IgM Autoantibodies to Complement Factor H in Atypical Hemolytic Uremic Syndrome

Significance Statement

Atypical hemolytic uremic syndrome (aHUS) is often related to complement dysregulation, but its pathophysiology remains unknown in at list 30% of patients. Anti-factor H autoantibodies of the IgG class are responsible for 10% of patients with aHUS; autoantibodies of IgM class have not been reported. The authors found anti-factor H IgM autoantibodies in seven of 186 patients with aHUS, with a frequency six-fold higher in patients with a history of hematopoietic stem cell transplantation. The purified IgM autoantibodies recognize the active site of the factor H molecule and inhibit its binding to C3b. These findings indicate that some forms of aHUS of unknown origin could be placed within the setting of autoimmune diseases, stemming from the presence of IgM autoantibodies specific for factor H’s active site.

Visual Abstract

Abstract

Background

Atypical hemolytic uremic syndrome (aHUS), a severe thrombotic microangiopathy, is often related to complement dysregulation, but the pathomechanisms remain unknown in at least 30% of patients. Researchers have described autoantibodies to complement factor H of the IgG class in 10% of patients with aHUS but have not reported anti-factor H autoantibodies of the IgM class.

Methods

In 186 patients with thrombotic microangiopathy clinically presented as aHUS, we searched for anti-factor H autoantibodies of the IgM class and those of the IgG and IgA classes. We used immunochromatography to purify anti-factor H IgM autoantibodies and immunoenzymatic methods and a competition assay with mapping mAbs to characterize interaction with the target protein.

Results

We detected anti-factor H autoantibodies of the IgM class in seven of 186 (3.8%) patients with thrombotic microangiopathy presented as aHUS. No association was observed between anti-factor H IgM and homozygous deletions involving CFHR3-CFHR1. A significantly higher proportion of patients with bone marrow transplant–related thrombotic microangiopathy had anti-factor H IgM autoantibodies versus other patients with aHUS: three of 20 (15%) versus four of 166 (2.4%), respectively. The identified IgM autoantibodies recognize the SCR domain 19 of factor H molecule in all patients and interact with the factor H molecule, inhibiting its binding to C3b.

Conclusions

Detectable autoantibodies to factor H of the IgM class may be present in patients with aHUS, and their frequency is six-fold higher in thrombotic microangiopathy forms associated with bone marrow transplant. The autoantibody interaction with factor H’s active site may support an autoimmune mechanism in some cases previously considered to be of unknown origin.

Ample evidence exists on the pathogenetic role of complement activation in atypical hemolytic uremic syndrome (aHUS), a thrombotic microangiopathy (TMA) characterized by nonimmune-mediated hemolysis, thrombocytopenia, and acute renal damage.^1,2 Disease manifestations are due to dysregulated or inappropriate complement activation related to genetic or acquired factors. Genetic factors, including mutations of genes encoding complement regulators (factor H [FH], factor I, and membrane cofactor protein) or gain-of-function mutations of genes encoding complement activators (C3 and factor B), render the complement system hyperactive.¹ Acquired factors lead to excessive complement activation mainly by FH inhibition via autoantibodies.³ These autoantibodies usually develop on a genetic background characterized by a homozygous deletion involving complement factor H–related 1 (CFHR1) and CFHR3 genes, and the consequent disease is named deficiency of complement factor H–related plasma proteins and autoantibody-positive form of hemolytic uremic syndrome (DEAP-HUS). However, in approximately 30% of patients, aHUS arises from unknown mechanisms.² Complement involvement has been also demonstrated in the pathogenesis of transplant-associated thrombotic microangiopathy (TA-TMA),⁴ a severe complication of hematopoietic stem cell transplantation that may affect 10%–20% of recipients⁵ with a survival of 16.7% at 1 year in untreated patients.^4–6

Autoantibodies to FH of the IgG class have been described in complement-driven TMAs, both in aHUS^7–9 and in TA-TMA,⁴ whereas anti-FH autoantibodies of the IgM class have never been reported in these conditions. On the basis of our experience in acquired angioedema due to C1 inhibitor deficiency, in which we have found autoantibodies to C1 inhibitor not only of the IgG and IgA classes but also, of IgM,¹⁰ we assessed all of the three classes of autoantibodies to FH in our case list of patients with aHUS. Anti-FH IgM autoantibodies were purified and characterized with regard to their interaction with the target protein.

