4.2 The complement system 315

4.2 The complement system 315

ESSENTIALS The complement system, consisting of soluble and membrane-​ bound proteins, is a major effector mechanism of host defence against infection and inflammatory responses. It has an important role in the removal of immune complexes and dying cells, and also modulates humoral and cell-​mediated immune responses. Complement activation and regulation Complement activation occurs through three pathways, each generating enzyme complexes, termed C3 convertases. These cleave native C3 to form C3b and C3a. C3b can covalently attach to sur- faces (e.g. pathogen surfaces) where it triggers biological responses following interaction with membrane-​bound receptors and can also trigger cleavage of native C5 to C5b and C5a. C5b triggers the for- mation of the membrane attack complex, which disrupts target cell membrane integrity and may result in cell lysis. The physiological role of complement requires that activation specifically occurs upon non​host or altered host surfaces, which is achieved by recognition proteins (e.g. C1q, mannose-​binding lectin) within the activation cascades (recognition proteins). Complement in disease Disease may arise from:  (1) impaired complement activation, re- sulting in immunodeficiency, autoimmune disease, and/​or develop- mental abnormalities; (2) defective complement regulation, resulting in abnormal complement activation on host surfaces, leading to con- ditions including paroxysmal nocturnal haemoglobinuria, atypical haemolytic uraemic syndrome, and C3 glomerulopathy. Hereditary angio-​oedema is associated with deficiency of C1 inhibitor. Measurement of complement in clinical practice The most useful and widely available complement assays are anti- genic measurements of serum C4 and C3: hypocomplementaemia is typical in systemic lupus erythematosus. Functional tests are im- portant when immunodeficiency is suspected (e.g. recurrent pyo- genic infections in childhood). Complement therapeutics Eculizumab is a monoclonal antibody that blocks C5 activation pre­ venting formation of both C5a and the membrane attack complex. It is licensed for the treatment of transfusion-​dependent haemolytic anaemia in paroxysmal nocturnal haemoglobinuria, generalised my- asthenia gravis and atypical haemolytic uraemic syndrome. Introduction Complement was discovered as a heat-​labile plasma component that augmented (‘complemented’) antibody-​mediated killing of bacteria. It consists of a large number of plasma and cell-​bound proteins that interact with each other in an enzymatic cascade (Fig. 4.2.1). We limit our review of complement biology to aspects relevant for the clinician to understand complement-​associated diseases. There are three activation pathways: the classical pathway, the al- ternative pathway, and the lectin pathway. The classical pathway is a predominantly antibody-​dependent pathway. It is activated by ag- gregated IgG or IgM in immune complexes. The lectin pathway is antibody-​independent and is initiated by the binding of pattern rec- ognition molecules to, for example, mannose-containing structures on bacteria. The pattern recognition molecules include mannose- binding lectin (MBL), ficolins (ficolin 1, -2 and -3), collectin kidney 1 (CL-K1, collectin 11) and collectin liver 1 (CL-L1, collectin 10). The alternative pathway is constitutively active. Complement activation results in the generation of enzymatically active complexes (termed convertases) that cleave C3 and C5. Activated C3 (termed C3b) can be rapidly amplified through a posi- tive feedback cycle, termed the C3b amplification loop. Amplification of C3b can occur irrespective of which pathway generated the C3 convertase. Activation of C5 triggers the formation of the membrane attack complex (MAC), A large array of regulatory proteins (com- plement regulators) prevent inappropriate complement activation and limit host surface damage. Complement biological roles A wide range of roles has been ascribed to the complement system, but they can be categorized into two main activities:  (1) the de- struction/​removal of anything that is recognized as foreign or is not 4.2 The complement system Marina Botto and Matthew C. Pickering

316 SECTION 4  Immunological mechanisms adequately protected by soluble or cell-​associated complement regu- lators; and (2) the modulation of humoral and cell-​mediated immune responses. There is evidence of unrelated or alternative roles of the complement system outside immunity, mainly in regenerative and developmental processes (see following paragraphs regarding defects in the lectin components). The most important function of complement is the host defence against infectious disease. Complement provides mechanisms for the killing and clearance of microorganisms. It does this by the co- valent binding to their surface of complement fragments (e.g. C3b) that are ligands for receptors on phagocytic cells that ingest and kill the microorganisms. The activation of complement also causes the generation of anaphylatoxins (e.g. C5a and C3a), which have chemo- tactic activity and recruit leucocytes to sites of infection and inflam- mation. These small complement fragments signal through specific receptors (C5aR and C3aR) and attract neutrophils and monocytes to the site of complement activation. A further role of complement in host defence against infections is the generation of the MAC, which may disrupt the cell membrane and kill the microorganism. As part of its clearance role the complement system promotes the non​inflammatory disposal of dying cells and immune complexes. It is in this role that complement may prevent the development of systemic lupus erythematosus (SLE). Activation of complement by immune complexes facilitates the clearance of antigen and thereby helps to prevent immune complexes from causing inflammatory damage to tissues, although, as outlined to follow, complement may also contribute to inflammatory tissue injury in circumstances when immune complexes persist. The other main activity of complement is bridging innate and adaptive immunity. Activation of complement augments antibody responses and thereby enhances host defence against pathogens. The binding of complement to antigens reduces the threshold for B-​cell activation and enhances antigen presentation and B-​cell memory. Similarly, complement can modulate the T-​cell responses control- ling T-​cell homeostasis and activation. Complement activation pathways and regulation Classical pathway The initial step in classical pathway activation (Fig. 4.2.2) is the binding of C1q, the first