Activated protein C ameliorates coagulopathy but does not influence outcome in lethal H1N1 influenza: a controlled laboratory study

Influenza A infection is a major cause of morbidity and mortality: Seasonal influenza A infection causes over 200,000 hospitalizations and approximately 41,000 deaths in the United States annually, being the seventh leading cause of mortality [1]. Besides its regular seasonal character, influenza A, due to the introduction and adaptation of novel hemagglutinin subtypes from other mammals or birds resulting in antigenic shifts, has the potential to cause pandemics, as the pandemics in 1918, 1957 and 1968 have shown [2]. Currently, a novel influenza A (H1N1) strain from swine origin has evolved to a pandemic, now worldwide causing major concern for the near future [3]. Although the greatest proportion of mortality caused by influenza A infection is due to secondary bacterial pneumonia and cardiovascular complications, influenza itself is also an important cause of community-acquired pneumonia (CAP), causing 5 to 10% of CAP-cases [4‐7]. As such, influenza is a major concern for pulmonologists and intensive care physicians [8].

Severe infection and inflammation have been closely linked to activation of coagulation and downregulation of anticoagulant mechanisms and fibrinolysis [9]. In bacterial pneumonia, pulmonary activation of coagulation as well as downregulation of the anticoagulant protein C (PC) pathway and fibrinolysis have been demonstrated [10‐12]. Beside anticoagulant properties, activated (A)PC has been shown to have profibrinolytic, anti-inflammatory, anti-apoptotic and other cytoprotective properties [13]. Downregulation of the PC pathway has been correlated to disease severity and mortality in severe bacterial pneumonia and sepsis [14, 15] and continuous intravenous administration of recombinant human (rh-) APC for four days (Human Activated Protein C Worldwide Evaluation in Severe Sepsis (PROWESS) trial) has been shown not only to downregulate activation of coagulation, but also to reduce inflammation and improve survival in patients with severe sepsis [16]. The benefical effect of rh-APC in this trial seemed especially prominent in patients with severe sepsis due to pneumonia [17]. While much research has been done on coagulation activation during severe bacterial infection, data on coagulation activation in viral infection like influenza are sparse. Evidence that influenza can be associated with coagulation activation comes from a clinical study in pediatric patients hospitalized for severe influenza [18] and from a recent study showing elevated plasma levels of thrombin-antithrombin complexes (TATc) in mice infected with a non-lethal dose of influenza A [19]. Interestingly, and as mentioned above, many elderly patients with influenza infections suffer from cardiovascular complications.

At present it is unknown whether APC can influence the procoagulant and inflammatory response to lethal influenza A infection. Therefore, in the present study we sought to establish the effect of recombinant mouse (rm)-APC treatment on local and systemic activation of coagulation and fibrinolysis during lethal H1N1 influenza A in mice and moreover determined the effect of rm-APC on lung inflammation, pulmonary viral loads and survival. We here show, that lethal H1N1 influenza A infection is associated with both pulmonary and systemic activation of coagulation and inhibition of fibrinolysis. Moreover, we show that rm-APC treatment, started 24 hours after the onset of infection, partially prevents these hemostatic derangements, but does not impact on lung inflammation or survival.

Influenza is an important cause of pneumonia, causing 5 to 10% of all CAP cases [4, 7]. While bacterial pneumonia has been linked to activation of coagulation and downregulation of anticoagulant mechanisms and fibrinolysis [10‐12], knowledge of the impact of influenza on hemostasis is limited. We here studied alterations in local and systemic activation of coagulation and fibrinolysis together with induction of inflammation during lethal influenza A infection. In addition, considering that previous investigations in patients and animals have especially pointed to beneficial effects of APC treatment in the lungs [17, 31‐33], we determined the effect of APC on the procoagulant and inflammatory response to and the outcome of lethal influenza. We show that lethal H1N1 influenza A infection is associated with extensive pulmonary and systemic activation of coagulation accompanied by inhibition of fibrinolysis. Systemic administration of APC, started 24 hours after infection, mimicking a possible clinical scenario, strongly attenuated coagulation activation and partially reversed inhibition of fibrinolysis, but did not influence lung inflammation or survival.

Concurrent alterations in coagulation and fibrinolysis during influenza have not been studied in detail thus far. One clinical study in children has indicated that severe influenza can be associated with disseminated intravascular coagulation [18]. In addition, mice with non-lethal influenza A infection displayed a rise in plasma TATc and PAI-1 levels; although fibrinolytic activity (such as measured by PAA) was not determined in this previous investigation, these data point to concurrent activation of coagulation and inhibition of fibrinolysis at the systemic level during mild influenza [19]. Our own preliminary data have suggested that lethal influenza not only results in systemic coagulation activation, but also in induction of the coagulation system in the lungs (Schouten et al, XXIst Congress of the International Society of Thrombosis and Haemostasis, Boston, July 2009, abstract no. 3065). Our current results confirm and expand these previous data. First, we demonstrate local and systemic activation of coagulation, as evidenced by increased lung and plasma TATc and FDP levels in influenza infected mice at 48 and 96 hours. Moreover, we show that activation of coagulation is accompanied by local as well as systemic downregulation of fibrinolysis, as reflected by elevated PAI-1 and reduced PAA levels in lung homogenates and plasma, which probably further contributes to the influenza-induced procoagulant state. Most likely, the downregulation of fibrinolytic activity can be explained at least partially by upregulation of PAI-1, the main inhibitor of the fibrinolytic system. As such, severe influenza appears to cause similarly opposite changes in pulmonary coagulation and fibrinolysis as previously reported for bacterial pneumonia and acute respiratory distress syndrome [34‐37].

