Acute lung injury (ALI) is one of the most frequent organ dysfunctions evolving from severe sepsis 1. In Gram-negative sepsis, endotoxin (lipopolysaccharide (LPS)) released from bacteria stimulates transcriptional factors to activate the inflammation and coagulation systems 2. Proinflammatory cytokines, tissue factor, eicosanoids and endothelins act in concert to promote the disintegration of lung endothelial and epithelial barriers that characterize ALI 34567. Recent research has revealed that thrombin, a blood coagulation protease, is a key trigger of both the coagulation and the inflammation system in ALI 8.

Protein C is an inactive precursor of the serine protease activated protein C (APC), which is catalyzed by the action of thrombin/thrombomodulin complex when protein C is bound to its endothelial receptor (endothelial cell protein C receptor). Acting together with protein S, APC precludes generation of thrombin in a negative feedback loop by proteolytic cleavage of activated factors V and VIII. APC also escalates fibrinolysis by inhibiting plasminogen activator inhibitor 1 and thrombin-activatable fibrinolysis inhibitor 9. Activated protein C attenuates inflammation by inhibiting translocation of NF-κB and of activator protein-1 in LPS-stimulated monocytes, thereby suppressing release of proinflammatory cytokines 1011. Moreover, APC decreases leukocyte rolling and adhesion to the endothelium by suppressing adhesion molecules, causing reduced migration and accumulation of leukocytes in the air spaces 10111213. Investigators showed recently that cleavage of the protease-activated receptor-1 by the APC-endothelial cell protein C receptor complex exerts antiapoptotic and enhanced barrier-protective effects in endothelial cells via transactivation of the sphingosine-1-phosphate receptor 1415.

In sepsis and ALI both the production and the activation of protein C decrease because of extended endothelial damage 16. Elevation of the plasma concentration of protein C, which normally occurs upon activation of the coagulation system, is hence reversed with depletion of protein C and APC in septic states. Investigators reported recently that infusion of recombinant human activated protein C (rhAPC) enhances the plasma concentration of protein C and lowers the mortality rate of patients with severe sepsis, simultaneously reducing circulatory dysfunction and respirator-dependent days 1718. Although many effects of APC on coagulation and inflammation are well established 919, those effects dealing specifically with respiratory functions have been sparsely investigated 2021. Whether APC protects efficiently against ALI in human sepsis and in animals exposed to endotoxin is therefore still undetermined.

In sheep, LPS enhances lung microvascular pressure and permeability, thereby increasing the extravascular lung water content and deteriorating gas exchange 22232425, which is consistent with most signs of ALI in humans 26. Activation of inflammatory mediators and coagulation factors accompanies the changes in pulmonary hemodynamics and gas exchange 56720242527. We speculated whether rhAPC, given its anti-inflammatory and anticoagulation properties, would antagonize these changes. Our aim was to investigate the effects of rhAPC on sustained LPS-induced ALI in sheep over 24 hours by assessing changes in pulmonary hemodynamics, in extravascular lung water content, in pulmonary vascular permeability index and in gas exchange. Moreover, we followed alterations in selected variables of coagulation and inflammation.

The Norwegian Experimental Animal Board approved the study according to the rules and regulations of the Helsinki Convention for Use and Care of Animals.

One sheep from the LPS group died of respiratory failure after 20 hours, and one animal from the LPS + rhAPC group died of circulatory failure after 16 hours. None of the animals suffered from bleeding. As displayed in Figure 1 and Table 1, the pulmonary hemodynamic and volumetric parameters remained unchanged with rhAPC alone. The infusion of rhAPC, however, decreased the LPS-induced elevations of PAP, Pmo and PVRIdwn with intergroup differences (P < 0.05). The PVRI_up and the PVRI also declined with rhAPC treatment (P < 0.05), but with no intergroup differences (P = 0.63 and P = 0.17). The pulmonary arterial occlusion pressure remained unchanged. Infusion of LPS led to rapid and sustained rises in the PVPI (P < 0.05) and the EVLWI (P < 0.001), which were both antagonized by infusion of rhAPC (P < 0.001) with differences between the groups (P < 0.01 and P < 0.05 for PVPI and EVLWI, respectively).