component of this pathway, to the Fc portion of antibodies complexed with antigens. The various IgG isotypes have different capacities to bind to and activate C1q. In humans, IgG3 is the most potent activator, followed by IgG1 and then IgG2. IgG4 does not bind C1q and cannot activate the classical pathway. C1q also binds to the CH3 domain of IgM that has adopted a staple configuration fol- lowing binding of antigen. The general order of complement-​fixing potential to human antibodies is thus IgM > IgG3 > IgG1 > IgG2 >> IgG4. IgA can activate the alternative pathway, whereas IgE is not an effective complement-​activating isotype. The classical pathway can also be triggered in an antibody-​ independent manner by the binding of C1q directly to the surface of certain pathogens or host proteins such C-​reactive protein bound to its ligand or amyloid fibrils. C1q is part of a complex, the C1 complex, which includes two serine proteases (C1s and C1r). Once activated, ALTERNATIVE LECTIN CLASSICAL PATHWAY TRIGGERS C3 CONVERTASE C3b C3a C5 CONVERTASE C5a C5b TERMINAL MAC C3b AMPLIFICATION Fig. 4.2.1  Schematic depiction of the complement system. There are three activation pathways: the classical, lectin, and alternative pathways. The classical and lectin pathways are activated by specific triggers; the alternative pathway is constitutively active. On activation the pathways result in the formation of an enzyme complex termed C3 convertase. The C3 convertase proteolytically cleaves C3 to form the anaphylatoxin C3a and the opsonin C3b. Further generation of C3b occurs through a positive feedback loop termed the C3b amplification pathway. The addition of another molecule of C3b to a C3 convertase results in a complex (C5 convertase) that can proteolytically cleave C5. The C5 convertase generates the anaphylatoxin C5a and C5b. The C5b molecule triggers formation of the terminal pathway which, through the sequential addition of C6, C7, C8, and multiple C9 molecules, results in the membrane attack complex (MAC, also denoted C5b-​9).

4.2  The complement system 317 the C1 complex acts on the next two components of the classical pathway, cleaving C4 and then C2, to generate the classical pathway C3 convertase (C4b2a). This surface-​bound convertase cleaves C3 resulting in the generation of C3b, which coats the activating sur- face, a phenomenon termed opsonization. The activation of C3 is followed by the formation of a multiprotein complex that cleaves C5. Alternative pathway The main feature of the alternative pathway (Fig. 4.2.2) is that is in a constant state of activation or ‘tick-​over’ that results in the generation of low levels of activated C3 fragment (C3b) in the fluid phase. Most of this fluid-​phase C3b is rapidly inactivated by hydrolysis, but a small amount can bind to surfaces in the immediate vicinity of the C3 acti- vation and initiate the amplification of the alternative pathway. The fate of surface-​bound C3b is controlled by two mechanisms. The first mechanism is the presence of membrane-​bound comple- ment regulatory molecules, which serve to protect host cells by inhibiting further C3 activation. The second mechanism depends on the affinity of the surface-​bound C3b for factor H. Factor H is the major fluid-​phase regulator of C3 activation but also is an important regulator of surface-​bound C3b, particularly along the renal endo- thelium. Factor H binds preferentially to C3b bound to vertebrate cells as it has a high affinity for the sialic acid residues present on these cells. In contrast, pathogen surfaces lack sialic acid residues, rendering the bound C3b resistant to inactivation by factor H and allowing amplification to proceed. The alternative pathway hence works through an amplification loop (Fig. 4.2.3) that can proceed efficiently on the surface of a pathogen, but not on a host cell. The same amplification loop enables the alternative pathway to amplify complement activation initially triggered through the classical or the lectin pathway (C3b amplification loop, Fig. 4.2.1). Lectin pathway The lectin pathway is initiated by the binding of MBL, ficolins 1–​3, or CL-L1 and CL-K1 Fig. 4.2.2) to targets on the surface of pathogens or damaged tissue. All these proteins belong to a family of collagenous lectins, named collectins, which are capable of rec- ognizing carbohydrate domains. After recognition, initiation pro- ceeds through the activation of serine proteases, known as MASPs (mannose-​binding lectin-​associated serine proteases). The activa- tion of the MASP system, which contains three different enzymes—​ MASP-​1, MASP-​2, and MASP-​3, also a protein with no proteolytic activity named MAP19 or sMAP—​results in cleavage of C4, C2, and then C3 in a similar way to the classical pathway. The relationship between the lectin pattern recognition molecules and the MASPs is complex and incompletely understood. MASP-3 contributes to alternative pathway activation since it is required for activation of factor D (i.e. its conversion from pro-Factor D to active Factor D). ALTERNATIVE LECTIN CLASSICAL C1q C1r, C1s C4 C2 Mannose-binding lectin (MBL) Ficolin-1, Ficolin-2, and Ficolin-3 CL-K1 and CL-L1 MBL-associated serine protease (MASP)-1, -2, -3 MAp19, MAp44 C4 C2 Hydrolysed C3 Factor B Factor D Properdin (P) Immune-complex (antibody-antigen complex) Carbohydrate or acetyl patterns Spontaneous Main pathway triggers: Pathway components: Recognition molecules: Other components: C4b2a C3bC4b2a C4b2a C3bC4b2a C3bBb C3bC3bBbP Final enzyme complexes: C3 convertase: converts C3 to C3b and C3a C5 convertase: converts C5 to C5b and C5a Fig. 4.2.2  Complement activation pathways. Examples of the triggers of the three activation pathways are shown. These triggers are recognized by specific proteins within the activation pathways. Examples include the interaction of the classical pathway recognition protein, C1q, with immune complexes; and the interaction of the lectin pathway recognition protein, mannose-​binding lectin (MBL), with carbohydrates on bacteria. There are multiple recognition proteins in the lectin pathway and the biological role of many of these is incompletely understood. Further activation is achieved through the actions of enzymes within the classical (C1r, C1s, C2), lectin (MASP-​1, MASP-​2, MASP-​3, C2), and alternative (factor B, factor D) pathways. The final enzyme complexes are able to proteolytically cleave C3 to C3b and C3a. These complexes are termed C3 convertases. The addition of a further C3b molecule to the C3 convertases enables the resultant complex to cleave C5. These complexes are termed C5 convertases.

