© AJC 2007.
| MicrobiologyBytes: Infection & Immunity: Complement | Updated: October 21, 2004 | Search |
BACTERIA
Early reports have shown that bacterial LPS (lipopolysaccharide) as well as gram-negative bacteria are able to activate the complement system via the alternative pathway leading to a consumption of the five terminal complement components (153, 154, 155). At this early time, it was unclear which components of the bacterial membrane are involved in the binding of C1. Therefore, it was of advantage to have bacterial strains of a defined surface composition. Serum-resistant smooth (S) and serum-sensitive rough (R) strains of S.minnesota and S.typhimurium were chosen. Various experiments reported that the binding of C1 to antigen-antibody complexes occurred via the C1q subcomponent of C1 (156). Further observations confirmed that C1q is bound only by the serum-sensitive rough form and not the serum-resistant smooth form of S.minnesota (157,158). Therefore, it is thought that the O antigenic sugar chains prevent binding of C1q and C1 to the bacterial surface of the smooth form. The membrane components such as lipid A and outer membrane proteins are accessible on rough forms, which may cause the tight C1 binding to these bacteria. Experiments such as the above, confirm reports that the S forms of Gram-negative bacteria tend to be less sensitive than the R forms; however, the R bacteria have been found to survive in the absence of any classical complement component.
Reports by Betz and co-workers (159,160) suggested that E.coli J5 activates complement in human serum via the classical complement pathway. In further studies, it was demonstrated that a binding and activation of C1 occurs in the absence of antibodies (161). Such membrane-bound C1 can initiate the serum-bactericidal reaction. Horwitz and Silverstein (162) suggest that unencapsulated E.coli strains can also undergo phagocytosis independently of antibody, since bacteriolysis also requires C3 activation and C3b binding. Thus direct binding of C1 to some bacteria may well contribute to host defence against gram-negative infection at both a cellular and a humoral level.
It has been hypothesised that many lesions formed on the bacterial surface might not directly induce killing unless they were formed at precise targets, such as where the inner membrane is bound tightly with the outer membrane (163). Consequently, the significance of certain observations, regarding the above hypothesis, depends upon the subsequent effect of activated complement components on the bacterium. Although specific antibodies (IgG, IgM) may enhance C1 binding by E.coli J5 (160), the ability of these additional C1 molecules to alter later events in the complement cascade may depend on the control of C1 activation and its subsequent activity when bound to different membrane components.
The lipopolysaccharide (LPS) of gram-negative bacteria has been reported to be a potent activator of the alternative complement pathway (164,165,166). However, it has also been demonstrated that the first component of complement (C1) interacts directly via its subcomponent C1q with LPS and lipid A (167,168). In addition, a direct interaction with C1 has been shown for E.coli, K.pneumoniae (169), Mycoplasma pneumoniae (170), and retroviruses (171).
In the late 1960's, it was of interest to find out which bacterial surface structures are involved in C1 and C1q binding. Lipopolysaccharides (LPS) are the most accessible molecules in the outer membrane of gram-negative bacteria. Besides its endotoxic activity, LPS had also been shown to interact with the complement system in several ways. Some investigations have shown that incubation of cell wall LPS of gram-negative bacteria with whole serum led to a decrease of the complement activity of this serum (171,172). It has been demonstrated that the loss of whole complement activity in serum was due to a consumption of C3 to C9, but LPS had a minimal effect on C1, C2, and C4 (173); the consumption of C3 to C9 by LPS was shown to be a result of activation of the alternative pathway (174). Besides the LPS induced activation of the alternative pathway, it is reasonable that LPS-antibody complexes activate the classical complement sequence via C1, C4, and C2. In addition to these possibilities, there may exist a direct interaction of LPS with the first three components of complement resulting in binding and activation or inhibition of these components. Therefore, the influence of different LPS preparations on purified C1, C4, and C2 was tested.
