Abstract
Autoantibodies targeting factor H (FH), which is a main alternative complement pathway regulatory protein, have been well characterized in atypical hemolytic uremic syndrome (aHUS) but have been less well described in association with alternative pathway–mediated glomerulopathies (GP). In this study, we studied 17 patients presenting with GP who were positive for anti-FH IgG. Clinical data were collected and biological characteristics were compared with those of patients presenting with anti-FH Ab-associated aHUS. In contrast to the aHUS patients, the GP patients had no circulating FH-containing immune complexes, and their anti-FH IgG had a weaker affinity for FH. Functional studies demonstrated that these Abs induced no perturbations in FH cell surface protection or the binding of FH to its ligand. However, anti-FH IgG samples isolated from three patients were able to affect the factor I cofactor activity of FH. Epitope mapping identified the N-terminal domain of FH as the major binding site for GP patient IgG. No homozygous deletions of the CFHR1 and CFHR3 genes, which are frequently associated with the anti-FH Ab in aHUS patients, were found in the GP patients. Finally, anti-FH Abs were frequently associated with the presence of C3 nephritic factor in child GP patients and with monoclonal gammopathy in adult GP patients, who frequently showed Ig Lchain restriction during reactivity against factor H. These data provide deeper insights into the pathophysiological differences between aHUS and GP, demonstrating heterogeneity of anti-FH IgG.
Introduction
Factor H (FH) is the main regulatory protein of the complement alternative pathway. This protein is composed of 20 short consensus repeat (SCR) domains. FH acts by competing with factor B (FB) to bind to C3b, enhancing the dissociation of the C3bBb complex and serving as a cofactor for the factor I (FI)–mediated proteolytic inactivation of C3b to iC3b (1). These functions are carried out by the N-terminal domain of FH (SCR1–4) (2, 3). The last two C-terminal domains bind to glycosaminoglycans as well as to C3b and its cleavage fragment C3d to ensure the anchoring of FH to the cell surface (4–6). FH dysfunction leads to severe renal diseases, including atypical hemolytic uremic syndrome (aHUS) and C3 glomerulopathy (C3G). In contrast with aHUS, C3G belongs to a group of chronic glomerular diseases frequently diagnosed by nephrotic-range proteinuria that result in progressive renal failure after several years. Histologically, C3G is characterized by predominant C3 glomerular deposition. According to the characteristics of deposits observed using immunofluorescence (IF) and electron microscopy, two types of C3Gs have been described to date, that is, C3 glomerulonephritis (C3GN) and dense deposit disease (DDD), which is characterized by the intramembranous deposition of materials, appearing very dense via electron microscopy (7–10). An additional glomerulopathy (GP), type 1 membranoproliferative glomerulonephritis (MPGN I), is also characterized by the deposition of C3 together with Ig components (11, 12). MPGN I is thought to be mediated by immune complexes and arise secondary to infection or autoimmune disorders, but it can occur spontaneously. For these three diseases, dysregulation of the alternative pathway has been reported in association with a C3 convertase stabilizing Ab, which is termed C3 nephritic factor (C3Nef), present in 45–85% of patients or with FH mutations in 10–17% of patients depending on the histological group (12, 13).
aHUS is characterized by acute hemolytic anemia, thrombocytopenia, and endothelial cell damage, leading to acute renal failure. In ∼30% of cases, it is associated with FH mutations located mainly in the C-terminal domain (14), whereas mutations in the N-terminal domain are associated with both aHUS and GP (13, 15–17).
An autoimmune form of aHUS (AI-aHUS), which is caused by the development of autoantibodies directed against FH, has been well described since it was first identified in 2005 (18). This form is present in 6–56% of aHUS cases according to ethnicity, and it occurs mainly in children (18–23). AI-aHUS is highly correlated with a homozygous deletion of two largely homologous complement FH–related (CFHR) genes, CFHR1 and CFHR3 (24–26). Previous studies have shown that anti-FH autoantibodies bind to the FH C-terminal domain (20, 23, 27, 28) and also to the N-terminal domain (23, 27, 28), forming FH-containing immune complexes (CIC-FH) that induce the neutralization of FH activity (27, 28).
Abs targeting FH were first described 20 y ago in a patient presenting with atypical GP with features of both MPGN I and DDD. This patient was found to possess a monoclonal λ L chain dimer able to bind to FH SCR3 (29, 30), impairing the regulatory functions of FH. More recently, complete anti-FH IgG has been reported in seven patients with C3G and in association with monoclonal gammopathy (MG) in two patients (31–36). GP-associated anti-FH autoantibodies have been partially studied in only three patients. In one individual, anti-FH IgG recognized SCR1–4 of FH and was able to decrease FH cofactor activity (35). In two other patients, anti-FH IgG was able to recognize SCR1–5, with additional binding sites localized to SCR7–8 and SCR8–15 detected in one individual (36), but their functional consequences were not studied. An association of anti-FH Abs with C3Nef has been reported in one of these cases (36).
