Plasma-derived human antithrombin attenuates ventilator-induced coagulopathy but not inflammation in a Streptococcus pneumoniae pneumonia model in rats

Authors

  • H. ASLAMI,

    1. Interdepartmental Division of Critical Care Medicine, University of Toronto, Keenan Research Center, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, ON, Canada
    2. Laboratory of Experimental Intensive Care and Anesthesiology (LEICA), Academic Medical Center
    Search for more papers by this author
  • J. J. HAITSMA,

    1. Interdepartmental Division of Critical Care Medicine, University of Toronto, Keenan Research Center, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, ON, Canada
    Search for more papers by this author
  • J. J. HOFSTRA,

    1. Laboratory of Experimental Intensive Care and Anesthesiology (LEICA), Academic Medical Center
    2. Department of Intensive Care, Academic Medical Center
    Search for more papers by this author
  • S. FLORQUIN,

    1. Department of Pathology, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands
    Search for more papers by this author
  • C. DOS SANTOS,

    1. Interdepartmental Division of Critical Care Medicine, University of Toronto, Keenan Research Center, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, ON, Canada
    Search for more papers by this author
  • C. STREUTKER,

    1. Department of Pathology, University of Toronto, Keenan Research Center, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, ON, Canada
    Search for more papers by this author
  • H. ZHANG,

    1. Interdepartmental Division of Critical Care Medicine, University of Toronto, Keenan Research Center, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, ON, Canada
    Search for more papers by this author
  • M. LEVI,

    1. Department of Internal Medicine, Academic Medical Center, University of Amsterdam, Amsterdam, the Netherlands
    Search for more papers by this author
  • A. S. SLUTSKY,

    1. Interdepartmental Division of Critical Care Medicine, University of Toronto, Keenan Research Center, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, ON, Canada
    Search for more papers by this author
  • M. J. SCHULTZ

    1. Laboratory of Experimental Intensive Care and Anesthesiology (LEICA), Academic Medical Center
    2. Department of Intensive Care, Academic Medical Center
    Search for more papers by this author

Hamid Aslami, Laboratory of Experimental Intensive Care and Anesthesiology, H0–127 Academic Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands.
Tel.: +31 20 5668224; fax: +31 20 6979441.
E-mail: h.aslami@amc.uva.nl

Abstract

Summary.  Background: Mechanical ventilation exaggerates pneumonia-associated pulmonary coagulopathy and inflammation. We hypothesized that the administration of plasma-derived human antithrombin (AT), one of the natural inhibitors of coagulation, prevents ventilator-induced pulmonary coagulopathy, inflammation and bacterial outgrowth in a Streptococcus pneumoniae pneumonia model in rats. Methods: Forty-eight hours after induction of S. pneumoniae pneumonia rats were subjected to mechanical ventilation (tidal volume 12 mL kg−1, positive end-expiratory pressure 0 cmH2O and inspired oxygen fraction 40%). Rats were randomized to systemic treatment with AT (250 IU administered intravenously (i.v.) before the start of mechanical ventilation) or placebo (saline). Non-ventilated, non-infected rats and non-ventilated rats with pneumonia served as controls. The primary endpoints were pulmonary coagulation and inflammation in bronchoalveolar lavage fluid (BALF). Results: Pneumonia was characterized by local activation of coagulation and inhibition of fibrinolysis, resulting in increased levels of fibrin degradation products and fibrin deposition in the lung. Mechanical ventilation exaggerated pulmonary coagulopathy and inflammation. Systemic administration of AT led to supra-normal BALF levels of AT and decreased ventilator-associated activation of coagulation. AT neither affected pulmonary inflammation nor bacterial outgrowth from the lungs or blood.Conclusions: Plasma-derived human AT attenuates ventilator-induced coagulopathy, but not inflammation and bacterial outgrowth in a S. pneumoniae pneumonia model in rats.