The intrathoracic blood volume index, pulmonary blood volume index, right ventricle end-diastolic volume and right ventricle ejection fraction did not change significantly between the groups (data not shown). Moreover, infusion of rhAPC had no significant effects on the LPS-induced changes in right atrial pressure, heart rate, mean systemic arterial pressure, systemic vascular resistance index, CI, stroke volume index, LAP, and left and right ventricle stroke work indices in any of the groups (Table 1).

Figure 2 and Table 2 show no changes in oxygenation after rhAPC alone. Infusion of LPS decreased the PaO₂ abruptly in concert with a rise in the alveolar–arterial oxygen tension difference (P < 0.01), which improved significantly in response to administration of rhAPC (P < 0.05). Mixed venous oxygen saturation and PaCO₂ increased in the LPS + rhAPC group (P < 0.05), but with no significant intergroup differences. Exposure to LPS increased the body temperature and oxygen delivery index (P < 0.001), and tended to increase the oxygen consumption index (P = 0.051). Treatment with rhAPC, however, had no significant effect on these changes. The hemoglobin concentration initially increased during LPS infusion (P < 0.001) and decreased during rhAPC treatment (P < 0.001), but with no significant intergroup difference.

With rhAPC alone, the protein C and antithrombin III levels remained unchanged while the International Normalized Ratio and the fibrinogen level increased by nearly 10% (P < 0.05; Table 2 and Figure 3). Infusion of LPS reduced protein C by 49% and 33% in the LPS group and in the LPS + rhAPC group, respectively (P < 0.001), with an intergroup difference at 24 hours (P < 0.05). In both LPS groups, fibrinogen decreased to a nadir approximately 30% below baseline and then increased in rhAPC-treated animals (P < 0.05), but with no difference between groups (P = 0.08). The antithrombin III concentration fell and the International Normalized Ratio increased gradually (P < 0.001), with no significant intergroup difference.

As depicted in Figure 3, TNFα peaked at 12 hours (P < 0.05) in the LPS group and subsequently declined to significantly above baseline at 24 hours. In the LPS + rhAPC group, TNFα decreased beyond 4 hours, with an intergroup difference after 12 hours (P < 0.05).

In lung tissue from sheep exposed to LPS, we found a 65% to 75% reduction of the α and ε isoforms of PKC in the cytosolic fraction of tissue homogenate, as compared with samples from sham-operated animals (P < 0.05; Figure 4). In the LPS + rhAPC group, the cytosolic fraction of PKCα and PKCε remained significantly higher than in the LPS group (P < 0.05).

Table 3 presents the individual scores constituting the lung injury score. In animals exposed to LPS the lung injury score was elevated (P < 0.01), but the score was 34% lower in the LPS + rhAPC group (P < 0.01). Figure 5 shows photomicrographs of typical lung sections, with low-power fields to the left. The marked squares are displayed as high-power fields to the right. In sheep subjected to rhAPC alone (Figure 5a), nearly normal histological conditions were found. Lungs exposed to LPS (Figure 5b) displayed edematous, expanded and disrupted alveoli with interstitial neutrophile infiltration. All the latter changes were less pronounced in the LPS + rhAPC group (Figure 5c).

The present study demonstrates that continuously infused rhAPC alleviates endotoxin-induced lung injury in awake sheep. The enhanced EVLWI, a major characteristic of this model, decreased significantly upon administration of rhAPC. Dampening of pulmonary vascular pressures in parallel with a reduction of PVRI_dwn and improvement of gas exchange concerted the reduction of EVLWI. Administration of rhAPC also prevented elevation of the plasma concentration of TNFα, and normalized the plasma level of protein C. Moreover, rhAPC reduced the PVPI and counteracted the depletion of the α and ε isoforms of PKC from the cytosolic fraction of lung tissue homogenate, simultaneously maintaining pulmonary histological integrity.