Considering the lack of characterization of these autoantibodies in C3G, we assessed patients with C3G or MPGN I (GP patients) and anti-FH IgG using a larger cohort than in previous studies. Clinical data were collected and the biological characteristics of the anti-FH autoantibodies were compared with those observed in aHUS patients. We detected significant differences between the anti-FH IgG from the aHUS patients and that from the GP patients in terms of binding site localization, functional avidity of the autoantibodies to FH, perturbation of FH activity, and genetic composition.
Materials and Methods
Patients
Samples collected from 17 patients presenting with the clinical and biological characteristics of GP were studied. Complete clinical data were available for 14 patients. Patients presenting with MG underwent bone marrow examination, which was consistent with MG of unknown significance in five patients (not determined in one).
None of the MPGN I patients had a history of infectious disease, MG, or autoimmune disorder. Evaluation for hepatitis B and C was negative for all patients. Antinuclear Abs were positive in three C3G patients without any specificity or anti-DNA Abs.
Complete histological data pertaining to the native kidney were available for 14 of 17 patients. DDD, which was diagnosed in one patient, was defined as a thickening of capillary walls due to both intraglomerular basement membrane (GBM) deposits and double contours (as shown by light microscopy), predominant C3 deposits as shown by IF, and the thickening of the GBM due to electron-dense deposits within the lamina densa (as visualized by electron microscopy). C3GN, which was diagnosed in 10 patients, was defined as mesangial and epimembranous deposits that stained with C3 only in the absence of intramembranous deposits within the GBM as shown by IF. In 60% of the C3GN patients, an MPGN pattern of glomerular changes as defined by mesangial proliferation, an increase in the mesangial matrix, a double contour aspect, and subendothelial and mesangial deposits was observed following staining for C3 as shown by IF. Mild to severe endocapillary proliferation was observed in seven cases and crescent formation in three cases.
MPGN I, which was diagnosed in five patients, was identified via light microscopy by mesangial hypertrophy and hypercellularity associated with subendothelial deposits following staining together with polyclonal anti-Ig and anti-C3 conjugates as shown by IF. The intensities of the C3 deposits were at least twice those of the Ig deposits according to the actual consensus C3G/MPGN classification (7, 10).
Regardless of the underlying GP subtype, sclerotic glomeruli represented <5% of the total glomeruli, and no differences were observed between the two groups. Mild to severe endocapillary hypercellularity was found in 75% of patients, and no differences were detected between the two groups. Epithelial crescents were found in four patients (diffuse in two). Tubular and interstitial fibrosis were absent or mild in 13 patients and moderate to severe in 1.
Samples collected from 19 patients (17 children and 2 adults) presenting with clinical and biological characteristics of aHUS and previously reported patients (28) comprised the comparative group.
Additional samples collected from a group of 28 patients presenting with both GP and MG who were negative for anti-FH IgG and from a group of 11 adult AI-aHUS patients who also tested positive for anti-FH κ and λ L chain reactivity.
Informed consent was obtained from each patient or the parents of the children, and the study was approved by the Ethics Committee (CPP Ile de France V, IDRCB2008-A00144-51).
Complement evaluation
All blood specimens were collected in EDTA tubes, transported at cold temperatures, centrifuged, and stored in 200-μl plasma aliquots at −80°C until use.
The antigenic levels of C3, C4, FB, FH, and FI were determined as previously described (28). The presence of C3 nephritic factor was determined using a hemolytic assay as previously reported (12).
All patients were screened for mutations and polymorphisms in the CFH, CFI, CD46, C3, and FB genes, and the numbers of CFHR1 and CFHR3 genes were determined by multiplex ligation–dependent probe amplification, as previously described (26).
Determination of Ab titers and circulating FH–anti-FH IgG complexes
Anti-FH IgG titers were determined by an ELISA as previously described (37).
The FH–anti-FH IgG circulating immune complexes were quantified using a sandwich ELISA as previously described (28).
IgG purification
Total IgG was purified using protein G–coated Sepharose beads (GE Healthcare, Vélizy-Villacoublay, France) as previously described (28). When precised in some experiments, a Melon gel IgG purification kit (Thermo Scientific, Illkirch, France) was employed, according to the manufacturer’s instructions. To avoid the risk of aggregation, the IgG samples were used extemporarily after the verification of their purity on acrylamide gels and by Coomassie blue staining.
Epitope mapping of anti-FH Abs
The SCR19–20 and SCR1–4 recombinant fragments were expressed and produced as previously described (28). Microtiter plates (Nunc microtiter microplates, Thermo Scientific, Courtaboeuf, France) were coated overnight at 4°C at equal molarities (67 nM) with FH (CompTech, Tyler, TX) or FH fragments (0.01 mg/ml FH, 0.001 mg/ml SCR19–20, and 0.002 mg/ml SCR1–4). After wells were blocked with 1% BSA in PBS, the patients’ samples were diluted 1:50 in 0.1% Tween 20/PBS in the wells and incubated for 1 h at 37°C. Following a washing step with 0.1% Tween 20/PBS, anti-human IgG labeled with HRP (SouthernBiotech, Montrouge, France) diluted to 1:2000 was added to the wells, and bound Abs were visualized with tetramethylbenzidine. Each sample was tested in duplicate.
