found to ensue from the use of the other membranes with regard either to quantitative binding or improved capping. Among the IF studied, VIM, GFA, LMW, MMW, and HMW bound well to NC; optimal mass/dot was 1 mug. Filtered, non-fat dry milk is a better capping agent than either albumin or fetal calf serum, but interferes with ab binding to GFA. Binding of CYK and DES is weak at neutral pH. Standard densitometric techniques provide the possibility of quantitation. We conclude that dot and slot blot assays may be practical methods to detect ab to IF antigens.

Immunocytochemical localization of cytoskeletal antigens in KB and HEp-2 cells.

KB and HEp-2 are cell lines that are commonly used for autoantibody detection. We have used conventionally produced and monoclonal antibodies as probes for cytoskeletal antigens in these cells. The presence and distribution of microfilament, microtubule, and intermediate filament antigens (vimentin and keratin) have been established. We have also confirmed that anti-vimentin antibodies of one clone do crossreact with DNA on an immunoadsorbent column, although this antibody did not appreciably stain the nucleus of either cell type.

Autoantibodies and circulating immune complexes in subjects infected with human immunodeficiency virus.

We wish to report the results of a retrospective study of patients with acquired immunodeficiency syndrome (AIDS) and AIDS-related complex (ARC) as well as a group of asymptomatic human immunodeficiency virus (HIV)-infected individuals, in whom we examined markers of systemic autoimmune disease including anti-cell antibodies, immunoglobulin (Ig) and complement (C3 and C4) profiles as well as circulating immune complexes (CIC). Antibodies to cytoplasmic (but not nuclear) antigens were significantly elevated in the AIDS and ARC groups when compared with a population of normal individuals (for AIDS, P = 0.0002 vs. control; for ARC, P = 0.0004 vs. control). The non-AIDS/ARC, HIV+ group did not demonstrate significance (P = 0.054). Each of the three study groups also demonstrated increased Ig (P less than 0.05 for each immunoglobulin class). CIC, as determined by a Clq-binding enzyme immunoassay, were higher in all three study groups (P less than 0.05) when compared with controls. C3 and C4 were not significantly lower than control subjects (P greater than 0.05), but C4 did demonstrate a significant (P = 0.01) inverse correlation with CIC. These findings strengthen the hypothesized autoimmune aspect of AIDS and ARC, and extend this concept to include HIV-infected individuals without frank disease.