Role of complement and perspectives for intervention in ischemia-reperfusion damage

Abstract

Reperfusion of an organ following prolonged ischemia instigates the pro-inflammatory and pro-coagulant response of ischemia / reperfusion (IR) injury. IR injury is a wide-spread pathology, observed in many clinically relevant situations, including myocardial infarction, stroke, organ transplantation, sepsis and shock, and cardiovascular surgery on cardiopulmonary bypass. Activation of the classical, alternative, and lectin complement pathways and the generation of the anaphylatoxins C3a and C5a lead to recruitment of polymorphonuclear leukocytes, generation of radical oxygen species, up-regulation of adhesion molecules on the endothelium and platelets, and induction of cytokine release. Generalized or pathway-specific complement inhibition using protein-based drugs or low-molecular-weight inhibitors has been shown to significantly reduce tissue injury and improve outcome in numerous in-vitro, ex-vivo, and in-vivo models. Despite the obvious benefits in experimental research, only few complement inhibitors, including C1-esterase inhibitor, anti-C5 antibody, and soluble complement receptor 1, have made it into clinical trials of IR injury. The results are mixed, and the next objectives should be to combine knowledge and experience obtained in the past from animal models and channel future work to translate this into clinical trials in surgical and interventional reperfusion therapy as well as organ transplantation.

Key words::

• Complement
• ischemia
• reperfusion

Abbreviations
C1-INH	=	C1-esterase inhibitor
C5b-9	=	terminal complement complex
CABG	=	coronary artery bypass graft
CRP	=	C-reactive protein
CR1	=	complement receptor 1
DXS	=	dextran sulfate
IR injury	=	ischemia / reperfusion injury
MASP	=	MBL-associated serine protease
MBL	=	mannose-binding lectin
sCR1	=	soluble recombinant complement receptor 1 (TP10)
sCR1-sLex	=	sCR1 conjugated to sialyl Lewis x (TP20)

Key messages

• Complement activation and complementmediated damage play an important role in ischemia / reperfusion injury.

• Complement inhibition may be generalized or pathway-specific, systemic or local, and target the fluid phase complement components, their receptors, or influence complement regulatory proteins.

• Future work must concentrate on bringing together all knowledge and data from preclinical work to push forward clinical trials on complement regulation in ischemia / reperfusion injury.

The complement system

The complement system is composed of more than 30 plasma and cell membrane proteins, many of which interact with one another in enzyme activation cascades. As part of the innate immune defense, complement is not able to differentiate between self and non-self. Rather, it is designed to recognize ‘danger’ signals, such as bacterial cell walls and altered cell surfaces. The three complement activation pathways (classical, alternative, and lectin routes), the role of complement receptors, and the anaphylatoxins C3a and C5a, as well as the natural regulation of complement activation by membrane-bound and soluble inhibitors, have been reviewed extensively in the past (1,2). The present review therefore focuses on the role of complement in ischemia / reperfusion (IR) injury and novel, complement-based therapeutic strategies for the prevention of IR injury.

Ischemia / reperfusion injury

Reperfusion of tissues or organs following prolonged ischemia is a double-edged sword. Although reperfusion is critical for tissue salvage, it may itself initiate a local and systemic inflammatory response, further aggravating ischemic damage. IR injury is therefore defined as cellular damage following reperfusion of previously viable ischemic tissue. IR injury is observed in many clinically relevant situations, including myocardial infarction, stroke, transplantation, sepsis and shock, and cardiovascular surgery on cardiopulmonary bypass.

The pathophysiology of IR injury is characterized by a complex cascade of pro-inflammatory and pro-coagulatory events. Whilst some authors consider radical oxygen species to be the key to reperfusion injury, and inhibition of their formation the solution to its prevention, others favor the role of leukocytes, particularly polymorphonuclear cells. Radical oxygen species cause tissue injury via direct damage to cellular membranes through lipid peroxidation (3) as well as through leukocyte activation and chemotaxis. Much of the data obtained on the role of radical oxygen species is derived from in-vitro studies and ex-vivo cardiac perfusion experiments with oxygenated buffers (Krebs-Henseleit buffer), suggesting that oxygen free radicals are the prime triggers of IR injury (see Hess et al. for a review (4)). Indeed, the use of superoxide dismutase was associated with a positive outcome in animal models (5). However, clinical trials to prevent IR injury in transplantation using recombinant human superoxide dismutase have so far not proven very successful (6). Release of radical oxygen species, proteases, and elastases by activated leukocytes leads to increased microvascular permeability, edema, thrombosis, and cell death (7). Inhibition of leukocyte recruitment and rolling and expression of adhesion molecules by anti-adhesion molecule therapy has been shown to reduce tissue injury in animal models (8) as well as in human studies, e.g. using the K-ATP channel opener Nicorandil (9). More recent research efforts have also concentrated on the vascular endothelium, in particular its luminal protective glycocalyx layer, a carbohydrate-rich lining connected to the endothelium through several so-called ‘backbone’ molecules, mainly proteoglycans and glycoproteins. For a recent comprehensive review see Reitsma et al. (10). Reperfusion appears to be critical for injury to the endothelial glycocalyx (11), and inhibition of glycocalyx modifications upon reperfusion attenuates IR damage (12). Whilst all of these components are relevant to IR injury, it is likely that the part played by each varies from one organ system to the next.