The results of such tests showed that different LPS preparations had no significant inhibitory effect on C4 and C2. In contrast, LPS inhibited C1 up to 98% depending on the preparation used. The inhibition of C1 by LPS is a quick, temperature-independent reaction (172,175).
Evidence suggests that porins, present on the surface of gram-negative bacteria, bind C1q and C1. The association of LPS and porins may potentiate the C1q and C1 binding. This antibody-independent binding of C1 to LPS and porins is a prerequisite for the activation of the classical pathway of complement leading to the killing of serum-sensitive gram-negative bacteria (176) .
Neisseria meningitidis is an encapsulated gram-negative bacterium and a major causative agent of bacterial sepsis and meningitis. The meningococcal capsules, which consist of acidic polysaccharides, are assumed to be important virulence factors. Most capsular polysaccharide serotypes found to be associated with meningococcal disease are composed of neuraminic acid (sialic acid),which mediates resistance to phagocytosis and complement-mediated bacteriolysis (177). In N.meningitidis the capsule is composed of thick polysialic acid which is assumed to be the crucial virulence factor in meningococci (177), because activation of the alternative pathway is known to be inhibited by sialic acids on cell surfaces. Inhibition of the alternative complement pathway occurs by an enhanced binding of complement factor H to cell-surface deposited C3b, resulting in the inhibition of the C3b convertase by dissociation of C3b and factor Bb (178). Furthermore, factor H is a co-factor for factor I, which cleaves cell-bound C3b to form inactivated C3bi (179). In 1994 Hammerschmidt et al (180) showed that expression of the polysialic acid capsule alone is not sufficient to prevent meningococci from being killed via alternative complement pathway activation. In contrast, capsule-deficient meningococci, which are able to sialylate the meningococcal lipo-oligosaccharide (LOS), survive in C4-deficient guinea-pig serum. This indicates that the capsule does not form an impermeable layer around the meningococci, and that complement factors can penetrate into,
and be activated at, the membrane, thus causing cell lysis; although this only occurs if sialic acid is absent from the LOS. Therefore it seems that LOS sialylation may be more important than the polysialic acid capsule in conferring resistance to complement-mediated killing via alternative pathway activation on meningococci (180).
Because of the important role of the capsule in the pathogenicity of meningococcal disease, great efforts have been undertaken to use the capsular polysaccharide as a vaccine candidate. Present studies have indeed implied that the meningococcal LOS might well be an alternative candidate for the development of a vaccine against meningococcal disease. Future studies intend to show whether anti-LOS antibodies are able to prevent sialylation of the LOS. Alternatively, inactivation of the LOS-sialyltransferase by vaccination could prevent sialylation of the LOS, thus making the meningococcus susceptible to the host immune system.
Resistance to complement-mediated killing is an attribute of another Neisseria sp., Neisseria gonorrhoeae, especially those strains that are responsible for disseminated gonococcal infection (181). The terminal complement components are of particular importance in host defence against N.gonorrhoeae infection, as has been clearly shown by rare individuals with isolated terminal complement component deficiencies who have a marked propensity for disseminated gonococcal infections (182). Gonococcal infection of the human genital tract involves bacterial colonisation of the mucosal surface, invasion and disruption of the mucosal barrier, and in most cases of a symptomatic exudative inflammatory response (183). Phase variation of N.gonorrhoeae lipopolysaccharide (LPS) controls both bacterial entry into human mucosal cells, and bacterial susceptibility to killing by antibodies and complement. Many bacterial pathogens, including N.gonorrhoeae, are endowed with sophisticated mechanisms to adapt to a rapidly changing microenvironment of the host (184).
Evidence is provided (185) that structural phase variation of N.gonorrhoeae LPS may act as an adaptive mechanism enabling the pathogen both to enter into mucosal epithelial cells and to resist the initial human immune defence, both prerequisites for establishing infection. The biological significance of the phenotypic variation first became apparent in the presence of the natural sialic acid donor, CMP-NANA, which can be used by the gonococci to sialylate LPS. In the presence of this compound, intrinsic LPS variation allows a differential sialylation of the expressed LPS molecules, and thus a reversible switching between an invasive, barely sialylated phenotype that is susceptible to complement-mediated killing; and a non-invasive, highly sialylated phenotype that is neither killed by antibodies nor complement.
