Deletion of Complement Factor H–Related Genes CFHR1 and CFHR3 Is Associated with Atypical Hemolytic Uremic Syndrome

Atypical hemolytic uremic syndrome (aHUS) is associated with defective complement regulation. Disease-associated mutations have been described in the genes encoding the complement regulators complement factor H, membrane cofactor protein, factor B, and factor I. In this study, we show in two independent cohorts of aHUS patients that deletion of two closely related genes, complement factor H–related 1 (CFHR1) and complement factor H–related 3 (CFHR3), increases the risk of aHUS. Amplification analysis and sequencing of genomic DNA of three affected individuals revealed a chromosomal deletion of ∼84 kb in the RCA gene cluster, resulting in loss of the genes coding for CFHR1 and CFHR3, but leaving the genomic structure of factor H intact. The CFHR1 and CFHR3 genes are flanked by long homologous repeats with long interspersed nuclear elements (retrotransposons) and we suggest that nonallelic homologous recombination between these repeats results in the loss of the two genes. Impaired protection of erythrocytes from complement activation is observed in the serum of aHUS patients deficient in CFHR1 and CFHR3, thus suggesting a regulatory role for CFHR1 and CFHR3 in complement activation. The identification of CFHR1/CFHR3 deficiency in aHUS patients may lead to the design of new diagnostic approaches, such as enhanced testing for these genes.


Introduction
Atypical hemolytic uremic syndrome (aHUS) is characterized by a triad consisting of microangiopathic hemolytic anemia, thrombocytopenia, and acute renal failure in the absence of a preceding diarrheal illness. aHUS can be either sporadic or familial. Defective complement regulation occurs in both sporadic and familial aHUS. Disease-associated mutations have been described for the genes encoding the complement regulators complement factor H (CFH), membrane cofactor protein, factor I, and factor B [1][2][3][4]. In addition, autoantibodies to factor H have been reported in aHUS patients [5]. Recently, we showed in a family with aHUS that nonallelic homologous recombination [6] results in the formation of a hybrid gene derived from exons 1-21 of CFH and exons 5-6 of complement factor H-related 1 (CFHR1) [7]. The protein product of this hybrid gene is identical to the aHUS-associated CFH mutant S1191L/V1197A, which arises through gene conversion [8]. CFH and the genes encoding the five complement factor H-related proteins (CFHR1-CFHR5) reside in a centromeric 355-kb segment on Chromosome 1. Sequence analysis of this region provides evidence for multiple independent large genomic duplications, also known as low-copy repeats, resulting in a high degree of sequence identity between CFH and CFHR1-CFHR5 [9,10]. The secreted protein products of these genes are related in structure, as they are composed of repetitive units (;60 amino acids) named short consensus repeats (SCRs) [11]. In this study, we describe a novel form of nonallelic homologous recombination that results in the deletion of CFHR1 and CFHR3, but leaves CFH intact. This deletion is associated with an increased risk of aHUS.