Methods

Patients

One hundred and eighty-six patients with clinical presentation of aHUS were studied for the presence of autoantibodies against complement FH and for genetic mutations involved in complement-driven diseases. Patients with anti-FH autoantibodies were retested at least 1 month apart to confirm the positive findings. Ninety-nine patients were women, and 87 were men, with a median age of 32 years (range, 1–84 years). The diagnosis was on the basis of clinical and laboratory data. Twenty of the 186 patients had had a previous bone marrow transplant. As control groups, we used 40 healthy subjects (20 were men), with a median age of 29 years (range, 4–62 years).

Blood sampling was performed from an antecubital vein into plain tubes and processed within 2 hours by centrifugation at 2000×g for 15 minutes at room temperature. The aliquots of serum were immediately frozen and stored at −80°C before testing.

All adult patients or parents of pediatric patients signed a written consent for genetic tests, and all subjects agreed on the use of their blood samples in an anonymous form for research purposes. The local review boards approved the study, which was conducted following the ethical principles of 2013 revision of the Declaration of Helsinki and the code of Good Clinical Practice.

Anti-FH Antibody Assay

Anti-FH antibodies were assayed by an ELISA that used purified FH for capture and anti-human IgM, IgG, and IgA for detection.¹¹ Purified FH (10 μg/ml in PBS, pH 7.4; Calbiochem, EMD Chemicals, San Diego, CA) was coated overnight onto microtitration plates, and, after washing, the wells were coated with BSA to avoid nonspecific binding. After additional washes, a 1:20 dilution of the serum samples was added and incubated for 45 minutes at room temperature. After washing, the FH-bound Igs were identified by means of class-specific mouse monoclonal anti-IgM, -IgG, or -IgA (Sigma Aldrich, St. Louis, MO), which was detected by peroxidase-conjugated anti-mouse antibodies (Sigma Aldrich) and revealed with orthophenylenediamine. Absorbance was read at 490 nm. The results were expressed as units per milliliter and referred to an internal standard (serum collected from a patient with a high anti-FH antibody titer) arbitrarily fixed at 100 U/ml. In order to avoid the confounding effect of natural antibodies, we decided to use the maximum level observed in normal subjects as cutoff between normal and abnormal levels.

Purification of Anti-FH IgMs

Anti-FH IgMs were isolated from patients’ serum by a two-step affinity chromatography. At first, IgG and albumin were depleted from patients’ serum using albumin and IgG depletion SpinTrap columns (GE Healthcare, Buckinghamshire, United Kingdom) according to the manufacturer’s protocol. Depleted serum was then used to purify IgM by affinity chromatography on FH-conjugated Sepharose. FH purified from normal human serum as previously described¹² was coupled with cyanogen bromide–activated Sepharose 4B resin (Sigma Aldrich). To avoid crosscontamination, a single affinity column was prepared for each patient, and Micro Bio-Spin Columns (Bio-Rad, Hercules, CA) were filled with 200 µl of FH Sepharose resin. IgG- and albumin-depleted serum in binding buffer (20 mM sodium phosphate and 0.15 M NaCl, pH 7.4) was brought to a final concentration of 0.5 M NaCl and loaded onto the column. After a 30-minute incubation on a rotary shaker, columns were washed with five column volumes of PBS, and IgMs were then eluted in 0.1 M glycine, pH 3.0, and immediately neutralized with Tris-HCl 1.5 M, pH 8.8. Purity of eluted IgM was verified by SDS-PAGE under reducing and nonreducing conditions followed by silver staining.

Characterization of the Interaction between Anti-FH IgM and FH Domains

To overcome the lack of FH fragments for mapping FH epitopes that interact with IgM antibodies, an alternative approach was used. Autoantibody binding to full-length FH was assessed in a competition assay with known anti-FH mAbs that recognize specific short consensus repeats (SCRs) epitopes.¹³ Microplates were coated with 10 μg/ml FH overnight at 4°C. Plates were blocked with BSA for 1 hour at room temperature. Plates were incubated with 20 µg/ml of one of the known mAbs: OX23 (mapping at SCR1–4, Ab17928; Abcam), OX24 (mapping at SCR5, MA1–81868; Thermo Scientific), L20 (mapping at SCR19, GAU-020–03–02; Thermo Scientific), and C18 (mapping at SCR20, GAU-018–03–02; Thermo Scientific) for 15 minutes at room temperature. Finally, after washing, serum samples were tested at 1:20 dilution by incubation for 1 hour at room temperature. Signal was detected using a peroxidase-conjugated anti-human IgM antibody. The binding of the patient antibodies was considered specific for a particular epitope of the FH molecule when it was inhibited by the presence of the monoclonal directed against that specific epitope.

