Edinburgh Research Explorer Atypical Hemolytic Uremic Syndrome

Summary: Hemolytic uremic syndrome (HUS) is a triad of microangiopathic hemolytic anemia, thrombocy- topenia, and acute renal failure. The atypical form of HUS is a disease characterized by complement overactivation. Inherited defects in complement genes and acquired autoantibodies against complement regulatory proteins have been described. Incomplete penetrance of mutations in all predisposing genes is reported, suggesting that a precipitating event or trigger is required to unmask the complement regulatory deficiency. The underlying genetic defect predicts the prognosis both in native kidneys and after renal transplantation. The successful trials of the complement inhibitor eculizumab in the treatment of atypical HUS will revolutionize disease management.

T he hemolytic uremic syndrome (HUS) is characterized by the triad of thrombocytopenia, microangiopathic hemolytic anemia, and acute renal failure. 1 The most common form of HUS is secondary to shiga toxin (Stx)-producing bacteria, typically Escherichia coli O157:H7. Atypical HUS (aHUS) has been used to classify any HUS not caused by Stx. A variety of precipitating events have been associated with aHUS including infections, drugs, autoimmune conditions, transplants, pregnancy, and metabolic conditions (Table 1). These have frequently been called secondary aHUS. With the discovery of the role of complement gene mutations in aHUS, primary aHUS has been used to refer to those cases with documented complement dysregulation. Although a useful aide memoir, these terms do not account for the increasing recognition that patients with an underlying complement risk factor often require a secondary trigger for aHUS to manifest. Classifications that take account of both the genetic background and etiologic trigger are beginning to be introduced. 2 The best estimate of aHUS incidence is 2 of 10 6 in a North American population, 3 although the precise proportion with an underlying complement defect is not known.

PATHOLOGY
In acute aHUS, the pathologic picture is of capillary thrombosis. Glomerular capillary wall thickening is seen as a result of endothelial cell swelling and accumulation of flocculent material between the endothelial cell and the basement membrane. Platelet and fibrin thrombi result in occlusion of the glomerular capillaries. Fibrinoid necrosis of the afferent arteriole associated with thrombosis also may be seen. Mesangiolysis occurs early in the disease process and subsequently is replaced by sclerotic changes. Early arterial changes are variable, ranging from only mild endothelial swelling to fibrinoid necrosis with occlusive thrombus formation. Subsequently, there is mucoid intimal hyperplasia with narrowing of the vessel lumen. Deposition of fibrin or fibrinogen in the glomeruli and in the mesangium, as well as within the vessel walls, are seen on immunofluorescence. Complement and immunoglobulin deposits along the capillary loops of glomeruli may be seen. 4

