Characterization of autoantibodies from patients with Goodpasture's disease using a resonant mirror biosensor

Abstract

Goodpasture's disease is characterized by the binding of IgG autoantibodies to the glomerular basement membrane, leading to glomerular inflammation. The autoantigen has been identified as the noncollagenous domain of the alpha3 chain of type IV collagen (alpha3(IV)NC1). We have used the IAsys resonant mirror biosensor to analyse the extent and affinity of binding of anti-GBM antibodies from sera of patients to purified alpha3(IV) NC1. alpha3(IV) NC1 monomers were immobilized to a carboxylate cuvette, with the simultaneous use of a control well. The binding of serum from patients with Goodpasture's disease (n = 12), normal controls (n = 14) and disease controls with vasculitis (n = 14) was analysed. Antibody binding was detected in sera from all patients with Goodpasture's disease but not from controls. IAsys measurements of binding correlated with antibody levels assessed by the standardized ELISA used for clinical assays. Both ELISA and biosensor measurements showed declining antibody levels in serial serum samples from treated patients; however, the biosensor detected antibody recrudescence when ELISA remained negative. Autoantibodies from patients' serum had average affinity constants (Kd) of 6.5 x 10-11M to 52.07 x 10-10M, as determined by an inhibition assay, indicating high affinity. Sips analysis showed that the antibody response was relatively homogeneous (values of 0.46-1). Biosensor techniques can therefore be used to detect and characterize anti-GBM antibodies in serum from patients, with high sensitivity and without need for antibody purification. This technique may be useful in diagnosis and monitoring of patients with Goodpasture's disease, and may be applicable to other autoantibody mediated diseases.

Fig. 1

Biosensor profile of the association and dissociation of individual samples to α3(IV)NC1 (upper line) and bovine serum albumin (lower line). (a) Serum from three patients with anti-GBM antibody disease applied sequentially to the biosensor, with 5 min of association, dissociation and then regeneration of the cuvette with hydrochloric acid; (b) a single binding curve for the monoclonal antibody to α3(IV)NC1 (MoAb 17).

Figure 2

Binding of serum from patients with anti-GBM antibody disease, normal controls and patients with ANCA-associated vasculitis, to a3(IV)NC1 in the IAsys biosensor and by conventional ELISA. The graphs show: (top) biosensor binding quantified by measurement of initial slope; (middle) biosensor binding quantified by measurement of total binding response at 5 minutes; (bottom) binding in ELISA.

Figure 3

Comparison of antibody binding from patients with anti-GBM antibody disease in ELISA and biosensor assays: (a) correlation between the biosensor total binding response and ELISA (r = 0·77, P = 0·0031); (b) correlation between the two different biosensor measures − total binding response at 5 minutes and initial slope (r = 0·93, P = 0·0001).

Figure 4

Serial antibody samples obtained during and after therapeutic plasmapheresis from one patient with Goodpasture's disease assessed by IAsys biosensor (top) and ELISA (bottom). Antibody levels decrease throughout treatment with each measurement, but the biosensor detects recrudescence of antibody on day 21 which is not seen in the ELISA. Limits of detection of each assay shown by the horizontal line.

Figure 5

Affinity analysis and Sips plots of antibodies from patients with anti-GBM antibody disease. (a) and (b) show plots from two patients of rc againstc,whererc is the number of bound antigen molecules per antibody andcis the molar concentration of free antigen.Kd is obtained from the curve, where rc = c/(c + Kd).In (a)Kd = 5·2 × 10⁻⁹M (r2 = 0·98), and in (b) Kd = 1·93 ¥ 10⁻⁹M (r2 = 0·95). (c) and (d) show Sips plots of the distribution of the antibody affinities, and demonstrate in both cases a relatively homogeneous population: in (c), a = 0·7589 and (d), a = 0·69.