Introduction

Systemic disturbances in coagulation and fibrinolysis are common occurrences in sepsis [1]. Similar but pulmonary disturbances in fibrin turnover have been described in patients with acute lung injury (ALI) or pneumonia [2]. In addition, pulmonary coagulopathy has been found with forms of mechanical ventilation that have the potential to injure the lungs [3]. Under normal conditions, coagulation is under the control of natural inhibitors of coagulation, for example activated protein C (APC), tissue factor pathway inhibitor (TFPI) and antithrombin (AT) [4–6]. In sepsis, systemic levels of these natural inhibitors of coagulation are reduced because of consumption, diminished production and increased degradation by inflammatory proteases [7–11]. In ALI and pneumonia, local levels of natural inhibitors of coagulation are excessively low as well [12], presumably via similar mechanisms. It is unclear whether local levels of natural inhibitors of coagulation can also be affected by mechanical ventilation.

There is circumstantial evidence that anticoagulant treatment may benefit patients with ALI or pneumonia [13–15]. The lung-protective effects of systemic treatment with AT was demonstrated in a relatively limited number of patients with severe sepsis [16]. However, AT treatment failed to improve survival in a larger study of patients with severe sepsis, although it must be mentioned that no subgroup analysis of patients with pneumonia as the primary source of sepsis was performed [17]. In a model of non-ventilated rats with Streptococcus pneumoniae pneumonia, we recently demonstrated that treatment with plasma-derived human AT ameliorated pulmonary coagulopathy and inflammation, as well as having antimicrobial effects [18,19].

Mechanical ventilation may induce additional lung injury in patients with established ALI or pneumonia [20–22] or may even initiate lung injury in patients without lung injury at onset of mechanical ventilation [23,24]. One previous human study demonstrated ventilator-associated lung injury to be accompanied by pulmonary coagulopathy [3]. Several animal studies confirmed the existence of ventilator-induced coagulopathy [25–27]. One approach to prevent or treat mechanical ventilator-induced lung injury is to use natural inhibitors of coagulation. In the present study, we hypothesized systemic treatment with plasma-derived human AT to attenuate pulmonary coagulopathy, inflammation and bacterial outgrowth during mechanical ventilation in a well-established S. pneumoniae pneumonia model in rats.

Materials and methods

Animals

In accordance with the Canadian Council of Animal Care guidelines, the animal care committee at St. Michael’s Hospital approved the study. A total of 91 male Sprague–Dawley rats (weight 240–300 g; Charles River, QC, Canada) were included. A flow diagram of the study is presented in Fig. 1.

Figure 1.

 Schematic representation of the study groups including the numbers of rats.

Induction of pneumonia

As previously described [18,27], per rat ∼ 1 x 107 colony-forming units (CFU) S. pneumoniae serotype 3 (American Type Culture Collection 6303; Manassas, VA, USA) were aerosolized intratracheally using a trans-oral miniature nebulizer (Penn-Century, Philadelphia, PA, USA) [27] under light anesthesia (65% nitrous oxide/33% oxygen/2% isoflurane). Non-infected rats received sterile normal saline. Rats were allowed to recover from the anesthesia and returned to their cages with food and water ad libidum. Rats were monitored every 8 h and received supplemental fluids every 24 h by intraperitoneal injection (30 mL kg−1 lactated Ringer’s solution).

Experimental protocol

Forty-eight hours after administration of S. pneumoniae or saline, rats were anesthetized with intraperitoneal injection of 100 mg kg−1 ketamine (Ketalean, Bimeda-MTC Animal Health Inc., Cambridge, ON, Canada) and 6 mg kg−1 xylazine (Rompun, Bayer Inc., Toronto, ON, Canada). A sterile cannula was inserted into the trachea, and a polyethylene catheter was inserted into the carotid artery for hemodynamic monitoring. Anesthesia was maintained by continuous intravenous (i.v.) infusion of ketamine (15 mg kg h−1) and xylazine (3 mg kg h−1); paralysis was achieved by infusing pancuronium (0.35 mg kg h−1) (Sandoz, QC, Canada). All medications were infused via a tail vein catheter. The body temperature was maintained at 37 °C using a heating pad.

Mechanical ventilation

Rats were subjected to mechanical ventilation inducing ventilator-induced coagulopathy [27]. In brief, rats were ventilated in a volume-controlled mode, with a VT of 12 mL kg−1 body weight and a positive end-expiratory pressure (PEEP) of 0 cmH2O. The respiratory rate was set at 18 breaths min−1 at baseline and adjusted to maintain normocapnia. The inspiratory-to-expiratory ratio was set at 1 : 2 and the inspired oxygen fraction (FiO2) at 40%, and adjusted to prevent hypoxemia. Rats were mechanically ventilated for 3 h.

