The Alternative Pathway Is Necessary and Sufficient for Complement Activation by Anti-THSD7A Autoantibodies, Which Are Predominantly IgG4 in Membranous Nephropathy

Abstract

Membranous nephropathy (MN) is an immune kidney disease characterized by glomerular subepithelial immune complexes (ICs) containing antigen, IgG, and products of complement activation. Whereas proteinuria is caused by complement-mediated podocyte injury, the pathways of complement activation remain controversial due to the predominance of IgG4 in ICs, an IgG subclass considered unable to activate complement. THSD7A, a transmembrane protein expressed on podocytes, is the target autoantigen in ~3% of cases of primary MN. In this study, we analyzed sera from 16 patients with THSD7A-associated MN with regard to the anti-THSD7A IgG subclasses and their ability to fix complement in vitro. The serum concentration of anti-THSD7A IgG varied over two orders of magnitude (1.3-243 μg/mL). As a relative proportion of all IgG anti-THSD7A, IgG4 was by far the most abundant subclass (median 79%), followed by IgG1 (median 11%). IgG4 was the dominant subclass of anti-THSD7A antibodies in 14 sera, while IgG1 was dominant in one and co-dominant in another. One quarter of MN sera additionally contained low levels of anti-THSD7A IgA1. ICs formed by predominantly IgG4 anti-THSD7A autoantibodies with immobilized THSD7A were relatively weak activators of complement in vitro, compared to human IgG1 and IgG3 mAbs used as positive control. Complement deposition on THSD7A ICs was dose-dependent and occurred to a significant extent only at relatively high concentration of anti-THSD7A IgG. C3b fixation by THSD7A ICs was completely abolished in factor B-depleted sera, partially inhibited in C4-depleted sera, unchanged in C1q-depleted sera, and also occurred in Mg-EGTA buffer. These results imply that THSD7A ICs predominantly containing IgG4 activate complement at high IgG4 density, which strictly requires a functional alternative pathway, whereas the classical and lectin pathways are dispensable. These findings advance our understanding of how IgG4 antibodies activate complement.

Keywords: IgG4 antibodies; THSD7A (thrombospondin type 1 domain-containing protein 7A); alternative pathway of complement; complement activation; membranous nephropathy.

Figure 1

Analysis of IgG subclasses of anti-THSD7A autoAbs. (A) Standard curves show the concentration dependence for the binding of recombinant human anti-TNFα IgG1 (blue), IgG2 (orange), IgG3 (green) and IgG4 (red) to wells coated with TNFα antigen. (B) Sera from 16 patients with THSD7A-associated MN were analyzed by ELISA for IgG autoAb subclasses binding to immobilized THSD7A. The serum concentrations of anti-THSD7A IgG1 (blue), IgG2 (orange), IgG3 (green) and IgG4 (red) in each MN serum were interpolated from standard curves using recombinant human IgG1-IgG4 mAbs anti-TNFα. Sera were designated MN-1 to MN-16 in decreasing order of their levels of anti-THSD7A IgG, calculated as the sum of the concentrations of all four IgG subclasses. (C) The correlation between serum levels of anti-THSD7A IgG measured by ELISA and the anti-THSD7A titers measured by indirect immunofluorescence test (IIFT) was significant (Pearson r=0.849, p=0.0001). (D) The relative proportion of IgG subclasses in MN sera was calculated by dividing the serum concentration of each subclass to sum of concentrations for all four IgG subclasses of autoAbs. The error bars show the median and the interquartile range.

Figure 2

Identification of IgA anti-THSD7A autoAbs. (A) Quantitative analysis of anti-THSD7A IgA in MN sera (circles). The threshold for positivity, depicted by the dotted line, was defined as the mean + 4SD of control sera (tringles). B-C. IgA-positive MN sera (circles) show significantly higher binding to THSD7A of IgA1 (B) but not IgA2 (C), compared to IgA-negative MN sera (squares) or control sera (triangles). The significance of differences among groups was analyzed by one-way ANOVA follow by Bonferroni’s test for multiple comparison (ns, not significant; *p<0.05; **p<0.01).

Figure 3

Development of a functional assay for complement fixation by THSD7A ICs. (A) Schematic representation of the in vitro assay for complement activation. (B) Simplified diagram of the complement activation pathways. The complement proteins for which the binding to ICs was assayed in this study are shown in red.

