Deficiencies of the complement MAC II gene cluster (C6, C7, C9): is subtotal C6 deficiency of particular evolutionary benefit?

COMPLEMENT ACTIVATION

Complement is, apart from other key functions, an important humoral innate immune defence system against invading microorganisms. Its activation via the classical, lectin or alternative pathway leads to the formation of bioactive anaphylatoxins, opsonic fragments and the potentially lytic membrane attack complex (MAC) on the surface of target membranes [1–3].

TERMINAL COMPLEMENT PATHWAY

After cleavage of C5 by either the classical or the alternative C5 convertase the terminal complement components C6, C7, C8 and C9 are sequentially activated [1–3]. The ability of C5b to bind C6 decays rapidly, but once bound C5b6 forms a stable bimolecular complex. C5b6 binds C7 and, if local C7 concentrations are limiting, the stable bimolecular C5b6 complex dissociates from the C5 activating complex and accumulates in solution. In the presence of C7 fluid phase C5b-7 is formed which has a transient ability to secondarily attach to target membranes in the vicinity initiating ‘reactive lysis’[4].

Both membrane-bound C5b-7 and fluid-phase C5b-7 are capable of binding C8β via C5b. Polymerization of C9 is initiated by C5b-8 binding to C9 via C8α, enabling formation of an elongated molecule and insertion of further C9 molecules into the target membrane causing local distortion of the phospholipid bilayer resulting in ‘leaky patches’[5] or forming a hydrophilic channel (‘pore’) through the membrane [6].

The main biological functions of the terminal complement cascade extend far beyond those originally described. On host nucleated cells, complement activation is often sublytic [7], which offers some protection to the cell as it can withstand single (and erroneous) attacks, unlike erythrocytes which are readily lysed. Furthermore, previous sublytic effects exerted on nucleated cells protect from further, otherwise lytic, attack [8], favouring those host cells which are constantly in contact with complement. Sublytic attack not only protects host cells, it also stimulates their protein biosynthesis and acts in a proinflammatory [7] and procoagulant manner [9]. Even a cytolytically inactive TCC has been reported to be able to activate endothelial cells [10].

Formation of the terminal complement complex is controlled by a number of different cellular and plasma molecules, reviewed elsewhere [1–3,11]. It is often overlooked that C7 [11] and C8 [12] are important modulators as well, as their local presence or absence may have a more profound effect than the absence or presence of the specific complement inhibitors. This is probably especially true for C7, which can be primarily synthesized extrahepatically at the site of inflammation by granulocytes and endothelial cells, thereby modulating lytic or sublytic membrane attack [11].

DEFICIENCIES OF THE TERMINAL COMPLEMENT PATHWAY AND DISEASE

Looking at the multiple functions of the TCC, it is no surprise that the TCC has been implicated in an impressive number of diseases, evident via its detection in diseased tissues or its elevated levels in the blood [1–3,7], and that so many different clinically relevant microorganism have adopted numerous different techniques to escape the destructive action of complement [13]. On the other hand it is still a mystery why deficiency of a single terminal complement component is usually compatible with life and probably the majority of homozygous carriers remain undetected for most of their life [14,15]. The reason for this very probably lies in the high redundancy of the immune system which, however, is not complete.

Deficiencies of the terminal cascade predispose to meningococcal infections, indicating that its cytolytic properties are of particular importance in host defence against Neisseria [14,15], for which there appears to be no sufficient back-up. Neutrophils can only provide partial protection, although they can kill meningococci when incubated in terminal complement deficient serum, and this effect increases after vaccination [16] so that this may represent an alternative treatment option in addition to, or instead of, vigilance and antibiotic prophylaxis.

Typically, terminal complement deficient subjects present in adolescence or in young adulthood, and suffer from recurrent meningococcal infections with especially the rarer serogroups. By way of example, the index case of a report in this journal issue by Vazquez-Bermudez et al.[17] was ascertained because of three meningococcal disease episodes in a 15-year-old-boy.

Due to the clear association of autoimmune disease, in particular SLE, and deficiency of components of the classical pathway [1–3], several million patients have been assayed for complement deficiency and terminal complement deficiency cases have also been found. Given that roughly 1 in 10 000 individuals is deficient in one of these components, although there are huge geographical and ethnical differences (e.g. C9D in Japan: 1 in 1000 subjects [15,18]), it is now well accepted that there is no association between autoimmune disease (or non-Neisserial infections) and terminal complement deficiency [15,19]. Heterozygosity does not lead to disease [15].

PROTEINS, GENES AND POLYMORPHISMS OF C6, C7 AND C9

C6, C7 and C9 share not only functional similarities as members of a functional unit, the terminal complement cascade, they are also biochemically and structurally similar [20] and share a common ancestor, which was very likely a C6-like and not a C9-like molecule [21]. Despite gene duplications and deletions [21], the genes for C6, C7 and C9 remain located in the same chromosomal region (5p13) [20] in the MAC II gene cluster [22] (unlike the C8A and C8B genes which are located at 1p32 (MAC I) [22]).

Several polymorphisms have been detected for the MAC II gene cluster genes, listed in detail elsewhere [20] (note typing error: the C7 m/N polymorphism is due to 1759 A/C [23] not 1769 A/C as listed in [20], Dr B. Sanchez, personal communication), or published recently for C6 [24,25], C7 [26–28] and C9 [29]. The whole set of alleles can form informative haplotypes whereby complement deficiency genes can be traced within families [20,26], also applied for the case in this journal issue [17].