Results/Discussion
Two cohorts of patients with aHUS have been studied, one from Jena, Germany and one from Newcastle, United Kingdom. For the Jena cohort of 121 aHUS patients, we used Western blotting to determine the absence of CFHR1 and CFHR3 in serum, as demonstrated for three patients in Figure 1A-1C. Complete absence of both CFHR1 and CFHR3 but presence of factor H, factor H-like protein 1, CFHR2, and CFHR4A was detected in 19 aHUS patients (16%) compared to two out of 100 control participants (v 2 ¼ 10.4, p ¼ 0.0012, odds ratio ¼ 8.5). All 19 patients showed normal factor H serum levels. In three of these 19 patients, DNA analysis confirmed that the deficiency was caused by a homozygous genomic deletion. The CFH genes were normal, as determined by sequence analysis. Specific primers were designed which span the 113-kb region from the 39 exons of CFH to CFHR4 (Figure 2A). Failure of primers R2-R6 to amplify DNA of these patients is explained by a 84-kb deletion of a genomic fragment that includes CFHR3 and CFHR1 and is located downstream of CFH and upstream of CFHR4. This deletion is flanked by two duplicated segments, B and B9, which are 28,638 bp and 28,714 bp in length, respectively. B includes exons 21, 22, and 23 of CFH and is located 59 of CFHR3. B9 includes exons 3, 4, and 5 of CFHR-1 and is located ;60 kb further downstream. Both segments have the same orientation, harbor several truncated long interspersed nuclear elements, and their sequence identity is .98 % [12]. The position of the deletion was mapped by amplifying regions of sequence variation between the duplicated segments. Forward and reverse primers specific for B and B9, respectively, generated a 9.2-kb product from aHUS patients' DNA, but not from control DNA. Sequence analysis allowed the identification of nucleotides from either B or B9 ( Figure 2B), thus demonstrating fusion of B and B9 as a result of nonallelic homologous recombination. This was confirmed by amplification of the fused segment BB9 using primers that also generate amplification products from segment B and segment B9 of control DNA. Sequence analysis of a divergent region in B and B9 confirmed that amplification of DNA from the patient generated one product with a single sequence ( Figure 2C, upper panel), while the control DNA generated two products with divergent sequences ( Figure 2C, middle panel), of which one is derived from segment B and one from segment B9 ( Figure 2C, lower panel). These data demonstrate fusion of segments B and B9 in the patient's DNA, and confirm that this patient is homozygous for the deletion ( Figure 3A). The same results were obtained for the other patients (unpublished data). All three patients had identical breakpoints, which were mapped to a 1.9-kb region of perfect homology within or directly preceding a L1MA2 element ( Figure 3B). (A) Using a monoclonal antibody that reacts with the C terminus of CFHR1, complete deficiency of CFHR1 is seen in the three HUS patients. The normal control shows the two differently glycosylated forms of CFHR1 (CFHR1a and CFHR1b, respectively). (B) Using CFHR3 antiserum, complete deficiency of CFHR3 is seen in three HUS patients, but CFHR3 is detected in the control subject. CFHR4A is present in all samples. (C) The sera of the three HUS patients contain factor H-like protein 1 and the glycosylated and nonglycosylated forms of CFHR2 (CFHR2a and CFHR2, respectively). doi: 10

Author Summary
Hemolytic uremic syndrome (HUS) is a severe kidney disease, which is characterized by hemolytic anemia, thrombocytopenia, and acute renal failure. The nondiarrhea-associated form, also known as atypical HUS (aHUS), is rare, sometimes familial, often recurrent, and has a poor outcome. Several studies have shown that aHUS is associated with mutations in genes coding for complement regulators, which leads to defective regulation of complement activation, particularly at cell surfaces. We report a novel susceptibility factor for aHUS in the form of a chromosomal deletion of a large (;84 kb) genomic fragment in the regulators of complement activation gene cluster at Chromosome 1q32. This deletion is a result of nonallelic homologous recombination and leads to the loss of two genes, CFHR1 and CFHR3, which encode factor H-related proteins 1 and 3, respectively. We recommend diagnostic screening of aHUS patients for these susceptibility factors.
The Newcastle cohort of 66 aHUS patients was investigated using multiplex ligation-dependent probe amplification (MLPA) [13] to measure copy number of CFHR1 exons 2 and 3 ( Table 1). The data show that deletion of CFHR1 is strongly associated with aHUS. In the aHUS group, 28% of the patients had this deletion, compared to 6% of the control group (v 2 ¼ 33.2, p ¼ 1.0 3 10 À8 , odds ratio ¼ 6.3). The following copy numbers of CFHR1 were found in aHUS patients: zero copies, 10%; one copy, 35%; two copies, 55%. In control subjects, the copy numbers were: zero copies, 2%; one copy, 9%; and two copies 89% (v 2 ¼ 28.7, p ¼ 5.9 3 10 À7 ). The allele frequency of CFHR1 deletion was 30% in those patients known to carry a mutation and 27% in those without a mutation (v 2 ¼ 0. 16 The functional effect of complete deficiency of CFHR1 and CFHR3 proteins was investigated using a modified version of a hemolytic assay with factor H-depleted plasma and sheep erythrocytes [14,15]. Addition of heat-inactivated serum from the three patients caused increased erythrocyte lysis (Figure 4). In contrast, serum derived from a healthy individual showed dose-dependent protection. These data show that CFHR1/CFHR3-deficient plasma has reduced protective activity and suggest that the absence of CFHR1 and/or CFHR3 contributes to the defective regulation of complement activation on cell and tissue surfaces. This is underlined by the fact that these patients have normal serum levels of factor H and factor I. Similarly, we previously showed that mutant factor H derived from HUS patients displayed reduced cell binding and protection activities [16].