Complement in ischemia / reperfusion injury

The relevance of complement activation in IR injury was already proposed over four decades ago. In human subjects, complement deposition was revealed in autopsy samples from myocardial infarction, often in combination with C-reactive protein (CRP) (13). Immunohistochemical co-localization of CRP and IgM in human myocardial infarction suggests neoepitopes (see also below) exposed in the process of IR injury, leading to antibody-triggered activation of the classical complement pathway (14). Indeed, the classical pathway has been shown to be of importance in IR injury in several organs (15). On the other hand, the alternative pathway appears equally important in IR injury in a range of organs, including intestinal and renal IR injury (16,17). Here, it is also involved in amplifying the response mediated by the lectin pathway.

Indeed, whilst particularly the classical complement pathway was initially thought to be at the center of complement-mediated injury in IR, more recently the importance of the lectin pathway of complement activation has been recognized. This has been highlighted in several papers in experimental settings using rats and knock-out mouse models by the group of Stahl (18,19). In a model of gastrointestinal IR injury mannose-binding lectin (MBL) A/C and C2/factor B, but not C1q knock-out animals were protected from reperfusion-induced damage. In the same model pulmonary injury secondary to gastrointestinal IR injury was not prevented by MBL deficiency, and only the C2/factor B-deficient animals were also resistant to this type of injury. The authors therefore hypothesized that ficolins, proteins containing both a collagen-like domain and a fibrinogen-like domain, may play a role in initiating complement-mediated pulmonary injury (19). Finally, a double knock-out model with mice lacking MBL as well as secretory IgM provided evidence that both IgM as well as the lectin pathway are required for myocardial IR injury (20). The jury is still out as to whether the lectin pathway is of importance for the development of IR injury in species other than rodents, but increasing evidence suggests that MBL plays a key part in IR injury in other settings. Castellano et al. just published a study on renal IR injury in pigs in which this conclusion was drawn (21). Also, clinical studies show MBL deficiency to be advantageous for patients in cardiovascular surgery, myocardial infarction, and intestinal IR injury (22).

The exposure of neoepitopes in complement activation upon reperfusion has been recognized as being central to initiating tissue damage. The group of Michael Carroll presented a murine model of mesenteric IR injury in which natural IgM binds to neoepitopes induced on cell surfaces upon ischemia. In this model, the target epitope of IgM was identified as non-muscular myosin heavy chain type II A and C (23). Whilst these data are all based upon experiments in mice, the group nevertheless was able to show that human IgM as well as human peripheral blood mononuclear cells restore IR injury in severe combined immunodeficiency mice, which are otherwise protected (24). Other studies by Fleming et al. and Kulik et al. have identified additional neoepitopes exposed upon intestinal IR injury in mouse models, including negatively charged phospholipids and annexin V and the corresponding circulating natural antibodies (25,26). Similarly, a role for natural anti-ribonucleoprotein antibodies has been described in intestinal IR injury in RAG1 (-/-) animals (27).

The modes of action of complement in IR injury, following initiation of a specific pathway or pathways, are as diverse as the triggers kicking off its primary activation. Selective C1q accumulation is observed in ischemic-reperfused myocardium, where its concentration inversely correlates with regional myocardial blood flow (28). Fixation of C1q to subcellular fractions of myocardial cells, released upon ischemic damage, activates the complement cascade and stimulates infiltration of polymorphonuclear leukocytes (29).

The potent anaphylatoxin C5a is also crucially implicated in the pathogenesis of IR injury by influencing neutrophil-dependent and -independent pathways, including the regulation of CXC chemokines, but not TNF-alpha or apoptotic pathways (30). Furthermore, C5a induces up-regulation of adhesion molecules, such as α-integrin, β2-integrin, and Mac-1 in polymorphonuclear leukocytes, as well as expression of P-selectin in endothelial cells (31).