MUTATIONS LEADING TO COMPLETE DEFICIENCIES (D) OF C6, C7 AND C9

Five deletions on more than 70 independently ascertained chromosomes [20,24,25,30] have been reported leading to complete C6D and 13 different C7 mutations (on 30 chromosomes) to C7D (Fig. 1) [17,20,31]. Some of these deficiencies may actually turn out to be subtotal as discussed below. Whereas there appears to be a cluster for mutations for C7D at the 3′ end [20], those for C6D are more or less scattered around the gene with possibly a hot spot region in exon 6 [24,25]. For C9D four different mutations have been detected in three European subjects, one of which is the R95X stop codon common in the Japanese (Fig. 1) [18].

Fig. 1.

Locations of all known C6, C7 or C9 mutations leading to deficiencies (to June 2003) in relation to the modules (indicated in the box). Deletions (?), boundary mutations, generating nonsense intron sequence till a stop codon is reached (#), and amino acid substitutions (A111B) are indicated. In addition, mutations leading to subtotal deficiency are boxed.

SUBTOTAL DEFICIENCIES (SD) OF C6, C7 AND C9

Several cases of C6SD either alone (Germany), or combined with C7SD (UK) or as compound heterozygote with complete C6D (South Africa) have been described. In all cases the C6 was detectable at approximately 1% of the normal concentration, able to incorporate into the TCC, bactericidally and haemolytically active but structurally abnormal (about 20% smaller with the same abnormal isoelectric focusing (IEF) pattern) [32]. Not surprisingly, the same mutation was found in all subjects: a mutated 5′ splice donor site of intron 15, preventing splicing and thus ‘activating’ an in frame stop codon located 17 codons downstream, leading to a truncated polypeptide 13·5% shorter than normal C6 [20]. Most of the carboxyterminal two factor I modules are missing (Fig. 1), comparable to normal mouse C6A (which is functionally active). Most importantly, no meningococcal disease was reported for any of the 10 C6SD or C6SD/C7SD subjects.

The C7 in the C6SD/C7SD subjects was found to be functionally active at a concentration of approx. 5% of normal, and despite its normal size, also showed an abnormal IEF pattern, caused by a R499S substitution (Fig. 1) [20]. Interestingly, this mutation was also found in a subject from Russia who had had 2 meningococcal episodes. Because of constant complement activation due to chronic otitis media, in this individual, the high amounts of circulating C5b6 generated consumed his low C7 levels, so he was phenotypically completely C7D and only after a blood transfusion, leading to a marked clinical improvement, was his own C7 detectable in the circulation [33], as assessed via the C7 M/N polymorphism [23]. The lack of C5b6 in the unactivated transfused serum together with sufficient exogenous C7 to bind the free patient C5b6 allowed restoration and maintenance of a low but detectable level of his own de novo synthesized C7 for a few days post-transfusion until his C5b6 level rose again as a result of his chronic infection. In contrast, due to the additional low C6 – and thus C5b6 – concentration in C6SD/C7SD subjects their C7 was not consumed.

Interestingly, one subject from Malta with a R198Q exchange in C7 and another yet unknown C7 defect had low levels of possibly two different circulating C7 molecules (as judged by IEF) despite the presence of C6 and thus C5b6 (Fig. 1) [20]. There are two main explanations for this: (i) the C7 concentration is high enough to withstand constant consumption, or (ii) the C7 is dysfunctional (i.e. it does not bind to C5b6 or C8).

Functionally active C9 molecules have been found at 0·2% of the normal mean in a compound heterozygous C9SD subject, i.e. two C9 mutations were discovered [20]. Following enrichment and epitope-mapping using monoclonal antibodies it became evident that the circulating molecule is not lacking any major parts and thus is the product of C98G (exon 4), producing normal length C9 but lacking one cysteine-residue, and not of the S406X premature stop codon in exon 9, leading to a carboxyterminal truncation of about 25% (Fig. 1) [34]. In this abnormal C9, the two adjacent intramolecular disulphide bridges in the LDL-repeat probably stabilize the conformation even in the absence of the third. Interestingly, the Irish propositus and her affected sister (in their 50 s) suffered severe diabetes but a sister had died in youth of meningitis, probably meningococcal [34].

BENEFICIAL ASPECTS OF TERMINAL COMPLEMENT DEFICIENCY

The inability to generate a functionally active membrane attack complex may be of advantage, e.g. in autoimmune diseases [15]. Meningococcal disease may also be less severe in terminal complement deficient subjects [14], due to a lesser extent of microbial disease damage and thus less severe gram-negative endotoxin shock [35]. This has earlier been put forward as possibly being of selective advantage, especially in areas where infantile gastro-enteritis is relatively common [36], and may account for the high frequency of the C9 R95X mutation in Japan [18]. In contrast, the high incidence of deceased siblings with symptoms typical for meningococcal disease in affected families somewhat argues against this hypothesis. Unlike C6SD or C6SD/C7SD, C7SD or C9SD will, due to C5b6- or C5b8-triggered consumption, both phenotypically and clinically behave like a complete deficiency, with a likelihood of meningococcal disease (similarly, it cannot be fully excluded that the present C7 case [17] is actually subtotal). However, in view of the finding that less than 10% of the normal C6 level provides quite sufficient complement activity [32], and that none of the C6SD or C6SD/C7SD subjects suffered from meningococcal disease, C6SD may indeed be beneficial by providing enough but not too much lytic activity.