THE COMPLEMENT SYSTEM
The complement system is an ancient defense mechanism that stimulates the inflammatory response and destroys pathogens through opsonization and lysis 5 (Fig. 1). In addition to protecting the host against invading pathogens, it bridges innate and adaptive immunity and it disposes of immune complexes and injured tissues and cells. 6 The alternative pathway of complement (AP), which plays a key role in the pathogenesis of aHUS, is continually activated by a tick-over mechanism, and can also be triggered by the classic and lectin pathways.
In the AP, complement C3 undergoes spontaneous hydrolysis, depositing C3b onto the surface of foreign and host cells in the vicinity. On an activating surface such as a bacterium, C3b joins with factor B, which then is cleaved by factor D to form the C3 convertase, C3bBb. The binding of properdin stabilizes this enzyme. This enzyme complex then cleaves more C3 to C3b to initiate a feedback loop. Downstream of this amplification loop, C3b also may join with the C3 convertase to form the C5 convertase. C5 is cleaved to the anaphylatoxin C5a and C5b, which initiates formation of the lytic membrane attack complex (C5b-9) (Fig. 1). To protect host cells from collateral complement damage, many soluble and membrane-associated complement regulatory proteins function to inactivate complement on their surface. It is the imbalance between complement activation and regulation on host cell surfaces that underlies the pathogenesis of aHUS.
In addition to regulating complement in the fluid phase, CFH also can protect host surfaces by binding to polyanions such as the glycosaminoglycans. 18 CFH has two glycosaminoglycan binding domains in CCPs 6 to 8 and CCPs 19 and 20, 17 which have different sulfate specificities resulting in the C-terminal domains (CCP 19 and 20) being predominantly responsible for binding to kidney, and CCPs 6 to 8 being responsible for binding in the eye. 19 Other recent studies have shown that CFH also binds to the lipid peroxidation product malondialdehyde, 20 the acute phase proteins, C-reactive protein, [21][22][23] and pentraxin 3, 24 as well as necrotic cells. 21 The majority of mutations in CFH are heterozygous, are located in CCPs 19 and 20 (Fig. 2), and do not usually result in a quantitative deficiency of CFH. Structural and functional analysis of the C-terminal mutants has revealed variable consequences on binding to heparin, C3b, and endothelial cells; however, cell surface complement regulation is consistently impaired as measured using sheep erythrocyte lysis assays [25][26][27] (Table 3). Thus, these C-terminal mutants are predicted to fail to control complement activation at the glomerular endothelium. In keeping with this, renal biopsy data from an aHUS patient with a C-terminal mutant showed reduced CFH binding to renal endothelium compared with wild type. 27 C-terminal CFH mutants also have been shown to have impaired binding to platelets resulting in increased complement activation with consequent platelet activation, aggregation, and release of tissue factor-expressing microparticles. 28 Although CFH mutations cluster in the C-terminus, genetic changes are reported throughout the molecule. Oral contraceptives [193][194][195] Illicit drugs (eg, cocaine, heroin, ecstasy) 196 Autoimmune Anticardiolipin 197 C3Nef 198 Systemic lupus erythematosus 199 Vaccination Hepatitis B 10 Bone marrow transplantation 200 Malignancy (gastric, breast, prostate, lung, colon, ovarian, pancreatic, lymphoma) 201 Combined methylmalonic aciduria and homocystinuria 202 Mutations in the N-terminal region of CFH (CCP 1-4) have been reported and functional analysis has shown that they are defective in their ability to control the AP both in the fluid phase and on cell surfaces. 29 Sequence variants also have been described in the intervening CCPs of CFH, although for many of these changes the functional consequences and importance in pathogenesis remains uncertain. Tortajada et al 30 examined two of these variants of uncertain significance (VUS) and showed no functional consequences. They went on to show that these variants form part of an infrequent ancient CFH haplotype in Caucasians that is not enriched in the aHUS population, highlighting the importance of modeling genetic mutations before ascribing functional significance. CFH resides in the Regulators of Complement Activation (RCA) cluster and the architecture of this region makes it particularly prone to genomic rearrangements. The gene for CFH is in close proximity to the genes encoding the five factor H-related proteins that are thought to have arisen from several large genomic duplications and thus have a very high degree of sequence identity to CFH.
This homology predisposes to gene conversions and genomic rearrangements through nonallelic homologous recombination and microhomology-mediated end joining. The CFH mutations S1191L, V1197A, and combined S1191L/V1197A arose through gene conversion between CFHR1 and CFH. 31 A hybrid (fusion) gene comprising the 21 N-terminal exons of CFH and the 2 C-terminal exons of CFHR1 was shown to have arisen through nonallelic homologous recombination and resulted in aHUS. 32 More recently, a hybrid gene consisting of the 22 N-terminal exons of CFH and the 5 C-terminal domains of CFHR3 arising through microhomologymediated end joining was reported in aHUS. 33 As with C-terminal point mutations in CFH, these hybrid genes also result in loss of cell-surface complement regulation.
A transgenic mouse lacking the C-terminal domains of CFH (Cfh -/-Δ16-20) was generated to provide an in vivo model of aHUS. 34 These mice spontaneously developed aHUS, confirming the importance of local endothelial cell complement regulation 35

COMPLEMENT FACTOR I
Mutations in complement factor I (CFI) account for between 5% and 10% of aHUS (Tables 2 and 4 Figure 1. Complement activation and regulation. The AP is a positive amplification loop. C3b interacts with factor B (B), which then is cleaved by factor D to form the AP C3 convertase (C3bBb). This enzyme complex is attached to the target covalently via C3b while Bb is the catalytic serine protease subunit. C3 is the substrate for this convertase, thus creating a powerful feedback loop. Unchecked, this leads to activation of the terminal complement pathway with generation of the effector molecules, the anaphylatoxin C5a, and the membrane attack complex (C5b-9). To protect host cells from bystander damage the AP is down-regulated by complement regulators including CFH, CFI, and MCP. Activating mutations in C3 and CFB and loss-of-function mutations in CFH, CFI, and MCP, in addition to autoantibodies to CFH and CFI, result in overactivation of the AP with resultant aHUS.
mutations described in aHUS are all heterozygous. These mutations cluster in the serine protease domain ( Fig. 3 and Table 4). Functional analysis has been performed for a number of mutants and shows a loss of both alternative and classic pathway regulatory activity in the fluid phase and on cell surfaces (Table 4). 39,41,42 As seen in CFH, several VUS have been described in CFI in which analysis has failed to show any alteration in secretion or function (eg, G261D). Such variants may not be involved in the pathogenesis of disease.

