PERSPECTIVES OPINION
Cardiovascular complications in atypical haemolytic uraemic syndrome Marina Noris and Giuseppe Remuzzi Abstract | Haemolytic uraemic syndrome (HUS) is characterized by nonimmune haemolytic anaemia, thrombocytopenia and renal impairment—most incidents in childhood are caused by shiga toxin-producing bacteria. Atypical HUS (aHUS) accounts for 10% of cases and has a poor prognosis. About 60% of patients with aHUS have dysregulation of the alternative complement pathway (complementmediated aHUS). The kidney is the main target organ, but other organs might also be affected. Cardiac complications occur in 3–10% of patients with complementmediated aHUS, as a consequence of microangiopathic injury in the coronary microvasculature, and can cause sudden death. Emerging evidence also suggests that either thrombosis or stenosis of the medium and large arteries might complicate disease course, and such disorders occur even after renal function is lost. In this Perspectives article we discuss the impact of cardiovascular involvement in complement-mediated aHUS, the role of acute and chronic complement hyperactivation in such events and the implications for treatment. Noris, M. & Remuzzi, G. Nat. Rev. Nephrol. advance online publication 14 January 2014; doi:10.1038/nrneph.2013.280
Introduction Haemolytic uraemic syndrome (HUS) is a rare disorder that is characterized by haemolytic anaemia, thrombocytopenia and acute renal failure.1 The lesion that leads to HUS—thrombotic microangiopathy—is characterized by thickening of arterioles and capillaries, endothelial swelling and detachment, thrombosis and obstruction of the vessel lumina.2 Lesions typically affect the kidney, but the brain, heart, lungs, eyes, gastrointestinal tract, liver and pancreas might also be involved.1,3 Around 90% of childhood HUS cases are caused by E. coli strains that produce shiga-like toxins, which can cause thrombotic microangiopathy. 1 HUS that is not caused by shiga-like toxins and 60% of cases classified as atypical HUS (aHUS)1 are the result of defective alternative complement pathway regulation due to genetic or acquired abnormalities, which lead to endothelial damage and microvascular thrombosis. 1 Other causes of aHUS have been described, including Competing interests The authors declare no competing interests.
non-enteric bacterial and viral infections, drugs, malignant hypertension, transplantation, and pregnancy—these forms are frequently known as secondary aHUS. 1 However, secondary aHUS does not explain how many of the above conditions can trigger the disease in patients with a genetic background that leads to complement dysregulation.4 Another form of thrombotic micro angiopathy, thrombotic thrombocytopenic purpura, 5 results from a severe defici ency of ADAMTS13, a plasma metallo protease that cleaves von Willebrand factor multimers—leading to widespread hyaline thrombosis that affects the small vessels.4,6,7 Thrombocytopenic purpura lesions mainly localize in the brain,5 but the heart is also affected and cardiac injury is recognized as a leading cause of death in patients.8,9 Cardiac events do occur in patients with aHUS, although more rarely than in those with thrombocytopenic purpura, and might cause sudden death. aHUS is a rare condition and most available data on cardiac involvement in aHUS derive from case reports, 10–15 with the exception of a few series. 4,16,17 This Perspectives article
NATURE REVIEWS | NEPHROLOGY
discusses the emerging evidence of cardio vascular involvement in complementmediated aHUS and outlines the role of complement in such complications.
Pathogenetic mechanisms The complement system is part of the innate immune system, which helps protect individuals against invading organisms and maintains tissue homeostasis (Figure 1). The complement system functions via three activation pathways—classical, lectin and alternative pathways.18 The classical and lectin pathways rely on recognition of pathogens or damaged cell surfaces by anti bodies and pattern-recognition molecules.19 Both classical and lectin pathways lead to cleavage of C4, consequently forming the C4bC2a convertase complex on target surfaces, which cleaves C3 into C3a and C3b. The alternative pathway is spontan eously initiated in plasma by C3 hydrolysis responsible for covalent deposition of a low amount of C3b onto the majority of cellular surfaces exposed to plasma. On bacterial cell surfaces, C3b binds leucocyte receptors, which subs equently undergo phagocytosis. In addition, C3b, together with com plement factor B (CFB), forms the C3bBb convertase of the alternative pathway that cleaves additional C3 molecules. 20 When C3b associates with the C3 convertase, the complex becomes a C5 convertase that cleaves factor C5 into C5a and C5b. C5a and C3a can initiate phagocyte migration and activate endothelial cells.21 C5b forms a complex with complement C6, C7, C8 and C9, which culminates in the assembly on target surfaces of the C5b‑9 proteolytic membrane attack complex.22 Without regulation, a small initiating stimulus is quickly amplified; on mammalian cells this cascade is controlled by membrane-anchored and fluid-phase regulators, such as complement factor H (CFH), membrane cofactor protein and thrombomodulin (Figure 1). Such regulators favour the cleavage of C3b to inactive iC3b by the plasma serine- protease factor I (CFI, cofactor activity) and dissociate the C3 and C5-convertases, which halts the complement cascade. 23 About 60% of aHUS is attributed to abnormalities in components of the alternative complement pathway.1 ADVANCE ONLINE PUBLICATION | 1
© 2014 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES Membrane attack complex C5b-9 Target cell lysis
Chemotaxis and activation of inflammatory cells Endothelial activation C3a C5a
Classical pathway (antigen–antibody interaction)
Bb
C5b
Bb
C3b Alternative pathway (spontaneous C3 tick-over)
C3 convertase
Lectin pathway (mannose-binding lectin–pathogen interaction)
C5 convertase Pathogen/altered host surfaces THBD
Bb
FH
MCP CFI
iC3b
Bb
Healthy host surfaces
Figure 1 | Pathways of complement activation and regulation. All characterized pathways of complement activation cleave the central component C3 into C3a and C3b. C3b deposits on surfaces and together with factor‑D activated factor B–Bb, forms the C3 convertase C3bBb of the alternative pathway forming more C3a and C3b. C3a is a potent inflammatory mediator, whereas C3b opsonizes cell surfaces and contributes to assembly of the C5 convertase C3bBb3b, which subsequently cleaves C5 into C5a and C5b. The latter deposits on surfaces together with late complement components C6–C9, leading to cell lysis through the membrane attack complex. On normal host surfaces, complement activity is inhibited by an array of physiological regulators. The serine protease CFI cleaves C3b into an inactive form (iC3b), in the presence of CFH, and membrane-bound complement regulators (like MCP and THBD). CFH also directly accelerates dissociation of the C3 convertase C3bBb. Mutations in all these complement regulators have been detected in aHUS, predisposing the carrier to uncontrolled, destructive activation of complement alternative pathways on host cells. Abbreviations: CFH, complement factor H; CFI, complement factor I; MCP membrane cofactor protein; THBD, thrombomodulin.