Figure 4

In vitro complement activation by THSD7A ICs under conditions in which all activation pathways are active. Complement activation was performed using 10% normal human serum (NHS) diluted in GVB++ buffer containing Ca²⁺ and Mg²⁺ in wells coated with THSD7A (200 ng/well) and then incubated with sera MN-1 through MN-14 (solid black circles) or control sera (open blue circles), all diluted 1/25. For each sample, the x-coordinate indicates the concentration of anti-THSD7A IgG in the diluted MN serum. In parallel experiments, activation was performed on reference ICs prepared by incubating human IgG3 (1 μg/ml, solid red circle) and IgG4 (1 μg/ml, open red circle) anti-GFP mAbs in GFP-coated wells (200 ng/well). Graphs depict the binding of antibodies specific to C3b/iC3b (A), C5b-9/TCC (B), FP (C), FB (D), C4 (E), and C1q (F). A blank correction was performed by subtracting the OD value on immobilized antigen incubated with PBS from the OD value of immobilized antigen incubated with antibody. The threshold for positivity (blue dashed line) was calculated as the mean OD of the control sera + 4 SD. Labels indicate MN sera which showed complement deposition above the positivity threshold. The OD values for MN sera were fitted to the concentration of anti-THSD7A IgG in each sample (values taken from Figure 1 and divided by the dilution factor) by simple linear regression (black line). The red line indicates the baseline obtained by performing the complement activation assays using 10% heat-inactivated NHS in GVB++ as a source of complement.

Figure 5

The alternative pathway is essential for complement activation by THSD7A ICs. (A) Plot depicting the deposition of C3b/iC3b (expressed as OD) onto THSD7A ICs formed by MN1-MN4 sera (black symbols) or in THSD7A-coated wells incubated with buffer or control serum (blue symbols). Complement activation was performed in 10% normal human serum (NHS), C1q-depleted serum (ΔC1q), C4-depleted serum (ΔC4), and FB-depleted serum (ΔFB) in buffer containing Ca2+ and Mg2+, as well as 10% NHS in EDTA buffer. The complement activation pathway active under each condition are shown in red (CP, classical pathway; LP, lectin pathway, AP, alternative pathway). (B) The relative inhibition of C3b/iC3b deposition of onto THSD7A ICs formed by MN1-MN4 sera was calculated for C1q-, C4- and FB-depleted sera on a scale from 0% inhibition (10% NHS in buffer containing Ca2+ and Mg2+) to 100% inhibition (10% NHS in EDTA buffer). C3b/iC3b deposition was significantly inhibited in FB-depleted serum lacking a functional alternative pathway (p=0.0006, one-way repeated measures ANOVA followed by post-hoc Dunnet test). (C) Deposition of C3b/iC3b on human monoclonal IgG1 and IgG3 anti-GFP (1 μg/mL each) bound to TNFα was measured after incubation with 10% NHS or C1q-, C4 and FB-depleted sera in buffer containing Ca2+ and Mg2+ or in 10% NHS in EDTA-containing buffer as source of complement.

Figure 6

In vitro complement activation by THSD7A ICs in MgEGTA buffer, which inhibits the classical and lectin pathways. (A) Concentration dependence of C3b/iC3b deposition on ICs formed by anti-THSD7A antibodies (black circles) from MN-1, MN-2, MN-3 and MN-4 sera (all diluted 1/25) bound to immobilized THSD7A or by anti-GFP human IgG1 (red circles) and anti-GFP IgG4 (orange circles) mAbs bound to immobilized GFP. (B) C3b/iC3b deposition on ICs formed by IgG1 mAb (1 μg/mL) and IgG4 mAbs (3 μg/mL) bound to immobilized GFP. Complement activation was performed using 10% NHS diluted in buffer containing Ca²⁺ and Mg²⁺, MgEGTA, or EDTA. Inhibition of classical and lectin pathways by MgEGTA reduced C3b/iC3b binding to IgG1 ICs by >80% but had minimal effect on C3b binding to IgG4 ICs. (C) The deposition of C3b/iC3b (blue), factor B (FP, red) and properdin (FP, green) on ICs formed by anti-GFP IgG4 mAb (3 μg/mL) was completely abolished when using FB-depleted serum or normal serum containing EDTA as complement source, but largely unchanged in C4-depleted serum.