Genetic disorders in complement (regulating) genes in patients with atypical haemolytic uraemic syndrome (aHUS)

Abstract

Background. Atypical HUS (aHUS) is thought to be caused by predisposing mutations in genes encoding complement (regulating) proteins, such as Factor H (CFH), Factor I (IF), membrane co-factor protein (MCP) and Factor B (FB), or by auto-antibodies against CFH (αFH) in combination with a homozygous polymorphic deletion of the genes encoding Complement Factor H-related 1 and 3 (ΔCFHR1/3). The clinical impact of this knowledge is high, as it might be a prognostic factor for the outcome of renal transplantations and kidney donations.

Methods. Mutational screening, by means of PCR and DNA sequencing, is performed in the above-mentioned genes in a group of 72 aHUS patients. Also, the presence of αFH and ΔCFHR1/3 was tested in patients and controls.

Results. In 23 patients, a genetic aberration in at least one gene or the presence of αFH was found. A heterozygous mutation was observed in CFH in nine patients, in IF in seven patients and in MCP in three patients. No mutations were observed in FB. Seven patients presented αFH, of whom five also carried ΔCFHR1/3. Three patients carried a combined mutation (two patients: IF and MCP; one patient: IF, αFH and ΔCFHR1/3). A significant difference between patients and controls was detected for the presence of all three associated polymorphisms in CFH.

Conclusions. Genetic abnormalities or the presence of αFH were detected in 31.9% of the aHUS patients. Furthermore, bigenic mutations were present, indicating that routine DNA mutation analysis of all complement factors associated with aHUS is important.

atypical HUS, clinical impact, complement regulation, DNA analysis, genetic defects

Introduction

Haemolytic uraemic syndrome (HUS) is a rare and severe disease, which is characterized by thrombotic microangiopathy, haemolytic anaemia, thrombocytopaenia and acute renal failure [1]. In most cases, HUS is seen in childhood, is preceded by watery or bloody diarrhoea and is caused by Shiga-like toxin (Stx)-producing Escherichia coli (STEC) [2,3]. Non-Stx-associated HUS is seen in 5% to 10% of all HUS cases; these patients have a much poorer prognosis. Up to 50% of these so-called atypical cases progress to end-stage renal disease (ESRD), and 25% may result in death during the acute phase of the disease [4–7]. In recent years, a clear link has been established between atypical HUS (aHUS) and genetic abnormalities in regulator genes of the alternative pathway of the complement system. Mutations have been described in genes encoding complement factor H (CFH), complement factor I (IF) and membrane co-factor protein (MCP/CD46), three important regulatory proteins of the alternative pathway [7,8]. Up to now, mutations in aHUS patients are predominantly found in the CFH gene and in lesser amounts in IF and MCP [8–11]. Very recently, the presence of auto-antibodies against CFH (αFH) in combination with a homozygous polymorphic deletion of complement factor H-related genes CFHR1 and CFHR3 (ΔCFHR1/3) have been associated with aHUS [12], as well as mutations in genes encoding complement factor B (FB)[13], complement C3 [14] and thrombomodulin [15].

Knowing the genetic variations present in aHUS patients is important, because it can be of prognostic value for the outcome of renal transplantation and kidney donations. In general, ∼50% of the patients that underwent renal transplantation had a recurrence of the disease in the graft, and graft failure occurred in >90% of them [8]. However, several groups have shown that patients with a mutation in the CFH or IF gene have a worse outcome after kidney transplantation (recurrence: 80–100%) than patients with a MCP mutation (recurrence 0–20%) [8,16,17].

In the present study, mutations in genes encoding (regulating) proteins of the alternative pathway of the complement system (CFH, IF, MCP and FB) and the presence of αFH in combination with ΔCFHR1/3 were studied in a cohort of 72 Dutch and Belgian patients diagnosed with aHUS. Of 19 patients, EDTA serum was available to identify the presence of αFH.

Subjects and methods

Study population

The research population consisted of 72 aHUS patients (33 children and 39 adults; age 5 months to 55 years at onset), referred to the Paediatric Nephrology Centre of the Radboud University Nijmegen Medical Centre (RUNMC). All patients were diagnosed with atypical, non-STEC HUS. The patients were of Dutch or Belgian origin, mostly with a Caucasian background. In ten patients from four families, the familial form of aHUS was identified; the other 62 patients were diagnosed with sporadic aHUS. A permission to study the DNA material was given by all patients or their parents.