318 SECTION 4  Immunological mechanisms Terminal pathway The final phase of complement activation is the assembly and for- mation of the membrane attack complex (Fig. 4.2.4). The end result is a pore in the lipid bilayer membrane, firstly identified in electron micrographs as membrane ‘pores’ and ‘hollow cylinders’ that des- troy the membrane integrity. Structural studies have shown that the MAC has a ‘split-washer’ configuration which partially penetrates the lipid bilayer, resulting in an irregular β-barrel pore. The first step of the terminal pathway is the enzymatic cleavage of C5 to release C5a and a larger fragment, C5b, that binds sequentially and non​enzymatically to the plasma proteins C6, C7, C8, and C9. The polymerization of C9 produces a hydrophobic complex, com- monly denoted as C5b-​9n, where n represents the number of poly- merized C9 molecules that forms ‘pores’ in lipid bilayers resulting in target lysis. The lytic effect of MAC is particularly evident in condi- tions in which red cells are targeted by complement activation and in host defence against Neisseria. Regulation of complement activation Given the destructive effects of complement and the way in which its activation is rapidly amplified through a triggered-​enzyme cascade, it is not surprising that its activation is tightly regulated both in the fluid phase and on cell surfaces, serving not only to prevent tissue damage from autologous complement activation but also to prevent the depletion of complement proteins. The molecular mechanisms utilized to block complement activa- tion include: (i) inhibiting activating proteases such as C1s (see later paragraph on C1 inhibitor deficiency); (ii) acting as competitor/​ decay factor for enzyme complexes; (iii) performing cofactor activ- ities in proteolytic cleavage; and (iv) working as a substrate-​specific protease. The evolution of the complement system has therefore been accompanied by the development of a sophisticated regulatory system consisting of fluid-​phase and membrane-​bound regulatory proteins that act at many steps of the pathway (Fig. 4.2.5). Complement deficiency and disease Disease may arise when there is (1) impaired complement activation or (2) defective complement regulation (Table 4.2.1). Impaired complement activation and disease Increased susceptibility to infections Increased susceptibility to childhood pyogenic bacterial infections is frequently seen in many of the homozygous deficiencies, while heterozygous deficiency is usually asymptomatic. Homozygous defi- ciencies are rare, with the exception of C2, MBL, and (in the Japanese population, where about 1/​1000 are affected) C9. C2 deficiency occurs in 1:20 000 white individuals and is usu- ally asymptomatic. MBL deficiency occurs in 5–​10% of individuals, but increased infection risk is only seen if there are immunosup- pressive comorbidities. Examples include cancer patients receiving chemotherapeutic treatment and organ-​transplant recipients on immunosuppressive medication. Impaired complement activation should be suspected in an individual with recurrent pyogenic child- hood infections, a family history of infection and consanguinity. Patients with hereditary C3 deficiency or with mutations in mol- ecules leading to C3 consumption show increased susceptibility to recurrent and severe bacterial infections, particularly those C3b AMPLIFICATION Factor B Factor D C3 Properdin C3b Pathway trigger: Pathway components: Recognition molecules: Other components: C3bBb Pathway enzymes: C3 convertase: converts C3 to C3b and C3a Fig. 4.2.3  C3b amplification pathway. C3b on a surface can be rapidly amplified through a feedback cycle. This pathway utilizes the same components as the alternative pathway, hence it is sometimes referred to as the amplification loop of the alternative pathway. Surface C3b can interact with factor B to form a pro-​C3 convertase (C3bB). Factor B in complex with C3b can be cleaved by the enzyme factor D. This converts factor B to two fragments termed Bb (an enzymatically active fragment which remains attached to C3b) and Ba, which is released from the complex. This results in the conversion of the pro-​convertase (C3bB) to a convertase (C3bBb). The C3 convertase now cleaves C3 generating further C3b. Note that C3b can be amplified via this pathway irrespective of how it was formed. TERMINAL C6 C7 C8 C9 C5b Pathway trigger: Pathway components: Recognition molecules: Other components: C5b-9 Final complex: Membrane attack complex: Fig. 4.2.4  Terminal pathway. This is triggered by the generation of C5b and the sequential addition of a single molecule of C6, C7, C8, and multiple (n) C9 molecules. This results in the formation of a macromolecular complex (C5b, C6, C7, C8, C9n, typically denoted C5b-​9) termed the membrane attack complex (MAC). The MAC damages cell membranes by creating membrane pores. In cells that are unable to repair the damaged membrane, cell death or lysis may occur. Erythrocytes are particularly susceptible to MAC-​induced lysis since, unlike many nucleated cells, they are unable to actively remove MAC
that has been incorporated into their membranes.