Experimental groups

Infected and non-infected rats, either ventilated or not ventilated, were randomized to i.v. treatment with plasma-derived human AT or placebo (Fig. 1).

Antithrombin treatment

Plasma-derived human AT (Baxter, Vienna, Austria) was administered through the tail vein catheter before the start of mechanical ventilation. For this, 250 IU kg−1 plasma-derived human AT was diluted in 1 mL of sterile normal saline and administered as a bolus. A bolus of 250 IU kg−1 has been proven to be an effective dose in achieving supraphysiological AT levels in the lungs [28,29]. Rats not receiving AT received 1 mL of sterile normal saline.

Hourly measurements

The carotid arterial line was used to continuously measure blood pressure and to obtain hourly blood for arterial blood gases, using a blood gas analyzer (Ciba-Corning Model 248 blood gas analyzer; Corning Medical, Medfield, MA, USA). Boluses of normal saline solution were administered i.v. to keep the systemic blood pressure > 70 mmHg.

Sampling

At the end of the experiment, blood (5 mL) was taken from the arterial line, followed by bronchoalveolar lavage (BAL) with saline (2 × 30 mL kg−1 body weight) in seven rats. For each study group, the lungs of three rats were used for histopathology.

Sample handling

The number of viable bacteria in BAL fluid (F) was determined by plating 10-fold dilutions on sheep-blood agar plates; the plates were incubated at 37 °C in 5% CO2; CFU were counted after 24 h. Neutrophil counts were performed using conventional techniques immediately at the end of the experiment in BALF. Subsequently, BALF was centrifuged at 4 °C at 400 g for 10 min to remove cells and cellular debris. BALF was snap-frozen in liquid nitrogen and stored at −80 °C until further processing. Blood samples of 100 μL were taken hourly and cultured undiluted onto sheep-blood agar plates, incubated at 37 °C in 5% CO2 and counted after 24 h.

Coagulation and fibrinolysis

Levels of TATc and plasminogen activator inhibitor (PAI)-1 and fibrin degradation products (FDP) were measured by ELISA according to the manufacturer’s instructions (TATc; Behring, Marburg, Germany; PAI-1, Biopool, Umeå, Sweden; FDP, Asserachrom D-Di, Diagnostica Stago, Asnieres-sur-Seine, France). Levels of PAA and AT were measured using an automated amidolytical assay [30,31]. Measurements were done simultaneously in BALF and plasma. FDP was only measured in BALF.

Pulmonary inflammation and injury

Levels of tumor necrosis factor (TNF)-α, interleukin (IL)-6 and macrophage inflammatory protein (MIP)-2 in BALF and plasma were measured using rat-specific ELISA kits according to the manufacturer’s instructions (BioSource, Camarillo, CA, USA). Total protein as a determinant for lung permeability was determined in BALF using the Bradford method as described previously [32].

Histopathology and immunostaining

Lung histopathology was scored as described previously [27] by a pathologist who was not aware of randomization groups. Briefly, lungs were removed en bloc without preceding BAL and fixed in 4% paraformaldehyde under a mean pressure of 10 cmH2O. Scoring categories of none, mild, moderate and severe (0, 1, 2 and 3, respectively) were assigned for each of the following criteria: alveolar collapse, alveolar hemorrhage, perivascular edema, neutrophil infiltration, alveolar membranes and alveolar edema. Fibrin deposition was scored as categories of none (0, 0% involved), mild (1, 0–25% involved), moderate (2, 25–50% involved) and severe (3, > 50% involved) at an alveolar, pleural and vascular level using immunostaining (α-fibrinogen, goat, α-mouse; Accurate Chemical & Scientific Corp., Westburry, NY, USA) [33] on lung slides. The lung injury score as well as the fibrin deposition score were calculated as sum of the categories.