Interactions between Anti-FH Antibodies and FH in Fluid Phase

After incubating the patients’ serum for 1 hour at 37°C with increasing FH concentrations (0.015, 0.031, 0.063, 0.125, 0.250, and 0.5 mg/ml), we evaluated the free anti-FH antibodies still capable of binding to the microplate-immobilized FH.

Effect of Anti-FH IgM Autoantibodies on FH Functional Activity

The capacity of IgM autoantibodies to inhibit FH activity was investigated by means of the C3b binding assay.¹⁴ Microtiter plates were coated with 5 μg/ml C3b (Calbiochem) in coating buffer (0.05 M carbonate-bicarbonate, pH 9.6) overnight at 4°C. Plates were blocked with 3% BSA for 1 hour at room temperature. We preincubated 100 ng of FH (Calbiochem) with different dilutions of purified IgM anti-FH (starting from 100 U/ml), purified IgM anti–C1 inhibitor (starting from 100 U/ml) as negative control, and C18 mAb (starting from 1000 ng/ml) as positive control in Tris-buffered saline (20 mM Tris and 150 mM NaCl) for 10 minutes at 20°C. The samples were subsequently added to the wells and incubated for 1 hour at 37°C. A standard curve with 200, 100, 50, and 25 ng of FH was included. Binding was detected with mouse anti-FH mAb 5H5 (in house; 2.5 μg/ml)^12,15 and rabbit anti-mouse IgG-HRP (DAKO) followed by TMB development. After stopping with 2 M H₂SO₄, the absorbance was measured at 450 nm on a Spectra Max 190 photometer (Molecular Devices, Eugene, OR). Percentage of bound FH was calculated from the standard curve.

Genetic Studies

Genomic DNA extraction was performed on the QIAsymphonySP automated platform (Qiagen GmbH, Hilden, Germany). Detection of nucleotide variations was assessed by next generation sequencing (NGS) on the MiSeq platform (Illumina) by using the “targeted sequencing” technique (HaloPlex Kit; Agilent Technologies) on a multiple gene custom panel comprising CFH (NM_000186.3), MCP/CD46 (NM_002389.4), CFI (NM_000204.4), C3 (NM_000064.3), CFB (NM_001710.5), THBD (NM_000361.2), DGKE (NM_003647.2), CFHR1 (NM_002113.2), CFHR3 (NM_021023.5), and CFHR5 (NM_030787.3) at 100× coverage. Bioinformatics analysis of NGS data with filtering to identify putative causative variants was performed with the SureCall application. All variants identified by NGS analysis were then confirmed by the standard Sanger sequencing method. The potential effect of amino acid changes was assessed by in silico analysis (SIFT, PolyPhen-2, Mutation Taster, and VarSome; https://varsome.com) to predict the functional significance of unpublished and/or uncommon variants that were also evaluated as frequency of the variant compared with the general population from the Exome Aggregation Consortium database (https://gnomad.broadinstitute.org/). We also took into account the genetic variant classification from the database of complement gene variants (https://www.complement-db.org).¹⁶ Multiplex ligation-dependent probe amplification (MLPA kit P236; MRC-Holland, Amsterdam, The Netherlands) was used to identify CFHR3/CFHR1 copy number and macrorearrangements, such as CFH/CFH-Related hybrid genes. Raw data were analyzed by Coffalyser.net (https://www.mlpa.com), and relative dosage ratio was calculated.

FH Antigen Measurement

Levels of FH antigen were measured in serum from the seven patients with IgM anti-FH autoantibodies, 12 patients with DEAP-HUS, and 34 patients without anti-FH autoantibodies by a radial immunodiffusion method (Human factor H “NL” BINDARID; The Binding Site, Birmingham, United Kingdom). Blood samples were collected and handled according to the manufacturer’s instructions. Intra- and interassay CVs were <12%.