4.2  The complement system 319 caused by encapsulated organisms (e.g. Streptococcus pneumoniae and Staphylococci). Similar infections are seen among individuals lacking antibodies or normal phagocytic function. This indicates that the normal pathway for eradication of these bacteria requires antibody, complement, and phagocytes. In patients with C3 defi- ciency infections become less frequent during adulthood when the protective antibody repertoire has expanded following repeated in- fectious challenges. Eradication of Neisseria requires alternative pathway C3 activa- tion on the surface of the pathogen and subsequent formation of the MAC through the terminal pathway. Homozygous deficiencies of the alternative and terminal pathway components are associ- ated with increased susceptibility to Neisserial infections. Patients treated with the anticomplement C5 antibody (eculizumab) are also at risk of Neisserial infections because eculizumab prevents terminal pathway activation. Susceptibility to meningitis is also influenced by genetic poly- morphism across the complement factor H gene family. This is because some Neisserial strains bind the alternative pathway negative regulator (factor H) and evade complement-​mediated eradication. Autoimmunity There is a strong association between classical pathway deficiency and SLE. Over 95% of homozygous C1q deficient individuals de- velop SLE. The reasons for this association are complex and include abnormal immune complex processing and impaired tolerance to autoantigens through defective clearance of apoptotic cells (see Chapter 19.11.2). The association between C2 deficiency and SLE is much less strong and suspected to be approximately 10%. Developmental abnormalities Mutations in the lectin pathway components, MASP1 and CL-​K1, are associated with a developmental syndrome, termed 3MC (Mingarelli, Malpuech, Michels, Carnevale) syndrome. This condition includes multiple developmental defects including severe growth retardation, facial dysmorphism, and skeletal abnormalities. The contribution of the lectin pathway components to its pathogenesis was unexpected and remains incompletely understood. Defective complement regulation and disease C1 inhibitor deficiency (Also see Chapter 4.5.) Aetiology and pathogenesis The disease hereditary angio-​oedema (OMIM 106100) is caused by deficiency of C1 inhibitor. This is inherited as an autosomal dom- inant disorder with partial penetrance. The disease is dominantly inherited because the production of C1 inhibitor from a single, normal allele is insufficient to maintain normal homeostasis of the complement and kinin pathways. The mutations may have two ef- fects on protein production. In type I  hereditary angio-​oedema, which accounts for approximately 85% of cases of the disease, the mutant prevents any expression of protein from the mutant allele and hence there are reduced levels of C1 inhibitor. Type II heredi- tary angio-​oedema is caused by a series of point mutations in the C1 inhibitor gene that alter one of the amino acids at the active centre of the protein and abolish its activity as a serine proteinase inhibitor. These mutations allow expression of normal amount of protein, ALTERNATIVE LECTIN CLASSICAL PATHWAY TRIGGERS C3 CONVERTASE C3b C3a C5 CONVERTASE C5a C5b TERMINAL MAC C3b AMPLIFICATION CD59 Factor H CD46 (MCP) Factor I Factor H CD55 (DAF) C1INH Fig. 4.2.5  Complement regulation. There is a complex network of proteins in plasma and within cell membranes that negatively regulate complement activation. These proteins act at different steps in the pathways. Key examples are shown. These include the membrane-​bound protein CD59 that prevents the assembly of the MAC. Factor H and CD55 (also termed decay-​accelerating factor, DAF) enhance the disassembly of the C3 convertases. C3b can be proteolytically cleaved by the enzyme factor I to form a product (termed iC3b) that can no longer bind to factor B. The conversion of C3b to iC3b therefore prevents further C3b amplification. Factor I requires cofactors to mediate the cleavage of C3b to iC3b. These cofactors are factor H and CD46 (also termed membrane cofactor protein, MCP), a membrane-​bound cofactor. Factor I and factor H are the key inhibitors of the alternative pathway (see text). C1 inhibitor (C1INH) negatively regulates both the classical and lectin pathways (see text).

320 SECTION 4  Immunological mechanisms which is non​functional, or even abnormally high C1 inhibitor levels, because the mutant protein is not consumed by normal interaction with activated serine proteinases. It is easy to miss the diagnosis of this variant of hereditary angio-​oedema if it is not appreciated that levels of C1 inhibitor can be normal or high in patients with the dis- ease: functional C1 inhibitor assays are required. In normal circumstances, C1 inhibitor binds to and inactivates enzymatically active C1r and C1s. It also inhibits plasmin, kallikrein, and activated coagulation factors XIIa and XIa. Deficiency results in uncontrolled fluid-​phase classical pathway activation and conse- quently reduced levels of both C4 and C2. Acute angio-​oedema attacks are characterized by increased vas- cular permeability at the affected sites. The swellings are believed to be caused by the action of small peptides, called kinins, in par- ticular bradykinin, that induce increased vascular permeability by their actions on vascular endothelium and smooth muscle. These kinins are produced by the action of serine proteinases that are ineffectively regulated in the presence of reduced activity of C1 in- hibitor. Plasmin activation may be important in the precipitation of attacks by consuming the reduced amounts of available C1 in- hibitor in individuals with only half normal functional expression of the protein. Individuals exhibiting clinical