MEMBRANE COFACTOR PROTEIN
Mutations in MCP (CD46) are found in approximately 10% of patients with aHUS 11,40,[43][44][45][46][47] (Table 2). MCP is a surface-bound complement regulatory protein that acts as a cofactor for the CFI-mediated cleavage of C3b and C4b that are deposited on host cells. 48 The majority of mutations described in aHUS are found in the extracellular four CCP domains that are responsible for C3b and C4b binding (Table 5 and Fig. 4). Most MCP mutations described to date have resulted in a quantitative defect in MCP ( 75%). The remaining mutations have been shown to result in a secreted, nonfunctional protein (Table 5).

ACTIVATING MUTATIONS IN COMPLEMENT COMPONENTS
Mutations have been described more recently in the complement components C3 and complement factor B (CFB). C3 is cleaved to form the anaphylatoxins C3a and C3b, which are highly reactive, and can bind to cell surfaces via their reactive thioester. C3b then can interact with CFB in the presence of factor D to form the AP C3 convertase (C3bBb), which cleaves further C3, introducing a positive-feedback loop.
Mutations in C3 have been reported in several cohorts of aHUS patients 9-11,49-54 at a frequency of 2% to 10% (Table 2). Initial functional analysis showed that MCP was unable to bind to mutant C3, preventing its cleavage to iC3b. 50 Two C3 mutations have been described that resulted in decreased secretion and their role in pathogenesis remains uncertain. More recently, two mutations in C3 have been reported that bind to CFB with higher affinity, resulting in increased C3 convertase formation. 52,55 These mutations result in increased complement activation on platelets 52 and glomerular endothelium 55 ( Fig. 5 and Table 6).
Gain-of-function mutations also have been reported in CFB although these appear to be rare. 9,10,51,56-58 Goicoechea de Jorge et al 56 described two mechanisms through which these separate mutations led to increased complement activation. One mutant (F286L) showed Table 2. Type I mutation A type I mutation results in a quantitative deficiency of the protein, and a type II mutation results in normal levels of a nonfunctional protein.
*Includes CFH/CFHR hybrid genes. ‡ Five-year outcome data. § Three-year outcome data. † Estimate was based on mutation, not antigenic levels.
enhanced formation of the C3bB proenzyme that resulted in a more active enzyme in vivo. The other mutant (K323E) formed a C3bBb enzyme more resistant to decay by the complement regulators decay accelerating factor (CD55) and CFH. This also caused increased enzyme activity. 56 Subsequent analysis of two further mutants located in the von Willebrand type A domain (D279G and K350N) showed increased convertase formation and resistance to CFH dissociation, ultimately resulting in increased complement deposition on endothelial cells. 55 THROMBOMODULIN Thrombomodulin (THBD) facilitates the activation of protein C by thrombin 59 and enhances thrombinmediated activation of plasma procarboxypeptidase B (CPB2), an inhibitor of fibrinolysis that also inactivates complement-derived anaphylatoxins C3a and C5a. It has also been shown to down-regulate the AP by accelerating CFI-mediated inactivation of C3b in the presence of cofactors. 60 Mutations in THBD recently were described in aHUS. The variations reported were heterozygous missense mutations, with no mutations resulting in a loss of secretion. These mutations resulted in a loss of cofactor activity. 60 Maga et al 11 also have reported THBD genetic variants in aHUS, although they suggested that at least one of the previously reported mutations was a polymorphism present in 3% of the population and that several THBD variants were present with an additional mutation. No isolated THBD mutations were described in a large French cohort of patients (n ¼ 214), although a few individuals carried THBD genetic variants in addition to a mutation in a known complement gene. 49 Several large cohorts of aHUS have yet to report on the frequency of genetic variants in THBD in aHUS.  Figure 2. CFH and aHUS-associated mutations. CFH is composed of 20 CCP modules. The N-terminal modules (CCP1-4) bind to C3b and act as a cofactor for the CFI-mediated cleavage to the inactive iC3b. The C-terminal modules (CCP19 and 20) bind to C3b and glycosaminoglycans on host cells to mediate cell surface protection. Genetic variants described in aHUS cluster in CCPs 19 and 20, but can be seen throughout the molecule. Functional analysis of aHUS-associated variants has focused predominantly on the C-terminal variants (Table 3).

OTHER GENETIC VARIANTS
limited understanding of the functional role of CFHR5 further adds to the uncertainty of its role in aHUS pathogenesis and replication studies are required.
A functionally significant mutation (Q433P) in the membrane attack complex regulator, clusterin, has been reported in a family with aHUS. 63 The affected individual also carried a functionally significant mutation in MCP; therefore, it is unclear whether mutations in clusterin are sufficient alone to cause aHUS.