Genetic analysis of genes encoding CFH, MCP, IF and FB

Genomic DNA, isolated from peripheral blood leukocytes as described by Miller et al. [18], was amplified for CFH [National Center for Biotechnology Information (NCBI) RefSeq NM_000186], IF (NM_000204), MCP (NM_172361) and FB (NM_001710) by means of PCR. The primer data are available on request. Fragments included both DNA sequences of the individual exons, and the splice donor and acceptor site. The amplimers were subjected to double-stranded DNA sequence analysis on an ABI 3130 xl GeneticAnalyzer (Applied Biosystems). Sequence analysis was performed using Sequencher 4.8 software. The genomic DNA from 82 healthy ethnically matched control individuals was used to confirm novel mutations. A mutation is defined as any change in the sequence of a DNA molecule produced in mitosis or meiosis; an unknown change that is present in more than one of the 82 controls is considered as a polymorphism. Potential pathogenicity of genetic alterations was checked in literature, the HUS database (http://www.fh-hus.com), splice site prediction programmes (http://www.fruitfly.org/seq_tools/splice.html and http://www.cbs.dtu.dk/services/NetGene2), evolutionary conservation and Sorting Tolerant From Intolerant (SIFT) (http://sift.jcvi.org).

Auto-antibodies against factor H

Nineteen patients and controls were tested for the presence of αFH by means of ELISA, as described before [19]. A positive control sample was obtained via Dr. Dragon-Durey (Paris, France). Test results were considered positive if they were above twice the standard deviation calculated from controls, and samples were tested at least three times.

Deletion of CFHR1 and CFHR3 genes

To identify ΔCFHR1/3, genomic DNA was amplified by PCR using specific primers for two regions located in a 100-kb region downstream of the gene encoding complement factor H (fragment P1 and P2 in Figure 1A; R5 and R8 in Zipfel et al. [20]). The amplification of fragment P1 fails in case of ΔCFHR1/3 (Figure 1B).

Fig. 1

(A) Location of amplimers used for identification of deletion of CFHR1 and CFHR3. The last three SCRs of CFH, the five SCRs of CFHR3, the five SCRs of CFHR1, and the first SCR of CFHR4 are indicated by vertical bars. The horizontal bar indicates the location of ΔCFHR1/3. The position of the two amplimers (P1 and P2) is shown; amplification of fragment P1 fails in case of a homozygous polymorphic deletion of CFHR1 and CFHR3. (B) PCR results of representative patients with deletion of CFHR1 and CFHR3. The top electrophoresis shows the PCR product of fragment P1, the bottom part of the figure displays the PCR product of fragment P2. M indicates marker; + indicates positive control; − negative control; numbers indicate patient subject codes.

Statistical analyses

Differences between allele frequencies of the strongly associated polymorphisms in the CFH gene among patients, controls and the European population [mean from NCBI single-nucleotide polymorphism (SNP) database], and the presence of αFH and ΔCFHR1/3among patients and controls were analysed by calculating the 95% confidence intervals (95% CI). For analysis between patients and controls, a 95% CI that did not include zero was considered statistically significant.

Results

Complement factor H

The total open reading frame of the CFH gene was analysed in 72 aHUS patients. Seven potential pathogenic heterozygous mutations were found in nine patients, as summarized in Table 1 and shown in Figure 2A. Seven of these patients display sporadic aHUS, while two of them are diagnosed with the familial form of the disease. Four aberrations found have not been described before (g.−315c > t, g.IVS19 + 1g > a, p.Arg1203Trp and p.Arg1206Cys); the remaining three are known disease-causing mutations. Five independent mutations cluster between short consensus repeat (SCR) 16 and 20 (Figure 2A). Five of the detected mutations are missense mutations, one mutation affects the donor site of intron 19, causing a sequence that is not recognized as a splice site, and one mutation is located near the binding site of nuclear factor kappa beta (NFκβ), a region probably involved in the transcription of CFH during inflammation and infection [21]. None of the mutations described were found in more than one of 82 healthy controls.