Statistical Analyses

Because of non-normal distribution, results were reported as medians and ranges (minimum to maximum), and nonparametric methods were used to assess statistical significance of differences between groups. Categorical variables were reported as counts and percentages. Differences in proportions were assessed by using the chi-squared test. The significance level was set at P=0.05. The associations between parameters were evaluated by logistic regression. Odds ratios and 95% confidence intervals were reported. The Spearman correlation coefficient was calculated to assess relationships between the variables. The data were analyzed using the SPSS PC statistical package, version 25 (IBM SPSS Inc., Chicago, IL).

Results

Anti-FH Autoantibodies Serum Levels

Of 186 patients with TMA, IgG anti-FH autoantibodies were elevated in 12 patients with aHUS who also presented a homozygous deletion in the genes of CFHR proteins and thus, were considered to have DEAP-HUS. High levels of anti-FH IgG autoantibodies were also found in four patients with TA-TMA. IgM anti-FH autoantibodies were elevated in four of the 154 patients with primary aHUS, in none of the 12 patients with DEAP-HUS, and in three of the 20 patients with TA-TMA (Figure 1). A significantly higher frequency of IgM anti-FH autoantibodies was found in patients with TA-TMA (15.0%) compared with the remaining patients (2.4%; P=0.005). Patients with IgM anti-FH autoantibodies had a median age of 38.5 years (range, 18–61), whereas patients with DEAP-HUS had a median age of 13.5 years (range, 1–21 years; P=0.003). Patients with abnormally high levels of anti-FH IgM autoantibodies had a titer of 5–27 times the upper limit of normal, whereas in patients with high levels of anti-FH IgG, the titer was 4–140 times the upper limit of normal. None of the patients in this patient list developed autoantibodies of both classes. Anti-FH autoantibodies of the IgA class were not found in any patient. Demographic and clinical characteristics of patients with anti-FH antibodies of the IgM class are reported in Table 1. The positive patients are identified by their sequential number in Figure 1. Table 2 reports serum levels of anti-FH autoantibodies of IgM, IgG, and IgA classes in patients with primary aHUS, DEAP-HUS, or TA-TMA and in normal controls.

Figure 1.

Anti-FH autoantibodies of IgM class are present in serum from 7 patients with TMA. Serum levels of anti-FH autoantibodies of IgM class were measured in patients with TMA presented as aHUS and in healthy subjects (normal controls). Patients were divided into primary aHUS, DEAP-HUS, and TA-TMA. The dashed line represents the upper limit of normal controls. Patients with elevated values of anti-FH IgM are identified with their sequential number reported in Table 1.

Table 1.

Demographic, clinical, and genetic data of patients with anti-FH of IgM class and TMA

Patients	Sex	Age, yr	Primary Disease	TMA Type	CFHR3-CFHR1 Deletion	Complement Gene Variant	Variant Clinical Significance	IgM Anti-FH, U/ml	C3, %	FH, %
1	W	48	Multiple myeloma	TA-TMA	No	No	n.a.	174	60	73
2	M	53	LLAC	aHUS	No	CFH c.3493+1G>C	Pathogenic	137	84	85
3	W	36	—	aHUS	Heterozygous	MCP p.(Thr383Ile)	Likely pathogenic	87	93	90
4	W	41	—	aHUS	No	CFI p.(Lys441Arg)	Benign	75	101	56
5	M	18	Ulcerative colitis	aHUS	No	No	n.a.	60	72	60
6	W	26	Acute lymphoblastic leukemia	TA-TMA	No	No	n.a.	54	111	102
7	W	61	Burkitt lymphoma	TA-TMA	No	No	n.a.	53	109	64
Normal range								0.3–6.7	76–117	70–120

W, woman; n.a., not applicable; M, man; LLAC, lupus-like anticoagulant; CFH, complement factor H; —, no primary disease; MCP, membrane cofactor protein; CFI, complement factor I.

Table 2.