Study of the specificity of the autoantibodies against FH
Different experiments were performed to assess the specificity of the anti-FH and FH interaction.
Fluid-phase FH-mediated inhibition.
Different concentrations of FH (0.1–10 μg) were incubated with plasma samples diluted with PBS to obtain appropriate OD in the range of 0.6–0.8 (based on anti-FH ELISA results). In parallel, 10 μg/ml FH was used for coating on ELISA plate and incubated at 4°C overnight. Later, blocking was done with 0.1% Tween 20 in PBS at room temperature for 60 min followed by three washes with 0.1% Tween 20/PBS. FH and plasma preincubated samples were then transferred to FH-coated plates and all further steps were carried out as described in the previous experiments.
Competition with anti-FH commercial Abs.
FH (10 μg /ml) was used for coating at 4°C overnight. Blocking was done with 0.1% Tween 20 in PBS at room temperature for 60 min and then three washes were given with 0.1% Tween 20/PBS. This was followed by incubation with goat polyclonal anti-FH Ab (1:100, Quidel) and Abs 90X (directed against SFH SCR1 at dilution 1:100, Quidel), OX24 (directed against FH SCR5 at dilution 1:25, Thermo Scientific), L20/3 and C18/3 (directed against FH SCR20 at dilution 1:100, Santa Cruz Biotechnology) for 30 min at 37°C. After washes, patients’ IgGs were incubated (at an appropriate dilution for OD to be in the range of 0.6–0.8). The further steps were done as previously described.
ELISA of avidity of anti-FH IgG binding to FH
A kinetic assay was performed in a time-dependent manner using an anti-FH IgG ELISA, in which the plasma samples were incubated for 5, 10, 15, 30, and 60 min. Avidity for FH was determined by the slope of the dose-response curve obtained between 5 and 30 min of incubation (absorbance versus time).
Formation of Ag–Ab complexes in the presence of NaCl
The stability of the capacity of the Abs to bind FH in the presence of increasing salt concentrations was assessed by an ELISA. Microtiter plates (Nunc microtiter microplates, Thermo Scientific) were coated overnight at 4°C with FH (CompTech) at 0.01 mg/ml in PBS. After blocking with 1% BSA in PBS for 1 h at 37°C, the plates were washed four times with 0.1% Tween 20/PBS. Samples from patients diluted to 1:50 in PBS/Tween 20 containing increasing concentrations of NaCl (150, 250, and 500 mM) were incubated for 30 min at room temperature. The plates were washed four times with 0.1% Tween 20/PBS, human IgG labeled with HRP (SouthernBiotech, Montrouge, France) diluted to 1:1000 was used, and IgG binding was visualized with tetramethylbenzidine.
Surface plasmon resonance evaluation of stability of FH-anti-FH complexes
The interaction of FH with anti-FH Abs was analyzed using surface plasmon resonance (SPR) technology with Biacore 2000 equipment (GE Healthcare). To study the binding of anti-FH Abs from the GP patients, total IgG was purified using Melon gel resin (Thermo Scientific). FH (CompTech, 1 mg/ml) was diluted in maleate buffer at pH 5 to a final concentration of 15 μg/ml, and it was coupled to a CM5 biosensor chip using standard amine-coupling technology according to the manufacturer’s instructions. IgG was used as an analyte, and serial dilutions were prepared in filtered and degassed PBS (Life Technologies). A total of 80 μl purified IgG was injected for 480 s at a flow rate of 10 μl/min, and dissociation was followed for 480 s on empty (treated with the same chemical procedure but without FH) and FH-coupled flow cells. The specific signal was obtained after subtraction of the signal obtained from the empty flow cell from that from the FH flow cell. The regeneration of the FH-coated surface after each injection of IgG was performed using 3 M NaCl. Because the precise concentrations of anti-FH IgG out of the total IgG in the samples were unknown, kinetic parameters could not be calculated. The stability of the formed complexes was independent of the initial anti-FH IgG concentration and, hence, they were measured by calculating the off-rate using BIAevaluation software.
Cell lysis assays
Nonsensitized sheep erythrocyte hemolytic assay.
This assay was performed as previously described (38, 39). Samples from 12 aHUS patients with anti-FH Abs collected during the acute phase and from 17 GP patients were tested. Plasma samples from healthy donors (n = 20) were used as controls. Diluted plasma (25%) was added to sheep erythrocytes (108) and incubated for 30 min at 37°C in buffer (7 mM MgCl2, 10 mM EGTA [pH 7.2–7.4], 2.5 mM barbital, 1.5 mM sodium barbital, and 144 mM NaCl). After the addition of 1 ml 0.9 M NaCl, sheep erythrocyte lysis was measured according to the absorbance at 414 nm.
Rabbit erythrocyte hemolytic assay.