C5b-9, the terminal complement complex, induces expression of the adhesion molecule P-selectin on platelets and endothelial cells, where, in the setting of IR on cardiopulmonary bypass, it appears to be the major determinant of platelet activation (32). Similar to iC3b, formed after cleavage of C3b, the terminal complement complex may also alter endothelial cell function and promote leukocyte adhesion by activating endothelial NF-kappa B (33). Indeed, the interaction of complement with the endothelium plays an important part in the pathogenesis of IR tissue damage (34).

Complement inhibition

Complement may be inhibited in a generalized manner, for instance by the administration of cobra venom factor leading to consumption-induced depletion. However, this is not a feasible option in the clinical setting. Alternatively, an individual pathway may specifically be targeted. Inhibition may be systemic or local and target the fluid phase complement components, their receptors, or influence complement regulatory proteins.

Currently most of the complement inhibitors used in research are essentially protein-based drugs. Often these substances prove expensive to produce, difficult to formulate, and may be hazardous due to their possible immunogenic potential. Low-molecular-weight drugs, including short peptides and synthetic small molecules, may attest more advantages in this regard. In the following section some chosen studies of complement inhibition in experimental settings followed by currently available clinical data will be discussed (for a summary see Table I and Figure 1).

Table I. Summary of selected references in preclinical and clinical trials of complement inhibition in ischemia / reperfusion injury.

Animal models
Inhibitor	Model	IR injury	References
sCR1	rat	heart	(35)
sCR1-sLe^x	mouse, baboon, rat	brain, brain, heart	(36–38)
Compstatin	pig-to-human xeno-tx	kidney	(39)
anti-C5a, minibody to C5	pig, rat allo-tx, rat	heart, heart	(40,41,82)
C5aR-antagonist, C5aR-siRNA	rat, rat, mouse syn-tx, mouse	kidney, kidney, kidney, brain, kidney	(42–46)
C3aR-antagonist	rat, mouse	intestine, brain	(47,48)
C1s-inhibitor	rabbit	heart	(50)
mimotope/anti-non-muscle myosin, anti-ribonucleoprotein	mouse, mouse	heart, intestine	(27,54)
GAGs	rabbit, rabbit	heart, heart	(56,57)
C1-INH	dog, pig, mouse, mouse, pig, pig	heart, heart, brain, brain, kidney, heart	(15,21,61–63,72)
	CD59^−/− mouse, CD55^−/− /CD59^−/mouse tx, C6^−/− mouse syn-tx, C6^−/− mouse	brain, kidney, liver, kidney	(34,64–66)
Clinical trials
Inhibitor	Setting	Clinical situation	References
Humanized anti-C5 (Eculizumab, Pexelizumab)	clinical	PNH, CABG, PTCA post MI, thrombolysis post MI, CABG, CABG/valve surgery, PTCA post MI, review	(67,75–81)
C1-INH (Cinryze), C1-INH	clinical, clinical rescue, clinical, clinical, emergency	PNH, emergency CABG post PTCA, heart, CABG post MI	(68–71)
sCR1 (TP10)	clinical	CABG, CABG/valve, lung-tx	(83–85)

sCR1 = soluble complement receptor 1; sCR1-sLe^x = sCR1 conjugated to sialyl Lewis x; C1-INH = C1-inhibitor; C5a/C3aR = C5a/ C3a receptor; siRNA = small interfering ribonucleic acid; GAGs = glycosaminoglycans; xeno-/allo-/syn-tx = xenogeneic, allogeneic, and syngeneic transplantation; PNH = paroxysmal nocturnal hemoglobinuria; CABG = coronary artery bypass surgery; PTCA = percutaneous transluminal coronary angioplasty; MI = myocardial infarction.

Figure 1. Summary of selected aspects of complement inhibition in ischemia / reperfusion injury. GAGs (glycosaminoglycans) are released from the cell surface upon reperfusion. Functional replacement of shed GAGs through DXS (low-molecular-weight dextran sulfate) that inhibits all three complement pathways. C5a/C3aR-antagonists = C5/C3a receptor antagonists. Members of the lectin pathway: MBL = mannose binding lectin; MASP = MBL-associated serine protease. Exposure of neoepitopes with subsequent binding of natural antibodies upon ischemia and reperfusion. C4b2a3b and C3bBb3b = classical- and alternative-pathway C3 convertases. TCC = terminal complement complex (C5b-9); CD55/59 = complement regulatory proteins for surface inhibition of the TCC.