Systemic administration of rm-APC strongly inhibited activation of the coagulation system, as indicated by markedly reduced plasma and lung concentrations of TATc and FDPs in rm-APC treated mice relative to vehicle treated animals. In addition, rm-APC had a modest but statistically significant effect on the fibrinolytic system, partially blunting the influenza-induced rise in plasma and lung PAI-1 levels and partially preserving plasma and lung fibrinolytic activity. The capacity of APC to attenuate systemic coagulation during severe bacterial infection has been demonstrated in several studies [13, 16, 38]. Our group previously reported on the effects of intravenous administration of recombinant APC on pulmonary coagulation in healthy humans intrabronchially challanged with lipopolysaccharide (LPS) [39] and in rats challenged with LPS systemically [40] or with viable bacteria via the airways [41, 42]. All of these previous studies [39‐42], in which APC treatment was started before the challenge with LPS or bacteria, revealed the capacity of APC to inhibit coagulation in the lungs. The current study adds to these earlier findings that APC is capable of inhibiting systemic and local coagulation during influenza-induced pneumonia and that this effect is present when APC administration is initiated 24 hours after infection, that is, in a clinically more relevant setting. Of interest, endogenous APC may also reduce influenza-induced coagulation, as indicated by studies in mice with a mutation in their thrombomodulin gene that results in a minimal capacity for endogenous APC generation: these mice demonstrated increased plasma levels of TATc (relative to wild-type mice) during non-lethal influenza [19]. Our finding that rm-APC stimulated fibrinolysis by inhibiting PAI-1 is supported by evidence derived from in vitro investigations [29, 30]. Of note, previous studies from our laboratory could not demonstrate an effect of recombinant APC on pulmonary fibrinolysis during LPS-induced lung injury [39, 40] or bacterial pneumonia [41, 42].

Besides anticoagulant and profibrinolytic properties, APC has been found to exert anti-inflammatory activity (reviewed in [13]). Previous studies have suggested that recombinant APC can inhibit LPS-induced neutrophil recruitment and activation in the lungs [31, 43]. Nonetheless, in the current study rm-APC did not have a major impact on lung inflammation during lethal influenza A infection, as indicated by similar histopathology scores of lung tissue, a similar influx of neutrophils to the site of infection and largely similar cytokine and chemokine concentrations in lung homogenates. Interestingly, rm-APC did reduce lung TNF-α and IL-12 levels 96 hours after infection; similarly, APC has been found to inhibit the LPS-induced production of TNF-α in vitro and in vivo [32, 44].

To our knowledge, the effect of APC on antiviral defense per se has not been studied. We here show that rm-APC temporarily lowers pulmonary viral loads about four-fold, as measured 48 hours after infection. These differences between rm-APC and vehicle treated mice had disappeared 96 hours post infection. The transiently reduced viral loads in rm-APC treated animals are surprising considering that APC is not known to impact on antiviral mechanisms and did not influence the inflammatory response to influenza A in a way that might have improved host defense. The difference in viral load between rm-APC and buffer treated mice did not result in a substantially changed inflammatory response or a delayed mortality. However, since we tested only one infectious dose of influenza A, we cannot exclude that rm-APC does impact on lethality after infection with different viral doses. The mechanism by which rm-APC reduces viral loads at an early stage of influenza infection needs further investigation. Besides anticoagulant and anti-inflammatory properties, APC has been described to influence the hemodynamic response to an inflammatory stimulus [13, 45]. The potential effect of rm-APC on hemodynamics was not measured in our current study and therefore warrants further investigation.

In order to mimic the clinical situation, APC should be administered by a continuous intravenous infusion. However, this is difficult to achieve in mice for a period of several days. In this study, we therefore administered rm-APC intraperitoneally every eight hours at a dose of 125 μg (a daily dose of approximately 15 mg/kg, that is, approximately 25 times higher than the daily dose administered to humans). This administration protocol resulted in plasma levels which were not dissimilar to the levels observed after intravenous administration of lower doses in previous studies in rodents in which anti-inflammatory effects of recombinant APC were demonstrated after LPS administration [31, 32, 46, 47] and which are in the same range as those achieved by continuous intravenous infusion in septic patients [48]. In light of these earlier rodent and patient investigations [31, 32, 46‐48] and considering that the APC dosing schedule used here caused profound anticoagulant effects, we consider it unlikely that higher APC doses would have had a significant effect on lung inflammation or survival. Such studies would be less clinically relevant and moreover would be associated with an increased risk for bleeding, which was not observed with the current dosing regimen. It would be of considerable interest, however, to study the effects of mutant forms of APC with reduced anticoagulant but enhanced cytoprotective properties in models of lethal influenza [46, 47, 49].

Lethal H1N1 influenza infection is associated with both pulmonary and systemic activation of coagulation and inhibition of fibrinolysis. Rm-APC treatment, started 24 hours after the onset of infection, partially prevents these hemostatic derangements, but does not impact on lung inflammation or survival.