The transition to a gonococcal phenotype that resists killing by complement and phagocytes, and is poorly immunogenic upon the development of an inflammatory response, may explain the lack of protective immunity against the bacterium in the natural disease. The current concept that LPS phase transitions are an essential feature for bacteria to cross the mucosal barrier and resist the host inflammatory response by differential sialylation, requires that the sialylation event occurs in vivo during the natural infection. Several lines of evidence suggest that this is indeed the case. Such studies (186,187,188) have shown that in the initial stage of infection, only non-sialylatable and thus potentially invasive and complement-killing susceptible LPS variants are isolated; whereas after the development of an inflammatory response, other highly sialylatable LPS phenotypes appear, which represent an immuno-resistant gonococcal phenotype. All such findings indicate that the present function of structural phase variation of gonococcal LPS may be only the first example of a more generally applicable virulence mechanism which allows bacteria to adapt to the rapidly changing host environment during the different stages of a natural infection.
LPS variation is a common characteristic of many other mucosal pathogens, including N.meningitidis, H.influenzae, and B.pertussis, and in vivo infection data suggests that this variation is indeed linked to virulence.
In chronic infections, such as Bronchopulmonary Pseudomonas aeruginosa Infection in Cystic Fibrosis (CF) patients, bacteria persist despite an intact host immune defence and frequent antibiotic treatment. An important reason for the persistence of the bacteria is their capacity for a biofilm mode of growth. The role of biofilms in the activation of complement has thus been investigated (189). Complement activation by P.aeruginosa, a gram-negative enteric bacterium, was inhibited by the chemical polymyxin B, this indicates that LPS is the main mediator of complement activation. The formation of immune complexes and the massive influx of neutrophils are known to cause inflammatory changes in the lungs of patients infected with P.aeruginosa It seems that the P.aeruginosa persisting in biofilms may contribute to the constant inflammation taking place in the lungs of CF patients.
Complement activation by the bacteria could have both positive and negative implications for the chronically infected patient. A positive consequence would be the eradication of the typically serum-sensitive P.aeruginosa from the CF patient. However, live and dead bacteria, as well as bacterial fragments, activate complement by themselves, and the concomitant active production of specific antibodies and immune complex formation would add to this complement-mediated inflammation. This constant inflammation means that the destruction of pulmonary tissues ensues. The data from various experiments (189) does seem to suggest that the bacteria persisting in biofilms mediate a constant low-grade complement activation and thereby contribute to the chronic inflammation seen in these patients. Actually, it is not the released LPS from the biofilm or the planktonic bacteria, but the LPS still within the biofilm matrix or associated with the bacterial surface that mediates the complement activation.
Various evidence suggests that other Pseudomonas products and strains may be implicated in complement activation, for example, P.elastase has been shown to enzymatically generate a chemotactic activity from C5, and to inactivate complement components.
Group B Streptococcus type III (GBS) is a gram-positive bacterium and is a major cause of neonatal death. The group B Streptococcus type-specific capsule possesses a terminal sialic acid moiety which has been shown to be an important virulence factor (190). As stated
earlier, sialic acid groups contribute to virulence by inhibiting alternative complement pathway activation. Thus, sialylation plays a critical role in the bacterial evasion of host defences particularly in newborns in whom specific antibodies are not yet developed (191).