Serum levels of anti-FH autoantibodies of IgM, IgG, and IgA classes in 186 patients with TMA presented as aHUS and 40 healthy subjects

Condition	Anti-FH Autoantibodies, U/ml
	IgM	IgG	IgA
Primary aHUS, n=154	1.9 (0.0–133.0) a	0.4 (0.1–4.8)	0.4 (0.0–4.2)
DEAP-HUS, n=12	1.8 (0.0–3.7)	40.0 (11.3–725.0)	0.5 (0.0–4.5)
TA-TMA, n=20	2.3 (0.0–174.0) b	0.3 (0.0–40.0) c	0.7 (0.1–3.9)
Normal controls, n=40	2.0 (0.3–6.7)	0.5 (0.1–5.1)	0.4 (0.0–4.8)

^aFour of 154 patients with primary aHUS had abnormally high levels of anti-FH IgM autoantibodies.

^bThree of 20 patients with TA-TMA had abnormally high levels of anti-FH IgM autoantibodies.

^cFour of 20 patients with TA-TMA had abnormally high levels of anti-FH IgG autoantibodies.

Five of seven patients with IgM autoantibodies were followed with serial determinations of anti-FH autoantibodies for a period ranging between 4 months and 4 years. During the follow-up, no patient developed anti-FH autoantibodies of the IgG class, and the levels of IgM anti-FH have fluctuated over time but have remained above the normal range in all patients both during periods of active disease and during remission.

Dialysis was necessary in four of seven patients. Four patients received plasma therapy alone, one patient received plasma therapy plus corticosteroids and azathioprine, and two patients were treated with hydration and corticosteroids. Remission was initially obtained in six of seven patients and maintained in five patients (two of these patients unfortunately did not recover renal function and remained in dialysis). One patient had one relapse and was successfully treated with eculizumab. The patient who did not respond initially underwent kidney transplant, but aHUS relapsed 2 months later, and a second transplant was performed; from that time on, the patient was successfully managed with eculizumab maintenance treatment.

Purification of IgM Anti-FH Autoantibodies

The two-step affinity chromatography on FH-conjugated Sepharose performed on serum samples of patients with IgM anti-FH antibodies allowed the recovery of about 70% of these antibodies. On SDS-PAGE under reducing conditions, the IgM heavy and light chains (75 and 25 kD, respectively) were visible, together with an additional band around 250 kD that may represent an incompletely dissociated IgM molecule. Under nonreducing conditions, a major band barely migrating in the resolving gel, representing total IgM (approximately 500 kD), was present (Figure 2).

Figure 2.

SDS-PAGE confirms the purification of IgM anti-FH from serum. SDS-PAGE was performer under reducing conditions (lanes 1–3) and nonreducing conditions (lanes 4–6) followed by silver staining of purified IgM anti-FH obtained from patient numbers 1–3. Under reducing conditions, two bands at 75 and 25 kD are visible (IgM heavy and light chains, respectively); moreover, there is an additional band at 250 kD due to an incompletely dissociated IgM molecule. Under nonreducing conditions, there is a major band barely migrating in the resolving gel, representing total IgM (approximately 500 kD). Pt, patient.

Interaction between Anti-FH IgM and FH Domains

In all seven patients positive for anti-FH IgM, the binding of these autoantibodies to FH molecule, expressed as OD (mean of two experiments), was not inhibited by OX23 (mapping at SCR1–4), OX24 (mapping at SCR5), and C18 (mapping at SCR20), whereas L20 (mapping at SCR19) did prevent the binding (Figure 3, left panels). The right panels of Figure 3 represent the binding of the specific mAbs and the binding of patients’ IgM anti-FH autoantibodies to the different SCR domains of the FH molecule. To confirm the data on serum, we tested the IgM anti-FH purified from patient number 1, obtaining high OD when FH was preincubated with OX23, OX24, and C18 (780±90 OD) and low OD after incubation with L20 (100±27 OD).

Figure 3.

IgM anti-FH autoantibodies bind to SCR 19 as shown by the competition with specific mAbs anti-FH. The binding of purified IgM anti-FH to FH (expressed as mean OD of two experiments) was inhibited only by the mAb L20 that interacts with the SCR domain 19, whereas mAb OX23 (interacting with SCR1–4), OX24 (interacting with SCR5), and C18 (interacting with SCR20) did not modify the binding of IgM anti-FH to FH. The right panels of the figure represent the binding of the specific mAbs and the binding of patients’ IgM anti-FH autoantibodies to the different SCR domains of the FH molecule.