Purified IgG samples from patients and controls were dialyzed against GVB buffer (0.1% gelatin, 7 mM MgCl2, 10 mM EGTA [pH 7.2–7.4], 2.5 mM barbital, 1.5 mM sodium barbital, and 144 mM NaCl). Normal human plasma was diluted (50%) in Mg-EDTA-GVB, increasing quantities of purified IgG (from 100 to 400 μg/tube) were added, and the mixture was incubated with rabbit erythrocytes (108 cells) suspended in GVB MgCl2 for 1 h at 37°C. After the addition of 0.9 M NaCl, rabbit erythrocyte lysis was measured according to the absorbance at 414 nm.
FI cofactor activity assay
Purified IgG (4 μg) from each patient and control was incubated with 20 ng purified FH (CompTech) for 10 min at 37°C. Purified FI (20 ng; CompTech) and C3(H2O) (100 ng; Calbiochem, Fontenay sous Bois, France) were then added to the mixture and incubated for 30 min at 37°C. The reaction was stopped at 3, 6, 12, and 30 min by adding loading buffer, and a Western blot was performed to reveal the generated C3 fragments using a goat anti-C3 Ab (Calbiochem) followed by incubation with a labeled secondary Ab (rabbit anti-goat HRP; Santa Cruz Biotechnology, Heidelberg, Germany). C3 α43-chain generation was then quantified by densitometry (Thermo Scientific myImageAnalysis software), followed by the calculation of the α43/α′-chain ratio of C3. The ratio at each time point was normalized with that obtained in the presence of IgG purified from healthy controls.
ELISA to assess FH binding to C3(H2O), C3c, and C3d
IgG was purified from patients’ samples using a Melon gel IgG purification kit (Thermo Scientific) according to the manufacturer’s instructions and was used immediately to avoid any IgG precipitation. Purified human FH (0.01 mg/ml, CompTech) was preincubated with purified patient IgG (0.5 mg/ml) overnight at 4°C. Microtiter plates (Nunc microtiter microplates, Thermo Scientific) were coated with 2 μg/ml C3(H2O) (Calbiochem), C3c, or C3d (CompTech) overnight at 4°C.
After blocking with 0.1% BSA in PBS at 37°C for 1 h, preformed FH–anti-FH IgG complexes were added in duplicate to the coated plates. Anti-FH mAbs (OX23) were incubated for 1 h at 37°C (Santa Cruz Biotechnology, Yvelines, France) and diluted to 1:250 to assess FH binding, followed by incubation with HRP-labeled anti-mouse IgG Abs (28). FH-blocking mAbs (OX24 and L20/3, 0.02 mg/ml) (Santa Cruz Biotechnology, Heidelberg, Germany) directed against the N- and C-terminal domains of FH, respectively, were used as controls. Each experiment was performed at least three times.
Ig L chain restriction in anti-FH recognition
Wells of microtiter plates were coated overnight at 4°C with 0.01 mg/ml FH diluted in PBS. After blocking with PBS, 0.1% Tween 20, 1% BSA, and diluted plasma samples (1/20) from the aHUS and GP patients were added and incubated for 1 h at room temperature. After washing with 0.1% Tween 20/PBS, goat anti-L chain κ or anti-L chain λ Abs (SouthernBiotech, Birmingham, AL) were added for 1 h at room temperature. Additional washing steps were performed, followed by incubation with anti-goat Ab labeled with HRP (Santa Cruz Biotechnology, Yvelines, France). Each sample was tested in duplicate at least three times.
The ratios of the optical densities obtained with the anti-κ and anti-λ Abs (κ/λ) were calculated for all samples collected from the adult and children GP and AI-aHUS patients. Samples collected from 28 additional patients presenting with MG and GP without anti-FH IgG and 11 adult AI-aHUS patients were also tested.
Results
Patients
Seventeen patients (GP patients; 5 children and 12 adults) representing 11% of the French C3G/MPGN I cohort presenting with anti-FH Abs were studied (12).
All GP patients were positive for anti-FH IgG (see Fig. 2). Seven patients were positive for both anti-FH IgG and C3Nef, and 10 were positive for anti-FH IgG without C3Nef. Low C3 and FB levels were detected in 29 and 12% of the patients, respectively, but these levels did not significantly differ between the patients with and without C3Nef (Fig. 1 and data not shown). FH antigenic levels were normal in all patients.
Ages at disease onset according to the disease and simultaneous presence of C3Nef. (A) Although AI-aHUS occurred mainly in children, anti-FH Ab–associated GP was found in all age groups but was more frequently associated with C3Nef in children. (B) Distribution of C3 levels according to age and the presence of C3Nef. Dotted line represents the lower limit of the normal values of C3 levels calculated in the laboratory (18). **p < 0.01, ****p < 0.0001 unpaired t test.
Complete clinical data were available for 14 patients (Table I). Based on the simultaneous presence of anti-FH IgG and C3Nef, patients were classified into two groups as follows: those positive for both anti-FH IgG and C3Nef (n = 7) and those positive for anti-FH IgG only (n = 10). The patients with C3Nef were diagnosed at a younger age (median age, 11 y; range, 6–36 y) compared with the other patients (median age, 45 y; range, 8–67 y) (Fig. 1). Nephrotic-range proteinuria was less frequent in the C3Nef+ patients compared with the patients lacking C3Nef (p = 0.038). MG was identified in six patients, all of whom were C3Nef− (p = 0.015). These patients had serum monoclonal IgG (κ in three patients and λ in three patients), and L chain proteinuria was present in two patients.