FACTOR H AUTOANTIBODIES
In addition to the genetic abnormalities described in aHUS, autoantibodies to CFH also have been linked to disease in 4% to 14% of aHUS patients (Table 7). 10,51,[64][65][66][67][68] In a pediatric-only cohort, this figure was reported to be as high at 25%. 69 CFH autoantibodies in aHUS are strongly associated with an 80-kb-long genomic homozygous deletion of CFHR1 and CFHR3. 10,65,66,70 More recent analysis also has shown that some aHUS patients with CFHR1 deficiency resulting from point mutations in CFHR1 71 or from a deletion incorporating CFHR1 and CFHR4 71,72 have developed anti-CFH antibody (Ab). This may suggest that deficiency of CFHR1 is the predominant predisposing factor in the development of autoantibodies. Despite this, deficiency of CFHR1 is not a prerequisite for formation of autoantibodies because several aHUS patients have been reported with high titers of Ab with no evidence of deficiency of CFHR1 or CFHR3. 10,67,73,74 In many cases, aHUS patients with anti-CFH Ab also carried functionally significant mutations in other complement genes. 67 The anti-CFH antibodies reported to date have been predominantly of the IgG class, although a few IgA anti-CFH Ab have been described. 74 Mapping of the epitopes initially suggested that the anti-CFH Ab      bound predominantly to the C-terminus, 66,67,75,76 however, recently it was reported that the response was polyclonal to multiple epitopes throughout CFH. 77 There was also cross-reactivity with CFHR1 67,74,77 and CFHR2, 77 but it was not reported for CFHR3 or CFHR4A. 74 Several studies have shown various functional consequences of anti-CFH Abs. The antibodies have been shown to reduce binding to C3b 75,77 and other C3 fragments. 77 They perturb CFH-mediated cell surface protection 75,77 and in some individuals the autoantibodies also impair cofactor activity 77 or decay accelerating activity. 76 In keeping with this, low C3 levels frequently are seen in the autoimmune form (Table 7).
It also has been shown that CFH autoantibodies form immune complexes in the serum, 74,77 which may explain the low CFH levels seen in 28% of the cases. 77 In addition, these immune complexes correlated with disease activity. 77 In summary, these functional studies suggest a pathogenic role for CFH autoantibodies in aHUS.

FACTOR I AUTOANTIBODIES
Autoantibodies to CFI also have been described in aHUS but are much rarer than anti-CFH Abs (0%-2%). 68,78 Unlike anti-CFH Abs they are not associated with a deletion of CFHR1 and 3. Anti-CFI Abs were seen to form immune complexes in serum, however, functional analysis showed only a minor effect in fluid phase cofactor activity. 78 The co-existence of functionally significant mutants in the majority of patients, added to the lack of correlation of anti-CFI Ab titer and disease activity, raise the possibility that they are an epiphenomenon rather than a direct cause of disease. Large replication studies will be needed to confirm this initial observation.

GENETIC SUSCEPTIBILITY FACTORS
A number of single-nucleotide polymorphisms (SNPs) in CFH have been shown to be associated with aHUS in several studies. 34,64,[79][80][81][82] A haplotype in CFH (CFH-H3; tgtgt) composed of these SNPs increases this risk of aHUS two-to four-fold. 34,49 This haplotype contains a SNP in the region of CFH responsible for cofactor activity. Functional analysis has shown that the risk variant, CFH-Val 62 , has a subtle decrease in cofactor activity compared with the protective variant, 29,83,84 in keeping with the minor structural differences between these SNPs. 85 A haplotype block in MCP (MCPggaac) comprising 2 SNPs in the promoter region has been associated with a two-to three-fold increased risk of aHUS. 49,80,81 Some of these studies have suggested that this risk occurs exclusively in those patients already carrying complement mutations. 80,82 Reporter gene assays have suggested that this haplotype reduces transcriptional activity by 25%, 80 Figure 3. Location of aHUS-associated mutations within the crystal structure of factor I (protein database identification code: 2XRC). 160 Factor I is a heterodimer consisting of a noncatalytic heavy chain linked by a disulfide bond to a catalytic light chain. The domain structure of CFI is shown with the heavy chain comprising the FIMAC domain, light blue; SRCR domain, pale green; LDLr1, cyan; and LDLr2, magenta; and the light chain or serine protease domain, deep olive. aHUS-associated genetic variants are shown as red spheres. Yellow spheres mark the catalytic triad of the serine protease domain. Functional analysis of aHUS-associated CFI variants are described in Table 4. expression in vivo. 80,86,87 Further experiments on human umbilical vein endothelial cells (HUVECs) with risk and protective MCP haplotypes showed equal MCP expression whether in the resting state, after cytokine treatment, or free heme treatment. Likewise, no increase in complement deposition could be shown on HUVECs bearing the MCP risk haplotype. 87 A SNP in C4b binding protein (R240H) was associated with aHUS in cohorts from the United Kingdom and France. C4b binding protein is the predominant classic pathway fluid phase regulator but also has weak AP regulatory activity. Functional analysis showed inefficient AP regulation compared with wild type. 88 In a replication study in a Spanish cohort this association could not be confirmed. 89 A CFHR1 polymorphism (CFHR1*B) resulting from a gene conversion between CFH and CFHR1 is associated strongly with aHUS in the homozygous state, 64 an association replicated by Fremeaux-Bacchi et al. 49 It has been suggested that because the CFHR1*B is identical to CFH in CCP18 that it may compete with CFH at the glomerular endothelium, thus impairing complement regulation.
In a study examining SNPs in 47 complement genes in two separate cohorts, SNPs in CFHR2 and CFHR4 were also associated with aHUS. In this study there were no reproducible associations between SNPs and aHUS outside the RCA cluster. 82