Complement factor H CFH regulates the alternative complement pathway by acting as a cofactor for CFI and enhancing dissociation of C3 convertase (Figure 1). More than 100 different CFH mutations24,25 have been identified in patients with aHUS, which equates to a mutation frequency of ~30%.26–28 These mutations most commonly lead to normal levels of a protein that is unable to bind and regulate complement components on endothelial cells.29 A high degree of sequence identity between the CFH gene and the CFHR1– CFHR5 genes that encode five complement factor H‑related proteins (CFHR) located in tandem to CFH might predispose a patient to nonallelic recombinations.11 In 3–5% of patients with aHUS, a heterozygous hybrid gene, derived from an uneven crossover and containing the first 21 CFH exons and the last two exons of CFHR1, produces a gene product with decreased complement regulatory activity on endothelial surfaces.11 Other CFH–CFHR hybrid genes have been described in aHUS, such as a CFH–CFHR3 hybrid gene consisting of the first 22 CFH exons and exons 2–6 of CFHR3.30,31
Acquired immunity defects (in the form of inhibitory CFH antibodies) are also reported in 5–10% of all patients and 25–50% of paediatric patients with aHUS. 32–34 Analogous to the CFH gene mutations, these autoantibodies predominantly target the C‑terminus of CFH and disrupt complement regulation on host cell surfaces. 35 The development of CFH autoantibodies is strongly associated with a deletion of CFHR1 and CFHR3.36
Membrane cofactor protein Membrane cofactor protein (MCP) is a transmembrane complement regulator that is widely expressed on the surface of cells and is a cofactor of CFI (Figure 1). Mutations in MCP account for 10–15% of all aHUS,27,37,38 and about 90% of mutations are heterozygous.25 Approximately 75% of patients have mutations resulting in a quantitative defect and reduced expression of MCP on blood leucocytes, proportional to the level of MCP reduction.27 The remaining 25% of MCP mutations lead to a MCP protein with low C3b-binding capability and decreased cofactor activity.27
2 | ADVANCE ONLINE PUBLICATION
Complement factor I Heterozygous CFI mutations affect 4–10% of patients with aHUS,27,39 and 80% of these cluster in the serine-protease domain. Approximately 50% of mutations cause absolute low levels of CFI protein, some mutated forms are secreted but have impaired proteolytic activity.27,39 Complement factor B and C3 Gain-of-function mutations can affect genes encoding complement factor B (CFB) and C3—the alternative pathway C3 convertase components.20 In patients with aHUS, CFB mutations are rare (1–2%).40 Mutated CFB have excess affinity for C3b and form a hyperactive C3 convertase complex that is resistant to dissociation.40 About 10% of patients with aHUS have heterozygous mutations in C3,41 most of which reduce C3b binding to comp lement regulators, thereby severely impairing degradation of mutant C3b.41 Thrombomodulin Heterozygous mutations in the gene that encodes thrombomodulin (THBD) have been found in 3–4% of patients with aHUS. 42,43 THBD is a membrane-b ound glycoprotein with anticoagulant properties that modulates complement activation on cell surfaces. Cells expressing mutated THBD inactivate C3b less efficiently than cells expressing wild-type thrombomodulin.42 Diacylglycerol kinase ε Diacylglycerol kinase ε (encoded by DGKE) is expressed in the endothelium, platelets and podocytes. DGKE is a lipid kinase that catalyses the phosphorylation of diacyl glycerol substrates (DAGs) to phosphatidic acid. Loss of DGKE function results in enhanced signalling through arachidonic acid-containing DAGs and increased activation of protein kinase C (PKC).44 DGKE is apparently unrelated to complement and the mechanism by which DGKE mutations cause aHUS is unknown. However, both homozygous and compound heterozygous mutations in the DGKE gene co-segregated with aHUS in nine unrelated kindreds with autosomal recessive inheritance.45 Patients with mutations in DGKE developed aHUS before their first birthday and had persistent hypertension, haematuria and proteinuria.45 Disease penetrance determinants Two further factors are thought to determine the development of aHUS. Firstly, in most patients, aHUS is triggered by an unrelated factor, most frequently an www.nature.com/nrneph
© 2014 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES infection or pregnancy.4 Secondly, a further genetic variant (modifier) can increase the disease risk. Approximately 5% of patients have mutations in more than one aHUS risk gene.46 Common genetic variants (single nucleotide polymorphisms [SNPs] or haplo type blocks) in CFH, MCP and CFHR1 are all aHUS susceptibility factors.46
Clinical course and treatment aHUS has previously been associated with poor prognosis and elevated mortality, despite intensive treatment regimens with plasma therapy and life support measures. 1,2 More than 50% of patients with aHUS die of their disease, require dialysis or develop end-stage renal disease (ESRD) during the first year after diagnosis.1,3 Improved characterization of the genetic components of aHUS has led to the conclusion that the endothelial damage caused by the complement system is the critical factor for renal microangiopathic injury, which has led to the development of complement-tailored treatments. Several case reports and two phase II trials47,48 have documented the efficacy and safety of eculizumab, a monoclonal antibody that binds to human complement C5. Eculizumab can induce remission of haematologic and renal abnormalities in aHUS, which led to the drug’s regulatory approval in the USA and Europe.49
Cardiovascular complications Acute and chronic cardiovascular events have been reported in 3–10% of patients with adult-onset or paediatric-onset complement-related aHUS. 4,50 The risk of cardiovascular event varies according to the molecular defect. Patients with CFH mutations, anti-CFH antibodies, gain of function C3 or CFB mutations are particularly susceptible to cardiovascular disorders (Table 1).4
Heart injury Myocardial infarction, cardiomyopathy and heart failure have all been reported in patients with aHUS4,10 and cardiac complications seem to be linked to severe disease. 51 In a study reported in 1997, 16 cardiomyopathy—defined as heart failure requiring the use of inotropes—was recorded during the acute phase of aHUS in 10 out of 23 children (43%) and two patients died within 3 months of onset. Unfortunately, the genetic status of the individuals was unknown because genetic screening was not available at that time.
Table 1 | Studies describing cardiovascular and macrovascular complications in aHUS Reference
Study outline
Findings
Noris et al. (2010)4
273 patients (adults and children)
Seven patients (five with CFH mutations) had cardiovascular disease during acute HUS episodes
Sallee et al. (2010)10
Single adult with aHUS and CFH mutation
Patient died of myocardial infarction 15 days after HUS onset
Venables et al. (2006)11
Eight patients with familial aHUS and a CFH–CFHR1 hybrid gene
One patient died of myocardial infarction 10 years after HUS onset, another died of cardiac arrest 8 weeks after onset
Sellier-Leclerc et al. (2007)50
46 children with aHUS
Four children (one with CFH mutation) had cerebrovascular events at HUS onset
Neuhaus et al. (1997)16
23 children with aHUS (no genotyping)
10 children had cardiomyopathy at discharge, two died
Vilalta et al. (2012)12
Single child with aHUS and CFH mutation
Patient had dilated cardiomyopathy at HUS onset, and myocardial dysfunction during follow-up monitoring Eculizumab improved cardiac function
Dragon-Durey et al. (2010)17
45 patients with aHUS and anti-CFH autoantibodies
Three patients developed cardiac insufficiency on long-term, one died
AbarrateguiGarrido et al. (2009)52
Seven children with aHUS and anti-CFH autoantibodies
One child died of myocarditis
Roumenina et al. (2012)53
14 patients with a C3 mutation
Seven patients had dilated cardiomyopathy (during acute HUS in five, delayed in two patients) One died at onset following a cardiovascular event
Le Quintrec et al. (2013)54
57 adults with aHUS and kidney transplant
Four patients died following cardiovascular events during follow-up No genotype data
Malina et al. (2013)13
Two children with aHUS (one with anti-CFH autoantibodies and one with a C3 mutation)
Both patients had ischaemic changes and gangrene in fingers and toes during acute HUS In one patient, eculizumab induced remission of HUS and of ischaemic changes
Ozel et al. (2003)55
Single child with aHUS and low C3 levels
Patient had ischaemic necrosis of distal phalanges during acute HUS episode No genetic screening
Kaplan et al. (2000)56
Three children with aHUS; Low C3 levels in two patients
The three children had necrotic fingers and toes at HUS onset followed by auto-amputation One child also had occlusion of carotid arteries 90 days after onset No genetic screening
Davin et al. (2010)14 Vergouwen et al. (2008)60
An adolescent with aHUS and a CFH mutation, on dialysis
Patient had stenosis in cerebral arteries, which manifested brain infarcts during kidney transplant No event occurred during post-transplant follow-up under eculizumab
Loirat et al. (2010)15
Single child with aHUS and a CFB mutation, on dialysis
Patient had emiparesis and stenoses of carotid, cerebral, left subclavian, vertebral and pulmonary arteries
Bekassy et al. (2013)61
Single child with aHUS and combined CFB and CFI mutations, on dialysis
Patient had occlusion of carotid arteries, and received a kidney transplant and eculizumab Uneventful follow-up and no progression of vascular occlusion
Ažukaitis et al. (2013)71
Single child with aHUS and a C3 mutation
Patient developed cerebral artery stenoses leading to death due to stroke 9 days after transplant under eculizumab
Abbreviations: aHUS, atypical HUS; CFB, complement factor B; CFH, complement factor H; CFI, complement factor I; HUS, haemolytic uraemic syndrome.