DDD, C3GN, and MPGN I were diagnosed in 1 (with C3NeF), 11, and 5 patients, respectively.
Characterization of GP anti-FH Abs
Determination of anti-FH IgG and CIC-FH titers.
No significant difference in anti-FH IgG titers was observed between samples collected from the GP patients with a mean of 3004 arbitrary units (AU)/ml (positive threshold, 100 AU/ml; extreme values, 400–15,000 AU/ml) and from the AI-aHUS patients (mean, 6973 AU/ml; range, 1,000–20,830 AU/ml) (Fig. 2). However, no or very low levels of FH–anti-FH IgG immune complexes (CIC-FH) were detected by an ELISA in the samples collected from the GP patients, whereas all AI-aHUS samples exhibited high levels of these complexes (median, 7110 versus 67; p < 0.0001) (Fig. 2). The specificity of the autoantibodies binding was assessed by fluid-phase FH-mediated and by FH polyclonal Ab inhibitions (Supplemental Figs. 1, 2, Table II).
Anti-FH IgG and FH–anti-FH IgG immune complex titers. (A) No significant difference in anti-FH IgG titer between the AI-aHUS patients and the GP patients included in this study was observed. (B) The levels of FH–anti-FH immune complexes as measured by a sandwich ELISA were significantly higher in the AI-aHUS patient samples compared with the GP patient samples. *p < 0.05, ***p < 0.001, ****p < 0.0001 unpaired t test.
Binding characteristics of anti-FH IgG.
The relative avidity of the anti-FH Abs observed in the samples from the GP patients were significantly lower than those obtained from the AI-aHUS patients as measured by a kinetic ELISA (Fig. 3A). The samples from the GP patients also possessed fewer complexes in the presence of high concentrations of NaCl compared with the AI-aHUS patients (p = 0.04 at 1 M NaCl) (Fig. 3B). Finally, IgG purified from three patients (G1, one MG+ and C3Nef− adult; G3 and G4, two adults negative for C3Nef and MG) were studied using SPR. The observed off-rates were 1.64 × 10−3 s−1, 15.8 × 10−3 s−1, and 8.98 × 10−3 s−1 for the three tested patients (G3, G1, and G4, respectively, Fig. 4). No SPR measurement could be performed for the AI-aHUS patients owing to the presence of CIC-FH in the purified IgG fractions.
Characteristics of FH recognition by anti-FH. The relative avidity of the anti-FH IgG toward FH in the GP patients was compared with that from the AI-aHUS patients. (A) A kinetic ELISA allowed for the determination of the slopes of the dose-response curve obtained between 5 and 30 min of incubation (absorbance versus time), which were significantly lower in the GP patient group compared with the AI-aHUS patient group. (B) In the presence of increasing concentrations of NaCl, the GP patient samples showed a significantly lower capacity to form immune complexes compared with the AI-aHUS samples. The experiments were performed at least three times, and the results of a representative experiment are shown. *p < 0.05, ***p < 0.001, unpaired t test.
FH binding activity of purified IgG from GP patient by SPR. Representative data from one patient (G3) are shown.
Determination of antigenic binding domains.
The binding sites of the anti-FH Abs were determined using FH SCR1–4 and SCR19–20 recombinant proteins (Fig. 5). All plasma IgG except one showed significant binding to the N-terminal portion of FH (13 of 14, 93%), which was accompanied by binding to the C-terminal portion in 3 patients (3 of 15, 20%; one child presenting with C3Nef and two adults lacking C3Nef and MG). These results were in accordance with the results of inhibition by different monoclonal FH Abs in selected patients (Supplemental Fig. 2, Table II)
Binding sites of anti-FH IgG in GP patients. Binding to FH (A), SCR1–4 (B), and SCR19–20 (C) were studied. FH or recombinant FH fragments were coated to an ELISA plate at equal molarities and incubated with plasma from the aHUS patients or normal donors. Bound IgG was determined using anti-human IgG HRP. Results represent the mean and SD of three different experiments. All patient samples except two bound to the FH SCR1–4 fragment, and three also bound to the FH SCR19–20 fragment. *p < 0.05, **p < 0.01, ***p < 0.001, unpaired t test.
Functional consequences of FH activity
In the presence of IgG from GP patients, no decreased binding of FH to C3(H2O), C3c, or C3d was observed. In contrast, we observed a significantly increased binding of FH to C3d with IgG in 3 of 14 of the tested patients (21.4%; one adult with MG, one child with C3Nef, and one adult negative for C3Nef and MG) (Fig. 6). Alternatively, in some patients, we observed a dose-dependent inhibition of the IgG and FH interaction by C3b (Supplemental Fig. 3, Table II).