INCOMPLETE PENETRANCE
Incomplete penetrance has been reported for all the genes associated with aHUS. Penetrance has been reported at around 50% for individuals carrying CFH, CFI, MCP, and CFB mutations, 40,90 and slightly lower for C3 mutations, albeit with small numbers. 53 In index cases, the age-related penetrance is significantly higher than their mutation-positive relatives regardless of gene. 91 This suggests that the penetrance is altered by other genetic and environmental modifiers.
Patients have been reported with mutations in more than one complement gene 11,86,[92][93][94] or mutations in one complement gene in addition to autoantibodies to complement components. 67,78 In a study of 795 aHUS patients the European Working Party on Complement Genetics showed that at least 3.4% of aHUS cases will have more than one mutation. Eight percent to 10% of patients with mutations in CFH, C3, or CFB had combined mutations whereas 25% of patients with mutations in CFI or MCP had combined mutations. 46 The penetrance increased as the number of mutations in a patient increased. 46 As described, haplotypes and SNPs act together with mutations and inhibitory autoantibodies to increase the penetrance of disease.
Even in the situation in which a patient has multiple genetic risk factors, disease may not manifest until middle age, suggesting a triggering stimulus is required for disease to manifest. In such cases, it is likely that aHUS results from an otherwise innocuous stimulus that triggers the AP and sets off a self-amplifying cycle that cannot be controlled adequately in susceptible individuals.

TRIGGERING EVENTS
Historically, many agents have been linked to aHUS (Table 1). Although many of these associated conditions are common, aHUS is rare, suggesting an underlying predisposition.
Recent analysis of cohorts of aHUS patients with complement mutations have identified upper respiratory tract infections, fevers, pregnancy, drugs, and non E coli diarrheal illnesses as potential triggers. 40  Following the CCPs is an alternatively spliced region, rich in serine, threonine, and proline (STP region). The STP region is followed by a group of 12 amino acids of unknown function, a hydrophobic domain, a charged transmembrane anchor, and the alternatively spliced cytoplasmic tail (CT). Mutations associated with aHUS are clustered in the four extracellular CCPs of the molecule. Functional analysis of aHUS-associated MCP variants are described in Table 5.
diarrhea preceded aHUS in 23% of a French cohort 94 and in 28% of an Italian cohort. 51 Infectious events, mainly upper respiratory tract infections or diarrhea/gastroenteritis, trigger onset of aHUS in at least half of patients. 49,51 Likewise, in pregnancy-associated aHUS, Fakhouri et al 98 showed that 86% of patients for whom this was a trigger carried a complement gene mutation. Drugs also may unmask latent complement defects resulting in aHUS. 40,99