Patients with genetic or acquired CFH defects seem to be particularly at risk of cardiac complications, which mostly occur during, or close to, disease onset. Anti-CFH autoantibody-associated aHUS has the highest frequency of cardiovascular events (~10% of all cases). In two independent
NATURE REVIEWS | NEPHROLOGY
studies, cardiac failure was reported in three (two adults, one child) out of 45 patients17 and in one of seven children with anti-CFH autoantibodies.52 In a series of 49 patients with mutations in CFH, five (7.7%) had cardiovascular disease at aHUS onset.4 A 43-year-old woman with ADVANCE ONLINE PUBLICATION | 3
© 2014 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES a Cys623Ser CFH mutation died of sudden cardiac arrest 15 days after aHUS onset.10 An autopsy revealed the patient had a myocardial infarction with no obstruction of the coronary arteries. However, all small coronary vessels had features of thrombotic microangiopathy with endothelial swelling, vessel wall thickening and intense antibody staining for C5b‑9 was observed in the small coronary vessels and infarcted cardiomyocytes.10 In this patient, myocardial tissue was, therefore, affected by the same anomalies of microvascular circulation as those observed in the kidney. A second patient, a 19-year-old woman with aHUS associated with a CFH– CFHR1 hybrid gene, died from cardiac arrest 8 weeks after onset of the disease.11 However, no details regarding the nature and cause of cardiac insufficiency were reported. Another study described a 4-year-old girl with aHUS owing to an Asp1119Asn mutation in CFH who presented with acute kidney injury, dilated cardiomyopathy and cardiorespira tory arrest.12 Intensive plasma-exchange was unable to halt disease progression or improve cardiac dysfunction and improvements in renal, haematological and cardiac values were only achieved after chronic treatment with eculizumab.12 Taken together, these studies indicate that the heart is affected by CFH deficiency. Gain-of-function mutations in the C3 gene also seem to predispose the carrier to cardiovascular complications. Among 14 patients with sporadic aHUS who carry an Arg139Trp mutation in C3, eight patients (60%) experienced cardiac events.53 Dilated cardiomyopathy, with reduction in left ventricular function, was recorded in seven patients, whereas the type of myocardial lesion was not determined in one patient. Heart failure occurred at aHUS onset in six patients, but two experienced a cardio myopathy 2 months and 6 months after haematological remission, respectively.53 For other forms of aHUS, the frequency of cardiovascular events has yet to be assessed.
Post-transplant cardiovascular events In some patients with complement-related aHUS who have undergone kidney transplantation, chronic complement activation combined with the adverse-effects of immunosuppressive drugs might harm the cardiovascular system. In 57 adult patients with aHUS who received a kidney transplant, four (7%) aged 21–50 years and without major cardiovascular risk factors died after cardiovascular events (acute cardiac failure, n = 1; myocardial infarction, n = 2;
and extensive stroke, n = 1).54 Three patients in this study died with a functioning graft and one died of myocardial infarction after restarting dialysis.
Small peripheral artery thrombosis In some children with complement-related aHUS, severe perfusion defects of the arteries supplying the fingers and toes led to gangrenous lesions of distal phalanges, which developed either during the acute phase of the disease or weeks after onset. A 4-year-old girl with anti-CFH autoantibodies developed gangrene of the finger tips 2 days after initial presentation of aHUS, and died 3 weeks later because of dialysis-related complications.13 The same report describes a second girl who developed ESRD at 4 months of age owing to aHUS associated with a C3 mutation. At 9 months of age, she developed ischaemia in her fingers and toes, and several finger tips became gangrenous despite intensive plasma-exchange therapy. However, after administration of eculizumab, all nonnecrotic digits rapidly regained perfusion and disease activity subsided completely.13 These observations add to previous reports in four children who developed peripheral gangrene of fingers and toes as a complication of aHUS.55,56 Notably, all patients with perfusion problems in the peripheral arteries were children with very severe forms of aHUS. Considering that the luminal diameter of the arteries supplying phalanges is 100-fold larger than that of the glomerular arterioles typically affected in HUS,57,58 the findings from these patients suggest that under certain conditions HUSspecific thrombotic lesions can also affect small and medium arteries. Indeed, one paediatric patient with aHUS and peripheral gangrene began to have seizures 90 days after disease onset.56 MRI studies showed thrombotic occlusion of both internal carotid arteries and patent vertebral arteries.56 Low levels of C3 were documented in all four patients with peripheral gangrene in which complement was measured.13,55,56 In two patients, low C3 levels were associated with anti-CFH autoantibodies and a C3 mutation respectively, and indicate a role for uncontrolled activation of the alternative complement pathway in vascular damage. Large artery stenosis Under normal circumstances, haematological recurrences of aHUS cease when renal function is lost.1 This clinical observation has led to the theory that viable renal tissue is required for overt aHUS relapses.
