Protein C anticoagulant system—anti-inflammatory effects

The protein C system is best known for its anticoagulant activity seen most clearly in the clinical observation that patients born with a total protein C deficiency exhibit massive neonatal thrombosis that is usually lethal unless treated, reviewed in [1]. Indeed this is also one aspect of the anti-inflammatory functions of the pathway since coagulation, particularly thrombin generation, can trigger a wide variety of pro-inflammatory events including expression of adhesion molecules like P-selectin and activating the Nf-κB pathway [2]. While this is an important aspect of the anti-inflammatory function of the pathway, it does not distinguish the pathway from other anticoagulants. Indeed, heparin has long been noted to have apparent anti-inflammatory functions, in part likely due to its anticoagulant activity.

Some of the first suggestions that this pathway might have additional anti-inflammatory activity came from the treatment of newborns with protein C deficiency. The thrombotic lesions that developed in the newborns were surrounded by an intense red area that retracted rapidly following the administration of protein C, suggesting that protein C was preventing the inflammation in addition to decreasing the thrombosis, reviewed in [1].

These studies were followed by examination of the roles of thrombosis in sepsis. In an early study, Hinshaw and colleagues [3] observed that heparin could prevent the consumptive coagulopathy associated with Escherichia coli-induced sepsis in baboons but did not rescue the animals. Later we demonstrated that an active site blocked form of factor Xa could prevent the disseminated intravascular coagulation (DIC) but again failed to protect against sepsis [4]. Subsequently, Hinshaw and colleagues showed that extracorporeal perfusion without exogenous anticoagulation was protective against endotoxin-induced sepsis [5]. They also observed that an associated anticoagulant was being generated during these studies. Interestingly, the pump could be removed subsequently and the animals were still protected from subsequent bacterial challenge. With the identification of thrombomodulin [6] and demonstration of thrombin-dependent protein C activation in vivo [7], it was possible to test whether the anticoagulant might be activated protein C (APC) generated by thrombin formed by the pump. Indeed, thrombin infusion into dogs challenged with endotoxin was protective [8] despite the fact that the animals would develop DIC without the thrombin infusion. Thrombin infusion decreased both the DIC and inflammation. With the advent of a rapid means for purification of protein C from human plasma [9], it was possible to test the ability of APC to protect baboons from E. coli-induced sepsis. When APC was administered with the E. coli, the animals survived a normally lethal dose and exhibited reduced coagulation, protection from shock, and decreased inflammation [10]. These older studies highlight that APC could protect against an inflammation-induced disease like sepsis when other comparable anticoagulants could not. In contrast, inhibition of the pathway in the E. coli sepsis model, in this case with C4 binding protein, elevated cytokine production in response to E. coli challenge [11].

Either reducing protein C levels [12, 13] in mice or blocking protein C activation [10] in baboons increased a sublethal to a lethal challenge with bacteria or endotoxin. In order to perform its full anti-inflammatory functions, the APC must bind to the endothelial protein C receptor (EPCR) [14]. Mice overexpressing EPCR are resistant to endotoxemia [15], whereas those with low-level expression are sensitized [16, 17]. Furthermore, mice with low levels of EPCR have cardiac dysfunction from the challenge [16]. These studies illustrate the important role of the pathway in regulating the host response to acute inflammatory challenges.

One of the major mechanisms that augment inflammation is mediated through Nf-κB activation and nuclear translocation from the cytosol [18, 19]. This turns on synthesis of a variety of inflammatory mediators including cytokine production. APC can decrease the synthesis of Nf-κB components [19, 20] and decrease Nf-κB nuclear translocation [18]. Together these activities probably constitute the major mechanisms by which APC downregulates inflammatory cytokine production in inflamed endothelium in culture [21] and in animal models of sepsis [22, 23].

These effects are dependent on APC, EPCR, and protease-activated receptor 1 (PAR-1) [14, 24]. Activation of PAR-1 by the APC–PAR-1 complex leads to different cellular signaling than when thrombin activates PAR-1 despite cleaving the same site on the receptor [25] (Fig. 1). The mechanisms for this change in signaling are currently being elucidated. In one model, protein C binding to EPCR leads to migration of EPCR out of the lipid rafts at which time it interacts with PAR-1 coupled to a different G protein than when it was in the lipid rafts, thus resulting in the altered signaling profile [26‐28]. In support of this model, EPCR did appear to migrate from rafts in the presence of protein C [26] and recombinant mutant molecules containing the protein C Gla domain that could elicit signaling similar to that of APC [26].

Endothelial barrier function is compromised in a number of diseases resulting in edema. Thrombin is known to decrease endothelial barrier function, a process that is reversed by APC [26, 34, 35]. APC accomplishes this, at least in part, through the generation of shingosine 1-phosphate receptor transactivation [34, 35]. Improving endothelial barrier function is likely to provide anti-inflammatory effects since it should reduce leukocyte trafficking into the extravascular space. While not directly related to inflammation, one of the features of APC is that it diminishes both endothelial cell and neuronal apoptosis [36‐38]. Excessive apoptosis or cellular necrosis leads to release of relatively large amounts of nuclear material in the form of nucleosomes and also the release of mitochondrial contents. Both of these events will trigger inflammation. The histones on nucleosomes induce leukocyte migration into the tissue, platelet activation, and thrombosis [39] and induce cytokine formation, and the mitochondria induce leukocyte activation [40].

Extracellular histones are cytotoxic [39], and APC can cleave and neutralize this activity of histones. The importance of the latter observation was apparent in studies that demonstrated that histones were much more toxic in mice where the protein C pathway was blocked and that blocking histone function was protective in endotoxemia [39].

Mutants of APC have been developed that retain signaling activity but have very low anticoagulant activity [41]. These mutant forms of APC (5A- aPC and other similar mutants) were effective in preventing mortality in mouse models of sepsis [24, 42].

The original signaling studies were done in endothelium [20]. More recent studies have detected EPCR on leukocytes, particularly CD8+ dendritic cells [43]. Mice with low levels of EPCR (EPCR low) were studied and were less effectively protected from endotoxin-induced sepsis toxicity by 5A aPC than wild-type mice [43]. When bone splenic CD11c^hi dendritic cells from wild-type mice were transplanted into EPCR-low mice, they supported protection from endotoxin by 5A- aPC whereas similar cells from EPCR-low mice did not. In vitro, 5A-aPC inhibited the inflammatory response of dendritic cells which appeared to be independent of a requirement for normal levels of EPCR [43]. Thus, protective function seems EPCR dependent but there are cell populations that are responsive to APC in suppressing inflammation that do not appear to require EPCR. A likely receptor for APC on macrophages is CDllb/CD18, also known as Mac-1. APC administration in wild type, but not CD11b null mice, reduced mortality. CD11b was also required for suppression of the endotoxin-induced macrophage inflammatory response [44]. These results indicate that the cellular signaling mechanisms play a dominant role in protection from endotoxemia [45].