features of hereditary angio-​ oedema who have normal C1 inhibitor concentration and function have also been described. This type of hereditary angio-​oedema (OMIM 610 618) has been termed hereditary angio-​oedema type III or, more comprehensively, oestrogen-​related hereditary angio-​ oedema or oestrogen-​sensitive hereditary angio-​oedema. In contrast to hereditary angio-​oedema types I and II, hereditary angio-​oedema type III has been observed exclusively in women, where it appears to be correlated with conditions of high oestrogen levels (e.g. preg- nancy or the use of oral contraceptives). The aetiology of type III hereditary angio-​oedema appears to be heterogeneous, with some patients having gain-​of-​function mutations in F12, the gene Table 4.2.1  Complement deficiency and disease Complement deficiency Phenotype Comments ACTIVATION PROTEINS Classical pathway deficiency: C1q, C1r, C1s, C2, C4 SLE Recurrent encapsulated bacterial infections All extremely rare except C2 deficiency where estimated prevalence is 1:20 000 Association with SLE weakest for C2 deficiency Alternative pathway deficiency: Factor B, factor D Recurrent meningococcal infections Recurrent encapsulated bacterial infections All extremely rare Lectin pathway deficiency: MBL, ficolins 1–​3, MASP-​1, MASP-​2, and MASP-​3, CL-​K1 Increased infection among immunocompromised individuals (MBL deficiency) Ficolin-​3 deficiency associated with necrotizing enterocolitis Mutations in the genes encoding CL-​K1 (COLEC11),
CL-L1 (COLEC10) and MASP-​3 and MASP-​1 (MASP1) associated with an autosomal recessive developmental syndrome termed ‘3MC syndrome’a All extremely rare except MBL deficiency where estimated prevalence is 5% in white populations Terminal pathway C5, C6, C7, C8, and C9 Recurrent meningococcal infections All rare except C9 deficiency in Japanese where estimated prevalence is 1:1000 C3 Recurrent encapsulated bacterial infections Membranoproliferative glomerulonephritis (rare) SLE-​like illness (rare) Extremely rare REGULATORY PROTEINS C1INH Negative regulator of: classical and lectin pathways contact system coagulation system fibrinolytic system Hereditary angioedema Estimated prevalence 1:50 000 Angioedema results from uncontrolled production of bradykinin due to dysregulation of the contact system and does not arise from uncontrolled complement activation. Associated with low C4 due to uncontrolled classical pathway activation Factor H, Factor I, and CD46 Negative regulators of the alternative pathway and C3b amplification loop Atypical haemolytic uraemic syndrome (aHUS) C3 glomerulopathy All rare aHUS manifests in heterozygous deficiency states Complete factor H deficiency associated with C3 glomerulopathy (e.g. dense deposit disease) Factor H-​related protein 5 Putative regulator of C3 processing within the kidney—​biological function incompletely understood C3 glomerulopathy (‘CFHR5 nephropathy’) Rare Predominantly individuals with Cypriot ancestry CD59 Negative regulator of terminal pathway activation Paroxysmal nocturnal haemoglobinuria Rare Acquired somatic mutation Renders CD59-​deficient erythrocytes susceptible to complement-​mediated intravascular haemolysis Properdin Positive regulator of C3 activation Recurrent meningococcal infections Rare X-​linked deficiency MBL, mannose-​associated lectin; MASP-​2, MBL-​associated serine protease; C1INH, C1 inhibitor; CD46, also known as membrane cofactor protein; CL-​K1, also known as collectin-​11. a 3MC syndrome is a term used to describe clinically identical syndromes that were independently described: the Mingarelli, Malpuech, Michels, and Carnevale syndromes.

4.2  The complement system 321 encoding human coagulation factor XII (FXII, or Hageman factor), as a possible cause. Angio-​oedema may be acquired due to the development of auto- antibodies to C1 inhibitor. This may occur in association with lymphoproliferative disease, particularly B-​cell lymphoma. The clin- ical features are similar to hereditary angio-​oedema but with later age of onset. Typically, C4 and C1 inhibitor levels are low. Complement C1q level is reduced in patients with acquired angio-​oedema, but normal in patients with hereditary angio-​oedema. Clinical features and diagnosis Allergy is much more common than hereditary angio-​oedema as a cause of angio-​oedema. In hereditary angio-​oedema, the swelling is not itchy and is not accompanied by other features of allergy such as asthma and urticaria. Oedema can affect any part of the integu- ment but is most common in the extremities. Classically, the oedema and swelling develop gradually over several hours and then subside over 2 to 3 days. Involvement of the upper airways (including the tongue, pharynx, and larynx) may result in life-​threatening airway obstruction. Swelling of the bowel mucosa may produce severe ab- dominal pain, mimicking common surgical emergencies. In general, women have a more severe course of the disease than men. Patients with early onset of clinical symptoms are affected more severely than those with late onset. Diagnosis of hereditary angio-​oedema is made on the basis of the clinical findings described earlier, the presence of family his- tory, and blood tests. A family history of angio-​oedema makes diag- nosis much easier but is not always present because some cases are due to new mutations in the C1 inhibitor gene. In other families, other members with C1 inhibitor deficiency may have no clinical symptoms. All patients who are suspected of having hereditary angio-​oedema should have serum C4 levels measured, which is a good screening test for hereditary angio-​oedema as it is invariably low in untreated patients with the condition. This is because the reduced C1 inhibitor activity allows C1s to cleave C4 and C2 in an unregulated fashion. In patients with type I hereditary