Complement receptor 1/C3 inhibition

CR1 functions by displacing the catalytic subunits from the C3 and C5 convertases and equally serves as a co-factor for the degradation of C3b and C4b by factor I. Twenty years ago, Weisman et al. developed recombinant, soluble human CR1 (sCR1), which lacks the transmembrane and cytoplasmic domains, and demonstrated its complement-inhibitory and anti-inflammatory properties in a rat model of myocardial IR injury (35). In a murine model of ischemic brain injury, sCR1 and its sialyl Lewis x coupled variant sCR1-sLe^x conferred neuroprotection (36). However, in a further study by the same group, sCR1-sLe^x was evaluated in a primate stroke model with no difference in neurological score between the treatment and placebo groups. Indeed infarct size was larger in the treated group, and animals reacted with a hypotensive response, despite significant reduction of classical pathway complement activation (37). No such detrimental effects were observed in other animal studies, where sCR1-sLe^x exclusively proved beneficial (38).

Similarly to sCR1, Compstatin, a 13-residue cyclic peptide, potently inhibits activation of C3. This compound has already been shown to be effective in several models of complement-mediated tissue injury. For example, it attenuates hyperacute rejection in a pig-to-human kidney ex-vivo perfusion xenotransplantation model (39). However, to date no data have been published on the effect of Compstatin on IR injury.

Inhibition of the anaphylatoxins C3a and C5a and their receptors

The anaphylatoxins C3a and C5a are potent inflammatory mediators and possess marked hemodynamic effects. Inhibition at this level offers an excellent strategy to influence a multitude of harmful consequences of complement activation. Using a monoclonal antibody against C5a, Amsterdam et al. revealed early on that C5a inhibition and subsequent reduction of neutrophil aggregation, chemotaxis, and superoxide generation decreased myocardial infarct size in pigs (40). Also in IR injury in the setting of transplantation surgery, inhibition of C5a release and C5b-9 formation by a neutralizing minibody to C5 effectively prevented IR-associated graft injury in a model of cardiac allotransplantion (41).

The cyclic compound AcF-[OPdChaWR], a specific C5a complement receptor antagonist, administered prior to ischemia, substantially reduced IR-induced hematuria, vascular leakage, and serum creatinine levels as well as TNF-alpha tissue levels (42). Similarly, in a murine transplantation model, kidneys were significantly protected from IR-mediated damage when A8D71-773, a C5a receptor antagonist that targets both receptors, i.e. CD88 and C5L2 (43), was added to the University of Wisconsin solution during cold ischemia (44). Also in cerebral IR injury in a murine model, infarct volume was significantly decreased in animals treated with C5a receptor antagonist (45).

The technique of gene silencing using small interfering RNA is a valid alternative to fluid phase complement inhibition by antibodies. In a recent study in mice by Zheng et al., silencing of the C5a receptor efficiently inhibited C5aR gene expression both in vitro and in vivo and preserved renal function following IR injury (46).

In a rat model of intestinal IR injury both C3a and C5a receptor antagonists proved beneficial. However, the C3a receptor antagonist probably functioned through inhibition of global neutrophil sequestration in tissue rather than direct C3a receptor antagonism (47). C3a receptor inhibition improved neurological outcome and reduced brain edema and inflammatory cell infiltration in a mouse model of hemorrhagic stroke (48).

Serine protease inhibition

Other possible targets include the serine proteases involved in pathway initiation (C1r, C1s, MASPs, and C2), amplification (factors B and D), and regulation (factor I) (49). Both the alternative (C3bBb) and classical (C4b2a) C3 convertases generate the potent pro-inflammatory anaphylatoxins C3a and C5a implicated in IR injury. Buerke and colleagues report of the use of a highly selective small molecule C1s inhibitor (C1s-INH-248, Knoll) that blocks the classical complement pathway to preserve ischemic myocardium from IR injury in a rabbit model of myocardial infarction (50).

Inhibition of IgM-induced complement activation

Work in complement receptor 2-deficient mice, which only have a restricted repertoire of natural IgM (51), as well as recombination activation gene-deficient (RAG-/-) mice, deficient of antibody, has revealed protection from IR damage (52,53), suggesting a role for circulating natural IgM in reperfusion-mediated damage. Monoclonal and polyclonal antibodies against negatively charged phospholipids were able to reconstitute IR injury in complement receptor 2-deficient mice (25). Just of late Haas et al. proved this point by showing a 47% reduction in infarct size in a murine model using a synthetic peptide mimotope (N2) or monoclonal antibodies directed against the self-antigen non-muscle myosin heavy chain II (54). The mode of action relates to reduced IgM binding and complement activation. Similarly, natural anti-ribonucleoprotein antibodies titer-dependently reconstituted tissue damage in intestinal IR injury in RAG1(-/-) animals (27). Such results may be useful in the future to offer other possible novel targets for therapeutic intervention.

