Chapter 5 The Complement System

The Regulation of Complement Activity

All biological systems with the potential to damage the host are subject to rigorous regulatory mechanisms, and the complement system is no exception to this rule. Especially in light of the potent positive feedback mechanisms and the absence of antigen specificity of the alternative pathway, mechanisms must exist to ensure that the destructive potential of complement proteins is confined to the appropriate pathogen surfaces and that collateral damage to healthy host tissues is minimized.

Here, we discuss the different mechanisms by which the host protects itself against inadvertent complement activation. Protection of vertebrate host cells against complement-mediated damage is achieved by both general, passive regulatory mechanisms and specific, active regulatory mechanisms.

Complement Activity Is Passively Regulated by Short Protein Half-Lives and Host Cell Surface Composition

The relative instability of many complement components is the first means by which the host protects itself against extended periods of inadvertent complement activation. For example, the C3 convertase of the alternative pathway, C3bBbC3b, has a half-life of only 5 minutes, unless it is stabilized by reaction with properdin. A second passive regulatory mechanism depends on the difference in the cell surface carbohydrate composition of host versus microbial cells. For example, fluid-phase proteases that destroy C3b bind much more effectively to host cells that bear high levels of sialic acid, than to microbes that have significantly lower levels of this sugar. (We encountered this mechanism in the section “Complement and T Cell–Mediated Immunity,” in which we described complement’s role in ensuring the destruction of improperly developed T cells). Hence, any C3b molecule that happens to alight on a host cell is likely to be degraded before it can cause significant damage.

In addition to these more passive environmental brakes on inappropriate complement activation, a series of active regulatory proteins work to inhibit, degrade, or reduce the activity of complement proteins and their fragments on host cells. The stages at which complement activity is subject to regulation are illustrated in Figure 5-16, and the regulatory proteins are listed in Table 5-6.

An illustration shows five stages of regulation of complement activity. — FIGURE 5-16 Regulation of complement activity. The various stages at which complement activity is subject to regulation are shown (see text for details).

"Part a, Dissociation of C 1 complements, shows 18 polypeptide chains in six collagen-like triple helices bound by a ring with multiple spheres, labeled C 1 r subscript 2 s subscript 2. It reacts in the presence of C 1 I N H to disintegrate into polypeptide chains (labeled C 1 q) and a ring with spheres (labeled C 1 r 2 s 2).

Part b, decay-accelerating activity of C 3 convertases, shows two processes. In the classical pathway, a pair of connected spheres, labeled C 4 b 2 a, react in the presence of D A F (C D 55), C R 1 (C D 35), and C 4 B P to break apart into C 4 b and C 2 a. In the alternative pathway, a pair of connected spheres labeled C 3 b B b reacts in the presence of D A F (C D 55), C R 1 (C D 35), and Factor H to break apart into C 3 b and B b.

Part c, Factor I cofactor activity, shows two processes. In one process, a sphere labeled C 3 b reacts in the presence of Factor I, M C P (C D 46), C R 1 (C D 35), and Factor H to yield three fragments: C 3 c, i C 3 b, and C 3 d g. In the other process, a sphere labeled C 4 b reacts in the presence of Factor I, M C P (C D 46), C R 1 (C D 35), and C 4 B P to yield two fragments: C 4 c and C 4 d.

Part d, Inhibition of lysis, shows M A C (consisting of C 5 b, C 6, C 7, and C 8) reacting in the presence of C D 59(Protectin) and Vitronectin/S protein to form a protective structure on the cell membrane; M A C inhibition prevents C 9 binding and polymerization.

Part e, Cleavage of the anaphylatoxins, shows C 3 a and C 5 a reacting in the presence of carboxypeptidases N, B, and R to yield C 3 a des A r g, and C 5 a des A r g."