Another gram-positive Streptococcus is Streptococcus pyogenes Group A. This organism is a major cause of cellulitis and pharyngitis, and an occasional cause of bacteremia. Sequelae of infection include rheumatic fever and glomerulonephritis, which are thought to be related to a vigorous immune response to the bacterium. The virulence of this organism is dependent on the presence of M protein in the bacterial cell wall (192). This protein apparently confers on the Group A Streptococcus the ability to resist phagocytosis. Strains of Group A Streptococci lacking M protein are readily phagocytosed after opsonisation by C3 deposited by the alternative pathway (193). Thus, M protein interferes with complement opsonisation of these organisms. In fact, it appears, from studies by Jack-Weis et al (194,195) that M protein interferes with complement opsonisation both by inhibiting C3 binding to the organism and by inhibiting the interaction between bound C3b and phagocyte C3 receptors.
Haemophilus influenzae, a gram-negative bacterium, accounts for a broad range of infections among young children. The encapsulated organisms, predominantly H.influenzae type b, cause diseases associated with bacteremia (196). The unencapsulated, nontypeable strains cause localised respiratory tract diseases, such as otitis media, sinusitis, and pneumonia (197) . Although much remains to be learned about host defence against H.influenzae type b and nontypeable H.influenzae, the importance of complement has become increasingly clear, for example, patients with complement deficiencies are more susceptible than others to H.influenzae type b infection (198). It appears that the opsonic rather than the bactericidal activity of complement mediates the clearance of H.influenzae type b bacteremia (199,200). The participation of antibody to outer membrane proteins in host defence and complement activation is well known (201,202). In fact it has recently been demonstrated that such antibody is more important than anticapsular antibody in promoting complement component 3 (C3) binding to H.influenzae type b in the unimmunized adult (203).
Hetherington et al (204) have recently identified the major outer membrane proteins, P1 and P2, as the sites of C3 fragment binding during opsonisation of H.influenzae type b. The significance of C3 binding to these proteins, plus the demonstration by others that anti-P1 and anti-P2 antibodies are bactericidal, raise the possibility that either outer membrane protein could be useful as a vaccine for nontypeable H.influenzae (205,206). Several other protein bands have been identified as C3b covalently bound to other as-yet-unidentified H. influenzae type b proteins . Thus, other epitopes important to host defence against H.influenzae have yet to be discovered. Further studies of complement activation on the H.influenzae type b and nontypeable H.influenzae surfaces could identify additional protein candidates for vaccines as well as expand knowledge of complement-mediated host defence.
Listeria monocytogenes is a gram-positive facultative intracellular pathogen that can cause severe infections in humans and animals. The outcome of infection depends on many properties of both host and bacteria, and some of the latter have recently been unravelled by transposon mutagenesis and gene cloning (207,208,209). Although protective immunity against this organism is mediated by Listeria-specific T-cells and activated macrophages, serum factors including the complement system may play a role in the early stages of listeriosis. Components of the Listeria cell wall activate complement via the alternative pathway, thereby releasing chemotactic peptides for phagocytic cells (210) and possibly favouring their invasion via complement receptors for C3 fragments (211). However, the extent and regulation of complement activation by various strains of L.monocytogenes and related strains are poorly documented.
Direct evidence was presented by Croize et al (212) that L.monocytogenes activates the alternative pathway of human complement, leading to the binding of C3b and its cleavage products, C3b and C3bi, to the outer membrane of the organism, through both ester and amide linkages. This is in agreement with reports on other gram-positive bacteria; and those quantitating the two forms of C3, C3b and C3bi, covalently bound to the bacterial surface have generally shown that C3b is the predominant form (213).