Table 1

Characteristics of mutations found in CFH, MCP and IF in 17 out of 72 aHUS patients

Exon/intron, subject code	Mutation	Effect	SCR	Known/unknowne	Subgroup
Complement factor H
Promoter
40	g.−315c > t	Transcription	NA	Unknown	Sporadic
Exon 9
21	c.1198C > A	p.Gln400Lys	7	Known27	Sporadic
Exon 19
47	c.2850G > T	p.Gln950His	16	Known6,21	Sporadic
61	c.2850G > T	p.Gln950His	16	Known6,21	Sporadic
Intron 19
29a	g.IVS19 + 1g > a	Splice site not efficiently recognizedd	NA	Unknown	Sporadic
Exon 23
66	c.3607C > T	p.Arg1203Trp	20	Unknown	Sporadic
68	c.3616C > T	p.Arg1206Cys	20	Unknown	Familial
69	c.3616C > T	p.Arg1206Cys	20	Unknown	Familial
62	c.3628C > T	p.Arg1210Cys	20	Known28,33,34	Sporadic
Membrane co-factor protein
Intron 2
50	g.IVS2 + 2t > g	Splice site not efficiently recognizedd	NA	Known25	Sporadic
Exon 6
23b	c.811-816delGACAGT	p.delAsp271-Ser272	4	Known24	Sporadic
46b	c.811-816delGACAGT	p.delAsp271-Ser272	4	Known24	Sporadic
Complement factor I
Exon 3
23b	c.454G > A	p.Val152Met		Unknown	Sporadic
55	c.454G > A	p.Val152Met		Unknown	Sporadic
Exon 9
45	c.1019T > C	p.Ile340Thr		Known40	Sporadic
Exon 10
2c	c.1071T > G	p.Ile357Met		Unknown	Sporadic
Exon 11
64	c.1420C > T	p.Arg474Stop		Known9	Sporadic
Intron 12
1	g.IVS12 + 5g > t	Splice score decrease from 0.93 to 0.861		Known8	Sporadic
46b	g.IVS12 + 5g > t	Splice score decrease from 0.93 to 0.861		Known8	Sporadic

Patients are numbered according to an individual number. NA indicates not applicable.

a

Patient carrying a FH mutation and a homozygous ΔCFHR1/3 (αFH not tested).

b

Patients carrying both IF and MCP mutations.

c

Patient carrying an IF mutation and αFH in combination with ΔCFHR1/3.

d

According to Fruitfly and NetGene2.

e

According to www.fh-hus.com.

Fig. 2

(A) Schematic overview of the locations of described CFH mutations in aHUS patients of this study. The white blocks indicate SCRs that are not involved in binding processes; the darker grey SCRs bind to C3b; the lighter grey SCRs bind to sialic acids and heparine sulphate; the black SCRs bind to C3b and sialic acids and heparine sulphate. (B) Schematic overview of the locations of described MCP mutations in aHUS patients of this study. SCR indicates short consensus repeat; STP, serine–threonine–proline-rich domain; TM, transmembrane domain; CT, cytoplasmatic tail. (C) Schematic overview of the locations of described IF mutations in aHUS patients of this study. FIMAC indicates factor I membrane attack complex domain; SCRC, scavenger receptor cysteine rich domain; LDLR, low-density lipoprotein receptor domain; SP, serine protease domain.

The presence of all three strongly associated polymorphisms in CFH [21–23] (g.−331c > t, c.2016A > G and c.2808G > T) was found with a significant difference in 52.8% of the patients, in contrast to 31% of the controls (95% CI: 0.006–0.370; Table 2A). Allele frequencies of these polymorphisms (Table 2B and Table 2C) in this cohort were not significantly increased compared to those in the European population. Compared to the Dutch controls we investigated, the allele frequencies of c.2016A > G and c.2808G > T were significantly increased in patients (Table 2C).

Table 2A

Incidence and 95% CI of the three strongly associated polymorphisms in CFH in 72 aHUS patients and 82 controls

	Patients	Controls	95% CI (patients–controls)
Three polymorphisms	52.8%	31.0%	0.006–0.370a
Two polymorphisms	1.4%	8.3%	−0.133–0.004
One polymorphism	13.9%	22.6%	−0.206–0.033
No polymorphisms	31.9%	38.1%	−0.212–0.088

The percentages in the patient population are compared to the percentages in the control group by calculating 95% CI.

a

Statistically significant.