Glycosaminoglycan derivatives

Early data from the mid 1970s show inhibition of the enzymatic activity of the first complement component C1 by heparin (55). However, heparin alone is not sufficient to prevent IR injury in the clinical setting. Furthermore, the danger of possible hemorrhage may be real, particularly if high doses are used to achieve the desired complement-inhibition. Heparin derivatives (N-acetyl heparin) and glycosaminoglycan analogs (pentosan sulfate, sulodexide) with limited anti-coagulant properties were therefore developed and shown to limit IR injury in vivo (56,57).

C1-inhibitor in preclinical models of ischemia/reperfusion injury

A large amount of preclinical work in animal models, particularly in the field of myocardial IR injury, has been done—mostly with notable success—using C1-esterase inhibitor (C1-INH). C1-INH, initially described as the only physiological inhibitor of the classical complement pathway (58), has been shown to inhibit differentially all three complement pathways at high physiologic concentrations—a fact that may have therapeutic implications (59,60).

In a feline model of acute myocardial IR injury systemic C1-INH administered 10 minutes prior to reperfusion improved recovery of cardiac contractility, preserved coronary endothelial function, and significantly reduced cardiac polymorphonuclear leukocyte accumulation within the ischemic area, resulting in decreased myocyte necrosis (61). Two years later it was shown that also local, intracoronary C1-INH reduced myocardial damage in a porcine model of acute myocardial infarction (62). In a mouse model of acute cerebral ischemia and reperfusion, C1-INH significantly decreased infarct volume (15). The inhibitory effect of C1-INH on the lectin and alternative pathways may be of direct relevance to cerebral IR injury, where the classical complement pathway is possibly less prominently involved, as shown by the study of C1q-deficient mice that are not protected from ischemic damage and where C1-INH nevertheless exerted a positive effect (15).

Recently, human recombinant C1-INH was found restricted to cerebral vessels, co-localizing with MBL and reducing complement activation in an acute model of cerebral ischemia in mice (63). With its distinct glycosylation pattern, it proved superior in its protective effect as compared to the corresponding plasmatic protein, with a clearly improved time window for application, probably because it bound to intravascularly localized MBL deposited on the endothelial cell surface. Also in a porcine model of renal IR injury, C1-INH inhibited the lectin and classical complement pathways, attenuating complement-mediated damage and tissue deposition (21).

Inhibition of the terminal complement complex

Recent data suggest that the terminal complement complex plays an important role in IR injury. In a murine model of middle cerebral artery occlusion, male mice null for CD59a (the terminal complement complex inhibitor) experienced a significant increase in tissue damage, neurological deficits, and death as compared to their wild-type counterparts following 30 minutes of ischemia and 72 hours of reperfusion (64). However, the observed effects were dependent upon the exact model and gender of the animals used. Also in renal transplantation, deficiency of both membrane complement regulatory proteins CD55 and CD59 greatly exacerbated IR injury, through terminal complement complex-induced microvascular injury (65). In hepatic IR injury in orthotopic liver transplantation, livers from C6-deficient rats showed decreased vascular congestion and necrosis, as compared to C6 wild-type livers, suggesting an important role for direct inhibition of the terminal complement complex (66). Following reconstitution of the C6 complement component, C6-deficient mice revealed severe tissue damage in renal IR, independently of the classical complement pathway activation (34).

Clinical trials of complement inhibition in ischemia/reperfusion injury

Only few complement inhibitors have been approved by the FDA for clinical use. Eculizumab (Soliris, Alexion Pharmaceuticals), a humanized anti-C5 monoclonal antibody (67) is approved for the treatment of paroxysmal nocturnal hemoglobinuria, and more recently human C1-INH, (Cinryze, ViroPharma) for hereditary angioedema (68). As yet, no complement inhibitor has been approved for use in the setting of IR injury, e.g. for the treatment of myocardial infarction or in organ transplantation. However, several complement inhibitors, including C1-INH, sCR1, and Pexelizumab, have been used either as rescue therapies or in clinical studies.