Infusion of LPS provides a possibility of mimicking human ALI in different species, including sheep 2223. The fact that two animals died during the course of the 24-hour experiments illustrates the severity of this lung injury model, which is associated with a lethality of nearly 30% without supportive therapy 725. The LPS-induced increase in the EVLWI and the changes in lung hemodynamics are consistent with observations made in previous studies employing the same dose of endotoxin over 24 hours 2425. Despite the fact that we treated sheep with rhAPC specifically designed for the human, it markedly reversed the LPS-induced procoagulant and proinflammatory states. This might indicate a degree of cross-reaction between the inflammation and the coagulation systems of the two species, in agreement with other recent investigations 2021.

Following administration of rhAPC dosed as recommended for patients with severe sepsis 17, we noticed that the enhanced EVLWI declined towards baseline immediately after commencement of the drug. Furthermore, the PAP, Pmo and PVRI_dwn, all rising during infusion of LPS, decreased markedly during rhAPC infusion. We believe that the reduction of PVRI_dwn prompted the falls in PAP and Pmo, and therefore partly explains the rhAPC-induced reduction of EVLWI. The decrease in PAP is consistent with recent findings of Maybauer and colleagues, reporting that rhAPC, administered 1 hour subsequent to smoke inhalation and sepsis, improves pulmonary function in sheep 20. Correspondingly, Wang and colleagues, also using an ovine model of severe sepsis, noticed that rhAPC antagonizes lactic acidosis, derangements of hemodynamics and gas exchange as well as coagulation abnormalities 2 hours subsequent to peritoneal instillation of feces 21. The latter investigators also observed that rhAPC reduces the lung wet/dry weight ratio.

In the present study, we display the course of changes in EVLWI over 24 hours. Not obvious to us, however, was whether the decrease in EVLWI was a result of reduced Pmo, of reduced permeability, or of both reductions combined. Relating EVLWI to the lung blood volume, the PVPI supposedly indicates evolvement of edema due to increased leakiness of the lung vasculature 293132. We observed that PVPI increased and stabilized at a ratio of nearly 2.5 in the LPS group (Figure 1), which is slightly higher than that noticed by Kuzkov and colleagues in nonsurvivors of severe sepsis and ALI 31. Moreover, the ratio of approximately 1.5 that we observed in the LPS + rhAPC group was not significantly different from baseline. Although the PVPI is not unequivocally accepted as a measure of vascular permeability 33, this finding could indicate an APC-dependent preservation of vascular integrity.

Deterioration of gas exchange accompanied the accumulation of extravascular lung water, as shown in previous investigations of ALI after LPS treatment in sheep 52334. The impaired oxygenation (Figure 2) improved significantly in the LPS + rhAPC group in parallel with a normalization of PaCO₂, consistent with recent animal reports 2021 and with the PROWESS study 17. The latter workers demonstrated faster resolution of cardiovascular and respiratory dysfunctions in rhAPC-treated patients 18.

In the present investigation, administration of LPS led to depletion of protein C and fibrinogen. The increase in the plasma concentration of fibrinogen during treatment with rhAPC is consistent with the findings of Taylor and colleagues, who concluded that APC prevents both the coagulopathy and the lethal effects of infusion of live E. coli in baboons 19. Moreover, the better maintained concentration of protein C following administration of rhAPC agrees with corresponding findings of Wang and colleagues in septic sheep treated with rhAPC 21. According to a recent investigation, a decrease in protein C is a regular finding in patients with severe sepsis and an early and independent predictor of sepsis mortality 35.

