In all patients with sepsis there is abundant activation of inflammatory pathways, which results in demonstrable circulating levels of inflammatory cytokines and chemokines, activated inflammatory cells and other markers of increased inflammatory activity. Virtually all septic patients exhibit coagulation abnormalities as well. These abnormalities range from subtle activation of coagulation, which can only be detected by sensitive markers of coagulation factor activation; to more marked activation, which may be detectable based on a small decrease in platelet count and subclinical prolongation of global clotting times; and finally to fulminant disseminated intravascular coagulation (DIC), which is characterized by simultaneous widespread microvascular thrombosis and profuse bleeding from various sites 1. Septic patients with severe forms of DIC may present with thromboembolic disease or clinically less apparent microvascular failure that predominantly presents as multiple organ dysfunction 23. Interestingly, there is a tight, bidirectional relationship between activation of inflammation and coagulation, in which inflammatory activity results in activation of coagulation but activated coagulation proteases can also affect inflammatory pathways 4.

Activated protein C (APC) appears to play a central role in the pathogenesis of sepsis and associated organ dysfunction. There is ample evidence that insufficient functioning of the protein C pathway contributes to the derangement of coagulation observed in sepsis 56. The protein C system is summarized in Figure 1. The circulating zymogen protein C is activated by the endothelial cell bound thrombomodulin once this is activated by thrombin 7. APC acts in concert with its co-factor, protein S, and can proteolytically degrade the cofactors Va and VIIIa, which are essential for coagulation; APC is therefore an effective anticoagulant. The endothelial protein C receptor (EPCR) not only accelerates activation of protein C several fold but it also serves as a receptor for APC, and binding of APC to EPCR may amplify its anticoagulant and anti-inflammatory effects 8. A recent study 9 demonstrated that exposure of cultured endothelial cells to APC results in release of microparticles that contain EPCR, but the relevance of that observation to coagulation or inflammation is not yet clear.

In patients with sepsis the APC system malfunctions at virtually all levels. First, plasma levels of the zymogen protein C are low or very low because of impaired synthesis, consumption and degradation by proteolytic enzymes such as neutrophil elastase 101112. Furthermore, significant downregulation of thrombomodulin caused by pro-inflammatory cytokines such as tumour necrosis factor-α and interleukin-1 has been demonstrated, resulting in diminished protein C activation 1314. Low levels of free protein S may further compromise the functioning of the protein C system. In plasma, 60% of the co-factor protein S is complexed to the complement regulatory protein C4b-binding protein (C4bBP). Increased plasma levels of C4bBP, which occur as a consequence of the acute phase reaction in inflammatory disease, may result in relative protein S deficiency, which further contributes to a procoagulant state during sepsis. Although it has been shown that the β-chain of C4bBP (which mainly governs binding to protein S) is not much affected by the acute phase response 15, support for this hypothesis comes from studies conducted in baboons 16 in which infusion of C4bBP in combination with a sublethal dose of Escherichia coli resulted in a lethal response, with severe organ damage due to DIC. Finally (but importantly), in sepsis EPCR has been shown to be downregulated, which may further adversely affect the function of the protein C system 17. Apart from these effects, sepsis may induce resistance to APC by other mechanisms that are partly dependent on a sharp increase in factor VIII levels (released from endothelial cells) and partly due to as yet unidentified mechanisms 18.

Administration of APC in a baboon model of intravenous E. coli administration resulted in survival of all animals, whereas all control animals in the same experiment died 19. A similar beneficial effect was observed in rabbits with meningococcal endotoxin shock. In a rat model of septic shock, administration of APC prevented tumour necrosis factor-α mediated hypotension, probably caused by modulation of the nitric oxide response 20. In patients with severe sepsis, administration of recombinant human activated protein C (rhAPC) resulted in remarkable improvement in microcirculatory perfusion 21. Conversely, experiments conducted in baboons in which the protein C pathway was blocked with monoclonal antibodies resulted in complete lethality in an otherwise sublethal model of bacteraemia. In this same model, blockade of the EPCR also resulted in a more severe response to sublethal E. coli bacteraemia 17. In addition to these observations in experimental sepsis models, APC was shown to have antithrombotic properties in experimental thrombosis models in dogs, rabbits and baboons 2223. Interestingly, APC may also affect fibrinolysis by inhibiting plasminogen activator inhibitor type 1 (a fibrinolytic inhibitor). In a rat model of DIC, APC was shown to block activity of plasminogen activator inhibitor type 1, and other experiments demonstrated the ability of APC to enhance clot lysis in vivo 24.