Table 2B

Allele frequencies of the three strongly associated polymorphisms in CFH in 72 patients with aHUS, 82 controls and the European population

Genetic variant	Alleles	Patients	Controls	European populationa
Promoter
g.−331c > t	C	0.583	0.690	0.857
	T	0.417	0.310	0.143
Exon 14
c.2016A > G	A	0.681	0.792	0.720
	G	0.319	0.208	0.280
Exon 19
c.2808G > T	G	0.681	0.804	0.838
	T	0.319	0.196	0.162

a

Mean from SNP Database (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=snp).

Table 2C

95% CI of individual allele frequencies of the three strongly associated polymorphisms in CFH in 72 patients with aHUS

	Patients–controls	Patients–European population	Controls–European population
g.−331c > t	−0.002–0.213	−0.847–1.680	−0.778–1.390
c.2016A > G	0.014–0.213a	−0.787–1.426	−0.686–1.102
c.2808G > T	0.027–0.222a	−0.787–1.426	−0.672–1.065

The frequencies in the patient population are compared to the frequencies in the control group, as well as to those in the European population.

a

Statistically significant.

Membrane cofactor protein

As shown in Table 1 and Figure 2B, two heterozygous mutations were found in MCP in three patients with sporadic aHUS. None of the mutations were found in any of 82 healthy controls, and both mutations have been described before [24,25]. The deletion of six nucleotides in exon 6 causes a deletion of two amino acids (Asp271 and Ser272) and was observed in two unrelated patients, who also carried an IF mutation (c.454G > A and g.IVS12 + 5g > t, respectively). The other mutation affects the splice donor site of exon 2 and results in a splice site that is not recognized [20,21]. Both mutations are located in the amino-terminal region of MCP, confirming the previously described importance of this region for complement regulation [8,26].

Complement factor I

Five independent heterozygous mutational events in seven patients, all diagnosed with sporadic aHUS, were found in IF (Table 1). Two patients carried a mutation in both IF and MCP. Two of the observed mutations have not been described before (p.Val152Met and p.Ile357Met). Four mutations are located in the serine protease (SP) domain of factor I [27]: two are missense mutations, one mutation introduces a premature stop codon and one affects the donor site of intron 12 (Figure 2C). p.Val152Met is located in exon 3, causing an amino acid change in the scavenger receptor cysteine-rich (SCRC) domain [27]. The mutation at the donor site of intron 12 was detected in one of the 82 healthy controls; the other mutations were not found in controls.

Complement factor B

No potential pathogenic mutational events were found in the FB gene. Several polymorphisms were observed, but none of the allele frequencies differed from the European population (data not shown).

Auto-antibodies and ΔCFHR1/3

We found ΔCFHR1/3 in seven of the aHUS patients, compared to three of the 82 controls; a difference that is not statistically significant (95% CI: −0.019–0.140). EDTA plasma of 19 patients was investigated for the presence of αFH. In total, seven patients, all children, displayed αFH, of which five had ΔCFHR1/3 as well (Table 3). Two patients only presented αFH. In one patient (p.60) with ΔCFHR1/3, αFH could not be detected. Of Patient 29, in which ΔCFHR1/3 was observed, EDTA plasma was not available to test for auto-antibodies; this patient carried a CFH mutation as well. In one patient (p.2), an additional mutation in IF was found previously. In two controls (2/19; 10.5%) αFH were present. The difference between the patients and controls for the presence of αFH was considered significant (95% CI was 0.006–0.520).

Table 3

Clinical features of patients with genetic aberrations in CFH, IF, MCP and/or the presence of αFH