Statistical analysis

Statistical analysis was performed using GraphPad Prism version 5 (GraphPad software, La Jolla, CA, USA). To investigate the effect of systemic treatment with AT on ventilator-induced coagulopathy, groups were compared using one-way analysis of variance (anova) with Bonferroni’s multiple comparison test or a non-parametric test according to the data distribution. We first compared each group with its control. Then we compared ventilated-infected animals treated with plasma-derived human AT with ventilated-infected animals treated with placebo. A P-value of < 0.05 was considered statistically significant. The data are shown as mean ± standard deviation (SD) or median with interquartile range (IQR) where appropriate.

Results

Pneumonia and mechanical ventilation

The mortality rate of infected rats before the start of mechanical ventilation was 21% (Fig. 1). All rats survived 3 h of mechanical ventilation. During mechanical ventilation, mean arterial pressure and PaO2/FiO2 were significantly lower in infected rats compared with non-infected rats (Table 1, < 0.05). The arterial carbon dioxide pressure and respiratory rate, adjusted over time to maintain normocapnia, as well as administered fluids were not different between infected and non-infected rats (Table 1). Infected rats had bilateral macroscopic lung infiltrates with histopathological analysis showing evident signs of pneumonia (Table 3 and Fig. 2, < 0.05). Numbers of S. pneumoniae CFU in BALF and blood did not change with mechanical ventilation (Table 1).

Table 1.  Mean arterial pressure (MAP), ratio between arterial oxygen pressure (PaO2) and inspired oxygen pressure (FiO2), arterial carbon dioxide pressure (PaCO2) and respiratory rate (RR) in infected and non infected rats mechanically ventilated and treated with placebo or antithrombin (AT). The number of colony-forming units (CFU) was determined in brochoalveolar lavage fluid (BALF) and the blood of infected rats at the end of the experiment and hourly, respectively
 Non-ventilated ratsVentilated rats
Non-infected ratsInfectedNon-infected ratsInfected rats
PlaceboATPlaceboATPlaceboATPlaceboAT
  1. Data: mean ± standard deviation (SD). *Non-infected vs. infected; P < 0.05. AT, antithrombin; RR, respiratory rate; BALF, brochoalveolar lavage fluid; CFU, colony-forming units.

Time (h)T = 3T = 3T = 3T = 3T = 0T = 1T = 2T = 3T = 0T = 1T = 2T = 3T = 0T = 1T = 2T = 3T = 0T = 1T = 2T = 3
MAP (mmHg)134 ± 13117 ± 19117 ± 13112 ± 9129 ± 21111 ± 9126 ± 18107 ± 997 ± 29*84 ± 12*85 ± 14*83 ± 16*92 ± 887 ± 1684 ± 1771 ± 14
PaO2/FiO2 (mmHg)438 ± 54451 ± 45471 ± 65458 ± 56476 ± 44469 ± 31463 ± 39451 ± 39411 ± 72354 ± 59*305 ± 39*308 ± 66*355 ± 34305 ± 39311 ± 51288 ± 64
PaCO2 (mmHg)42 ± 543 ± 739 ± 1037 ± 638 ± 643 ± 644 ± 541 ± 538 ± 839 ± 1047 ± 941 ± 544 ± 545 ± 1441 ± 740 ± 6
RR (breath min−1)18 ± 019 ± 222 ± 323 ± 418 ± 018 ± 021 ± 523 ± 418 ± 018 ± 022 ± 323 ± 318 ± 120 ± 320 ± 322 ± 5
BALF (×105 CFU/mL−1)5.9 ± 4.93.0 ± 4.119 ± 19.016 ± 2.7
Blood (×103 CFU mL−1)0.9 ± 0.80.8 ± 1.70.6 ± 1.00.9 ± 1.01.8 ± 3.60.7 ± 0.80.4 ± 0.20.5 ± 0.40.6 ± 0.50.4 ± 0.3
Figure 2.

 Representative hematoxylin and eosin-stained lung slides from non-ventilated (A–D) and ventilated rats (E–H) with (C, D) and without (E, F) pneumonia either treated with bolus plasma-derived human antithrombin (B, D, F, H) or placebo (A, C, E, G). Magnification ×20 and inserts ∼×22.