Just over 10 years ago, Bauernschmitt et al. presented their first experiences with C1-INH as a rescue therapy in three patients undergoing emergency surgical revascularization after failed percutaneous transluminal coronary angioplasty. C1-INH resulted in rapid restoration of myocardial function during reperfusion, with significant improvement of hemodynamics (69). In 2002 de Zwaan and colleagues showed that an intravenous bolus of C1-INH, given no earlier than 6 hours after acute myocardial infarction, followed by 48 hours of continuous infusion, dose-dependently reduced circulating C4 fragments, troponin T and creatine kinase-MB in 22 patients (70). C1-INH bolus administration during reperfusion and postoperatively in patients undergoing emergency coronary artery bypass surgery for acute ST-elevation myocardial infarction proved safe, was effective in inhibiting complement activation, and reduced myocardial infarct size as measured by cardiac troponin I—however, only in patients undergoing emergency coronary artery bypass surgery within the first 6 hours of onset of symptoms (71). Despite the generally favorable outcome with C1-INH, the dose is of utmost importance. Whilst the 40 IU/kg administered in the study by Horstick et al. proved beneficial (72), tragically in a clinical report of 13 newborns receiving up to 500 IU/kg to prevent capillary leakage after cardiac surgery on cardiopulmonary bypass for congenital heart disease, great vein thrombosis occurred in all, and 9 died of embolic events (73). C1-INH not only inhibits classical complement pathway activation but also inactivates kallikrein and factor XIIa, two major activators of the fibrinolytic system (74), and prevents bradykinin release. Currently no further clinical trials have been published with C1-INH in IR injury in other organs, e.g. brain or kidneys. However, a recent study by Castellano et al. (21) in a large animal model (pig) of acute renal IR revealed significant reduction in peritubular capillary and glomerular C4d and C5b-9 deposition and significant inhibition of tubular damage by inhibiting the classical as well as lectin complement pathways (21). However, whether the results of this large animal model can directly be applied to the situation in humans for the improvement of delayed graft rejection in renal allotransplantation remains to be confirmed in human trials.

The central role of C5 and the possibility to limit tissue damage by its inhibition has been proven in various animal models. Currently only few clinical trials have implemented C5-complement inhibition in humans with more or less success.

Most of the recently published work of complement inhibition in cardiovascular reperfusion strategies and surgery in large randomized clinical trials has focused on the C5 complement inhibitor Pexelizumab and its predecessors. Used intravenously in a trial of cardiopulmonary bypass, this recombinant, humanized, single-chain antibody was proven safe for clinical use and led to a 40% reduction in myocardial injury and 80% reduction in postoperative cognitive deficits (75). In the COMMA trial (COMplement inhibition in Myocardial infarction treated with Angioplasty), a phase II trial in patients with ST-elevation myocardial infarction undergoing percutaneous coronary intervention, Pexelizumab did not measurably influence infarct size, but significantly reduced 90-day mortality (76). In the COMPLY trial (COMPlement inhibition in myocardial infarction treated with thromboLYtics), Pexelizumab blocked complement activity but neither reduced infarct size nor adverse clinical outcomes (77). In a further clinical trial in cardiac surgery on cardiopulmonary bypass Pexelizumab did not significantly affect the primary end-point; however, it did reduce death or myocardial infarction in the isolated coronary artery bypass grafting subpopulation (78). In a study of IR injury on cardiopulmonary bypass for coronary artery bypass graft (CABG) surgery, Pexelizumab was not associated with a significant reduction in the risk of death or myocardial infarction in the patients undergoing CABG-surgery alone but did provide a statistically significant risk reduction 30 days after the procedure among all patients undergoing CABG with or without valve surgery (79). Disappointingly, a further phase III study of intravenous administration of Pexelizumab in conjunction with primary percutaneous coronary intervention in ST-elevation myocardial infarction (APEX-AMI) failed to show a significant difference in mortality through day 30 as compared to the placebo treatment group (80). A meta-analysis by Testa et al. including seven trials with a total of 15,196 patients revealed no benefit of Pexelizumab in addition to therapies currently available for ST-elevation myocardial infarction. However, Pexelizumab appears to reduce the risk of death in patients undergoing CABG (81). The mechanisms by which Pexelizumab asserts its clinical benefits may relate to modification of pathways not directly related to complement inhibition such as reduction of apoptosis, as has already been shown in studies with rats (82), and anti-inflammatory effects on CRP and interleukin-6, as shown in the COMMA trial.