Study of interaction between FH and C3 fragments in the presence of anti-FH IgG as measured by ELISA. C3(H2O) (A), C3c (B), or C3d (C) were coated to a microtiter plate and incubated with FH in the presence of 0.5 mg/ml IgG purified from the patients’ plasma samples. Bound FH was determined, and the percentages of perturbation were calculated using ODs obtained in the presence of the normal donor IgG, which was considered as 100% unperturbed. The monoclonal anti-FH Abs OX24 (against the FH N terminus) and L20/3 (against the FH C terminus), which are known to inhibit FH function, were used as positive controls. All experiments were performed three times independently. *p < 0.05, **p < 0.01, ***p < 0.001, unpaired t test.
We assessed FI cofactor activity by incubating purified IgG from GP patients or normal donors with purified FH, C3b, and FI at 37°C and by examining C3 cleavage at different time points by Western blot. We observed a delay in the generation of the cleaved C3 band in three of eight tested patients (one child with C3Nef and two adults negative for MG and C3Nef) (Fig. 7).
Study of FI cofactor activity in the presence of purified IgG. Purified C3, FH, and FI were incubated at 37°C in the presence of IgG from GP patients or normal donors. The enzymatic reaction was stopped at different time points by the addition of reducing sample buffer. C3 cleavage was revealed by a Western blot using anti-C3–specific antiserum. (A) The result of one representative experiment is shown. (B) The generation of the C3 α43-chain over time was quantified by densitometry using the scanned bands (Thermo Scientific myImageAnalysis software), followed by the calculation of the α43/α′-chain ratio. The ratio at each time point was normalized to that obtained in the presence of IgG purified from healthy controls. For three of eight tested patients, the delayed generation of the C3 α43 band was observed.
No significant lysis of nonsensitized sheep erythrocytes was observed in the presence of plasma-containing anti-FH Abs in 12 GP patients compared from those observed in control samples, whereas lysis induced by the samples from the AI-aHUS patients was significantly higher (Fig. 8).
Sheep erythrocyte lysis assay. Sheep erythrocytes were incubated with diluted plasma (25%) from the GP patients and AI-aHUS patients collected during the acute phase and with healthy donor plasma. No significant lysis was observed in the GP samples containing anti-FH IgG. All experiments were performed three times independently, and the representative data are shown. ***p < 0.001, ****p < 0.0001, unpaired t test.
Rabbit erythrocyte assays were normal for all studied patient plasma samples, and the addition of increasing amounts of IgG purified from the patients to normal human plasma samples did not affect the results (data not shown).
L chain restriction of anti-FH reactivity
Using an ELISA, we studied Ig L chain restriction associated with reactivity against FH. We then calculated κ/λ ratios of the OD values obtained with the anti-κ and anti-λ Abs for all samples in this study as well as for the samples collected from 28 patients presenting with MG and GP that were negative for anti-FH IgG and in the samples collected from 11 AI-aHUS adults.
The Igκ/Igλ ratios of the OD values were more widely distributed in the samples collected from the adult GP patients (0.0006–942.5; mean, 88.7) than from the child patients (mean, 1.15) and the AI-HUS adult and child patients (means, 1.11 and 0.89, respectively) (Fig. 9). This distribution was related to λ L chain restriction (ratio, <0.1) in three GP adult patients and κ L chain restriction (ratio, >3) in three other GP adult patients.
Study of Ig L chain restriction of anti-FH recognition by ELISA. The binding of anti-FH IgG was revealed using an anti-κ or anti-λ human Ig L chain. The ratio of the OD obtained using each Ab is represented. The results are classified according to the disease and patient age. A larger distribution of ratios was observed in the adult GP patient group due to a λ (ratio, <0.1) or κ (ratio, >3) L chain restriction. The dotted lines represent the normal values of the ratios of the total serum-free L chains.
We evaluated 38 patients in total presenting with MG and GP and found an imbalance in the Igκ/Igλ ratios of the OD values in 14 of 38 patients (38%) together with the absence of anti-FH IgG in 4 of them (4 of 38, 10.5%), suggesting the presence of anti-FH monoclonal L chains (2 κ and 2 λ) (data not shown).
Discussion
A subgroup of aHUS (the autoimmune form AI-aHUS, which comprises 10% of European aHUS cases) and C3G and MPGN I (11% in the French cohort) are characterized by the presence of Abs against the major complement regulator FH. The anti-FH Abs found in AI-aHUS patients have been well studied in contrast with those found in GP patients. In the present study, we report the substantial differences existing between these anti-FH Abs according to the clinical context of development in terms of the epitopes present and mechanisms of action. Despite similar titers, anti-FH Abs from the GP patients differ from those from the AI-aHUS patients as follows: 1) they possessed weaker interactions with FH (Fig. 3), 2) they did not form circulating immune complexes (Fig. 2), and 3) they bound to different epitopes (Fig. 5). They also 4) perturbed fluid-phase regulation in one third of cases (Fig. 7), and 5) they did not affect cell surface complement control (Fig. 8). Additionally, no association with the CFHR1/R3 deletion was found in the GP patients in contrast with the aHUS patients. Finally, anti-FH Abs in the GP patients were frequently associated with an additional Ab, such as C3Nef or MG. These differences may explain why Abs directed against the same protein may be associated with two distinct diseases.