angio-​oedema C1 inhibitor pro- tein levels are typically low (usually <30% of normal levels), but they may be normal or high in the 15% of patients with type II disease, and functional assays of C1 inhibitor are necessary to make the diag- nosis. Genetic tests are not indicated routinely and are usually not necessary to confirm the diagnosis of hereditary angio-​oedema. Management Acute attacks of angio-​oedema may be stopped by infusion of puri- fied C1 inhibitor concentrate. Three preparations are available for clinical use: two are derived from plasma (Berinert P and Cinryze) and the third is a recombinant molecule (Rhucin). All are adminis- tered intravenously. An alternative is fresh frozen plasma, but this is less satisfactory because plasma not only contains C1 inhibitor but also kallikrein, C1r, and C1s, which may generate further kinin production. Acute attacks of hereditary angio-​oedema do not respond to adrenaline (epinephrine), though if there is any cause to suspect al- lergic rather than hereditary angio-​oedema, then administration of epinephrine is unlikely to cause any harm and may be life-​saving. C1 inhibitor is rapidly effective in acute attacks of oedema. In one study of laryngeal oedema, the most feared complication, administration of the inhibitor reduced the median duration of the attack from 100+/​–​26 h to 15+/​–​9 h. C1 inhibitor should be given rapidly at the first sign of an attack and prophylaxis considered if attacks are frequent or patients are undergoing procedures that may trigger an attack (e.g. dental procedures). C1 inhibitor levels originating from the single normal allele in- crease in response to treatment with attenuated androgens, such as danazol, stanozolol, and oxandrolone. These are moderately ef- fective treatments, although these compounds retain some virilizing activity. An alternative is the proteinase inhibitor tranexamic acid, which may reduce the consumption of C1 inhibitor by blocking the activity of the serine proteinases that interact with C1 inhibitor. With the knowledge that the pathogenesis of the angioedema results from dysregulated bradykinin production, two other thera- peutic approaches have been utilized:  bradykinin B2 receptor blockade (icatibant, a 10 amino acid peptide) and inhibition of kallikrein (ecallantide, a 60 amino acid protein). Both are effective in acute attacks, can be administered subcutaneously and—​unlike C1 inhibitor preparations—​are effective in type III angioedema. Advice on use of contraceptives and hormone replacement therapy should emphasize avoidance of oestrogen. Angiotensin-​ converting enzyme (ACE) inhibitors need to be avoided because of their effects on the kallikrein–​bradykinin pathway. Angiotensin-​II receptor antagonists may be used with caution in patients with her- editary angio-​oedema. Atypical haemolytic uraemic syndrome (aHUS) (See also Chapter 21.10.6.) aHUS is characterized by renal failure due to thrombotic microangiopathy (TMA). This condition, which is distinct from HUS associated with Shiga toxin-​producing strains of E.  coli, is linked with defective regulation of the alternative pathway along the renal endothelium. The defective regulation in aHUS can be inherited or acquired. Inherited causes include (1) loss-​of-​function mutations in the nega- tive regulators of the alternative pathway—​factor H, factor I, and CD46; and (2) gain-​of-​function mutations in the activation proteins of the alternative pathway—​C3 and factor B. Acquired causes in- clude autoantibodies to factor H. Antifactor H autoantibodies, like aHUS-​associated factor H mutations, impair the ability of factor H to regulate C3 activation along the renal endothelium. Alternative pathway activation triggers C5 activation, which is now known to be critical for the development of the renal TMA. Eculizumab (an anticomplement C5 antibody) is highly effective in treating patients with complement-​associated aHUS and is now licensed for this indication. C3 glomerulopathy (C3G) (See Chapter 21.8.6.) C3 glomerulopathy is characterized by predominant or isolated accumulation of C3 within glomeruli. This condition is associated with impaired regulation of the alternative pathway within plasma or along the glomerular basement membrane (GBM). The path- ology is thought to develop as a consequence of either accumulation of serum-​derived C3 metabolites along the GBM or activation of C3 directly on the GBM. The abnormal alternative pathway activation may be genetic or acquired. Loss of function of regulators and gain-​ of-​function of activation proteins within the alternative pathway can

322 SECTION 4  Immunological mechanisms result in familial C3G. Acquired factors include autoantibodies that stabilize the alternative pathway C3 convertase, termed C3 nephritic factors (C3NeF). C3NeF is frequently associated with low plasma C3 levels. However, since C3NeF does not commonly affect the clas- sical or lectin pathways, plasma C4 levels are usually normal. C3 nephritic factor is also associated with partial lipodystrophy (loss of fat from the face and upper part of the body). C3NeF potentiates alternative pathway activation on, or in the vicinity of, adipocytes resulting in complement-​mediated damage. Adipocytes are suscep- tible since they produce alternative pathway components: C3 and factor D (also termed adipsin). Loss-​of-​function mutations in factor H are associated with sus- ceptibility to both aHUS and C3G. The association with two distinct renal phenotypes (renal TMA versus glomerulonephritis) can be ex- plained at the molecular level. aHUS-​associated factor H mutations target domains of the protein that are required for its interaction with surface ligands (e.g. polyanions) along the renal endothelium. These mutations do not impair the ability of factor H to regulate C3 