In a study with 564 high-risk patients undergoing cardiac surgery on cardiopulmonary bypass an intravenous bolus of sCR1 (TP10) administered immediately before cardiopulmonary bypass significantly inhibited complement activity up to 3 days postoperatively. The primary end-point of the study, the composite incidence of death, myocardial infarction, prolonged ventilation (24 hours), and prolonged (24 hours) intra-aortic balloon pump support, was not significantly altered with the use of TP10. However, in the male population only, TP10 significantly reduced incidence of death and myocardial infarction (83). This gender-specific positive effect was later confirmed in a second trial with female patients (84). In a further trial TP10 was administered prior to reperfusion in lung transplantation. Complement inhibition significantly reduced the time to extubation. However, the operative deaths, incidence of infection and rejection, and overall length of hospital stay did not differ between the treatment and placebo groups (85).

Clinical application of complement inhibitors in ischemia/reperfusion injury: problems and pitfalls

Although much of the preclinical work and data accumulated in the past years has been positive and seeded hope for a ‘quick and easy fix’ to the problem of complement activation and complement-mediated damage in IR injury, the subsequent step of clinical application has been somewhat disappointing.

The reasons for the in part disappointing results of complement inhibitors in clinical trials are likely to be manifold. More specifically, the reasons for the mixed results with Pexelizumab in the clinical trials may lie in the fact that complement activation represents only one part of the whole battery of ‘assault’ directed towards the vasculature and the parenchymatous organ in ischemia and reperfusion. Beranek discusses the possible reasons for this ‘failure’ (86). In part thrombolytic agents, used for non-interventional treatment of myocardial infarction, themselves activate complement and cause hemorrhage into infarcted myocardium. Hence in the setting of thrombolysis for acute (myocardial) infarction, the whole potential of the complement inhibitors may not be as apparent.

Locally targeted complement inhibition

Another interesting path to follow to augment the efficacy of the administered drug and conceivably reduce unwanted side-effects is the possibility of local delivery of the complement inhibitor to the actual site of injury through selective (pre-)reperfusion of the affected tissue. Physiologically, complement inhibition is targeted to cell surfaces, rather than the fluid phase. This is mainly achieved by membrane-bound complement regulatory proteins like the decay-accelerating factor (DAF, CD55), membrane co-factor protein (MCP, CD46), and the membrane inhibitor of reactive lysis (CD59), all of which limit complement activation on host cells. However, the two main physiological ‘fluid phase’ complement inhibitors, C1-INH and factor H, have also been shown to bind to the endothelial glycocalyx and as such act as local, site-specific inhibitors of complement activation.

In the case of myocardial infarction, local targeting of complement inhibition may be achieved by administering the drug into the ischemic area at risk using an over-the-wire balloon to overcome the stenotic intracoronary lesion. Alternatively, in organ transplantation, the organ may directly be pre-perfused with or transported in preservation solution containing the complement inhibitor, thus concentrating the active agent to the site of injury and reducing systemic side-effects.

Indeed, our own work has concentrated on localizing complement inhibition in cardiovascular ischemia and reperfusion to the site of reperfusion damage. Targeting a complement-inhibitory and anti-inflammatory agent directly to the site of injury offers several obvious advantages over systemic application. A lower dose may suffice to realize a relevant effect, thus avoiding or reducing the chance of complications associated with systemic complement inhibition. Unwanted binding of the drug to organs other than the target organ, especially excretory organs (e.g. kidney), may be largely preventable. Additionally, site-targeted application may reduce the time for the drug to reach the site of injury. This may be of relevance, as several studies have shown that the first few minutes of reperfusion importantly affect tissue damage (87). A range of approaches has been used in the past to target the drugs to cell membranes or to prolong the half-life of systemically applied substances, e.g. linking these to albumin and immunoglobulin domains or to sialyl Lewis x moieties (88,89). In an earlier study by Pratt et al., the practicability of cell membrane-targeted cytoprotection to attenuate IR injury in kidney transplantation in rats was shown using Mirococept (known then as m-pSCR1-3) (90). In a porcine model of acute myocardial infarction we then demonstrated that Mirococept (91), an engineered fragment of sCR1, which, through coupling a peptide motif with a lipophilic tail (92), enables tethering of the inhibitor to the surface of cells, significantly reduced IR injury (93). By tethering soluble Crry, a mouse inhibitor of all complement activation pathways, to a fragment of complement receptor 2 to create a novel fusion protein (CR2-Crry), its bioavailability and therapeutic efficacy could be enhanced significantly in a model of intestinal IR injury. Unlike Crry-Ig, CR2-Crry only minimally affected serum complement activity, therefore preserving complement-mediated host resistance to infection (94). Likewise, in a model of acute cerebral infarction and reperfusion, infusion of CR2-Crry 30 minutes post-reperfusion into wild-type and C3-deficient mice inhibited complement activation at the C3 stage and resulted in similar levels of protection (95). In a further study by the group of Tomlinson, CR2-fH and CR2-fHfH, alternative pathway-specific inhibitors comprising a single or dimeric N-terminal region of mouse factor H (fH; short consensus repeats 1-5) linked to the same complement receptor 2 fragment, fully protected mice from local intestinal and remote lung injury (96).