TABLE 5-6 Proteins involved in the regulation of complement activity
Protein	Soluble or membrane bound	Pathway(s) affected	Function
C1 inhibitor (C1INH)	Soluble	Classical and lectin	Induces dissociation and inhibition of C1r₂s₂ from C1q; serine protease inhibitor
Decay-accelerating factor (DAF; CD55)	Membrane bound	Classical, alternative, and lectin	Accelerates dissociation of C4b2a and C3bBb C3 convertases
CR1 (CD35)	Membrane bound	Classical, alternative, and lectin	Blocks formation of, or accelerates dissociation of, the C3 convertases C4b2a and C3bBb by binding C4b or C3b Cofactor for factor I in C3b and C4b degradation on host cell surface
C4BP	Soluble	Classical and lectin	Blocks formation of, or accelerates dissociation of, C4b2a C3 convertase Cofactor for factor I in C4b degradation
Factor H	Soluble	Alternative All pathways	Blocks formation of, or accelerates dissociation of, C3bBb C3 convertase Cofactor for factor I in C3b degradation
Factor I	Soluble	Classical, alternative, and lectin	Serine protease: cleaves C4b and C3b using cofactors shown in Figure 5-16
Membrane cofactor of proteolysis, MCP (CD46)	Membrane bound	Classical, alternative, and lectin	Cofactor for factor I in degradation of C3b and C4b
S protein (vitronectin)	Soluble	All pathways	Binds soluble C5b67 and prevents insertion into host cell membrane
CD59 (protectin)	Membrane bound	All pathways	Binds C5b678 on host cells, blocking binding of C9 and the formation of the MAC
Carboxypeptidases N, B, and R	Soluble	Anaphylatoxins produced by all pathways	Cleave and inactivate the anaphylatoxins C3a and C5a

The C1 Inhibitor, C1INH, Promotes Dissociation of C1 Components

C1INH, the C1 inhibitor, is a plasma protein that binds in the active site of serine proteases, effectively poisoning them. C1INH belongs to the class of proteins called serine protease inhibitors (serpins), and it acts by forming a complex with the protease C1r₂s₂, causing it to dissociate from C1q and preventing further activation of C4 or C2 (see Figure 5-16a). C1INH inhibits both the classical pathway serine protease and that of the lectin pathway, MASP-2. It is the only regulatory protein capable of inhibiting the initiation of both the classical and lectin complement pathways, and its presence in plasma serves to limit the time period during which they can remain active.

Decay-Accelerating Factor Promotes Decay of C3 Convertases

Since the reaction catalyzed by the C3 convertase enzymes is the major amplification step in complement activation, the generation and lifetimes of the two C3 convertases, C4b2a and C3bBb, are subject to particularly rigorous control. The membrane-bound decay-accelerating factor, or DAF (CD55), accelerates the decay of the C4b2a C3 convertase on the surface of host cells. In order to complete its job, DAF requires the cofactors CR1 and C4BP (C4-binding protein) (see Figure 5-16b). These decay-accelerating proteins cooperate to accelerate the breakdown of the C4b2a complex into its separate components. C2a, inactive in the absence of C4b, diffuses away, and the residual membrane-bound C4b is degraded by another regulatory protein, factor I (see Figure 5-16c).

In the alternative pathway, DAF and CR1 function in a similar fashion. However, in place of C4BP they are joined by factor H in separating the C3b component of the alternative pathway C3 convertase from its partner, Bb (see Figure 5-16b). Again, inactive Bb diffuses away, and residual C3b is degraded (see Figure 5-16c).

Whereas DAF and CR1 are membrane-bound components and their expression is therefore restricted to host cells, factor H and C4BP are soluble cofactors of regulatory complement components. Host-specific function of factor H is ensured by its binding to negatively-charged cell surface carbohydrates such as sialic acid and heparin, which are essential components of eukaryotic, but not prokaryotic, cell surfaces. Similarly, C4BP is preferentially bound by host cell membrane proteoglycans such as heparan sulfate. In this way, host cells are protected from the deposition of complement components; in contrast, microbial invaders that lack DAF and CR1 expression and fail to bind factor H or C4BP are completely vulnerable to complement-mediated attack. However, as we will see, sometimes microbes hijack these mechanisms that are designed to ensure specificity of protection for host cells, and use them instead to protect themselves.

Factor I Degrades C3b and C4b

Factor H, C4BP, and CR1 also figure as co-factors in a second mechanism of complement regulation: that catalyzed by factor I. Factor I is a soluble, constitutively (always) active serine protease that can cleave membrane-associated C3b and C4b into inactive fragments (see Figure 5-16c).

However, if factor I is indeed soluble, constitutively active, and designed to destroy C3b and C4b, one might wonder how the complement cascades ever succeed in destroying invading microbes? The answer, once again, is that factor I requires the presence of these same, host cell–specific cofactors in order to function. Hence, cleavage of membrane-bound C3b on host cells is conducted by factor I in collaboration with the membrane-bound host cell proteins MCP and CR1, and the soluble cofactor factor H. Similarly, cleavage of membrane-bound C4b is achieved again by factor I, this time in collaboration with membrane-bound MCP and CR1 and soluble cofactor C4BP. Since these membrane-bound, or membrane-binding, cofactors are not found on microbial cells, C3b and C4b are thus destroyed if they alight on host cells, but are allowed to remain on microbial cells and exert their specific functions.