COMPLEMENT SCREENING IN AHUS
Once the initial diagnosis of a thrombotic microangiopathy (TMA) has been made, the initial management involves differentiating between Stx-HUS, thrombotic thrombocytopenic purpura, and aHUS (for a diagnostic algorithm see the article by Loirat and Fremeaux-Bacchi 100 ). Rapid exclusion by analysis of a disintegrin and metalloproteinase with a thrombospondin type 1 motif, member 13 (ADAMTS13) activity and microbiological analysis for Stx-producing E coli can lead to a diagnosis of aHUS. After exclusion of Stx-HUS and thrombotic thrombocytopenic purpura, precipitating events and the underlying genetic defects predisposing to aHUS should be sought (Table 8).
Complement analysis in cases of aHUS should include serum levels of C3, C4, CFH, and CFI before plasma exchange. Low C3 levels are seen commonly in patients with mutations in CFH, CFI, MCP, C3, and CFB and may point to a complement-mediated process, however, normal C3 levels do not exclude the presence of a mutation in, or autoantibodies against, the complement system (Table 2). 101 Fluorescenceactivated cell sorter analysis of peripheral blood mononuclear cells provides a quick and relatively inexpensive screening option for MCP mutations, although genetic analysis still is required to detect all changes.
Genetic screening in aHUS is challenging because most of the disease-associated mutations are individually rare. In the case of nonsense mutations, large gene rearrangements, and frame shift mutations, the functional consequences are certain. In aHUS, missense mutations or splice-site variations in complement genes comprise a high proportion of the variants found and the changes may be of uncertain biologic or clinical relevance. These VUS pose a challenge when reporting the genetic results. 102 Although predictions can be made as to the likely significance of a VUS, ultimately, functional assays are required to assess the importance of a variant. Such analysis often has led to the reclassification of a previously described mutation as an irrelevant polymorphism. 30 Care always should  (Table 6).
be taken when interpreting these VUS and screening laboratories should revisit past genetic data in light of new evidence. In addition to direct DNA sequencing, the not-insubstantial number of gene conversions and genomic rearrangements found in aHUS makes copy number analysis essential in aHUS screening (see www.genetests.org for national screening laboratories).

PROGNOSIS
The overall prognosis for patients with aHUS has been poor. Initial mortality has been reported to be higher in children (6.7% versus 0.8% at 1 year), 49 although adults progress to end-stage renal disease (ESRD) more frequently at initial presentation (46% versus 16%). 49 At 3 to 5 years after onset, 36% to 48% 49,51 of children and 64% to 67% 49,51 of adults had died or reached ESRD. Prognosis varies with genotype, with MCP mutations carrying the best prognosis, 49,51,94 although in one study this was true only if the first presentation of aHUS occurred in childhood 49 (Table 2). No patient with an MCP mutation from either the French 49 or Italian cohorts 51 died at first episode and none of the children and only 25% of adults with an MCP mutation developed ESRD at first episode. At 3 years only 6% of all patients with MCP mutations 51 had developed ESRD and by 5 years only 35% had developed ESRD. 49 Individuals with mutations in CFH, CFI, or C3 all had poor outcomes. In patients with a CFH mutation the initial mortality rate was 30% in children and 4% in adults, 49 and evolution to ESRD at first episode in survivors was 19% to 33% in children and 48% in adults. 49,51 At 3 to 5 years of follow-up evaluation, up to 77% of patients with CFH mutations had developed ESRD or had died. Only 30% to 40% of individuals with CFI and C3 mutations will be alive with native kidney function at 3 to 5 years. 49,51 The prognosis of aHUS with CFB mutations also is poor. 51,55,56 A proportion of patients will have combined mutations. In patients with either CFH or CFI mutations, the presence of mutations in other genes did not modify prognosis. In contrast, the prognosis for individuals with an MCP mutation was worse if they also carried an additional mutation. 46 In those patients with CFH autoantibodies, 36.5% to 63% die or reach ESRD in the long term 49,51,[64][65][66][67] (Table 7).

EXTRARENAL MANIFESTATIONS
Extrarenal manifestations are reported in only 10% to 20% of patients with aHUS. The most frequently reported symptoms ( 10%) are neurologic, with symptoms ranging from irritability to coma. It is unclear how many of these symptoms are the direct result of a cerebral TMA, with severe hypertension and posterior reversible encephalopathy syndrome possible Plasma amino acid chromatography Urine organic acid chromatography Genetic analysis differential diagnoses that may be differentiated on magnetic resonance imaging. 103 Many case reports of other organ involvement occurring concurrently with aHUS have been described (Table 9), but in the majority, definitive biopsy evidence of TMA in the organ was lacking. Extrarenal manifestations have been reported more commonly in CFH-autoantibodymediated disease, with 23.5% having seizures and 23.1% developing pancreatitis. 65

Plasma Exchange
Until the beginning of this decade, plasma exchange (PE) was considered the gold standard for management of aHUS. The replacement of nonfunctioning complement proteins and removal of CFH autoantibodies and hyperfunctional complement components (eg, gain-offunction mutations) made PE a logical choice (reviewed in European 104 and UK 105 guidelines on aHUS treatment). The consensus-based guidelines recommended that PE should be commenced as soon as possible after diagnosis of aHUS, using 1 to 2 plasma volumes per session in adults and 50 to 100 mL/kg in children. Initially, PE should be performed daily and the dose titrated to control hemolysis. Once hemolysis has been controlled, PE can be withdrawn slowly, although individuals with genetic defects in the complement system are frequently plasma dependent and require long-term plasma therapy (weekly/biweekly) to maintain remission. Only once ADAMTS13 deficiency is excluded should eculizumab be considered. The price of eculizumab will mean that PE will remain the only currently available option in many countries.