Interactions between Anti-FH Antibodies and FH in Fluid Phase

FH added (up to 0.5 mg/ml) to the serum of patients with the highest anti-FH IgM titer (patients 1–3) clearly inhibited the binding of autoantibodies to microplate-immobilized FH in a concentration-dependent manner as shown in Figure 4. The inhibition was expressed as percentage of the OD obtained by adding only buffer, which was considered 100%. The values represent the mean of two experiments.

Figure 4.

Inhibition of the binding of anti-FH to microplate-immobilized FH by soluble FH. When added to the serum of patients (numbers 1–3), FH (at concentrations of 0.015, 0.031, 0.063, 0.125, 0.250, and 0.5 mg/ml) inhibited the binding of anti-FH antibodies to the FH immobilized on microplates in a dose-dependent manner. The results are expressed as the percentage binding (mean of two experiments) recorded in the absence of added soluble proteins (buffer).

Effect of Anti-FH IgM Autoantibodies on FH Functional Activity

The C3b binding test was performed in the three patients whose serum was sufficient to obtain a high yield of purified antibodies (i.e., patient numbers 1, 4, and 5). We obtained samples with anti-FH IgM concentration of 200 U/ml in patient number 1, 100 U/ml in patient number 4, and 150 U/ml in patient number 5, irrespective of their serum levels, which were 174 U/ml, 77 IU/ml, and 60 U/ml, respectively. All purified Ig samples were brought to a concentration of 100 U/ml and then tested. Figure 5 shows that FH binding to C3b is inhibited by increasing amounts of IgM anti-FH antibodies from the maximum binding of 100% to a minimum of 40%–60% in a dose-dependent manner, whereas the negative control anti–C1 inhibitor IgM exerted no effect (Figure 5A). Figure 5B shows the inhibitory effect on C3b binding by mAb C18 from the maximum binding of 100% to a minimum of 50%.

Figure 5.

FH binding is inibited by IgM anti-FH autoantibodies. When added to the serum of patients (numbers 1–3), FH (at concentrations of 0.015, 0.031, 0.063, 0.125, 0.250, and 0.5 mg/ml) inhibited the binding of anti-FH antibodies to the FH immobilized on microplates in a dose-dependent manner. The results are expressed as the percentage binding (mean of two experiments) recorded in the absence of added soluble proteins (buffer).

Genetic Analyses

The analysis of genes involved in complement-regulating components showed three different variants in patients with IgM anti-FH autoantibodies (Table 1). CFHR3-CFHR1 homozygous deletion was detected in 12 patients with anti-FH IgG autoantibodies and aHUS (thus considered DEAP-HUS) and in none of the four patients with TA-TMA and anti-FH IgG autoantibodies nor in patients with anti-FH autoantibodies of the IgM class (Table 1). The association between anti-FH autoantibodies and CFHR3-CFHR1 homozygous deletion was confirmed in patients with IgG autoantibodies (odds ratio, 27; 95% confidence interval, 6.35 to 114.85), whereas it was not observed in patients with IgM autoantibodies (Table 1).

FH Antigen Measurement

Serum levels of FH antigen were significantly lower in patients with anti-FH autoantibodies of the IgM class (median, 73%; range, 56%–102%) and in patients with anti-FH autoantibodies of the IgG class (DEAP-HUS; 71%; range, 32%–115%) than in patients with TMA without anti-FH autoantibodies (108%; range, 58%–180%), with P values of 0.005 and 0.003, respectively (Figure 6).

Figure 6.

Serum levels of FH antigen are reduced in patients with IgM or IgG anti-FH autoantibodies. Serum levels of FH antigen were measured in 34 patients with TMA without anti-FH autoantibodies, in seven patients with TMA and anti-FH autoantibodies of the IgM class, and in 12 patients with anti-FH autoantibodies of the IgG class (DEAP-HUS).