activation in plasma. Consequently, plasma C3 levels are typic- ally normal in affected patients. The result is defective regulation of C3 activation specifically along the renal endothelium. In contrast, C3G-​associated factor H mutations result in impaired C3 regulation in plasma and low plasma C3 levels. Pigs and mice with complete factor H deficiency develop impaired plasma C3 regulation (with severe falls in C3 levels) and accumulation of C3 along the GBM. Distinct entities within the C3G classification include dense deposit disease (DDD, formerly termed membranoproliferative glomerulo- nephritis type II) and complement factor H-​related 5 (CFHR5) neph- ropathy. DDD is associated with ocular drusen, similar to that seen in age-​related macular degeneration (AMD). The pathogenesis of DDD-​ associated ocular drusen is unknown, but it is interesting that gen- etic susceptibility to AMD is strongly influenced by polymorphisms within complement genes. These include genes encoding factor H, factor I, C3, and factor B. These findings suggest a role for alternative pathway activation in the pathogenesis of ocular drusen. CFHR5 nephropathy is a C3G that is endemic in Cyprus and as- sociated with heterozygous mutations in the gene encoding CFHR5. The role of the factor H-​related proteins is unclear. CFHR5 might act as a competitive antagonist with factor H. CFHR5, unlike factor H, does not have C3 regulatory functions. When it interacts with activated C3 (C3b) it prevents factor H binding and enables further complement activation. The mutant CFHR5 protein in CFHR5 neph- ropathy is thought to potentiate C3 activation along the GBM even in the presence of normal factor H. Unlike complement-​associated aHUS, the contribution of C5 activation to kidney damage is less clear. Eculizumab has not been licensed for C3G, but there are case reports of its effectiveness in this condition, particularly where there is crescentic glomerulonephritis. Paroxysmal nocturnal haemoglobinuria (See also Chapter 22.5.3.) Paroxysmal nocturnal haemoglobinuria (PNH; OMIM 311770) il- lustrates the critical role of membrane-​bound complement regula- tory proteins in protection against complement-​mediated lysis. In PNH a somatic mutation results in a clone of erythrocytes that lack glycosylphosphatidylinositol-​linked proteins. These include the membrane-​bound complement regulators CD59 and CD55 (also termed decay-​accelerating factor, DAF). Affected cells (PNH erythrocytes) are susceptible to complement-​mediated lysis within the circulation. Activated C3 (C3b) accumulates on PNH erythrocytes through alternative pathway activation. Subsequent terminal pathway acti- vation results in cell lysis by the MAC. CD59 is an inhibitor of the terminal pathway and its deficiency appears to be responsible for PNH since isolated CD59 deficiency, but not CD55 deficiency, is as- sociated with a PNH-​like phenotype. Prevention of terminal pathway activation with eculizumab (a monoclonal antibody against complement C5) prevents complement-​mediated intravascular lysis of PNH erythrocytes. Eculizumab is licensed for the treatment of transfusion-​dependent PNH. During treatment C3b may accumulate on PNH erythrocytes since, unlike terminal pathway activation, the alternative pathway is unaffected by eculizumab. PNH erythrocytes coated with C3b may be prematurely removed from the circulation by phagocytosis in the liver and spleen (extravascular haemolysis). Complement investigations Standard and specialized complement investigations are listed in Table 4.2.2. The most frequently available complement assays are antigenic measurements of serum C4 and C3, but the results of such assays need to be interpreted cautiously. The normal ranges are wide because there is substantial genetic variation in the levels of these proteins. Furthermore, proteins levels are a product of both synthetic and catabolic rates, and both of these may vary in health and disease. Both C3 and C4 are acute phase reactants and concen- trations of these proteins may rise, in the case of C3 by as much as 0.5 g/​litre, in response to acute phase stimuli. Several approaches have been devised to assess the presence of complement activation in vivo. Many assays have been developed which identify the product of activation of the complement system (e.g. C3a, C3d). Although these assays are attractive in principle, the products of complement activation are only present in plasma very transiently and, in routine clinical practice, measurement of total C4 and C3 levels have not been supplanted as the best ‘rough and ready’ estimates of complement activation. A fall in serum C4 with or without a fall in C3 usually indicates classical pathway activation (Table 4.2.2). Formation of immune complexes, either in the circulation or in tissues, may result in clas- sical pathway activation sufficient to cause circulating C4 levels to fall. This is most likely to indicate SLE, and in many patients with this condition low complement levels (fall in C4 with or without a fall in C3) are a useful marker of active disease. A low C4 level may be due to the presence of a C4 null allele. C4 null alleles are more common in SLE patients. In this case the low C4 level may remain low during disease quiescence. Other important causes of classical pathway activation in- clude C1 inhibitor deficiency, chronic infections (e.g. subacute bacterial endocarditis, mixed essential cryoglobulinaemia, and hypocomplementaemic urticarial vasculitis syndrome). Cryoglobulins should be tested in patients presenting with unex- plained renal disease or peripheral neuropathy and low C4. A low C3 level in the setting of a normal C4 level indicates alterna- tive pathway activation. This is rare and is associated with the acute phase of postinfectious glomerulonephritis and C3 glomerulopathy.