4 | ADVANCE ONLINE PUBLICATION
Increased shear stress and platelet activation in the damaged renal microvasculature might sustain the microangiopathic process. For example, complete remission of aHUS was observed after bilateral nephrectomy in patients with severe renal involvement, refractory hypertension and signs of hypertensive encephalopathy.59 However, vascular lesions might develop in patients with ESRD or in whom renal tissue was removed. Specifically, stenosis of large arteries has been described in children with complement-related aHUS on long-term haemodialysis. A 15-year-old girl with a Ser1191Leu mutation in CFH, who was on long-term dialysis after loss of two renal grafts, began to experience intradialytic neurologic symptoms (both sensory and motor).14 Angiography revealed severe stenosis of the middle and anterior cerebral arteries,60 and radiological evaluation revealed no signs of calcifications or other loci of arterial wall narrowing.60 The patient received a third kidney transplant 2 years later, but postoperatively developed infarcts in the right frontal and frontoparietal regions. Owing to the potential cerebrovascular risk, and to prevent disease relapse in the graft, she received eculizumab therapy and had no major clinical events during the 6-month follow-up period.14 A second child with neonatal onset of aHUS owing to a Lys350Asp mutation in CFB developed progressive stenosis of multiple large arteries (including carotid arteries, left subclavian and vertebral arteries, intracranial, right humeral, coronary and all pulmonary arteries) after several years of dialysis.15 Unfortunately, a carotid siphon angioplasty was complicated by dissection leading to death. Another child who presented with aHUS at 17 months of age developed ESRD within 3 months. 61 She underwent a deceaseddonor kidney transplantation, but developed recurrence of aHUS in the graft with severe hypertensive crisis that led to removal of the transplant. At 10 years of age, and after 7 years of chronic haemodialysis, the patient developed a transient ischaemic attack but without the haematological symptoms of aHUS. Imaging revealed that the carotid arteries were occluded and had undergone stenosis. Two mutations, one in CFB and one in CFI gene, were identified and measurements of C3 and its degradation product indicated that complement activation was continuous.61 The study’s authors argued that advanced occlusion and stenosis in the carotid arteries were caused by ongoing complement-induced vascular injury www.nature.com/nrneph
© 2014 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES and treatment with eculizumab was initiated despite no evidence of aHUS activity. Eculizumab treatment blocked progression of vascular injury and the patient underwent a successful second kidney transplant. During 18 months follow-up monitoring, no recurrence of aHUS occurred and vascular stenosis did not progress.61 These cases suggest that progression of extrarenal vascular lesions can occur even in the absence of overt recurrences, a process that might be exacerbated by long-term haemodialysis.
Mechanisms Mutations in CFH or anti-CFH auto antibodies that cause complement-related aHUS disrupt the CFH surface recognition activity that underlines complement regulation on endothelial cells.19 Similarly, aHUS associated gain-of-function mutations in CFB or C3 induce a hyperactive C3 convertase with C3b deposition on endothelial cells, enhanced formation of C5b‑9 complexes and deposition of C3 fragments at cell surfaces.40,41 Complement activation disrupts the physiologically thromboresistant phenotype of endothelial cells, which might contribute to thrombotic occlusions in renal and extrarenal vasculature, including coronary arterioles and capillaries, and lead to cardiac dysfunction and myocardial infarction.62 This outcome is mediated by C3a and C5a binding to their cognate receptors and C5b‑9 deposition on cell surfaces. Endothelial damage caused by recurrent or continuous complement activation in patients with genetic complement disorders might not be limited to arterioles and capillaries, although thrombus formation might rarely reach a critical size to cause acute obstruction. The cases of children with peripheral gangrene of the finger and toes owing to aHUS, accompanied by bilateral carotid artery thromboses,56 indicate that macrovascular thrombosis might occur in severe cases of complement-related aHUS. Complement activation has been implicated in vascular stenosis, 63,64 but the mechanisms involved have not been fully elucidated. Complement components exert proinflammatory effects on the endo thelium and vessel wall by upregulating expression of adhesion molecules 65 and thereby release prostanoids, leukotrienes and cytokines that enhance leucocyte recruitment, activation and transendo thelial migration.62,65 Complement components, C3a and C5a act as stimuli for neutrophil, monocyte and macrophage
Defective regulation of the alternative pathway of complement Continuous complement activation C3a, C5a, C5b-9 Endothelial injury
Endothelial detachment Leukocyte transmigration Reactive cell proliferation
Thrombosis
Large artery stenosis
Ischaemia Peripheral arteries Gangrene of fingers and toes
Ischaemia Coronary microvessels
Heart injury and dysfunction
Multiorgan failure
Figure 2 | Potential pathophysiological mechanisms of cardiovascular complications in aHUS. In patients with aHUS, uncontrolled complement activation leads to endothelial injury and loss of antithrombogenic properties. This event might cause thrombosis of coronary arterioles and capillaries leading to ischaemic heart injury and dysfunction, or of small and medium peripheral arteries, leading to ischaemia. Endothelial injury might also initiate stenosis of large arteries in the brain, lung, heart, gut and other organs, due to endothelial cell swelling, detachment, inflammatory leucocyte infiltration, subendothelial expansion and reactive smooth muscle cell proliferation. Thrombus formation on injured endothelium also contributes to the stenotic process leading to multiorgan failure. Abbreviation: aHUS, atypical haemolytic uraemic syndrome.
migration.66 All these events culminate in accumulation of primed leucocytes at sites of complement activation, which lead to vessel wall inflammation and contraction due to the release of cytokines, reactive oxygen radicals and vasoactive medi ators. In addition, C5b‑9 stimulates platelets, macrophages and endothelial cells to secrete platelet-derived growth factor-BB, a major growth factor for vascular smooth muscle cells, which might lead to vessel wall t hickening and stenosis.67 Multiple stenoses of large arteries observed in children on chronic dialysis, and who have genetic complement dysregulation, indicate that chronic subclinical complement-mediated endothelial damage might indeed lead to stenosis of the arteries. Complement activation occurs both in human and experimental atherosclerosis and C1q, C3 and C5b-9 deposits have been found in atherosclerotic vessel lesions.68 In addition, children with renal failure have risks factors for atherosclerosis, and cardiovascular disease accounts for most deaths in young adults with childhood-onset chronic renal failure.69 However, in children with complement-related aHUS and stenosing arteriopathy, none had left ventricular hypertrophy or dysfunction, as observed in systemic atherosclerosis in patients on dialysis, and arterial abnormalities did not include calcifications.14,15,61 The authors hypothesized that a local complement attack
NATURE REVIEWS | NEPHROLOGY
rather than atherosclerosis, might be the main factor causing large artery damage.15 Complement-induced endothelial injury might result in stenosis due to endothelial cell swelling and detachment, followed by inflammatory leucocyte recruitment, subendothelial expansion and reactive smooth-muscle cell proliferation (Figure 2). Notably, in all reported case studies, those patients who developed vascular stenosis were receiving haemodialysis. Complement can be activated during haemodialysis, as demonstrated by the generation of C3a and C5a after exposure of blood to haemo dialysis filter.70 Such dialysis-related complement activation might be exacerbated in patients with complement-related aHUS and genetically determined complement dysregulation, which might contribute to advanced v ascular injury.