High titers of anti-FH Abs were found in both the GP and AI-aHUS patients. Nevertheless, these Abs did not share the same properties. The high titers of FH-containing circulating immune complexes present during the acute phase of AI-aHUS but not in the GP patients may be explained by differences in the binding characteristics of these Abs. Stable immune complexes are only formed when high-ffinity Ag–Ab binding is present. Indeed, we demonstrated the strong binding of the FH Abs in the aHUS patients, as measured by a time-dependent kinetic ELISA (28). Using the same test, the GP Abs showed a weaker binding avidity (Fig. 3A). SPR analysis also revealed the rapid off-rate of the GP anti-FH Abs (Fig. 4), indicating the weak stability of the complexes. Taken together, these phenomena can explain the lack of stable circulating FH–anti-FH immune complexes observed in the GP samples. Additionally, the GP Abs bound less frequently to FH in the presence of high concentrations of NaCl compared with the AI-aHUS Abs (Fig. 3B). A predominantly electrostatic protein–protein interaction is sensitive to increased ionic strength, but a largely hydrophobic interaction is less sensitive (40). Therefore, the processes of Ag binding by anti-FH Abs that occur in these two pathologies seem to be governed by different forces. Our results suggest that in AI-aHUS patients, anti-FH Abs rely mainly on hydrophobic interactions, whereas in GP patients, anti-FH Ab interactions are mostly electrostatic. Affinity maturation has been found to improve the binding of protein-specific Abs through the burial of increased proportions of the hydrophobic surface at the interface with the Ag at the expense of the polar surface (41). This may suggest that anti-FH Abs in AI-aHUS patients result in the advanced maturation of developing B cells in contrast with those in GP patients.
The anti-FH Abs in GP patients also differed from those in AI-aHUS patients by the localization of their binding epitopes. Anti-FH GP Abs bound predominantly to the N-terminal domain in accordance with recent reports in the literature (35, 36). Interestingly, 20 y ago, a patient presenting with MG and features of both MPGN I and DDD was reported (30). This disease was explained by the presence of λ L chains in the patient’s plasma that were able to bind to the SCR3 of FH and impair its regulatory activity (29).
The localization of binding epitopes to the regulatory portion of FH may partially explain the perturbation of the fluid-phase cofactor activity of FH observed in 37% of the tested cases. Nevertheless, we did not observe any functional defect in FH in the other patients, as measured in vitro. Considering the proven specificity of the autoantibodies (Supplemental Figs. 1, 2), this absence of detection of any FH dysfunction in some patients may be due to technical limitations such as the loss of IgG functionality during the purification process. For all tested GP plasma samples, cell surface protection proceeded as normal, suggesting the preserved binding of FH to the cell surface. This is in contrast with the anti-FH Abs found in the AI-aHUS samples, which bound to both the regulatory N-terminal and cell surface–anchoring C-terminal domains and markedly perturbed cell surface protection (20, 23, 27, 39). In individuals with aHUS-associated FH mutations, a disturbance in cell protection against complement attack is observed similar to what is observed in individuals with anti-FH Abs (27, 28, 38, 39, 42, 43). In contrast, in GP patients, FH mutations are predominantly located in the N-terminal region of FH, and they mainly affect fluid-phase control (15, 16, 44).
GP-associated anti-FH IgG was not able to decrease the binding of FH to the C3b-like form of C3, C3c, and C3d, in contrast with what was observed in the AI-aHUS patients (28). However, the significantly increased binding of FH to C3d in the presence of purified IgG was observed in ∼30% of the GP patients. This intriguing enhancement of FH-C3d binding may be explained by the interaction of IgG with two FH molecules. The binding of one of these molecules to C3d could bring a second molecule to the C3d-coated surface. Because FH circulates in a back-folded hairpin-like form, and the C-terminal domain interacts with the proximal portion of the molecule (45, 46), a potential alteration in the conformation of FH upon IgG binding to the N-terminal domain may also lead to the increased exposure of the C3d binding site at the C terminus, contributing to the increased C3d binding. In contrast, we observed a decreased binding of the IgG to FH when the latter was preincubated with increasing doses of coated C3b. These results suggest that the binding properties of the IgG may be dependent on the conformational states of FH, which are different when the protein is in fluid phase or bound to C3b or on cell surface.
Taken together, our results indicate that the presence of anti-FH IgG in GP patients has moderate functional consequences and distinct pathogenicity than those in AI-aHUS patients. This may explain why they are associated with chronic diseases leading to progressive renal insufficiency rather than an acute pathology such as aHUS. Similar situations have been reported in other diseases. In antiphospholipid syndrome, the presence of high-avidity autoantibodies directed against β2-glycoprotein I is associated with circulating immune complexes and is correlated with a higher pathogenicity (47).