More definitive proof of the beneficial effect of rhAPC in severe sepsis comes from clinical studies 2526. These studies are reviewed in detail in other reviews included in this supplement 272829 and are not discussed in detail here. However, it may be of interest to review the evidence that rhAPC acts as an anticoagulant agent in these studies.

First, it should be noted that the selected dose of rhAPC was based on the effect on D-dimer levels in a phase II clinical trial 25. Indeed, in the pivotal PROWESS (Recombinant Human Activated Protein C Worldwide Evaluation in Severe Sepsis) trial 26, septic patients who were treated with APC exhibited a significant decrease in D-dimer levels as compared with placebo control individuals. D-dimer levels dropped by 25% after 2 days of APC administration, which was in contrast to a 10% increase in placebo-treated patients. In a more detailed analysis of coagulation activation upon administration of rhAPC, it was clearly demonstrated that markers of thrombin generation sharply dropped almost immediately after initiation of the APC infusion 30. Second, subgroup analyses of trials including patients with severe sepsis 3132 demonstrated that patients with the most extreme coagulation abnormalities benefit the most from treatment with rhAPC. The relative risk reduction in mortality among patients with sepsis and DIC who received APC was 38%, as compared with a relative risk reduction of 18% observed among patients with sepsis who did not have DIC. Interestingly, the dynamics of coagulation abnormalities in the first days after intensive care unit admission for severe sepsis, including the response of the protein C system, is a strong predictor of outcome 33.

It is difficult to assess whether this anticoagulant effect of rhAPC translates into an antithrombotic effect. In the recently concluded XPRESS (Xigris [drotrecogin alfa] and Prophylactic Heparin in Severe Sepsis) study 34, all patients with severe sepsis received rhAPC but were also randomly assigned to receive prophylactic heparin or placebo. The main conclusion of this study was that heparin was not equivalent to placebo and might have a beneficial effect on 28-day mortality. Of note, this advantageous effect of heparin was completely due to a greater incidence of death and thrombotic adverse events in the patients receiving heparin but who were randomly assigned to placebo (in other words, those who stopped heparin during the trial). There was a markedly low incidence of venous thromboembolism in this study compared with previous reports (the incidence did not differ between placebo patients and patients receiving heparin), even though all patients underwent screening ultrasound for venous thrombosis at around day 6 of admission. In addition, a report of small series of patients with severe sepsis who were treated with rhAPC 35 demonstrated a lack of thrombotic obstruction of haemofiltration circuits, even in the absence of heparin or other anticoagulants.

Apart from the systemic response, there may be a differential localized effect of rhAPC on coagulation. The localized effect of APC appears to be particularly marked in the pulmonary compartment. In experiments involving unilateral instillation of endotoxin into healthy individuals, systemic administration of rhAPC resulted in a marked reduction in bronchoalveolar activation of coagulation 36. This observation may be relevant because the vast majority of patients with severe sepsis in the various trials had a pulmonary source of infection. An interesting novel finding is the modulatory influence of alveolar epithelial cells on the protein C pathway 37. It is likely that the underlying mechanism involves shedding of EPCR and thrombomodulin by metalloproteinases.

Evidence that APC acts as an important mediator in the systemic inflammatory response in sepsis comes from experiments showing that blocking the protein C pathway in septic baboons exacerbated the inflammatory response 38. In contrast, administration of APC ameliorated the inflammatory activation that occurred upon intravenous infusion of E. coli 38. Similar experiments in rodents yielded identical results and demonstrated a beneficial effect on inflammatory effects in various tissues 39. Support for the notion that APC has anti-inflammatory properties comes from in vitro findings, demonstrating an APC-binding site on monocytes that may mediate downstream inflammatory processes 4041. It also received support from experiments showing that APC can block nuclear factor-κB nuclear translocation, which is a prerequisite for increased levels of pro-inflammatory cytokines and adhesion molecules 42. These in vitro findings are supported by in vivo studies in mice with targeted disruption of the protein C gene 4344. In these mice with genetic deficiencies of protein C, endotoxaemia was associated with more marked increases in pro-inflammatory cytokines and other inflammatory responses than in wild-type mice.

