Lupus-specific kidney deposits of HSP90 are associated with altered IgG idiotypic interactions of anti-HSP90 autoantibodies

Abstract

Previous studies have shown that autoantibodies to heat shock protein 90 (HSP90) are elevated in a significant proportion of patients with systemic lupus erythematosus (SLE) who are more likely to have renal disease and a low C3 level. Using samples from 24 patients, we searched for glomerular deposits of HSP90 in renal biopsy specimens from seven patients with lupus nephritis and 17 cases of glomerulonephritis from patients without SLE. Positive glomerular immunofluorescent staining for HSP90 was observed in six of seven cases of SLE and positive tubular staining in two of seven SLE patients. The staining for HSP90 was granular in nature and was located in subepithelial, subendothelial and mesangial areas. None of the non-SLE renal biopsies revealed positive staining for HSP90 deposition. Further we showed the presence of anti-HSP90 IgG autoantibodies in IgG from sera of patients with SLE as well as in normal human IgG (IVIg). In normal IgG this autoreactivity could be adsorbed almost completely on F(ab')2 fragments from the same IgG preparation, coupled to Sepharose and could be inhibited by the effluent obtained after subjecting normal IgG to HSP90 affinity column. These findings indicate that anti-HSP90 natural autoantibodies are blocked by idiotypic interactions within the IgG repertoire. Unlike natural autoantibodies, anti-HSP90 IgG from SLE patients' sera were only moderately adsorbed on F(ab')2 fragments of normal IgG. These results demonstrate that immunopathogenesis of lupus nephritis is associated with HSP90 (as an autoantigen) and that the pathology is associated with altered idiotypic regulation of the anti-HSP90 IgG autoantibodies.

Fig. 1

Indirect immunofluorescene. Cryostat sections of renal biopsy material from patients with glomerulonephritis were treated consequently with anti-HSP90 monoclonal antibody (MoAb 3F12) and antimouse IgG conjugated with FITC. (a) Glomerulus from SLE patient. Mostly subendothelial and subepithelial glomerular pattern of staining. Original magnification 250×. (b) Glomerular capillary loop from SLE patient. Intense mesangial and subepithelial granular staining. Original magnification 1250×. (c) Tubular staining and tubular atrophy in SLE patient. Original magnification 500×. (d) Glomerulus from patient with non-SLE associated glomerulonephritis. No staining for HSP90 can be seen. Original magnification 250×.

Fig. 2

Reactivity with human HSP90 of anti-HSP90 antibodies, affinity-purified from different IgG preparations in cross-blot. The antigen is deposited on PVDF membrane in series of dilutions. All antibodies are applied at concentration 0·1 mg/ml of IgG. The figure shows graphically the results from the densitometry of the cross-blot. Thin lines: binding of IgG samples from SLE sera; thick lines: binding of IgG from healthy donors (filled triangles: pooled IgG from five healthy donors; filled squares: IVIg), dashed line with open circles: flow-through of IVIg through HSP90 column. On the abscissa: the concentration of HSP90 during deposition on PVDF membrane. On the ordinate: staining intensity, measured by densitometry on a grey scale of 256 discrete levels. The actual data are obtained by subtraction of the background reactivity of the antibodies to the blocked membrane.

Fig. 3

Adsorption of anti-HSP90 reactivity of anti-HSP90 antibodies (filled symbols) from IVIg (a), from two lupus sera (b) and from a single lupus serum (c) by passing through affinity column prepared from IgM. (a) IVIgM, open circles; (b) autologous IgM, open circles and crosses. (c) Effect of both columns (open circles, IVIgM; crosses, autologous) assayed in cross-blot. For the conditions of the cross-blot see Fig. 2.

Fig. 4

Anti-HSP90 autoreactivity in normal (IVIg) (circles) and SLE patient's IgG (squares) before (filled symbols) and after (open symbols) passing through IVIgM F(ab′)2 affinity column. Densitometry data from cross-blot. The conditions are the same as in the experiment presented in Fig. 2.

Fig. 5

Effect of reducing the content of anti-idiotypic antibodies in IgG preparations on the antiself HSP90 reactivity assayed by cross-blot. IVIg (a) or SLE patients’ IgG samples ((b) shows a typical case) were passed through affinity column containing F(ab′)₂ fragments prepared from IVIg. The anti-HSP90 reactivity of the effluents (dashed line) was compared to that of the untreated IgG sample (solid line). The conditions of the cross-blot are as in the experiment described in Fig. 2. The differences in the reactivity of untreated and treated IgG to antigen deposited at 0·05 mg/ml on PVDF membrane were quantitatively compared between IgG from SLE patients and IVIg (c). Results are presented for IgG samples from eight SLE sera and IVIg, all tested in two experiments at 0·1 mg/ml in cross-blot. The fall in anti-HSP90 reactivity of IVIg after passing through F(ab′)₂ column is significantly greater than that of SLE IgG (P < 0·05, Mann–Whitney U-test).

Fig. 6

Specificity of the anti-idiotype interactions leading to reduced anti-HSP90 activity in IVIg. Samples of IVIg were passed through F(ab′)₂ (dashed line) and F(ab′)₂-Id columns (filled diamonds) (see text for details) and their anti-HSP90 reactivity was compared to that of non-treated IVIg (filled triangles) in cross-blot. For the conditions of the cross-blot see Fig. 2. Only a column with F(ab′)₂ fragments, derived from normal IgG with intact repertoire, retains the anti-HSP90 reactivity of IVIg while a column with F(ab′)₂ fragments from the IgG pool, from which the anti-idiotypic antibodies have been removed, does not change this reactivity.

Fig. 7

Inhibition of anti-HSP90 reactivity of anti-HSP90 antibodies from IVIg (filled squares) by flow-through of HSP90 affinity column (dashed line, open circles) and by flow-through of HSP90 affinity column also passed through a F(ab′)₂ affinity column (open squares) assayed in cross-blot. For the conditions of the cross-blot see Fig. 2.