Pulmonary and systemic coagulopathy

Induction of pneumonia resulted in severe coagulopathy with signs of activation of coagulation (Figs 3A and 4A), decreased levels of AT (Fig. 3B and 4B) and inhibition of fibrinolysis (Figs 3C,D and 4C,D) in BALF and plasma (all, < 0.05). In addition, increased levels of FDP (Fig. 3E) and increased fibrin deposition (Table 3, Fig. 5C) were found (both, < 0.05). While mechanical ventilation only mildly induced coagulopathy in non-infected rats, it increased pneumonia-induced coagulopathy both locally and systemically (Figs 3 and 4, both < 0.05). Levels of AT were lower in ventilated rats, although the differences with non-ventilated rats did not reach statistical significance (Figs 3B and 4B). Mechanical ventilation did not increase levels of FDP (Fig. 3E) and fibrin deposition in the lung (Table 2).

Figure 3.

 Effect of induction of pneumonia, mechanical ventilation and administration of plasma-derived human antithrombin (AT) on pulmonary levels of thrombin antithrombin complex (TATc), antithrombin (AT), plasminogen activator activity (PAA) and plasminogen activator inhibitor (PAI-1). Horizontal lines indicate means. *Non-infected vs. pneumonia; #pneumonia vs. pneumonia and ventilated; pneumonia vs. pneumonia + AT; §pneumonia and ventilated + placebo vs. pneumonia and ventilated + AT; Фnon-infected vs. non-infected and ventilated; †non-infected and ventilated + placebo vs. non-infected and ventilated + AT, < 0.05.

Figure 4.

 Effect of induction of pneumonia, mechanical ventilation and administration of plasma-derived human antithrombin (AT) on systemic levels of thrombin antithrombin complex (TATc), antithrombin (AT), plasminogen activator activity (PAA) and plasminogen activator inhibitor (PAI-1). Horizontal lines indicate means. *Non-infected vs. pneumonia; #pneumonia vs. pneumonia and ventilated; pneumonia vs. pneumonia + AT; §pneumonia and ventilated + placebo vs. pneumonia and ventilated + AT; Фnon-infected vs. non-infected and ventilated; non-infected and ventilated + placebo vs. non-infected and ventilated + AT, < 0.05.

Figure 5.

 Representative fibrinogen-stained lung slides from non-ventilated (A–D) and ventilated rats (E–H) with (C, D) and without (E, F) pneumonia either treated with a bolus of plasma-derived human antithrombin (B, D, F, H) or placebo (A, C, E, G). Magnification.

Table 2.  Tumor necrosis factor (TNF)-α, interleukin (IL)-6 and macrophage inflammatory protein (MIP)-2 concentrations in bronchoalveolar lavage fluid of non-infected, infected and either non-ventilated or ventilated rats, treated with placebo or antithrombin (AT)
CytokineNon-ventilated ratsVentilated rats
Non-infectedInfectedNon-infectedInfected
PlaceboATPlaceboATPlaceboATPlaceboAT
  1. Data represent median with the interquartile range [IQR]. *Non-infected vs. infected, infected vs. infected ventilated rats, non-infected vs. non-infected ventilated rats, §non-ventilated infected vs. non-ventilated infected treated with AT, < 0.05.

TNF-α (ng mL−1)0.2 [0.2–0.3]0.1 [0.1–0.1]0.4 [0.1–0.8]*0.1 [0.1–0.4]0.1[0.1–0.1]0.1 [0.1–0.1]1.6 [0.4–1.3]1.7 [1.2–2.0]
IL-6 (ng mL−1)0.2 [0.1–0.5]0.2 [0.2–0.2]0.9 [0.6–1.2]*0.2 [0.2–1.5]0.4 [0.3–0.5]0.4 [0.2–0.4]2.0 [1.3–3.6]2.4 [0.8–3.9]
MIP-2 (ng mL−1)0.7 [0.7–0.9]0.0 [0.1–0.1]1.6 [0.7–4.6]*0.3 [0.2–0.6]§0.4 [0.2–0.5]0.2 [0.2–0.3]5.3 [2.9–9.6]6.4 [4.7–7.3]

The effect of AT on coagulopathy

Systemic administration of plasma-derived human AT resulted in supra-normal levels of AT in BALF and plasma (Figs 3B and 4B, both < 0.05). AT did not affect mild coagulopathy in ventilated but non-infected rats. However, in infected rats AT attenuated coagulopathy to levels seen in non-ventilated rats (Figs 3 and 4), with decreased fibrin deposition (Table 3 and Fig. 5H, < 0.05) correlating with a significant increase in FDP in the alveolar space (Fig. 3E, < 0.05). AT did not affect levels of PAA and PAI-1 in BALF or plasma.