It is likely that the effects of APC on inflammation are mediated by EPCR, which may mediate downstream inflammatory processes 45. Binding of APC to EPCR influences gene expression profiles of cells by inhibiting endotoxin-induced calcium fluxes in the cell and by blocking nuclear factor-κB nuclear translocation 4142. The EPCR-APC complex itself can translocate from the plasma membrane into the cell nucleus, which may be another mechanism of modulation of gene expression, although the relative contributions of this nuclear translocation and cell surface signalling are unclear 5. Some studies have also suggested that EPCR binding of APC can result in activation of protease activated receptor (PAR)-1 and thereby affect cytokine responses 46. In contrast, other experiments demonstrated that a significant physiological role for activation of PAR-1 by APC to be less probable 47. Like APC, EPCR itself may have anti-inflammatory properties. Soluble EPCR (the extracellular domain of the cell-associated EPCR shed from the cell surface by the action of an inducible metalloproteinase 48) can bind to proteinase 3, which is an elastase-like enzyme. The resulting complex binds to the adhesion integrin macrophage 1 antigen (Mac-1) 49. Of considerable interest is that the crystal structure of EPCR is remarkably similar to the structure of the MHC class 1/CD1 family of proteins, the majority of which are involved in inflammation 50. Blocking the EPCR with a specific monoclonal antibody aggravated both the coagulation and the inflammatory response to E. coli infusion 17.

Apart from the influence of APC on cytokine levels, remarkable effects of the agent on leucocyte chemotaxis and adhesion of circulating leucocytes to the activated endothelium have been demonstrated 5152. This was confirmed in a hamster endotoxaemia model at concentrations of rhAPC that preclude a significant anticoagulant effect 53. The localized effect of APC in the lung has also been shown to exhibit these anti-inflammatory properties 54. APC was shown to inhibit the expression of platelet-derived growth factor in the lung 55, which may reflect a potential mechanism underlying this localized effect. Also, APC was shown to protect against disruption of the endothelial cell barrier in sepsis, probably by interfering with EPCR and PAR-1 on endothelial cells 565758.

Finally, APC can inhibit endothelial cell apoptosis, which also appears to be mediated by binding of APC to EPCR and to require PAR-1 4659. Signalling through this pathway can affect Bcl-2 homologue protein, which can inhibit apoptosis, and further suppresses p53, which is a pro-apoptotic transcription factor 6061.

Inadequate functioning of the protein C system, and in particular APC, plays a central role in the pathogenesis of sepsis. Attempting to restore the function of this pathway in patients with sepsis appears to be a rational approach and is supported by the beneficial effect of APC in experimental models of sepsis and in clinical studies. Apart from its evident effect on the coagulation system and the ability of rhAPC to correct the deranged coagulation system in severe sepsis, a series of pleiotropic modulating effects of APC on inflammatory cytokines and cells as well as protective effects on disrupted endothelium have been reported. It should be noted that many of the effects on inflammatory cells and pathways, as well as the cytoprotective effect, have mostly been demonstrated in vitro and sometimes at inordinately high concentrations of APC. The relevance of these findings to the human in vivo situation and their importance to the treatment of sepsis remain to be established. Although the relative importance of the anticoagulant effect versus the inflammation-modulating effect of rhAPC is not clear, it is tempting to hypothesize that a combined effect is responsible for the benefit from rhAPC. Indeed, strategies aimed at restoring physiological pathways with less marked effects on inflammation (such as administration of antithrombin or recombinant tissue factor pathway inhibitor) were less successful. Further insight into the various mechanisms of action of rhAPC on the entangled processes of inflammation and coagulation may permit more detailed dissection of the relative importance of the various pathways that contribute to the pathogenesis of sepsis, potentially culminating in improved treatment strategies.

APC = activated protein C; C4bBP = C4b-binding protein; DIC = disseminated intravascular coagulation; EPCR = endothelial protein C receptor; PAR = protease activated receptor; rhAPC = recombinant human activated protein C.

ML and TvdP have participated in advisory boards of Eli Lilly and Company and have participated as investigators in experimental and clinical studies with rhAPC.