Another crucial point of complement activation in IR injury is triggered by endothelial damage leading to release of heparan sulfate proteoglycans (HSPG) (11) from the luminal surface of the endothelium. The shed HSPG expose a pro-coagulant and pro-inflammatory endothelial surface, as was initially described in xenotransplantation (97). Also in humans, heparan sulfate shedding into the plasma of patients undergoing cardiovascular surgery occurs in the early reperfusion phase. With the endothelium undergoing such changes on its surface during IR, and complement playing such a central role in this pro-inflammatory process, one strategy followed by our group is to ‘take advantage’ of this HSPG shedding and locally target complement inhibitors to the damaged endothelium. Low-molecular-weight dextran sulfate (DXS, MW 5000), a sulfated glycosaminoglycan analog which inhibits all three complement pathways (98,99) as well as the coagulation cascade (100), associates in vitro with endothelial cells cleaved of their glycocalyx heparan sulfate proteoglycans by heparinase. Binding of DXS directly correlates with complement inhibition (99). In a closed-chest acute model of acute myocardial IR injury in pigs, intracoronary DXS, applied directly to the myocardium at risk, significantly reduced infarct size and local complement activation and deposition, however, without significantly affecting hemodynamic parameters (101). Furthermore, addition of DXS to blood cardioplegia solution in part ameliorated post-cardiopulmonary bypass inflammation and features of cardiac function also in a porcine cardiac IR injury model (102). This specific localization of DXS to the damaged myocardium and reduced complement deposition correlated with diminished vascular HSPG staining observed in these areas.

Outlook and concluding remarks

As scientists in the field of IR research, we all should appreciate that, although we may want to believe that there is one solution to the problem of reperfusion injury, the pathophysiology of this phenomenon is highly complex. This fact will, more likely than not, require that a combination of efforts and interventions be implemented to tackle the problems encountered in reperfusion-induced damage. Another compelling idea, previously proposed by Menger and Vollmar in IR injury in transplantation surgery (103), may be the need to address the potential and possibility of using substances with pleiotropic effects in the fight against IR-mediated damage. Such candidate substances may not necessarily be directly derived from the field of complement research. One such candidate may include the statins, 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors. Known primarily for their lipid-lowering effects, they also possess immunomodulatory and anti-inflammatory properties. Of particular interest is the fact that the statins atorvastatin and simvastatin dose-dependently increased endothelial expression of the complement regulatory protein decay-accelerating factor (DAF) and reduced complement C3 deposition as well as complement-mediated lysis of antibody-coated endothelial cells in vitro (104). Also, cerivastatin reduces C5b-9-induced activation of the mitogen-activated protein kinase extracellular signal-regulated kinase and pro-inflammatory interleukin IL-6 (105).

In many clinical settings, as is the case for acute coronary intervention to treat myocardial infarction, standardized combination therapies have been developed and are now widely implemented to optimize outcome and reduce permanent tissue damage upon ischemia and reperfusion. It is highly unlikely that ‘the’ drug to combat IR injury will be found. On the contrary, it is probable that complement inhibitors will have to be integrated into and carefully used to supplement existing treatment regimens for targeted reperfusion and in organ transplantation. Indeed, ‘targeting’ may be the key to the future. This may include targeting of the complement inhibitor to the site of injury by localized intravascular application as well as modifications allowing for direct tethering to the (damaged) cell membranes within the damaged organ. And yet, although the potential of complement inhibition to reduce IR damage has been shown, a balance will have to be struck between inhibition in the acute phase of injury and allowing for subsequent post-IR reparatory processes. Indeed, the role of complement in tissue growth and regeneration (106), specifically in maintenance and repair of the adult brain (107) as well as in the liver following hepatic IR injury (108), has recently been demonstrated.

Our next goals should be to combine our knowledge and experience obtained in the past from animal models and channel future work to translate this into clinical trials in surgical and interventional reperfusion as well as organ transplantation.

Declaration of interest: The authors state no conflict of interest and have received no payment in preparation of this manuscript.