Conclusions In patients with complement-related aHUS, cardiovascular complications might occur acutely at disease onset and during relapses. These events affect cardiac microvasculature and peripheral arteries. In addition stenotic lesions might develop chronically due to continuous complementmediated inflammatory response, damaging endothelial cells and subendothelial vessel structures. Extrarenal vascular lesions progress even during symptom-free intervals, a process ADVANCE ONLINE PUBLICATION | 5
© 2014 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES that might be exacerbated by long-term haemodialysis. For this reason kidney transplant with eculizumab treatment should be considered as early as possible in patients with ESRD. Patients with CFH mutations, anti-CFH autoantibodies, or with CFB or C3 mutations, seem to be particularly at risk of developing cardiovascular complications. In such patients, and particularly those on chronic haemodialysis, echocardiographic screening, troponin monitoring and imaging for vascular changes should be performed at admission and during follow-up care. Eculizumab treatment during periods of haemodialysis might prevent serious vascular damage in complement-related aHUS patients with angiographic signs of vascular lesions.61 Why
Cardiovascular complications in atypical haemolytic uraemic syndrome Marina Noris and Giuseppe Remuzzi Abstract | Haemolytic uraemic syndrome (HUS) is characterized by nonimmune haemolytic anaemia, thrombocytopenia and renal impairment—most incidents in childhood are caused by shiga toxin-producing bacteria. Atypical HUS (aHUS) accounts for 10% of cases and has a poor prognosis. About 60% of patients with aHUS have dysregulation of the alternative complement pathway (complementmediated aHUS). The kidney is the main target organ, but other organs might also be affected. Cardiac complications occur in 3–10% of patients with complementmediated aHUS, as a consequence of microangiopathic injury in the coronary microvasculature, and can cause sudden death. Emerging evidence also suggests that either thrombosis or stenosis of the medium and large arteries might complicate disease course, and such disorders occur even after renal function is lost. In this Perspectives article we discuss the impact of cardiovascular involvement in complement-mediated aHUS, the role of acute and chronic complement hyperactivation in such events and the implications for treatment. Noris, M. & Remuzzi, G. Nat. Rev. Nephrol. advance online publication 14 January 2014; doi:10.1038/nrneph.2013.280
Introduction Haemolytic uraemic syndrome (HUS) is a rare disorder that is characterized by haemolytic anaemia, thrombocytopenia and acute renal failure.1 The lesion that leads to HUS—thrombotic microangiopathy—is characterized by thickening of arterioles and capillaries, endothelial swelling and detachment, thrombosis and obstruction of the vessel lumina.2 Lesions typically affect the kidney, but the brain, heart, lungs, eyes, gastrointestinal tract, liver and pancreas might also be involved.1,3 Around 90% of childhood HUS cases are caused by E. coli strains that produce shiga-like toxins, which can cause thrombotic microangiopathy. 1 HUS that is not caused by shiga-like toxins and 60% of cases classified as atypical HUS (aHUS)1 are the result of defective alternative complement pathway regulation due to genetic or acquired abnormalities, which lead to endothelial damage and microvascular thrombosis. 1 Other causes of aHUS have been described, including Competing interests The authors declare no competing interests.
non-enteric bacterial and viral infections, drugs, malignant hypertension, transplantation, and pregnancy—these forms are frequently known as secondary aHUS. 1 However, secondary aHUS does not explain how many of the above conditions can trigger the disease in patients with a genetic background that leads to complement dysregulation.4 Another form of thrombotic micro angiopathy, thrombotic thrombocytopenic purpura, 5 results from a severe defici ency of ADAMTS13, a plasma metallo protease that cleaves von Willebrand factor multimers—leading to widespread hyaline thrombosis that affects the small vessels.4,6,7 Thrombocytopenic purpura lesions mainly localize in the brain,5 but the heart is also affected and cardiac injury is recognized as a leading cause of death in patients.8,9 Cardiac events do occur in patients with aHUS, although more rarely than in those with thrombocytopenic purpura, and might cause sudden death. aHUS is a rare condition and most available data on cardiac involvement in aHUS derive from case reports, 10–15 with the exception of a few series. 4,16,17 This Perspectives article
NATURE REVIEWS | NEPHROLOGY
discusses the emerging evidence of cardio vascular involvement in complementmediated aHUS and outlines the role of complement in such complications.
Pathogenetic mechanisms The complement system is part of the innate immune system, which helps protect individuals against invading organisms and maintains tissue homeostasis (Figure 1). The complement system functions via three activation pathways—classical, lectin and alternative pathways.18 The classical and lectin pathways rely on recognition of pathogens or damaged cell surfaces by anti bodies and pattern-recognition molecules.19 Both classical and lectin pathways lead to cleavage of C4, consequently forming the C4bC2a convertase complex on target surfaces, which cleaves C3 into C3a and C3b. The alternative pathway is spontan eously initiated in plasma by C3 hydrolysis responsible for covalent deposition of a low amount of C3b onto the majority of cellular surfaces exposed to plasma. On bacterial cell surfaces, C3b binds leucocyte receptors, which subs equently undergo phagocytosis. In addition, C3b, together with com plement factor B (CFB), forms the C3bBb convertase of the alternative pathway that cleaves additional C3 molecules. 20 When C3b associates with the C3 convertase, the complex becomes a C5 convertase that cleaves factor C5 into C5a and C5b. C5a and C3a can initiate phagocyte migration and activate endothelial cells.21 C5b forms a complex with complement C6, C7, C8 and C9, which culminates in the assembly on target surfaces of the C5b‑9 proteolytic membrane attack complex.22 Without regulation, a small initiating stimulus is quickly amplified; on mammalian cells this cascade is controlled by membrane-anchored and fluid-phase regulators, such as complement factor H (CFH), membrane cofactor protein and thrombomodulin (Figure 1). Such regulators favour the cleavage of C3b to inactive iC3b by the plasma serine- protease factor I (CFI, cofactor activity) and dissociate the C3 and C5-convertases, which halts the complement cascade. 23 About 60% of aHUS is attributed to abnormalities in components of the alternative complement pathway.1 ADVANCE ONLINE PUBLICATION | 1
© 2014 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES Membrane attack complex C5b-9 Target cell lysis
Chemotaxis and activation of inflammatory cells Endothelial activation C3a C5a
Classical pathway (antigen–antibody interaction)
Bb
C5b
Bb
C3b Alternative pathway (spontaneous C3 tick-over)
C3 convertase
Lectin pathway (mannose-binding lectin–pathogen interaction)
C5 convertase Pathogen/altered host surfaces THBD
Bb
FH
MCP CFI
iC3b
Bb
Healthy host surfaces
Figure 1 | Pathways of complement activation and regulation. All characterized pathways of complement activation cleave the central component C3 into C3a and C3b. C3b deposits on surfaces and together with factor‑D activated factor B–Bb, forms the C3 convertase C3bBb of the alternative pathway forming more C3a and C3b. C3a is a potent inflammatory mediator, whereas C3b opsonizes cell surfaces and contributes to assembly of the C5 convertase C3bBb3b, which subsequently cleaves C5 into C5a and C5b. The latter deposits on surfaces together with late complement components C6–C9, leading to cell lysis through the membrane attack complex. On normal host surfaces, complement activity is inhibited by an array of physiological regulators. The serine protease CFI cleaves C3b into an inactive form (iC3b), in the presence of CFH, and membrane-bound complement regulators (like MCP and THBD). CFH also directly accelerates dissociation of the C3 convertase C3bBb. Mutations in all these complement regulators have been detected in aHUS, predisposing the carrier to uncontrolled, destructive activation of complement alternative pathways on host cells. Abbreviations: CFH, complement factor H; CFI, complement factor I; MCP membrane cofactor protein; THBD, thrombomodulin.