Table 3.  Neutrophils, protein levels, lung injury and fibrin deposition scores of non-infected, infected and either non-ventilated or ventilated rats, treated with placebo or antithrombin (AT)
ParameterNon-ventilated ratsVentilated rats
Non-infectedInfectedNon-infectedInfected
SalineATSalineATSalineATSalineAT
  1. Data represent mean ± SD. *Non-infected vs. infected, non-infected vs. non-infected ventilated rats, infected vs. infected ventilated rats. §non-infected ventilated vs. non-infected ventilated treated with AT, < 0.05. Infected ventilated vs. infected ventilated treated with AT.

Neutrophils (neutrophils ×105 mL−1)0.0 ± 0.00.0 ± 0.05.8 ± 3.8*8.2 ± 3.70.8 ± 0.60.4 ± 0.414 ± 6.416 ± 6.1
Protein (mg mL−1)0.1 ± 0.00.1 ± 0.10.2 ± 0.10.3 ± 0.20.3 ± 0.10.2 ± 0.10.3 ± 0.10.4 ± 0.1
Lung injury score (sum score)0.0 ± 0.01.0 ± 0.07.0 ± 5.7*6.9 ± 1.25.0 ± 2.02.6 ± 1.4§9.5 ± 2.110.5 ± 2.1
Fibrin (sum score)0.0 ± 0.01.0 ± 0.02.6 ± 1.6*2.5 ± 1.12.4 ± 1.12.1 ± 1.33.6 ± 2.31.5 ± 0.9

The effect of AT on pulmonary inflammation and injury

Induction of pneumonia resulted in increased pulmonary levels of pro-inflammatory cytokines including TNF-α, IL-6 and MIP-2 (Table 2, all < 0.05), increased neutrophil influx and a worsened lung injury score (Table 3 and Fig. 2C, all < 0.05). While mechanical ventilation caused only mild inflammation in non-infected rats, it increased pulmonary levels of cytokines (Table 2), pulmonary neutrophil influx (Table 3) and the lung injury score (Table 3 and Fig. 2G) in infected rats (all < 0.05).

AT reduced only MIP-2 concentrations in infected non-ventilated rats, other cytokines and markers for lung injury were not affected (Tables 2 and 3). Also, no differences in mean arterial pressure and PaO2/FiO2 were observed between AT- and placebo-treated rats (Table 1).

The effect of AT on bacterial outgrowth

Treatment with plasma-derived human AT did not affect outgrowth of bacteria in the lung or blood (Table 1).

Discussion

The present study demonstrates mechanical ventilation to exaggerate pneumonia-induced activation of pulmonary coagulation and inhibition of pulmonary fibrinolysis, increasing fibrin deposition in lungs. Systemic treatment with plasma-derived human AT attenuates ventilator-induced pulmonary coagulopathy, but failed to reduce a rise in pulmonary cytokines caused by S. pneumoniae pneumonia. AT also did not reduce bacterial outgrowth in the lung or blood.

Our study is in line with previous studies reporting on ventilator-induced pulmonary coagulopathy. Indeed, mechanical ventilation has been found to have the potential to induce pro-coagulant changes in the lung in preclinical studies [26,27] as well as clinical trials [3,34]. In a clinically relevant model of S. pneumoniae pneumonia in rats, mechanical ventilation exaggerated pulmonary coagulopathy [27]. Similarly, mechanical ventilation has been found to exaggerate pulmonary coagulopathy in patients suffering from pneumonia [34]. AT attenuates coagulation activation by neutralizing thrombin by forming TATc. AT also inhibits other coagulation factors in the coagulation cascade resulting in less thrombin generation and eventually leading to diminished TATc complexes [18,19]. Mechanical ventilation has the potential to inhibit fibrinolysis by increasing the production of local levels of PAI-1 [26]. We observed enhanced fibrinolysis in the control rats treated with AT. This is surprising as, the fibrinolytic effect of AT has not been described previously, suggesting an alternative unexplored pathway of AT. The present study expands on earlier work by showing that systemic treatment with a natural anticoagulant has the potential to reduce pulmonary coagulopathy induced by mechanical ventilator in pneumonia [27].