Eculizumab
Eculizumab, a recombinant humanized monoclonal antibody directed against C5, blocks the cleavage of C5 into its effector components C5a and C5b. 106 Since the initial use of this complement inhibitor in aHUS in 2009, 107 the effectiveness of eculizumab has been communicated in many case reports 107-119 , meeting abstracts [120][121][122][123][124][125] , and a recently published prospective trial. 125a Eculizumab appears highly effective, with approximately 85% of patients becoming disease-free in both plasma-resistant and plasma-dependent aHUS (reviewed by Wong et al 126 ). It is effective in patients with and without identified complement mutations. Similar to PE, the earlier eculizumab is commenced, the greater the preservation of kidney function. It has been suggested that eculizumab achieves better control of disease as witnessed by improvement in renal function after switching from PE and in rescuing plasmaresistant individuals. 126 It should be noted, however, that a randomized trial of eculizumab against PE was not, and is unlikely to be, performed.
Treatment with eculizumab should begin as soon as Stx-HUS and ADAMTS13 deficiency can be eliminated. Current protocols suggest life-long treatment with eculizumab will be required, however, as we gain more clinical experience it is likely that there will be certain subgroups in whom the treatment can be stopped (eg, those with isolated MCP mutations). Eculizumab has been used safely during pregnancy in patients with paroxysmal nocturnal hemoglobinuria. 126 Because host defense against encapsulated organisms is dependent on the ability to form a membrane attack complex (C5b-9), vaccination against Neisseria meningitides is required before treatment with eculizumab. However, because the current vaccine (tetravalent) does not cover the most prevalent strain (serogroup B), long-term prophylactic antibiotic cover also has been suggested. 126

RENAL TRANSPLANTATION IN aHUS
The outcome of renal transplantation in patients with aHUS is poor. In one adult series of 71 renal transplants, the 5-year death-censored graft survival was only 51%, with a 7% mortality rate at 5 years. 99 Graft failure is predominantly due to aHUS recurrence which occurs in 60% to 70% of patients, 99,127 and occurs early after transplant (70% in the first year after transplant 99 ).
The outcome of renal transplantation is predicted largely by the underlying genetic abnormality. In individuals with mutations in CFH the recurrence rate is greater than 80%. Similarly, activating mutations in C3 and CFB also have a high risk of renal recurrence. Initial studies all suggested that mutations in CFI carried a poor prognosis, although more recently one study failed to replicate these data. 99 It is likely that this variation in data reflects the functional Table 9. Extrarenal Manifestations Associated With aHUS

Extrarenal manifestation Reference
Digital gangrene [220][221][222] Cerebral artery thrombosis/stenosis [222][223][224] Extracerebral artery stenosis 223 Cardiac involvement/myocardial infarction 51,225 Ocular involvement 226 Pulmonary involvement 51,94 Pancreatic involvement 51 Neurologic involvement 51,94,103 consequences of genetic variants in the different populations. Unlike the complement proteins described earlier, MCP is membrane-tethered and is not a plasma protein. As such, a renal allograft will correct the complement defect and protect against aHUS. In keeping with this, the outcome after transplantation in those with mutations in MCP is much better, with a recurrence rate of only approximately 20%. 128 It has been suggested that in those who do recur, an additional genetic predisposition such as an at-risk haplotype may be present 47 or endothelial microchimerism may occur. 129 In keeping with this, outcomes were worse in individuals with combined MCP mutations compared with patients with an isolated MCP change. 46 Only one individual with a loss-of-function mutation in THBD has undergone transplantation to date and they had recurrent disease. 60 It is perhaps not surprising that individuals with underlying genetic defects have a high recurrence rate because the post-transplant milieu provides the necessary disease triggers (eg, viral diseases, ischemia reperfusion injury, donor-specific antibodies, immunosuppressive drugs, and so forth) to cause endothelial cell damage and activation of the complement cascade. Calcineurin inhibitors (tacrolimus and cyclosporin), although consistently linked as a trigger for aHUS, were not shown to be associated significantly in two recent studies of aHUS recurrence post-transplant. 99,127 Mammalian target of rapamycin inhibitors (eg, sirolimus), however, have been reported to increase the risk of recurrence. 99 Although plasma therapy has a low success rate in rescuing recurrent aHUS after renal transplantation, pre-emptive plasma exchange has been associated with a trend to decrease recurrence. 99 Such a regimen may now be superseded by pre-emptive eculizumab.