Complement factor H CFH regulates the alternative complement pathway by acting as a cofactor for CFI and enhancing dissociation of C3 convertase (Figure 1). More than 100 different CFH mutations24,25 have been identified in patients with aHUS, which equates to a mutation frequency of ~30%.26–28 These mutations most commonly lead to normal levels of a protein that is unable to bind and regulate complement components on endothelial cells.29 A high degree of sequence identity between the CFH gene and the CFHR1– CFHR5 genes that encode five complement factor H‑related proteins (CFHR) located in tandem to CFH might predispose a patient to nonallelic recombinations.11 In 3–5% of patients with aHUS, a heterozygous hybrid gene, derived from an uneven crossover and containing the first 21 CFH exons and the last two exons of CFHR1, produces a gene product with decreased complement regulatory activity on endothelial surfaces.11 Other CFH–CFHR hybrid genes have been described in aHUS, such as a CFH–CFHR3 hybrid gene consisting of the first 22 CFH exons and exons 2–6 of CFHR3.30,31
Acquired immunity defects (in the form of inhibitory CFH antibodies) are also reported in 5–10% of all patients and 25–50% of paediatric patients with aHUS. 32–34 Analogous to the CFH gene mutations, these autoantibodies predominantly target the C‑terminus of CFH and disrupt complement regulation on host cell surfaces. 35 The development of CFH autoantibodies is strongly associated with a deletion of CFHR1 and CFHR3.36
Membrane cofactor protein Membrane cofactor protein (MCP) is a transmembrane complement regulator that is widely expressed on the surface of cells and is a cofactor of CFI (Figure 1). Mutations in MCP account for 10–15% of all aHUS,27,37,38 and about 90% of mutations are heterozygous.25 Approximately 75% of patients have mutations resulting in a quantitative defect and reduced expression of MCP on blood leucocytes, proportional to the level of MCP reduction.27 The remaining 25% of MCP mutations lead to a MCP protein with low C3b-binding capability and decreased cofactor activity.27
2 | ADVANCE ONLINE PUBLICATION
Complement factor I Heterozygous CFI mutations affect 4–10% of patients with aHUS,27,39 and 80% of these cluster in the serine-protease domain. Approximately 50% of mutations cause absolute low levels of CFI protein, some mutated forms are secreted but have impaired proteolytic activity.27,39 Complement factor B and C3 Gain-of-function mutations can affect genes encoding complement factor B (CFB) and C3—the alternative pathway C3 convertase components.20 In patients with aHUS, CFB mutations are rare (1–2%).40 Mutated CFB have excess affinity for C3b and form a hyperactive C3 convertase complex that is resistant to dissociation.40 About 10% of patients with aHUS have heterozygous mutations in C3,41 most of which reduce C3b binding to comp lement regulators, thereby severely impairing degradation of mutant C3b.41 Thrombomodulin Heterozygous mutations in the gene that encodes thrombomodulin (THBD) have been found in 3–4% of patients with aHUS. 42,43 THBD is a membrane-b ound glycoprotein with anticoagulant properties that modulates complement activation on cell surfaces. Cells expressing mutated THBD inactivate C3b less efficiently than cells expressing wild-type thrombomodulin.42 Diacylglycerol kinase ε Diacylglycerol kinase ε (encoded by DGKE) is expressed in the endothelium, platelets and podocytes. DGKE is a lipid kinase that catalyses the phosphorylation of diacyl glycerol substrates (DAGs) to phosphatidic acid. Loss of DGKE function results in enhanced signalling through arachidonic acid-containing DAGs and increased activation of protein kinase C (PKC).44 DGKE is apparently unrelated to complement and the mechanism by which DGKE mutations cause aHUS is unknown. However, both homozygous and compound heterozygous mutations in the DGKE gene co-segregated with aHUS in nine unrelated kindreds with autosomal recessive inheritance.45 Patients with mutations in DGKE developed aHUS before their first birthday and had persistent hypertension, haematuria and proteinuria.45 Disease penetrance determinants Two further factors are thought to determine the development of aHUS. Firstly, in most patients, aHUS is triggered by an unrelated factor, most frequently an www.nature.com/nrneph
© 2014 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES infection or pregnancy.4 Secondly, a further genetic variant (modifier) can increase the disease risk. Approximately 5% of patients have mutations in more than one aHUS risk gene.46 Common genetic variants (single nucleotide polymorphisms [SNPs] or haplo type blocks) in CFH, MCP and CFHR1 are all aHUS susceptibility factors.46
Clinical course and treatment aHUS has previously been associated with poor prognosis and elevated mortality, despite intensive treatment regimens with plasma therapy and life support measures. 1,2 More than 50% of patients with aHUS die of their disease, require dialysis or develop end-stage renal disease (ESRD) during the first year after diagnosis.1,3 Improved characterization of the genetic components of aHUS has led to the conclusion that the endothelial damage caused by the complement system is the critical factor for renal microangiopathic injury, which has led to the development of complement-tailored treatments. Several case reports and two phase II trials47,48 have documented the efficacy and safety of eculizumab, a monoclonal antibody that binds to human complement C5. Eculizumab can induce remission of haematologic and renal abnormalities in aHUS, which led to the drug’s regulatory approval in the USA and Europe.49
Cardiovascular complications Acute and chronic cardiovascular events have been reported in 3–10% of patients with adult-onset or paediatric-onset complement-related aHUS. 4,50 The risk of cardiovascular event varies according to the molecular defect. Patients with CFH mutations, anti-CFH antibodies, gain of function C3 or CFB mutations are particularly susceptible to cardiovascular disorders (Table 1).4
Heart injury Myocardial infarction, cardiomyopathy and heart failure have all been reported in patients with aHUS4,10 and cardiac complications seem to be linked to severe disease. 51 In a study reported in 1997, 16 cardiomyopathy—defined as heart failure requiring the use of inotropes—was recorded during the acute phase of aHUS in 10 out of 23 children (43%) and two patients died within 3 months of onset. Unfortunately, the genetic status of the individuals was unknown because genetic screening was not available at that time.
Table 1 | Studies describing cardiovascular and macrovascular complications in aHUS Reference
Study outline
Findings
Noris et al. (2010)4
273 patients (adults and children)
Seven patients (five with CFH mutations) had cardiovascular disease during acute HUS episodes
Sallee et al. (2010)10
Single adult with aHUS and CFH mutation
Patient died of myocardial infarction 15 days after HUS onset
Venables et al. (2006)11
Eight patients with familial aHUS and a CFH–CFHR1 hybrid gene
One patient died of myocardial infarction 10 years after HUS onset, another died of cardiac arrest 8 weeks after onset
Sellier-Leclerc et al. (2007)50
46 children with aHUS
Four children (one with CFH mutation) had cerebrovascular events at HUS onset
Neuhaus et al. (1997)16
23 children with aHUS (no genotyping)
10 children had cardiomyopathy at discharge, two died
Vilalta et al. (2012)12
Single child with aHUS and CFH mutation
Patient had dilated cardiomyopathy at HUS onset, and myocardial dysfunction during follow-up monitoring Eculizumab improved cardiac function
Dragon-Durey et al. (2010)17
45 patients with aHUS and anti-CFH autoantibodies
Three patients developed cardiac insufficiency on long-term, one died
AbarrateguiGarrido et al. (2009)52
Seven children with aHUS and anti-CFH autoantibodies
One child died of myocarditis
Roumenina et al. (2012)53
14 patients with a C3 mutation
Seven patients had dilated cardiomyopathy (during acute HUS in five, delayed in two patients) One died at onset following a cardiovascular event
Le Quintrec et al. (2013)54
57 adults with aHUS and kidney transplant
Four patients died following cardiovascular events during follow-up No genotype data
Malina et al. (2013)13
Two children with aHUS (one with anti-CFH autoantibodies and one with a C3 mutation)
Both patients had ischaemic changes and gangrene in fingers and toes during acute HUS In one patient, eculizumab induced remission of HUS and of ischaemic changes
Ozel et al. (2003)55
Single child with aHUS and low C3 levels
Patient had ischaemic necrosis of distal phalanges during acute HUS episode No genetic screening
Kaplan et al. (2000)56
Three children with aHUS; Low C3 levels in two patients
The three children had necrotic fingers and toes at HUS onset followed by auto-amputation One child also had occlusion of carotid arteries 90 days after onset No genetic screening
Davin et al. (2010)14 Vergouwen et al. (2008)60
An adolescent with aHUS and a CFH mutation, on dialysis
Patient had stenosis in cerebral arteries, which manifested brain infarcts during kidney transplant No event occurred during post-transplant follow-up under eculizumab
Loirat et al. (2010)15
Single child with aHUS and a CFB mutation, on dialysis
Patient had emiparesis and stenoses of carotid, cerebral, left subclavian, vertebral and pulmonary arteries
Bekassy et al. (2013)61
Single child with aHUS and combined CFB and CFI mutations, on dialysis
Patient had occlusion of carotid arteries, and received a kidney transplant and eculizumab Uneventful follow-up and no progression of vascular occlusion
Ažukaitis et al. (2013)71
Single child with aHUS and a C3 mutation
Patient developed cerebral artery stenoses leading to death due to stroke 9 days after transplant under eculizumab
Abbreviations: aHUS, atypical HUS; CFB, complement factor B; CFH, complement factor H; CFI, complement factor I; HUS, haemolytic uraemic syndrome.