4.2  The complement system 323 Other standard assays include functional tests of the classical and alternative pathways. These are useful screening tests for assessing complement activation defects, but the most common cause of ab- normal functional activity is delays in ex vivo sample processing. However, if a persistent abnormality is demonstrable, then it is ap- propriate for specialist laboratories to proceed to measure individual components to identify the abnormal or missing component. C1 in- hibitor testing is routinely available and straightforward. Specialized assays include anti-​C1q antibodies, C3NeF testing, anti-factor H autoantibodies, and genetic screening for alternative pathway mutations associated with aHUS and C3G. These assays are useful in defined settings (see Table 4.2.2), but their clinical in- terpretation requires experience and care. Anti-​C1q antibodies are associated with glomerulonephritis in patients with SLE and are part of the defining criteria in hypocomplementaemic urticarial vasculitis syndrome (HUVS). These IgG autoantibodies are dir- ected against an epitope in the collagenous region of C1q, which becomes exposed when C1q is dissociated from the other proteins of the C1 complex, C1r and C1s. This type of epitope is known as neoepitope. Up to one-​third of patients with SLE develop anti-​C1q autoantibodies. These are associated with activation of the classical pathway, causing very low C4 levels and, to a lesser extent, reduced C3 levels. They are thought to amplify complement activation by immune complexes in tissues, by binding to C1q fixed to immune complexes enlarging the complexes and promoting further com- plement activation. Increases in anti-​C1q antibody titres have been shown to precede renal involvement in SLE and, in contrast to rises in anti-​DNA antibody titres, appear to increase specifically prior to renal relapse. In HUVS very high titres of anti-​C1q antibodies are typically found together with marked reduction in C4, C3, and C1q levels. HUVS is characterized by urticarial vasculitis (histologically is usually a leukocytoclastic vasculitis), polyarthritis/​polyarthralgia, membranoproliferative glomerulonephritis, angio-​oedema, neur- opathy, and obstructive pulmonary disease. Complement therapeutics Patients with chronic hypocomplementaemia are at particular risk of developing serious infection with encapsulated organisms Table 4.2.2  Complement investigations Standard assays Comments Serum C4 Reduced levels usually indicate classical pathway activation or presence of C4 null alleles. If C4 is low with normal C3 this indicates predominant classical pathway activation and typical causes include: • Active SLE • Hypocomplementaemic urticarial vasculitis (HUVS) • C1 inhibitor deficiency • Mixed essential cryoglobulinaemia • Rheumatoid vasculitis • Chronic infections associated with immune complex formation (e.g. subacute bacterial endocarditis) Serum C3 Reduced levels indicate complement activation. Commonly reduced in active SLE, usually in combination with reduced C4 level, indicating classical pathway activation Low C3 with normal C4 indicates predominant alternative pathway activation. This is rare, and causes include: • Poststreptococcal glomerulonephritis • C3 glomerulopathy (e.g. dense deposit disease) CH100 Functional test in which classical pathway activation is triggered in vitro and terminal pathway activation assessed by measurement of the membrane attack complex either directly (using anti-​C5b-​9 antibodies) or indirectly (using red cell lysis). Test calculates % of test sera required to cause 100% terminal pathway activation Normal CH100 requires an intact classical pathway, functional C3, and intact terminal pathway Important to screen for complement deficiency AP100 Functional test in which alternative pathway activation is triggered in vitro and terminal pathway activation assessed by measurement of the membrane attack complex either directly (using anti-​C5b-​9 antibodies) or indirectly (using red cell lysis) Normal AP100 requires intact alternative pathway, functional C3, and intact terminal pathway. Important test to screen for complement deficiency C1 inhibitor (C1INH) assays Standard assays test both antigenic and functional activity of C1INH Essential test to detect type I and type II hereditary angioedema Specialized assays Comments Anti-​C1q antibodies Associated with SLE nephritis but not routinely measured in clinical practice Indicated in suspected hypocomplementaemic urticarial vasculitis syndrome (HUVS) C3 nephritic factor Represents an autoantibody that enhances the C3 convertase of the alternative pathway Associated with enhanced alternative pathway activation Indicated in atypical haemolytic uraemic syndrome and C3 glomerulopathy Anti-Factor H autoantibodies Associated with enhanced alternative pathway activation Indicated in atypical haemolytic uraemic syndrome and C3 glomerulopathy Mutation screening for structural and sequence variation in complement genes encoding alternative pathway proteins and regulators Indicated in patients with atypical haemolytic uraemic syndrome prior to renal transplantation

324 SECTION 4  Immunological mechanisms such as Streptococcus pneumoniae and Neisseria meningitidis. The hypocomplementaemia, in addition to causing defective opsoniza- tion, results in reduced splenic clearance of these organisms. These patients can be thought of as functionally asplenic and we recom- mend prophylactic penicillin therapy together with pneumococcal and meningococcal vaccination. There is currently one complement inhibitor, eculizumab, that is in use in routine clinical practice. Eculizumab is a monoclonal antibody that blocks C5 activation, preventing the formation of the anaphylatoxin C5a and the generation of the MAC. It is licensed for the treatment of transfusion-​dependent haemolytic anaemia in PNH, generalised myasthenia gravis, and aHUS. It is admin- istered intravenously and its use is associated with increased sus- ceptibility to Neisserial infection. This is a predictable side-​effect since terminal pathway deficiencies are associated with increased susceptibility to this infection and eculizumab induces an acquired C5 (and therefore terminal pathway) deficiency state. Patients re- ceiving this agent must therefore be immunized against Neisserial strains and treated with prophylactic antibiotic therapy. C1 inhibitor deficiency can be treated by the intravenous admin- istration of C1 inhibitor preparations. Two are derived from plasma, Berinert P and Cinryze. A third is a recombinant C1 inhibitor mol- ecule, Rhucin. In some of the rare cases of complement activation protein deficiency, fresh frozen plasma has been used as a source of the missing component. For example, in complete C1q deficiency associated with a SLE-​like illness, short or long-​term plasma in- fusions have been beneficial. In severe cases, bone marrow trans- plantation has been used to achieve permanent restoration of C1q levels since C1q is synthesized by bone marrow-​derived cells. FURTHER READING Bordron A, et al. (2019). Complement system: a neglected pathway in immunotherapy. Clin Rev Allergy Immunol, doi: 10.1007/ s12016-019-08741-0. Conigliaro P, et al. (2019). Complement, infection, and autoimmunity. Curr Opin Rheumatol. doi: 10.1097/BOR.0000000000000633. Degn SE, Jensenius JC, Thiel S (2011). Disease-​causing muta- tions in genes of the complement system. Am J Hum Genet, 88, 689–​705. Lambris JD, Ricklin D, Geisbrecht BV (2008). Complement evasion by human pathogens. Nat Rev Microbiol, 6, 132–​42. Merle NS, et al. (2015). Complement system part I—​molecular mech- anisms of activation and regulation. Front Immunol, 6, 262. Merle NS, et al. (2015). Complement system part II—​role in immunity. Front Immunol, 6, 257. Thurman JM, Yapa R (2019). Complement therapeutics in autoimmune disease. Front Immunol, 10, 672. doi: 10.3389/fimmu.2019.00672. eCollection 2019.