					Biochemical analysis (g/l)
Patient	Genetic aberration	Sex	Age at presentation	Trigger	C3	C4	CFH	IF	Treatment first episode	Outcome	Transplantation history
1	IF: g.IVS12 + 5g > t	F	4 years, 1 months	1	N/A	N/A	N/A	N/A	1	1	None
2	IF: p.Ile357Met; αFH; ΔCFHR1/3	M	6 years, 6 months	1	0.61	0.20	N/A	N/A	2	2	None
18	αFH; ΔCFHR1/3	F	4 years, 4 months	1	0.85	0.23	Normal	Normal	2	3	None
21	CFH: p.Gln400Lys	M	1 year, 4 months	1	1.20	0.20	N/A	N/A	3	2	None
23	IF: p.Val152Met; MCP: p.delAsp271-Ser272	F	30 years	2	0.73	0.18	N/A	N/A	2	4	1
29b	CFH: g.IVS19 + 1g > a; ΔCFHR1/3a	F
32	αFH	M	8 years, 4 months	3	0.87	0.41	N/A	N/A	4	4	None
39	αFH; ΔCFHR1/3	F	6 years, 6 months	4	N/A	N/A	N/A	N/A	2	3	None
40	CFH: g.-315c > t	F	5 months	1	N/A	N/A	N/A	N/A	2	5	None
42	αFH; ΔCFHR1/3	M	4 years, 6 months	1	N/A	N/A	N/A	N/A	2	3	None
43	αFH	F	5 months	1	0.72	0.15	N/A	N/A	2	2	None
45	IF: p.Ile340Thr	M	1 year, 3 months	1	Low	N/A	Normal	Normal	1	4	2
46	IF: g.IVS12 + 5g > t; MCP: p.delAsp271-Ser272	M	13 years, 1 months	1	1.00	0.19	N/A	N/A	2	5	None
47	CFH: p.Gln950His	M	12 years, 2 months	1	0.90	0.20	N/A	N/A	2	5	None
50	MCP: g.IVS2 + 2t > g	F	37 years	4	1.25	0.55	Normal	Normal	1	5	None
54	αFH; ΔCFHR1/3	M	6 years, 5 months	1	0.69	0.25	N/A	N/A	2	2	None
55	IF: p.Val152Met	F	38 years	5	N/A	N/A	N/A	N/A	2	3	2
61	CFH: p.Gln950His	F	38 years	2	0.53	0.17	N/A	N/A	5	4	None
62	CFH: p.Arg1210Cys	M	12 years, 5 months	1	N/A	N/A	N/A	N/A	1	2	None
64b	IF: p.Arg474Stop	F
66	CFH: p.Arg1203Trp	M	39 years	4	0.76	0.68	Low	Normal	5	4	2
68	CFH: p.Arg1206Cys	M	22 years	4	0.83	0.22	N/A	N/A	2	2	None
69	CFH: p.Arg1206Cys	F	21 years	6	1.75	0.4	N/A	N/A	3	4	1

Patients are numbered according to an individual number. The polymorphic homozygous ΔCFHR1/3 was only reported in this table when occurring in combination with αFH or other genetic abnormalities in aHUS patients. For p.29, the presence of αFH has not been tested (indicated with ^a). For p.29 and p.64, it was not possible to obtain clinical features, which is indicated with ^b.

Explanation of the clinical features: ‘Triggers’: 1 indicates flulike, gastroenteritis, other infections; 2 unknown; 3 medication; 4 no trigger; 5 transplantation; 6 pregnancy . ‘Biochemical analysis’: N/A indicates not available; normal values C3: 0.90–1.80 g/l; normal values C4: 0.15–0.45 g/l. ‘Treatment’: 1 indicates supportive treatment; 2 plasmapheresis, infusion or exchange; 3 plasma and drugs acting on both coagulation cascade and immune system; 4 no treatment; 5 plasma and drugs acting on the immune system. ‘Outcome’: 1 indicates death; 2 complete remission; 3 chronic renal insufficiency; 4 ESRF; 5 partial remission. ‘Transplantation history’: 1 good renal function at 1 year; 2 disease recurrence in graft; 3 acute rejection.

Clinical information

Clinical information was obtained of all patients with a genetic aberration, except for two patients, and is shown in Table 3. Fourteen patients were younger than 18 years at onset of disease. Mutations in CFH, IF or MCP were found in patients of all ages; αFH were only found in children. Plasma C3 levels were measured in 20 patients and were low in 10. One patient died during the HUS episode. Five patients needed a transplantation (two with a mutation in CFH, one in IF and MCP, and two in IF); in three of them (two with an IF mutation, and one with a CFH mutation), the disease recurred in the graft.