Recently, six proteins related to factor H have been identified with varying levels of complement-regulatory activity. Their activities and regulation are currently under careful investigation. Interestingly, genetic variations of factor H and its related proteins have been associated with chronic inflammatory diseases such as age-related macular degeneration.

Variation in MCP expression has recently been implicated as a factor in the control of apoptosis followed by phagocytosis of dying T cells. When a T cell commits to apoptosis, it expresses DNA on its cell membrane that binds circulating C1q, as described earlier. It then begins to shed MCP from the cell surface. Only after MCP is lost can progression of the classical pathway occur, resulting in opsonization by C3b and eventual phagocytosis of the apoptotic T cells.

CD59 (Protectin) Inhibits the MAC Attack

In the case of a particularly robust antibody response, or of an inflammatory response accompanied by extensive complement activation, inappropriate assembly of MACs on healthy host cells can potentially occur, and mechanisms have evolved to prevent the resulting inadvertent host cell destruction. A host cell surface protein, CD59 (protectin), binds any C5b678 complexes that may be deposited on host cells and prevents their insertion into the host cell membrane (see Figure 5-16d). CD59 also blocks further C9 addition to developing MACs. In addition, the soluble complement S protein, otherwise known as vitronectin, binds any fluid-phase C5b67 complexes released from microbial cells, preventing their insertion into host cell membranes.

Both CD59 and DAF are membrane-associated molecules that are attached to the lipid bilayer via glycosylphosphatidylinositol anchors. Rare mutations in the gene PIGA prevent afflicted individuals from attaching these anchors, and so these patients suffer from dysregulation of complement activity. The development of drugs that specifically target complement components and help to address the symptoms of individuals suffering from complement-related diseases is described in Clinical Focus Box 5-3.

CLINICAL FOCUS BOX 5-3

The Complement System as a Therapeutic Target

The involvement of complement-derived anaphylatoxins in inflammation makes complement an interesting therapeutic target in the treatment of inflammatory diseases, such as arthritis. In addition, autoimmune diseases that potentially result in complement-mediated damage to host cells, such as multiple sclerosis and age-related macular degeneration, are also potential candidates for complement-focused therapeutic interventions.

The last three decades have seen the crystallization and molecular characterization of several complement components, a necessary precursor to the development of designer drugs targeted to specific proteins. At the time of writing, three therapies directed at interfering with complement components are approved for clinical use, with several more in either clinical or preclinical trials. The therapies already in the clinic are either antibodies or parts of antibodies that specifically target and regulate the complement proteins C5 (two different drugs) or factor D.

Until now, the most successful complement-related treatment is specifically directed toward a disease that stems from a dysregulation of the complement cascade. Paroxysmal nocturnal hemoglobinuria (PNH) manifests as increased fragility of erythrocytes, leading to chronic hemolytic anemia, pancytopenia (loss of blood cells of all types), and venous thrombosis (formation of blood clots). The name PNH derives from the presence of hemoglobin in the urine, most commonly observed in the first urine passed after a night’s sleep. The cause of PNH is a general defect in the synthesis of cell surface proteins, which affects the expression of two regulators of complement: DAF (CD55) and CD59 (protectin).

DAF and CD59 are cell surface proteins that function as inhibitors of complement-mediated cell lysis, acting at different stages of the process. DAF induces dissociation and inactivation of the C3 convertases of the classical, lectin, and alternative pathways (see Figure 5-16). CD59 acts later in the pathway by binding to the C5b678 complex and inhibiting C9 binding, thereby preventing formation of the membrane pores that destroy the cell under attack. Deficiency in these proteins leads to increased sensitivity of host cells to complement-mediated lysis and is associated with a high risk of thrombosis.

CD59 and DAF are attached to the cell membrane via glycosylphosphatidylinositol (GPI) anchors, rather than by stretches of hydrophobic amino acids, as is the case for many membrane proteins. Most patients with PNH lack an enzyme subunit, called phosphatidylinositol glycan class A (PIG-A), that attaches the GPI anchors to the appropriate proteins. The gene encoding PIG-A is found on the X chromosome, and this part of the X chromosome is silenced in females, meaning that each cell has only one copy of the gene. A defect in this copy therefore means that patients lack the expression of both DAF and CD59 needed to protect erythrocytes against lysis. The term paroxysmal refers to the fact that episodes of erythrocyte lysis are often triggered by stress or infection, which result in increased deposition of C3b on host cells. Episodes of hemoglobinuria result in dark red urine, and since urine is most concentrated first thing in the morning, the term nocturnal was applied to indicate that the release of hemoglobin into the urine was occurring overnight.

