THE MEMBRANE ATTACK COMPLEX (MAC)

Cell injury by complement occurs as a consequence of activation of either the classical or the alternative pathway on the surface of a cell. The "killer" molecule is the MAC (74,75). The MAC constitutes a supramolecular organisation that is composed of approximately twenty protein molecules and has a molecular weight of approx. 1.7 million. The fully assembled MAC contains one molecule each of C5b, C6, C7, and C8 and one or more molecules of C9. These are the five precursors present in the MAC, all of which are hydrophilic glycoproteins. When C5 is cleaved by C5 convertase and nascent C5b (C5b*) is produced, self-assembly of the MAC follows. C5b* and C6 form a stable and soluble bimolecular complex which binds to C7 and induces it to express a metastable site through which the nascent trimolecular complex (C5b-7*) can insert itself into membranes, when it occurs on or in close proximity to a target lipid bilayer. Insertion is mediated by hydrophobic regions on the C5b-7 complex that appear following C7 binding to C5b-6 (76). Membrane-bound C5b-7 commits MAC assembly to a membrane site and forms the receptor for C8. The binding of one C8 molecule to each C5b-7 complex gives rise to small transmembrane channels of less than 1nm functional diameter (77) that may perturb target bacterial (78) and erythrocyte membranes (76). Each membrane-bound C5b-8 complex acts as a receptor for multiple numbers of C9 molecules and appears to facilitate insertion of C9 into the hydrocarbon core of the cell membrane. The concentration of C9 in human serum is such that there are only about two molecules of C9 for every C8 molecule (79), therefore C5b-9 complexes generated on target membranes display a degree of heterogeneity with regard to C9 content (80,81). Binding of one molecule of C9 initiates a process of C9 oligomerisation at the membrane attack site, after at least 12 molecules (82) are incorporated into the complex, a discrete channel structure is formed. Therefore the end product consists of the tetramolecular C5b-8 complex (approximately 550kD) and tubular poly-C9 (approximately 1100kD). This form of the MAC, once inserted into the cell membranes, creates complete transmembrane channels leading to osmotic lysis of the cell. The transmembrane channels formed vary in size depending on the number of C9 molecules incorporated into the channel structure (83). Whereas the presence of poly-C9 is not absolutely essential for the lysis of red blood cells or of nucleated cells (84), it may be necessary for the killing of bacteria (85).

Evidence suggests that the proteins participating in the transmembrane channel formation are structurally interrelated. The notion of a complement supergene was first proposed in this context when a close linkage between loci for C6 and C7 was demonstrated through family studies of their genetic polymorphism (86), and C6 and C; were described as similar proteins (87). After further explorations of these relationships between the terminal components of the complement system it has become clear that all five proteins, C6, C7, C8, C9, and C9RP (also called "cytolysin", "perforin", or "pore-forming protein"), share certain antigenic properties (88,89). C9-related protein (C9RP) is the protein responsible for pore-forming activity and has been isolated from murine (90,91) and human (92,93) cells. Because the protein interacts with C9, it has been called C9-related protein, a term synonymous with cytolysin (94), perforin (an effector molecule of killer T-cells and NK cells) (95), and pore forming protein (96). Although C9 and C9RP are similar, probably homologous proteins, and may be analogous in their function, they differ in that isolated C9RP is cytotoxic by itself, whereas isolated C9 is not. Under conditions promoting homopolymerisation, C9RP kills cells without the participation of other proteins. For C9 to exert its cytotoxic effect it requires cell-bound C5b-8 (91).

An, as yet, unmentioned protein of the MAC is the S protein. S protein is the primary inhibitor of serum (97,98). It competes with membrane lipids for the metastable binding sites of C5b-7 and allows the binding of C8 and C9, but prevents C9 polymerisation; in fact, by binding to the complex, it prevents the attachment of C5b-7 to the cell surface (99). Its function appears to be to protect the cells adjacent to sites of complement activation from accidental attack. The hydrophilic complex, SC5b-7, binds C8 and three C9 molecules to form SC5b-8 and SC5b-9 (100,101). All three complexes contain neoantigens that are not detectable in the precursor proteins and that are distinct from the neoantigens of poly-C9. In addition to blocking the membrane binding site, S protein, as mentioned earlier, also prevents polymerisation of C9. These functions allow S protein to control the formation of the MAC. As well as S protein, the multifunctional protein, SP-40,40, has been discovered which modulates the soluble membrane attack complex (SMAC, SC5b-9) of complement, causes cell aggregation, and accelerates immune complex formation. SP-40,40 was discovered as a soluble protein present in serum, seminal plasma and cerebrospinal fluid. It was reported that SP-40,40 modulated the formation of the MAC of complement and was incorporated into SMAC at the stage of SC5b-7. When S protein or SP-40,40 binds to C5b-7 before C5b-7 binds to the membrane, SC5b-7, which is hydrophilic is formed. Finally, soluble SC5b-9 is formed via the bindings of C8 and C9. In the bindings of both S protein and SP-40,40 to C5b-7, the hydrophobic interactions seem to be a major force. In addition to S protein and SP-40,40, lipoproteins have been reported as inhibitors of MAC formation (102).

A recently discovered cell surface antigen, CD59, has been found to be an inhibitor of complement-mediated lysis. The function of CD59 was first suggested by the finding that the purified antigen inhibits complement-mediated lysis by binding in a glycosylation-dependent manner to C5b-8 and/or C5b-9 preventing the formation of the MAC. Comparisons of CD59 with other complement inhibitors, decay accelerating factor, and membrane cofactor protein, indicate that CD59 is the most potent inhibitor of complement-mediated lysis of human endothelial cells. In addition to inhibiting complement, CD59 can transmit activating signals to T cells and form a part of a signal-transducing complex on the surface of these cells. It has been proposed that CD59 is also a second ligand for the human T lymphocyte adhesion molecule, CD2, but this is controversial. In the short period since its discovery, considerable progress has been made in understanding the function of CD59. While it is clear that CD59 protects tissues from attack by the complement system, the molecular basis of this process remains to be elucidated (103).

CONTROL OF MAC CHANNEL FORMATION BY MEMBRANE-BOUND PROTEINS

One protein that exhibits a marked affinity for C9 has been isolated from the membranes of human red blood cells (RBCs) (104). It was capable of incorporating into the lipid bilayer of liposomes, and in this form was active in inhibiting channel formation by C5b-8, C5b-9, and polymerising C9. Antibody produced to this membrane protein caused a 20-fold increase in reactive lysis of human RBCs by isolated C5b-6, C7, C8, and C9. The antibody did not affect C5b-7 uptake, but enhanced C9 binding to the target cell membrane. The antibody effect was not seen when C8 and C9 of other species were tested. The protein was, therefore, termed "homologous restriction factor" (HRF) (104). HRF is a 65kD protein which, like DAF, is anchored to cell membranes via a glycan linkage to phosphatidylinositol (105) . Being a membrane protein, HRF is not soluble in aqueous solvents. HRF is detectable on human RBCs, platelets, neutrophils, monocytes, and lymphocytes (106).

Yet another protein that controls MAC channel formation has been isolated from human RBC membranes (107). Termed the "membrane inhibitor of reactive lysis" (MIRL), this 18- to 20kD protein was found capable of restricting the assembly of C5b-9 on target cell membranes (107,108). MIRL is identical with the leukocyte antigen CD59 (108), is expressed on endothelial cells (109), and is released, in part, from cells after treatment with phosphatidylinositol-specific phospholipase C (108,110). Antibodies to MIRL render normal cells susceptible to reactive lysis (107-109).