AUTOANTIBODIES
There is limited information regarding the outcome after transplantation in individuals with CFH autoantibodies. Two patients have been reported to have CFH autoantibodies and recurrent aHUS. 130,131 Two individuals had successful renal transplants using pre-emptive removal of CFH autoantibodies using rituximab and PE. 131,132 Five individuals with factor H autoantibodies have successfully undergone renal transplantation in the absence of specific therapy aimed at reducing CFH autoantibody titers with follow-up evaluation ranging from 2 to 17 years with no recurrence. 64,67 An additional confounding factor is that CFH autoantibodies frequently are found in association with other mutations. A pragmatic approach would be to use a regimen designed to remove CFH autoantibodies in those with a high titer.

LIVING RELATED TRANSPLANTATION
Live related transplantation is a particularly unattractive option in aHUS given the high recurrence rate and graft loss in the recipient. In addition, de novo aHUS has been recorded in four donors within a year of donation. 31,[133][134][135] In at least one of these cases, a CFH mutation in the donor subsequently was reported. 31 Although genotyping may reveal a known mutation in a family member, the additional presence of risk haplotypes and the fact that further genetic risk factors remain to be discovered makes it impossible to rule out subsequent aHUS for a donor.

LIVER/KIDNEY TRANSPLANTATION
Because CFH, CFI C3, and CFB are synthesized predominantly in the liver; combined liver/kidney transplantation has been viewed as a mechanism to correct the underlying genetic deficiency of complement regulation. Twenty liver/kidney transplants have been reported in the literature to date. 100,127,[136][137][138][139][140][141][142][143][144][145] The initial attempts were not successful because they did not include preoperative PE to correct the underlying complement defects. 136,137 In this setting, ischemiareperfusion injury triggered the complement cascade, resulting in primary liver nonfunction.
Subsequently, Saland et al 144 introduced perioperative PE in addition to aspirin and heparin, which has resulted in improved outcomes. Treatment guidelines for the use of liver-kidney transplantation in aHUS were recently proposed by a consensus conference. 146 Despite this, of the 14 patients who underwent the revised procedure, 2 died (14% surgical mortality), and the availability of eculizumab may change the riskbenefit profile of this type of surgery.

DE NOVO aHUS AFTER RENAL TRANSPLANT
The role of genetic defects in the complement system has also been shown in de novo aHUS after renal transplant. In a series of transplant patients whose initial diagnosis was not aHUS, 29% were seen to have a mutation in CFI or CFH. 147 This may be an underrepresentation of the genetic predisposition because the genes for CFB and C3 were not screened. The majority of these patients had chronic glomerulonephritis or malignant hypertension as their initial diagnosis. Mesangiocapillary glomerulonephritis/C3 glomerulopathy share many of the same complement risk factors as aHUS and the transformation from mesangiocapillary glomerulonephritis to aHUS is well recognized. Likewise, a patient presenting at ESRD with a biopsy picture of malignant hypertension is indistinguishable from aHUS. We would recommend that such patients also should be screened for complement mutations before transplantation.
More recently, a liver transplant recipient was shown to have developed de novo aHUS. Genetic analysis showed that the recipient carried a risk MCP haplotype but did not have a mutation. DNA from the donor liver, however, was shown to carry a mutant CFH. This is further evidence of the role of susceptibility genes in predisposing to post-transplant aHUS. 148

ECULIZUMAB USE IN RENAL TRANSPLANTATION
Eculizumab has also been used successfully in renal transplantation in those experiencing recurrent aHUS in a transplant [149][150][151][152][153][154] and as prophylaxis before transplantation in those with a known mutation. [155][156][157][158] In a recent review, Zuber et al 159 reported that of 9 patients with complement mutations, 8 had successful renal transplants under eculizumab prophylaxis and 13 patients had successful salvage treatment for recurrent aHUS after transplant. In those individuals with ESRD secondary to complement-mediated aHUS requiring renal transplantation, it is probable that prophylactic eculizumab will become the treatment of choice.

SUMMARY
Overactivity of the AP is central to the pathogenesis of aHUS. Many underlying genetic complement predispositions have been described but disease only manifests after an environmental trigger. The understanding of the role of complement in the pathogenesis of aHUS has facilitated the introduction of a successful treatment in the form of the complement C5 inhibitor, eculizumab.