Patients with genetic or acquired CFH defects seem to be particularly at risk of cardiac complications, which mostly occur during, or close to, disease onset. Anti-CFH autoantibody-associated aHUS has the highest frequency of cardiovascular events (~10% of all cases). In two independent
NATURE REVIEWS | NEPHROLOGY
studies, cardiac failure was reported in three (two adults, one child) out of 45 patients17 and in one of seven children with anti-CFH autoantibodies.52 In a series of 49 patients with mutations in CFH, five (7.7%) had cardiovascular disease at aHUS onset.4 A 43-year-old woman with ADVANCE ONLINE PUBLICATION | 3
© 2014 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES a Cys623Ser CFH mutation died of sudden cardiac arrest 15 days after aHUS onset.10 An autopsy revealed the patient had a myocardial infarction with no obstruction of the coronary arteries. However, all small coronary vessels had features of thrombotic microangiopathy with endothelial swelling, vessel wall thickening and intense antibody staining for C5b‑9 was observed in the small coronary vessels and infarcted cardiomyocytes.10 In this patient, myocardial tissue was, therefore, affected by the same anomalies of microvascular circulation as those observed in the kidney. A second patient, a 19-year-old woman with aHUS associated with a CFH– CFHR1 hybrid gene, died from cardiac arrest 8 weeks after onset of the disease.11 However, no details regarding the nature and cause of cardiac insufficiency were reported. Another study described a 4-year-old girl with aHUS owing to an Asp1119Asn mutation in CFH who presented with acute kidney injury, dilated cardiomyopathy and cardiorespira tory arrest.12 Intensive plasma-exchange was unable to halt disease progression or improve cardiac dysfunction and improvements in renal, haematological and cardiac values were only achieved after chronic treatment with eculizumab.12 Taken together, these studies indicate that the heart is affected by CFH deficiency. Gain-of-function mutations in the C3 gene also seem to predispose the carrier to cardiovascular complications. Among 14 patients with sporadic aHUS who carry an Arg139Trp mutation in C3, eight patients (60%) experienced cardiac events.53 Dilated cardiomyopathy, with reduction in left ventricular function, was recorded in seven patients, whereas the type of myocardial lesion was not determined in one patient. Heart failure occurred at aHUS onset in six patients, but two experienced a cardio myopathy 2 months and 6 months after haematological remission, respectively.53 For other forms of aHUS, the frequency of cardiovascular events has yet to be assessed.
Post-transplant cardiovascular events In some patients with complement-related aHUS who have undergone kidney transplantation, chronic complement activation combined with the adverse-effects of immunosuppressive drugs might harm the cardiovascular system. In 57 adult patients with aHUS who received a kidney transplant, four (7%) aged 21–50 years and without major cardiovascular risk factors died after cardiovascular events (acute cardiac failure, n = 1; myocardial infarction, n = 2;
and extensive stroke, n = 1).54 Three patients in this study died with a functioning graft and one died of myocardial infarction after restarting dialysis.
Small peripheral artery thrombosis In some children with complement-related aHUS, severe perfusion defects of the arteries supplying the fingers and toes led to gangrenous lesions of distal phalanges, which developed either during the acute phase of the disease or weeks after onset. A 4-year-old girl with anti-CFH autoantibodies developed gangrene of the finger tips 2 days after initial presentation of aHUS, and died 3 weeks later because of dialysis-related complications.13 The same report describes a second girl who developed ESRD at 4 months of age owing to aHUS associated with a C3 mutation. At 9 months of age, she developed ischaemia in her fingers and toes, and several finger tips became gangrenous despite intensive plasma-exchange therapy. However, after administration of eculizumab, all nonnecrotic digits rapidly regained perfusion and disease activity subsided completely.13 These observations add to previous reports in four children who developed peripheral gangrene of fingers and toes as a complication of aHUS.55,56 Notably, all patients with perfusion problems in the peripheral arteries were children with very severe forms of aHUS. Considering that the luminal diameter of the arteries supplying phalanges is 100-fold larger than that of the glomerular arterioles typically affected in HUS,57,58 the findings from these patients suggest that under certain conditions HUSspecific thrombotic lesions can also affect small and medium arteries. Indeed, one paediatric patient with aHUS and peripheral gangrene began to have seizures 90 days after disease onset.56 MRI studies showed thrombotic occlusion of both internal carotid arteries and patent vertebral arteries.56 Low levels of C3 were documented in all four patients with peripheral gangrene in which complement was measured.13,55,56 In two patients, low C3 levels were associated with anti-CFH autoantibodies and a C3 mutation respectively, and indicate a role for uncontrolled activation of the alternative complement pathway in vascular damage. Large artery stenosis Under normal circumstances, haematological recurrences of aHUS cease when renal function is lost.1 This clinical observation has led to the theory that viable renal tissue is required for overt aHUS relapses.