In this study we report that hetero-and homozygous deletion of CFHR1 and CFHR3 through nonallelic homologous recombination events downstream of CFH is associated (C) Similar PCR products were generated with primers common for segment B and segment B9. Amplification of both segments is confirmed by sequence analysis. Sequence of PCR products from the patient is identical to segment B of a control DNA (upper panel). Sequence of PCR products of a control person is identical to segment B and B9 (middle panel). Identification of sequence B9 of the control person (lower panel). Sequence differences between segment B and segment B9 are indicated (grey bar). doi:10.1371/journal.pgen.0030041.g002 with an increased risk of aHUS. aHUS patients with deficiency of CFHR1/CFHR3 are characterized by a relatively young age (1-21 y) at disease onset. Previously, we showed that deletion of CFHR1 and CFHR3 reduces the risk of agerelated macular degeneration [17]. It is fascinating that the same polymorphic variant is associated with opposite effects: an increased risk for HUS and a decreased risk for age-related macular degeneration. How could this deletion be acting? One possibility is that the absence or presence of either CFHR1 and/or CFHR3 has a disease-modifying action. Although each of the two proteins alone lacks cofactor and decay-accelerating activity, CFHR3 has a cofactor-enhancing activity [18]. In addition, both CFHR1 and CFHR3 bind C3b and heparin, thus suggesting a regulatory function in C3b processing [18].
Another possibility is that the deletion is in linkage disequilibrium with other susceptibility alleles in CFH or that it may affect CFH transcription. Understanding the functional effect of this disease-modifying deletion will help to extend our understanding of the role of complement in the pathogenesis of both aHUS and age-related macular degeneration.

Materials and Methods
Participants. Newcastle. Sixty-six patients from Newcastle with a clinical diagnosis of aHUS were included in this study. Within this group, 15 were known to carry a CFH mutation, two carried a membrane cofactor protein mutation, and three carried a factor I mutation. Screening with MLPA for exons 22 and 23 of CFH showed that none of these patients carried a hybrid CFH/CFHR1 gene. The allele frequency of CFHR1 deletion in the aHUS patients was compared to 119 samples of human control DNA from a randomly selected population (healthy blood donors) obtained from the European Collection of Cell Cultures (http://www.ecacc.org.uk). The study was approved by the Northern and Yorkshire Multi-Centre Research Ethics Committee.  Jena. A total of 121 aHUS patients from Jena were included in this study. Of these, 19 were found to be deficient in CFHR1 and CFHR3. Genomic deletion of CFHR1 and CFHR3 genes was determined in three patients ( Table 2). None of these three patients carried a hybrid CFH/CFHR1 gene. The study was approved by the Research Ethics Committee of the University of Cologne, Germany, and by the Research Ethics Board of the Hospital for Sick Children, Toronto, Canada.
Sera and Western blot analysis. Serum samples from 121 aHUS patients (Jena cohort) and 100 anonymous healthy blood donors were assayed for the presence and mobility of factor H and CFHR1. Serum was separated by SDS-PAGE, transferred onto a membrane, and incubated with polyclonal factor H antiserum (Calbiochem-Novabiochem, http://www.emdbiosciences.com) or monoclonal C18 antibody [19], which identifies an epitope in factor H and CFHR1. CFHR3 was detected with polyclonal CFHR3 antiserum [20].