Heterogeneous mechanisms of immunization could also explain the differences observed in the anti-FH Abs that developed in the patients with these two diseases. The homozygous deletion of CFHR1, which is strongly associated with AI-aHUS occurrence (22, 24, 26), was not found in any of the GP patients. This difference in CFHR1 genetic status may affect the mechanisms underlying the generation of anti-FH autoantibodies as well as their functional avidity and roles in the pathology of the associated disease.
Anti-FH Abs in GP patients may be associated with an additional autoantibody, such as C3Nef (36). In our cohort, 7 of 17 (41%) patients also had C3Nef. All but one of the patients were children, and 71% of the anti-FH+ children exhibited both anti-FH IgG and C3Nef. In this situation, the assessment of the specific contribution of each of these autoantibodies is particularly difficult. In some patients, Abs may have a cumulative or synergistic effect in the induction of alternative pathway overactivation. We did not observe a specific profile of the anti-FH IgG associated with C3Nef, but the limited number of patients assessed in this study did not allow us to address this question in detail. Recently, anti-C3 and anti-FB autoantibodies have also been reported in association with GP (48, 49). The association of these autoantibodies targeting different alternative pathway components may suggest that they induce similar types of kidney damage.
Interestingly, 6 of 17 (35%) of the GP patients who developed anti-FH autoantibodies also presented with MG. This particular association was found only in adult patients (50% of the adults) and confirms previous reports (29–31). The presence of MG was not associated with specific functional characteristics of anti-FH IgG. We studied L chain restriction implicated in anti-FH reactivity. With regard to the anti-C1 inhibitor leading to acquired deficiency (50, 51), we showed that the anti-FH autoantibodies possessed the same L chain as monoclonal Ig, suggesting that MG may be responsible for anti-FH reactivity. Thus, the ratio of κ/λ anti-FH reactivity could be a reliable marker of the evolution of GP in the context of MG (52).
In this work, we demonstrated that anti-FH Abs found in patients with GP display distinct properties compared with those found in AI-aHUS patients. Owing to the association of these Abs with MG and C3Nef in the GP patients, we strongly recommend screening for MG and the presence of other autoantibodies in patients presenting with glomerular pathologies associated with anti-FH Abs. This work aids in the elucidation of this rare example of autoantibodies directed against the same protein and the mechanisms underlying their association with two distinct diseases.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank N. Poulain, S. Ngo, P. Bordereau, and T. Rybkine for technical assistance. We thank the Société de Néphrologie Pédiatrique, the Société Française de Néphrologie, and all the clinicians who referred patients including J.-J. Boffa and E. Rondeau, Service de Néphrologie, Hôpital Tenon, Paris, France; H. Boulmerka, Service de Néphrologie, Montargis, France; D. Chauveau, Service de Néphrologie, Toulouse, France; S. Cloarec, Service de Néphrologie Pédiatrique, Tours, France; M. Fischbach, Service de Néphrologie Pédiatrique, Strasbourg, France; N. Godefroid, Service de Néphrologie, Brussels, Belgium; M. Hazzan, Service de Néphrologie, Lille, France; A. L. Lapeyraque, Service de Néphrologie Pédiatrique, Montreal, QC, Canada; P. Le Pogamp, Service de Néphrologie, Rennes, France; V. Moal, Service de Néphrologie, Marseille, France; S. Ohlmann, Service de Néphrologie, Strasbourg, France; C. Pietrement, Service de Néphrologie Pédiatrique, Reims, France; G. Touchard, Service de Néphrologie, Poitiers, France; and L. Vrigneaud, Service de Néphrologie, Valenciennes, France.
Footnotes
C.B., S.K.T., and S.C. performed research, analyzed data, and wrote the paper; M.-A.D.-D. designed the research, analyzed data, and wrote the paper; and M.L.Q., B.M., M.B., T.S.J., L.T.R., and V.F.-B. analyzed data and wrote the paper. All authors read and approved the submission of this manuscript.
This work was supported in part by Direction de la Recherche Clinique de l’Assistance Publique–Hôpitaux de Paris Grant CIRC 06037 and by European Union Seventh Framework Programme Grant 2012-305608 (EURenOmics). S.C. and S.K.T. are funded by fellowships from the Fondation pour la Recherche Médicale and the Indo-French Centre for the Promotion of Advanced Research, respectively.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- aHUS
- atypical hemolytic uremic syndrome
- AI-aHUS
- autoimmune form of aHUS
- AU
- arbitrary unit
- CFHR
- complement FH–related
- C3G
- C3 glomerulopathy
- C3GN
- C3 glomerulonephritis
- CIC-FH
- FH–containing immune complexes
- C3Nef
- C3 nephritic factor
- DDD
- dense deposit disease
- FB
- factor B
- FH
- factor H
- FI
- factor I
- GBM
- glomerular basement membrane
- GP
- glomerulopathy
- IF
- immunofluorescence
- MG
- monoclonal gammopathy
- MPGN I
- type 1 membranoproliferative glomerulonephritis
- SCR
- short consensus repeat
- SPR
- surface plasmon resonance.
- Received October 30, 2014.
- Accepted March 31, 2015.
- Copyright © 2015 by The American Association of Immunologists, Inc.
References
In this issue