4 | ADVANCE ONLINE PUBLICATION
Increased shear stress and platelet activation in the damaged renal microvasculature might sustain the microangiopathic process. For example, complete remission of aHUS was observed after bilateral nephrectomy in patients with severe renal involvement, refractory hypertension and signs of hypertensive encephalopathy.59 However, vascular lesions might develop in patients with ESRD or in whom renal tissue was removed. Specifically, stenosis of large arteries has been described in children with complement-related aHUS on long-term haemodialysis. A 15-year-old girl with a Ser1191Leu mutation in CFH, who was on long-term dialysis after loss of two renal grafts, began to experience intradialytic neurologic symptoms (both sensory and motor).14 Angiography revealed severe stenosis of the middle and anterior cerebral arteries,60 and radiological evaluation revealed no signs of calcifications or other loci of arterial wall narrowing.60 The patient received a third kidney transplant 2 years later, but postoperatively developed infarcts in the right frontal and frontoparietal regions. Owing to the potential cerebrovascular risk, and to prevent disease relapse in the graft, she received eculizumab therapy and had no major clinical events during the 6-month follow-up period.14 A second child with neonatal onset of aHUS owing to a Lys350Asp mutation in CFB developed progressive stenosis of multiple large arteries (including carotid arteries, left subclavian and vertebral arteries, intracranial, right humeral, coronary and all pulmonary arteries) after several years of dialysis.15 Unfortunately, a carotid siphon angioplasty was complicated by dissection leading to death. Another child who presented with aHUS at 17 months of age developed ESRD within 3 months. 61 She underwent a deceaseddonor kidney transplantation, but developed recurrence of aHUS in the graft with severe hypertensive crisis that led to removal of the transplant. At 10 years of age, and after 7 years of chronic haemodialysis, the patient developed a transient ischaemic attack but without the haematological symptoms of aHUS. Imaging revealed that the carotid arteries were occluded and had undergone stenosis. Two mutations, one in CFB and one in CFI gene, were identified and measurements of C3 and its degradation product indicated that complement activation was continuous.61 The study’s authors argued that advanced occlusion and stenosis in the carotid arteries were caused by ongoing complement-induced vascular injury www.nature.com/nrneph
© 2014 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES and treatment with eculizumab was initiated despite no evidence of aHUS activity. Eculizumab treatment blocked progression of vascular injury and the patient underwent a successful second kidney transplant. During 18 months follow-up monitoring, no recurrence of aHUS occurred and vascular stenosis did not progress.61 These cases suggest that progression of extrarenal vascular lesions can occur even in the absence of overt recurrences, a process that might be exacerbated by long-term haemodialysis.
Mechanisms Mutations in CFH or anti-CFH auto antibodies that cause complement-related aHUS disrupt the CFH surface recognition activity that underlines complement regulation on endothelial cells.19 Similarly, aHUS associated gain-of-function mutations in CFB or C3 induce a hyperactive C3 convertase with C3b deposition on endothelial cells, enhanced formation of C5b‑9 complexes and deposition of C3 fragments at cell surfaces.40,41 Complement activation disrupts the physiologically thromboresistant phenotype of endothelial cells, which might contribute to thrombotic occlusions in renal and extrarenal vasculature, including coronary arterioles and capillaries, and lead to cardiac dysfunction and myocardial infarction.62 This outcome is mediated by C3a and C5a binding to their cognate receptors and C5b‑9 deposition on cell surfaces. Endothelial damage caused by recurrent or continuous complement activation in patients with genetic complement disorders might not be limited to arterioles and capillaries, although thrombus formation might rarely reach a critical size to cause acute obstruction. The cases of children with peripheral gangrene of the finger and toes owing to aHUS, accompanied by bilateral carotid artery thromboses,56 indicate that macrovascular thrombosis might occur in severe cases of complement-related aHUS. Complement activation has been implicated in vascular stenosis, 63,64 but the mechanisms involved have not been fully elucidated. Complement components exert proinflammatory effects on the endo thelium and vessel wall by upregulating expression of adhesion molecules 65 and thereby release prostanoids, leukotrienes and cytokines that enhance leucocyte recruitment, activation and transendo thelial migration.62,65 Complement components, C3a and C5a act as stimuli for neutrophil, monocyte and macrophage
Defective regulation of the alternative pathway of complement Continuous complement activation C3a, C5a, C5b-9 Endothelial injury
Endothelial detachment Leukocyte transmigration Reactive cell proliferation
Thrombosis
Large artery stenosis
Ischaemia Peripheral arteries Gangrene of fingers and toes
Ischaemia Coronary microvessels
Heart injury and dysfunction
Multiorgan failure
Figure 2 | Potential pathophysiological mechanisms of cardiovascular complications in aHUS. In patients with aHUS, uncontrolled complement activation leads to endothelial injury and loss of antithrombogenic properties. This event might cause thrombosis of coronary arterioles and capillaries leading to ischaemic heart injury and dysfunction, or of small and medium peripheral arteries, leading to ischaemia. Endothelial injury might also initiate stenosis of large arteries in the brain, lung, heart, gut and other organs, due to endothelial cell swelling, detachment, inflammatory leucocyte infiltration, subendothelial expansion and reactive smooth muscle cell proliferation. Thrombus formation on injured endothelium also contributes to the stenotic process leading to multiorgan failure. Abbreviation: aHUS, atypical haemolytic uraemic syndrome.
migration.66 All these events culminate in accumulation of primed leucocytes at sites of complement activation, which lead to vessel wall inflammation and contraction due to the release of cytokines, reactive oxygen radicals and vasoactive medi ators. In addition, C5b‑9 stimulates platelets, macrophages and endothelial cells to secrete platelet-derived growth factor-BB, a major growth factor for vascular smooth muscle cells, which might lead to vessel wall t hickening and stenosis.67 Multiple stenoses of large arteries observed in children on chronic dialysis, and who have genetic complement dysregulation, indicate that chronic subclinical complement-mediated endothelial damage might indeed lead to stenosis of the arteries. Complement activation occurs both in human and experimental atherosclerosis and C1q, C3 and C5b-9 deposits have been found in atherosclerotic vessel lesions.68 In addition, children with renal failure have risks factors for atherosclerosis, and cardiovascular disease accounts for most deaths in young adults with childhood-onset chronic renal failure.69 However, in children with complement-related aHUS and stenosing arteriopathy, none had left ventricular hypertrophy or dysfunction, as observed in systemic atherosclerosis in patients on dialysis, and arterial abnormalities did not include calcifications.14,15,61 The authors hypothesized that a local complement attack
NATURE REVIEWS | NEPHROLOGY
rather than atherosclerosis, might be the main factor causing large artery damage.15 Complement-induced endothelial injury might result in stenosis due to endothelial cell swelling and detachment, followed by inflammatory leucocyte recruitment, subendothelial expansion and reactive smooth-muscle cell proliferation (Figure 2). Notably, in all reported case studies, those patients who developed vascular stenosis were receiving haemodialysis. Complement can be activated during haemodialysis, as demonstrated by the generation of C3a and C5a after exposure of blood to haemo dialysis filter.70 Such dialysis-related complement activation might be exacerbated in patients with complement-related aHUS and genetically determined complement dysregulation, which might contribute to advanced v ascular injury.
Conclusions In patients with complement-related aHUS, cardiovascular complications might occur acutely at disease onset and during relapses. These events affect cardiac microvasculature and peripheral arteries. In addition stenotic lesions might develop chronically due to continuous complementmediated inflammatory response, damaging endothelial cells and subendothelial vessel structures. Extrarenal vascular lesions progress even during symptom-free intervals, a process ADVANCE ONLINE PUBLICATION | 5
© 2014 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES that might be exacerbated by long-term haemodialysis. For this reason kidney transplant with eculizumab treatment should be considered as early as possible in patients with ESRD. Patients with CFH mutations, anti-CFH autoantibodies, or with CFB or C3 mutations, seem to be particularly at risk of developing cardiovascular complications. In such patients, and particularly those on chronic haemodialysis, echocardiographic screening, troponin monitoring and imaging for vascular changes should be performed at admission and during follow-up care. Eculizumab treatment during periods of haemodialysis might prevent serious vascular damage in complement-related aHUS patients with angiographic signs of vascular lesions.61 Why