In a report of work with strains of Streptococcus pneumoniae differing in capsular polysaccharides and hence serotype, Hotstetter (214) stressed the importance of the bacterial surface in determining the amount and the site of covalently bound C3b, as well as its degradative processing to C3bi and C3d. Studies on whether peptidoglycan or teichoic acid is responsible for alternative pathway activation has given controversial results, probably as a consequence of harsh conditions required to purify these cell wall components that may alter their native linking molecular conformations (215). In the case of L.monocytogenes, the peptidoglycan may be an important acceptor structure. Proteins from the Listeria surface do not seem to play a major role, as deduced from the absence of effect on C3 deposition and consumption of their enzymatic stripping, although the rough mutant form of L.monocytogenes, which is strongly altered in the protein structure of the cell wall (215), proved to be less effective than other strains in activating the alternative pathway, with higher susceptibility of deposited C3b to proteolytic processing. In keeping with previous reports concerning gram-positive bacteria, is the resistance of L.monocytogenes to complement-mediated lysis, occurring probably because the thick cell wall prevents access of complement components C5-C9 to the inner membrane (213). Several studies have suggested that heat-labile opsonic factors from serum, presumably complement, may participate with macrophages in the cleaving of L.monocytogenes from infected foci (216,217,218). In a study by Croize et al (212), the capability of this pathogen to activate solely the alternative pathway was examined in the setting of undetectable specific antibodies. This does not preclude a role for the classical pathway in Listeria-immune individuals. The relative contributions of antibodies and complement to L.monocytogenes opsonisation is unclear. This may be explained by variability in experimental conditions, especially serum concentration and animal origin, immune status of the serum donor, and the type of phagocytic cells used.
Moraxella (Branhamella) catarrhalis is a gram-negative diplococcus related to the neisseriae and frequently found as a commensal organism in the upper respiratory tract (219). The bacterium, however, can cause infections such as acute otitis media, sinusitis, and conjunctivitis, in otherwise healthy children and elderly people (220). Moreover, M.catarrhalis is an important cause of lower respiratory tract infections, particularly in adults with chronic obstructive pulmonary disease (221). In immunocompromised hosts, the bacterium can cause a variety of severe infections including pneumonia, endocarditis. septicaemia and meningitis (221,220,219).
It has been shown that complement is an important system in host defence against infection with neisseriae (222). This is reinforced by the notion that individuals with inherited complement deficiencies, particularly those with terminal complement component deficiencies, have a markedly increased risk (approx. 8000 fold) of acquiring neisserial infections (222,223). A second indication for the role of complement in antineisserial defence is the expression of complement resistance among strains involved in systemic infections (222). This strongly suggests that the complement resistance of neisseriae, which has been shown to be a multifactorial phenomenon (224,225,226,227), is an important virulence factor.
Because of the close relationship between M.catarrhalis and neisseriae, and the increasing appreciation of the pathogenic potential of the former, attention has been drawn to the complement-activating ability of M. catarrhalis Recent results of studies done on the resistance of M.catarrhalis strains show that only 10% of clinical isolates obtained from adults are actually complement sensitive (228). In contrast, most isolates (58%) from healthy school children are sensitive to complement-mediated killing. This strongly suggests that complement resistance is a virulence factor of M.catarrhalis.
Various experiments have been carried out to determine the differences in complement activation of complement-resistant and complement-sensitive M.catarrhalis strains (229). The observation that treatment of the bacteria with trypsin changed the phenotype from complement-resistant to complement-sensitive, indicated that complement resistance is mediated by a proteinaceous surface component, although this only occurred at higher bacterial numbers, which is in line with the idea that complement is not inhibited by either of the two activation pathways but at the stage of generation of the membrane attack complex. The finding that neuraminidase treatment of resistant M. catarrhalis does not alter its complement activation or resistance to killing, indicates that the mechanism behind complement resistance in M.catarrhalis does not involve sialic acid and is therefore different from that found in gonococci and meningococci. One of the intriguing possibilities is that the M. catarrhalis-associated protein binds a terminal complement inhibitor present in serum (vitronectin or clusterin) and in this way facilitates the functional inactivation of membrane attack complexes at the bacterial surface. This possibility is currently being studied.
Therefore the complement activation patterns of complement-sensitive and complement resistant M.catarrhalis strains differ greatly. The difference is located at the level of the membrane attack complex formation rather than in one of the activation pathways. Complement-resistant strains seem to express a protein that is capable of binding membrane attack complexes without causing harm to the bacterium. Whether this protein inhibits directly or functions as an acceptor for vitronectin, for example, remains to be elucidated.