Discussion

In 31.9% of the patients (23/72), a genetic alteration in one of the investigated genes (CFH, IF, MCP, FB) or the presence of αFH was observed. A heterozygous mutational events was found in nine patients in CFH (12.5%), in seven patients in IF (9.7%) and in three patients in MCP (4.2%). Six mutations were unknown genetic alterations, the remaining were known disease-causing mutations. Seven of 19 patients were tested positive for the presence of αFH (36.8%); five of these seven patients presented ΔCFHR1/3 as well. Remarkably, three of the patients (4.2%) carried a combined mutation (IF and MCP: two patients; IF, αFH and ΔCFR1/3: one patient). No mutations were found in FB.

Seventy-one percent of the independent mutations (5/7) in CFH cluster between SCR 16–20, which confirms the previously described importance of the C-terminus of CFH to the pathogenesis of HUS [21,28–31]. A mutant CFH protein with loss of binding sites for C3b and polyanions shows a loss of the capability to degrade endothelial-bound C3b [32].

The mutations p.Arg1203Trp and p.Arg1206Cys probably cause a lower binding to surface-bound C3b, as they are located in the same binding site for C3b and polyanions as the p.Arg1210Cys mutation that has been described before [28,33,34]. These two unknown mutations are not located at a highly conserved codon (Figure 3A), but SIFT (http://sift.jcvi.org) predicts that the substitution of a positively charged, hydrophilic arginine to a neutral, hydrophobic cystine at codon 1206 will not be tolerated. The patient with the p.Arg1210Cys mutation does not possess the CFH_tgtaat haplotype associated with increased risk to aHUS in combination with this variation [35]. The splice site mutation at the first nucleotide after exon 19 (g.IVS19 + 1g > a) results in a sequence that is not recognized as a splice site by the splice site prediction programmes. The p.Gln950His variation was detected previously and was reported as a mutation [21]. In another study, it was declared as polymorphism [6]. We did not detect this aberration in any of our 82 controls. The amino acid located at codon 950 is highly conserved, and a change into a polar, positively charged histidine could result in functional loss of the protein. The p.Gln400Lys mutation is located in SCR 7 and surrounded by positively charged residues that are important in the binding of CFH to GAGs and C-reactive protein [36]. The mutation was previously found heterozygous in the parents (first cousins) of a child whom died 15 days after birth, because of severe HUS [27]. The final mutation observed in CFH (g.−315c > t) is located near the binding site for NFκβ, an important transcription factor of CFH in case of infection and inflammation [21]. If NFκβ cannot bind properly to the gene, transcription occurs in lesser amount or does not occur at all.

Fig. 3

Evolutionary conservation in mammals of the amino acid at (A) codon 1203 and 1206 of CFH, (B) codon 152 of IF and (C) codon 357 of IF.

As we found a statistically significant difference between the occurrence of all three polymorphisms associated with aHUS in patients and controls, it is suggested that carrying all three disease-associated polymorphisms leads to a higher risk of developing aHUS.

Both mutations found in MCP were located in the extracellular SCRs with C3b binding and cofactor activity. The deletion in SCR 4 causes reduced MCP levels and a reduced binding of ∼50% to C3b compared to controls [24]. Reduced protein levels and reduced binding suggest that the mutant protein is less expressed on the cell membrane. The splice site mutation g.IVS2 + 2t > g leads in a homozygous form to a deletion of 144 base pairs and 48 amino acids, caused by a splicing of the first 45 base pairs of exon 2 onto exon 3 [25].

Eighty percent of the mutations found in IF are located in the light-chain SR domain, which cleaves the alpha-chains of C3b and C4b [37]. The only variation that is not located in this region is the change from a valine into a methionine at codon 152. This mutation is located in the SCRC domain: a domain for which the function is still unknown, but probably is involved in protein–protein interactions, important in the binding of IF to CFH and MCP, to ensure cofactor activity of the latter two [24]. Furthermore, codon 152 is highly conserved through evolution (Figure 3B). Remarkably, this unknown change was found in two patients that were adults at onset of the disease, while the other IF mutations were mostly found in children. The splice site variation found in the intron 12, resulting in a decrease of the splice score from 0.93 to 0.86 [8], was found in two unrelated patients as well as in one healthy control. This fact makes it doubtful whether the variation is a mutation or a rare polymorphism. The mutation found at codon 340 has been described before [38]. It is located at the start of the SP domain, where isoleucine is the first amino acid of a structurally conserved region of the catalytically active domain; C3b and C4b coactivity are reduced to 0% [39]. The nonsense mutation p.Arg474Stop has been associated with heterozygous factor I deficiency and normal C3 levels [9]. Finally, the unknown mutation detected at codon 357 is located at a highly evolutionary conserved area of the protein (Figure 93C), indicating that a transition of isoleucine into methionine might influence the function of the protein.