In the lungs, there is widespread cross-talk between coagulation and inflammation [35,36]. Treatment with plasma-derived human AT attenuated ventilator-associated pulmonary coagulopathy, and reduced pulmonary fibrin deposition. AT has the potential to inhibit leukocyte activation and migration, inhibit NF-kB production [8,9] and binds direct with neutrophils thus attenuating cytokine release [11]. This is partly supported by our data, as AT reduced the chemokine, MIP2, in infected but not ventilated rats. The lung injury score was also reduced by AT treatment in the ventilated but not the infected rat. In the present study, the anti-inflammatory effects of AT were, however, blunted in ventilated-infected rats. The present results also did not confirm earlier findings suggesting antimicrobial effects of AT therapy in non-ventilated S. pneumoniae pneumonia [18,19]. Importantly, AT has antimicrobial and anti-inflammatory effects when administered repeatedly and before bacterial installation in the lung [19]. Similarly, APC reduced inflammation and coagulopathy when administered early in the course of pneumonia [37]. In ALI, an early influx of mature circulating neutrophils to the injured lung is followed by a slower sustained recruitment of neutrophils from the bone marrow [10,38]. The rather short treatment with AT in the present study (3 vs. 48 h in previous studies) may have precluded anti-inflammatory and antimicrobial effects of plasma-derived AT observed in the previous studies. We believe that prolonged treatment and repeated administration could result in amelioration of lung injury in our model of ventilator-induced coagulopathy and possible outcome. Future studies should address this further.

Mechanical ventilation although necessary, is a double-edge sword, were the tradeoff between improved ventilation (oxygenation and normocapnia) sometimes puts the lungs at risk of additional injury [23,24]. If we consider pulmonary coagulopathy to be intrinsic to ALI, and lung-injurious mechanical ventilation to be one (additional) causative factor for developing pulmonary coagulopathy, there may be a role for anticoagulant strategies. This hypothesis is supported by data from studies in ALI that demonstrate decreased plasma levels of protein C and increased plasma levels of PAI-1 are independent risk factors for mortality and adverse clinical outcomes [39]. Although anti-inflammatory effects by anticoagulants may be beneficial, several risks associated with altering pulmonary coagulation need to be recognized. Clot formation may be helpful in preventing the spread of pathogens by containing them to the site of infection within locally formed clots, for example: heparin treatment facilitated the spread of bacteremia in models of pneumonia [18]. These findings underline that coagulation and inflammation are closely related in pivotal host defense mechanisms, and interference with these pathways should be performed with great care.

The thought that systemic coagulopathy is central to the pathogenesis of tissue injury is not new and has long been implicated in the pathogenesis of organ dysfunction in sepsis [40]. Restoring the balance in fibrin turnover has been tested in several trials of patients with severe sepsis with APC [41], TFPI [42] or AT [17]. Only APC proved to be beneficial in patients with severe sepsis [41]. There is an ongoing discussion on the exact mechanism by which APC prevents morbidity and mortality. In humans challenged with LPS, APC limits pulmonary coagulopathy [43] and in experimental acute lung injury, APC limits coagulation, neutrophil influx [44] and the prevention of endothelial barrier disruption, thereby improving gas exchange and reducing cytokine production [45]. However, in a recent randomized clinical trial, APC treatment did not reduce ventilator-free days or 60-day mortality in patients without severe sepsis possibly as a result of a lack of statistical power [46]. In the two trials testing TFPI and AT, heparin use was permitted as prophylaxis for deep vein thrombosis. Heparin causes release of TFPI from its binding sites on endothelium [47]. A survival benefit was observed in patients with severe sepsis treated with TFPI who were not simultaneously treated with heparin [42]. Similarly effects have been observed in the AT trial [17]. As heparin accelerates about 1000-fold the anticoagulant reaction of AT [48], an increase in hemorrhagic risk was observed in patients receiving both AT and heparin [17]. Importantly, a survival benefit was observed in patients who did not receive therapeutic heparin [17].