The defect identified in PNH occurs early in the enzymatic pathway leading to formation of a GPI anchor and resides in the PIGA gene. Transfection of cells from PNH patients with an intact PIGA gene restores the resistance of the cells to host complement lysis. Further genetic analysis revealed that the defect is not encoded in the germline genome (and therefore is not transmissible to offspring), but rather is the result of mutations that occurred within the hematopoietic stem cells themselves, such that any one individual may have both normal and affected cells. Those patients who are most adversely affected display a preferential expansion of the affected cell populations.

PNH is a chronic disease with a mean survival time post diagnosis of between 10 and 15 years. The most common causes of mortality in PNH are blood clots that affect veins in the liver, and progressive bone marrow failure.

A breakthrough in the treatment of PNH was reported in 2004; a humanized monoclonal antibody that targets complement component C5 was used to inhibit the terminal steps of the complement cascade and formation of the membrane attack complex. This antibody—eculizumab (Soliris)—was infused into patients, who were then monitored for the loss of red blood cells. A dramatic improvement was seen in patients during a 12-week period of treatment with eculizumab (Figure 1). Treatment of PNH patients with eculizumab relieves hemoglobinuria, reverses the kidney damage resulting from high levels of protein in the urine, and significantly reduces the frequency of thromboses. In 2007, eculizumab was approved for the treatment of PNH in the general population.

FIGURE 1 Treatment of PNH patients with eculizumab relieves hemoglobinuria. Shown are the number of days with paroxysms (sudden onsets) per patient per month in the month before treatment (left column) and for a 12-week period of treatment with eculizumab (right column). [Data from P. Hillmen, et al. 2004. Eculizumab in patients with paroxysmal nocturnal hemoglobinuria. New England Journal of Medicine 350:552.]

Since the control of infections with meningococcal bacteria (Neisseria meningitidis) relies on an intact and functioning membrane attack complex (MAC), patients being treated with eculizumab are routinely vaccinated against meningococcus.

The pathology of PNH underscores the potential danger to the host that is inherent in the activation of the complement system. Intricate systems of regulation are necessary to protect host cells from the activated complement complexes generated to lyse intruders, and alterations in the expression or effectiveness of these regulators have the potential to result in a pathological outcome.

The centrality of C3 within the complement cascades may suggest that it would be an excellent target for therapeutic intervention in complement-mediated disease. However, its very position at the crossroads of the three pathways means that interference with C3 activity places patients at increased risk for infectious diseases normally curtailed by the activities of the innate as well as the adaptive immune systems, and so systemic drugs specifically targeting C3 may prove to have too high a risk-benefit ratio.

However, scientists have recently succeeded in developing C3-targeted drugs that specifically address complement-mediated diseases that attack particular organs. One drug, compstatin, was developed as a result of experiments designed to discover compounds that bound specifically to C3. Chemical refinements based on structural and computational studies followed, and a compstatin derivative, POT-4, is now in phase 2 clinical trials for the treatment of age-related macular degeneration, a progressive and debilitating eye disease that leads inexorably and quickly to total blindness. Because POT-4 can be delivered directly into the vitreous humor of the eye, systemic side effects on the patient’s immune system are minimized. A second C3-directed drug, AMY-101, is now in preclinical trials and has been granted orphan status by both the European Medicines Agency and the U.S. Food and Drug Administration for treatment of C3 glomerulonephropathy (C3G), a rare disease that affects kidney glomeruli and is caused by dysregulation of the alternative pathway of complement activation.

Carboxypeptidases Can Inactivate the Anaphylatoxins C3a and C5a

Anaphylatoxin activity is regulated by cleavage of the C-terminal arginine residues from both C3a and C5a by serum carboxypeptidases, resulting in rapid inactivation of their anaphylatoxin activity (see Figure 5-16e). Carboxypeptidases are a general class of enzymes that remove amino acids from the carboxyl termini of proteins; the specific enzymes that mediate the control of anaphylatoxin activity are carboxypeptidases N, B, and R. These enzymes remove arginine residues from the carboxyl termini of C3a and C5a to form the so-called des-Arg (“without arginine”), inactive forms of the molecules. In addition, as mentioned above, binding of C5a by C5L2 also serves to modulate the inflammatory activity of C5a.