Discussion

aHUS associated with anti-FH autoantibodies was first reported by Dragon-Durey et al.,³ and to date, these autoantibodies have been described in up to 10% of cases of aHUS in cohort studies, patient series, and patient reports.^17,18 Anti-FH autoantibodies have also been described in some patients with TA-TMA by Jodele et al. ⁴ In all of the above-mentioned studies, autoantibodies to FH were of the IgG class.^4,17,18 To the best of our knowledge, our study on aHUS is the first one describing patients with anti-FH of IgM class. Interestingly, the median age at disease onset of our patients with aHUS and IgG anti-FH was 13.5 years, which is in agreement with previous studies.^8,9 In contrast, our patients with aHUS and IgM anti-FH were all adults, with a median age of 38.5 years. Three of them developed aHUS after bone marrow transplant; thus, the frequency of anti-FH IgM is six-fold higher in TA-TMA than in primary aHUS. The association between anti-FH IgG autoantibodies and CFHR3-CFHR1 homozygous deletion was confirmed in our patients list in agreement with previous studies,^19,20 whereas among our seven patients with aHUS and IgM anti-FH autoantibodies, no one had this gene abnormality. In these seven patients who do not have deletion, we expect FHR1 and FHR3 to be present in their serum; thus, the autoantibodies might act through a different mechanism compared with in patients with DEAP-HUS. In addition, because the presence of anti-FH IgM autoantibodies is significantly more common in patients with HSCT without a specific genetic background, it can be hypothesized that their development is related to the transplant itself in the setting of an autoimmune process. The competition assay with mapping mAbs (Figure 3) clearly shows that the IgM anti-FH autoantibodies of all patients interact with the SCR domain 19 that is located in the C-terminal part of FH and is important for the binding of FH to endothelial cells.^14,21 Moreover, because the anti-FH IgM autoantibodies inhibit the binding of FH to C3b in vitro, we can expect that this function is inhibited by the autoantibodies also in vivo. Such an inhibition, which mimics the effect of C-terminal FH mutations, has been previously demonstrated in patients with aHUS and anti-FH autoantibodies of the IgG class, and these autoantibodies were directed to SCR19–20 domains.¹⁴ Indeed, SCR19 is important for the binding of FH not only to endothelial cells but also, to C3b, as demonstrated by Kajander et al. ²²

Patients with IgM anti-FH autoantibodies showed a slight, but significant, reduction of FH antigen levels compared with patients without autoantibodies. Serum levels of FH have already been described as low in aHUS in a previous study¹⁷ and in 22% of patients in another study.¹⁸ The slight reduction may be due to a higher clearance of FH when it is complexed to the antibody; however, our antigenic method does not provide information on the activity of FH that potentially can be reduced due to the possible presence of neutralizing autoantibodies. In any case, because the IgM anti-FH autoantibodies of our patients bind the C-terminal domain of FH, which is a recognition region containing the binding sites for C3b, glycosaminoglycans, and endothelial cells,^14,21,22 a pathogenic effect of these antibodies is likely. Our in vitro data on inhibition of the FH binding to C3b by anti-FH autoantibodies further support this view. In our patients, the levels of IgM anti-FH have fluctuated over time but have remained above the normal range in all patients. Also, because anti-FH autoantibodies of the IgG class may fluctuate over time and asymptomatic patients with high anti-FH titers are at risk of relapse,²³ asymptomatic patients with anti-FH autoantibodies of the IgM class may also require cautious monitoring. Normal levels of C3 in most of our patients with anti-FH IgM are not surprising. Indeed, according to our experience and general experience, it is well known that C3 levels are expected to be normal in 50%–70% of patients with aHUS^11,24 and in 25% of patients with aHUS and anti-FH autoantibodies of the IgG class.²³

In conclusion, our data indicate that in patients with a clinical presentation of aHUS, the presence of autoantibodies of IgM class directed against the active site of FH is possible particularly in patients with TA-TMA. These autoantibodies may support an autoimmune mechanism in patients with aHUS previously considered of unknown origin.

Disclosures

G. Ardissino reports consultancy agreements with Alexion, Alnylam, and Chemo Research; honoraria from Alexion and Alnylam; and scientific advisor or membership with Alexion Inc. D. Cresseri reports speakers bureau from Alexion Pharma. F. Giglio reports honoraria from participation on advisory boards of Alexion, Amgen, and Pfizer. P. Messa reports consultancy agreements with Sandoz; honoraria from Sandoz and Vifor; scientific advisor or membership with the journals Blood Purification, Journal of Nephrology, and Nutrients; and speakers bureau for Vifor. F. Peyvandi reports scientific advisor or membership with Roche, Sanofi, and Sobi and speakers bureau for Bioverativ, Grifols, Roche, Sanofi, Sobi, Spark, and Takeda. All remaining authors have nothing to disclose.

Funding

None.