In the present study, the enhanced plasma concentration of TNFα and the depletion of PKCα and PKCε from the cytosolic fraction of lung tissue homogenate after LPS, as previously reported 24, vanished during treatment with rhAPC. Moreover, histological examination revealed less inflammatory changes in rhAPC-treated endotoxemic animals, as evaluated by reduced neutrophil infiltration and decreased edematous changes in lung tissue. Our results agree with an investigation in endotoxemic rats demonstrating an APC-induced decrease in the lung wet/dry weight ratio in parallel with reduced accumulation of leukocytes in the lungs and a fall in the plasma concentration of TNFα 1236. In volunteers subjected to instillation of LPS into the airways, Nick and colleagues demonstrated increased neutrophil migration into the air spaces that was prevented with rhAPC 13. The latter observation is consistent with our observation that the lung injury score decreased mainly as a result of less neutrophils in the lung sections of rhAPC-treated endotoxemic sheep. Moreover, Kirschenbaum and colleagues noticed that rhAPC prevents neutrophil-platelet interaction with endothelial cells from healthy subjects, when stimulated with plasma from patients in septic shock. These authors suggested that rhAPC preserves microvascular patency in severe sepsis 37, which is consistent with the regression of LPS-induced increases in the PVPI and the EVLWI after infusion of rhAPC in the present study. In this study, plasma concentrations of TNFα peaked 12 hours after administration of LPS. The sustained high levels of TNFα in the LPS group is consistent with observations in patients with sepsis, in whom persistent high levels of TNFα correlates with severity of illness 38.

Some limitations must be taken into account when evaluating the results of this study. Firstly, coagulation and inflammatory markers have in total 17% missing values that reduce the power of the study as these variables are concerned. Secondly, we used human kits for determination of TNFα and for coagulation tests, due to the lack of kits specifically designed for sheep. As evolutionary old systems, however, both the innate immune system and the coagulation system expectedly have a high degree of cross-reactivity in mammals 39. The change in plasma concentration of TNFα during exposure to LPS is consistent with previously reported observations from our group where we used human antibodies in combination with the same automated system for cytokine analysis 24.

The present study demonstrates that infused rhAPC ameliorates ovine LPS-induced ALI, as indicated by a decrease in the EVLWI associated with improvements of gas exchange and markers of inflammation and coagulation. We suggest that treatment with rhAPC reduces lung fluid extravasation by dampening both the pressure and the permeability of the lung microvasculature. Further studies are warranted to determine whether these beneficial effects also apply in patients suffering from sepsis – and in patients from nonsepsis-induced ALI.

• In ovine endotoxin-induced lung injury, rhAPC counteracts the increments in pulmonary vascular downstream resistance and pulmonary vascular pressures.

• rhAPC dampens inflammation and the pulmonary vascular permeability index, antagonizing the enhanced extravascular lung water and the deranged arterial oxygenation.

AaO₂ = alveolar-arterial oxygen tension difference; ALI = acute lung injury; APC = activated protein C; CI = cardiac index; EVLWI = extravascular lung water index; H&E = hematoxylin and eosin; LAP = left atrial pressure; LPS = lipopolysaccharide; NF = nuclear factor; PaCO₂ = partial tension of carbon dioxide in arterial blood; PaO₂ = partial tension of oxygen in arterial blood; PAP = pulmonary arterial pressure; PKC = protein kinase C; Pmo = pulmonary capillary micro-occlusion pressure; PVPI = pulmonary vascular permeability index; PVRI = pulmonary vascular resistance index; PVRI_dwn = pulmonary vascular downstream resistance index; PVRI_up = pulmonary vascular upstream resistance index; rhAPC = recombinant human activated protein C; TNF = tumor necrosis factor.

The present study was supported by Helse Nord (Project Number 4001.721.477) and by the Department of Anesthesiology, University Hospital of North Norway, Tromsø, Norway, by the Department of Anesthesiology, Institute of Clinical Medicine, University of Tromsø, Norway, and in part by Eli Lilly & Co., USA. The support from Eli Lilly & Co. was limited to a reimbursement of our expenses for experimental animals and free use of the Xigris^® (rhAPC) that was used in the study. MYK is a member of the Advisory Board of Pulsion Medical Systems (München, Germany).

KW analyzed the data and participated in the experiments and in writing the manuscript. VNK and MYK participated in the design of the study, in experimentation and in data analysis. MAS, BL, OCI and KY contributed with biochemical analyses and suggested textual improvements. LJB participated in the administration and design of the study, and drafted the manuscript. All authors have read and approved the final manuscript.