Seven out of 19 aHUS patients (of which five patients presented ΔCFHR1/3 as well) displayed αFH (7/19; 36.8%). These five patients were all younger than 7 years of age at the onset of the disease. The absence of CFHR1 and CFHR3 in the blood of a patient again seems to trigger the development of specific auto-antibodies that bind to the recognition region of CFH, and in this way, block the binding of CFH to C3 convertases, especially in young patients [12,19]. It is important to test all remaining patients for the presence of αFH as well.

The mutation frequencies reported here for CFH, IF and MCP are slightly lower than those previously described in other cohorts, especially for CFH [7,8]. This is probably due to the fact that most of our patients are diagnosed with sporadic aHUS, and the frequency of mutations in these patients is lower than in patients with familial aHUS (13–20% and 32–42%, respectively) [8,21]. Differences may also be explained by regional differences between the cohorts. At this time point, already 9.7% of the patients (7/72) show αFH. Therefore, in this aHUS group, about the same amount of patients possess αFH as in the cohort reported by Joszi et al. (11%; 16/147) [12]. The presence of ΔCFHR1/3 in this patient population (9.7%) is slightly lower than in other cohort studies, where a deletion (both homozygous as well as heterozygous) is identified in 16–28% of the patients with aHUS [12]. In this study, a PCR method is used that only identifies a homozygous deletion of CFHR1/3 and not heterozygous deletions or a hybrid CFH/CFHR1 gene [20], which may explain the difference in results.

Remarkably, we observed several bigenic abnormalities in our aHUS patients. It is therefore important to perform routine DNA mutation analysis of all complement (regulating) genes associated with aHUS before decisions concerning future treatment modalities like kidney donation and renal transplantation are made. Patients with a potentially pathogenic mutation or a deletion in one of the complement regulators should be discouraged to undergo a renal transplantation, especially when the mutation is located in CFH or IF. As shown by other research groups, patients with mutations in these genes have a worse outcome after kidney transplantation than patients with a MCP defect [8,16,17]. In our research population, only one patient has a defect in MCP alone; the other two patients with a MCP defect also possess a mutation in IF. It is not yet known what the effect of αFH in combination with ΔCFHR1/3 is on the outcome of a renal transplantation or donation, but it may be that these patients will have an increased risk of HUS recurrence in the graft, and plasma exchange in combination with rituximab is recommended prior to and after transplantation [40].

More than 68% (49/72; 68.1%) of the patients in this study presented no αFH or a genetic aberration in one of the investigated genes. Genetic disorders in other not yet examined genes involved in complement activation, like complement C3 and thrombomodulin, could have a role in aHUS as well [14,15,41]. Further research on genes involved in complement regulation is needed to increase the understanding of the pathogenesis of the disease. Finally, this might lead to better treatment tailored to the genetic profile of the patients suffering from atypical haemolytic uraemic syndrome.

We would like to thank Linda Boll, Kim Santegoeds, Lonneke van der Linden and Rachel van Swelm for technical assistance, and Dr. George Borm for statistical advice. Furthermore, we would like to thank the patients, their parents and the physicians for their participation in this research. Special thanks for the physicians whom provided clinical information: Dr. Becaus (Aalst, Belgium), Dr. Gamadia (Amsterdam, The Netherlands), Dr. Vande Walle (Gent, Belgium), Dr. Laverman (Groningen, The Netherlands), Dr. Kümhof (Groningen, The Netherlands), Dr. Wetzels (Nijmegen, The Netherlands), Dr. Schreuder (Nijmegen, The Netherlands), Dr. van Kransbergen (Rotterdam, The Netherlands), Dr. van der Vlugt (Sneek, The Netherlands) and Dr. Lilien (Utrecht, The Netherlands). This work was partially supported by the Dutch Kidney Foundation (KBSO 07.0004, KBSO 09.0008 and C09.2313).

Conflict of interest statement. None declared.