Genetic analysis. Genomic DNA was prepared from peripheral blood cells of three patients. Genomic DNA was amplified by PCR using specific primers R1-R8 that cover the 100-kb region downstream of the factor H gene ( Table 3). Amplification of the breakpoint region and sequence analysis was performed using primers P9.2 and B1 and B2. For amplification of segment B and B9, primer P was used. The sequence of the amplified products was determined using an ABI 3100 sequence analyzer (Applied Biosystems, www.appliedbiosystems.com). The sequences were compared to that of the genomic DNA of Chromosome 1. The sequences of segment B and segment B9 were compared and analyzed for repeat content using CENSOR (http://www.girinst.org).
Erythrocyte lysis assay. Factor H-depleted human plasma was prepared by immunoadsorbance of factor H from normal human plasma. Polyclonal factor H antiserum (Merck Biosciences, http:// www.merckbiosciences.co.uk) was covalently coupled to a 1-ml HiTrap NHS (N-hydroxy succimide) column (GE Health Care, http://www.gehealthcare.com). Factor H depletion was confirmed by Western blotting, and ELISA showed that about 66% of the protein was removed. Depleted plasma was directly used for hemolytic assays. Hemolytic experiments were performed in VBS buffer (veronal buffered saline, 144 mM NaCl, 7 mM MgCl 2 , and 10 mM EGTA, pH 7.4). Increasing amounts of heat-inactivated (10 min at 56 8C) CFHR1/ CFHR3-deficient serum or control serum were added to the depleted plasma and 2 3 10 7 sheep erythrocytes. Following incubation at 37 8C for 30 min, the mixture was cleared by centrifugation and the absorbance in the supernatants was measured at 414 nm. Using human complement active plasma sheep erythrocytes serve as nonactivators of complement. These cells are protected from lysis and no hemoglobin is released. However, when factor H is depleted from human plasma, this plasma has the ability to lyse sheep erythrocytes, as demonstrated by the release of hemoglobin and an increase in absorbance. Addition of purified factor H or normal human heatinactivated serum reconstitutes protection of erythrocytes [15].
Multiplex ligation-dependent probe amplification. The MLPA reaction has been previously described [13]. In this study, a completely synthetic probe set was used, obviating the need for a cloning step in the production of probes. Probes were designed to determine dosage for exons 2 and 3 of CFHR1, along with control probes for MSH2 exon 1 and MLH1 exon 19. Each probe pair hybridises to immediately adjacent targets at the sequence of interest. Hybridisation sequences are shown in Table 4. Probe pairs also contain binding sites for primers used in the MLPA reaction, as well as stuffer sequence to ensure that each amplified probe product is of a unique length. Oligonucleotides were obtained from TAG Newcastle, (http://www.vhbio.com). Righthand probes were 59-phosphorylated and purified by desalting.
Reagents for the MLPA reaction were purchased from MRC-Holland (http://www.mrc-holland.com). The ligation reactions were carried out according to the manufacturer's recommended protocol using 100-200 ng of genomic DNA and 2 fmol of probe. Incubations and PCR amplification were carried out on a DNA Engine Tetrad 2 thermal cycler (http://www.bio-rad.com). Amplified products were diluted 10-fold to give peak heights within the quantitative range (approximately 100-4,000 units) on the ABI PRISM 3130 Genetic Analyzer capillary electrophoresis system (Applied Biosystems, http://www.appliedbiosystems.com). Diluted product (1 ll) and 0.5 ll of ROX 500 internal size standard (Applied Biosystems) were made up to 10 ll using dH 2 O and samples were run on the ABI 3130. Peak areas for each sample were determined using the proprietary Genemapper software (Applied Biosystems) and dosage quotients calculated. . Defective Protective Activity with CFHR1/CFHR3 Deficient Plasma A novel type of hemolytic assay was used to analyze hemolytic activity of CFHR1/CFHR3 deficient human serum. Sheep erythrocytes were incubated with factor H-depleted human plasma. Sheep erythrocytes generally act as nonactivators of complement in human plasma. However, depletion of factor H from the plasma converts the sheep erythrocytes into activator surfaces (ctrl, empty circle). Addition of heat inactivated serum to depleted plasma as a source of factor H has a dosedependent protective effect (''control''). Compared to normal human serum, sera of patient P1 to P3 with CFHR1/CFHR3 deficiency showed less protective activity as indicated by the increased lysis of erythrocytes.