Administration of plasma derived human AT per se might evoke an inflammatory response in rats. To test this hypothesis, we treated non-infected rats with AT and sacrificed them 3 h later. Although the lung injury and fibrin score tended to be higher in the treated rats, more specific lung injury markers, for example cytokines and neutrophil influx, were similar between the groups. This suggests that there is not a direct immune response to high local human-derived plasma AT levels in rats.

We found no effect of AT on ventilator-induced coagulopathy and inflammation in non-infected rats. This might be surprising, as APC have been shown to be protective in healthy lungs subjected to ventilation [45,49]. In the present study, coagulation was only mildly activated by the ventilator, mechanical ventilation in combination with pneumonia, however, strongly increased TATc production. Similarly, lung injury parameters were only mildly increased by the ventilator in the non-infected rats. A more injurious strategy using larger tidal volumes may demonstrate an anti-inflammatory effect of AT.

The present study has several limitations. First, the duration of mechanical ventilation in our model was relatively short. Indeed, in critically ill patients with pneumonia the duration of mechanical ventilation usually last days instead of hours [50]. However, we previously demonstrated that 3 h of ventilation resulted in significant effects on ventilator-induced coagulopathy, and we observed these differences again within this short-time frame. Second, patients suffering from pneumonia are treated with additional supportive care, including antimicrobial therapy and i.v. fluids, which we did not apply in the present study. These limitations should be taken in consideration when interpreting the results of the present study.

We used higher tidal volumes than recommended in patients with ALI [51]. Unfortunately, these higher tidal volumes are currently still being used in mechanically ventilated patients [52], and others have argued they are essential to correct severe hypercapnia [53]. Although we agree that tidal volume should be limited, understanding the effects of higher tidal volumes as well as investigating possible treatments to ameliorate the injurious effects of high tidal volume ventilation could drive research forward in the field of mechanical ventilation-induced lung injury. Alternative supportive therapies including those that allow extra-corporal CO2 removal are presently available [53,54], but costs, availability and other complications limit their use. In addition, lung-protective mechanical ventilation has not (yet) been proven to benefit patients who do not (yet) meet the consensus criteria for ALI [55].

The present study has several strengths. Physiological models of pneumonia in combination with mechanical ventilation are limited [27]. Our rat model of S. pneumoniae pneumonia may better resemble the clinical situation than previously used models. Indeed, S. pneumoniae is the leading cause of community-acquired pneumonia, leading to hospitalization and eventually admission to the intensive care unit requiring mechanical ventilation [56]. Furthermore, the tidal volumes used in our model resemble clinical practice, more than any other study; 12 mL kg−1 per body weight compared with other animal studies using a tidal volume up to 45 mL kg−1 per BW. Similarly, timing (early before initiation of mechanical ventilation) and route (systemic infusion) of AT may well mimic the clinically situation.

In conclusion, in a S. pneumoniae pneumonia model in rats, ventilator-induced pulmonary coagulopathy was attenuated by systemic administration of plasma-derived human AT. Ventilator-induced pulmonary inflammation as well as bacterial outgrowth were not affected by AT treatment.

Addendum

H. Aslami and J.J. Haitsma; designed the research, performed the experiments, analyzed the data and wrote the paper, Jorrit J.Hostra; designed the research, measured coagulation and fibrinolysis parameters, S. Florquin; scored for fibrinogen deposition on lung slides, C. dos Santos; designed the research, C. Streutker; scored for lung injury score on lung slides, H. Zhang. designed the research, M. Levi; measured coagulation and fibrinolysis parameters, A.S. Slutsky; designed the research, M.J. Schultz; designed the research, performed the experiments, analyzed the data and wrote the paper. The co-authors contributed to the discussion of the paper.

Acknowledgements

We would like to thank Dr C. W. Wieland and A. M. Tuip from the Laboratory of Experimental Intensive Care and Anesthesiology, Academic Medical Center, Amsterdam, the Netherlands, for their contribution relating to immunostaining of the lung slides. Canadian Institutes of Health Research (CIHR). Jack J. Haitsma is supported by the Weston Foundation. Marcus J. Schultz is supported by a personal grant from the Netherlands Organization for Health Research and Development (ZonMW).

Disclosure of Conflict of Interest

The authors state that they have no conflict of interest.

Ancillary