Recombinant Human Thrombomodulin Reduces Mortality and Acute Lung Injury Caused by Septic Peritonitis in Rats

Sepsis is often associated with hemostatic changes ranging from subclinical activation of blood coagulation, which may contribute to localized venous thromboembolism, to acute disseminated intravascular coagulation (DIC), characterized by widespread microvascular thrombosis and subsequent consumption of platelets and coagulation proteins, eventually causing bleeding manifestations (1). The key event underlying this life-threatening complication is the overwhelming inflammatory host response to the infectious agent, leading to the overexpression of inflammatory mediators.

DIC is associated with high mortality in patients with severe sepsis. Excessive coagulation activation, inhibition of fibrinolysis, and consumption of coagulation inhibitors lead to hypercoagulation, resulting in fibrin deposition in microvessels, inflammatory reactions, and DIC (2). Although the effectiveness of anticoagulant therapy in septic patients is still controversial, previous studies suggested that rapid diagnosis and early treatment of DIC can improve the prognosis. In particular, therapeutic intervention against coagulation and inflammation in DIC caused by severe sepsis is relatively effective (3). Thus, early aggressive treatment of this disease is critical.

Thrombomodulin (TM), a type I transmembrane glycoprotein, was identified as an anticoagulant factor that activates protein C (4). TM is present on the surface of endothelial cells and readily binds to thrombin that is generated in the vicinity of intact endothelium (5). Soluble forms of TM, derived from proteolysis of the membrane-bound protein, also circulate in the blood and are excreted in the urine (6, 7). The highest concentration of TM is detected in the capillary bed, where the tissue surface-to-volume ratio is highest. In addition, TM is involved in several biological processes, including cell-to-cell adhesion and inflammation (8, 9). Because TM is a natural anticoagulant protein, recombinant human TM (rhTM) has been used clinically for DIC therapy in Japan since 2008 (10, 11). The clinical effects of rhTM on DIC were examined in patients diagnosed according to the Japanese Ministry of Health and Welfare criteria in a multicenter randomized clinical trial in Japan (12). Alternatively, the lectin-like domain in rhTM suppresses LPS-induced inflammation by binding directly to LPS and high-mobility group box chromosomal protein 1 (HMGB-1) (13, 14). Moreover, rhTM has additive effects in septic patients undergoing continuous hemodiafiltration caused by intestinal perforation (11).

Sepsis is initiated and perpetuated by the overproduction of inflammatory cytokines and chemokines, resulting in excessive tissue injury and death in ∼25–30% of patients (15, 16). Under these conditions, lung dysfunction is the first step in the development of multiple organ failure (17, 18). Furthermore, cecal ligation and puncture (CLP) is a frequently used experimental animal model of sepsis (19). This model closely resembles bowel perforation with tissue necrosis, leading to a polymicrobial infection that is similar to acute peritonitis. The CLP model results in an early hyperdynamic phase, followed by a hypodynamic phase, as in human septic peritonitis (20). In this animal model, the cause of death is acute lung injury (ALI) caused by activation of the tissue macrophages (Mφs), including the lung and the hepatic Mφs (21, 22). Accordingly, this study aimed to sequentially evaluate the effects of rhTM on ALI and cytokine expression induced by inflammation in a rat septic peritonitis model.

This study was performed according to protocols approved by the university review board (approval number A24-16). The experimental protocols followed our institutional criteria and the National Research Council criteria for the care and use of laboratory animals in research (Fig. 1). Male Sprague–Dawley rats (200 g body weight, purchased from Japan SLC, Shizuoka, Japan) were used in all experiments. All animals were housed in sterilized cages in a facility with a 12-h night/day cycle. The temperature and relative humidity were maintained at 23 ± 2°C and 50 ± 10%, respectively. The staff in the animal laboratory at the University of Yamanashi maintained these animal facilities, and veterinarians were always available to ensure animal health. All animals were given humane care in compliance with governmental regulations and institutional guidelines.

Septic peritonitis was induced by CLP, as described previously (17, 21). In brief, the rats were anesthetized with ketamine and xylazine (i.p. administration, 40–90 mg/kg ketamine and 5–10 mg/kg xylazine), and then a 15-mm midline lower abdominal incision was made to expose the cecum, which was then ligated below the ileocecal junction while maintaining intestinal continuity. The cecum was punctured twice with a 21-gauge needle, and a small amount of cecal content was discharged by the puncture wound. The incision was then closed, and 2 ml of sterile saline was administered s.c. for fluid resuscitation. After surgery, the rats were placed on a heating pad until they recovered from the anesthesia. In the sham operation group, animals were undergoing only laparotomy. The animals were given food and water ad libitum throughout the study, and survival was monitored for 7 d (n = 10 in each group). In the rhTM-treated [rhTM(+)] group, rhTM was administered through the tail vein (immediately after CLP, 1 mg/kg rhTM) (Fig. 1) (23). In the control group [rhTM(−) group], animals were injected with a saline vehicle (1 ml/kg body weight).

After CLP, mortality was investigated for 168 h (n = 10 in each group).

Blood samples were collected from the aorta at designated times after CLP (n = 5 at each designated time). The samples were centrifuged at 1200 × g for 10 min at 4°C, and the plasma was stored at −80°C until conducting the assay. Determination of the plasma TNF-α, HMGB-1, and IL-10 levels was performed using ELISA kits (TNF-α and IL-10 from R&D Systems [Minneapolis, MN] and HMGN-1 from SHINO-TEST [Tokyo, Japan]).

Tissue samples were collected at designated times after CLP or sham operation and stored at −80°C for further analysis. Tissue samples were also fixed in formalin, embedded in paraffin, and serially sectioned. Some sections were stained with H&E to assess inflammation and necrosis. The pathology was evaluated blindly by one of the authors and by an expert in rodent pathology.

After the animals were killed, pulmonary vasculature was flushed with 3 ml of cold PBS through the right atrium. The right upper lobe was fixed in formalin for 24 h before being embedded in paraffin. Sections of paraffin were stained with H&E. Histological lung injury was measured using an experimental lung injury scoring system approved by the American Thoracic Society (24). In brief, five random high-power fields (×340 magnification) were recorded for each sample. The lung sections were rated in a blind fashion from 0 (normal) to 1 (severe) based on the following findings: (1) neutrophils in the alveolar (AV) space, (2) neutrophil infiltration into the interstitial (IT) space, (3) hyaline membrane formation, (4) proteinaceous debris filling the air spaces, and (5) AV septal thickening. The scores were averaged for each experimental condition.

Nine hours after CLP, rats were anesthetized and laparotomized. The lungs were harvested and placed in a tared plastic Petri dish for weighing and then dried in a vacuum lyophilizer at −50°C and 10 mm Hg for 72 h (Refrigeration for Science, Island Park, NY). This process removed virtually all gravimetrically detectable water. The dry lung weight was determined, and the wet and dry weight ratios were calculated to assess pulmonary edema (21).

Three hours before sacrifice, rats were i.v. injected with Evans blue dye (EBD; 30 mg/kg in PBS) (Sigma-Aldrich, Burlington, MA) (21, 25). The right lower lobe was homogenized with formamide. Homogenized tissues were incubated for 24 h at 37°C and then centrifuged at 10,000 × g for 20 min. The OD of the supernatant was determined at 620 nm with a FluoStar Omega spectrometer (BMG Labtech, Ortenberg, Germany). The concentration of EBD leakage (mg/g tissue weight) was calculated against a standard curve of EBD solution (R² = 0.99).

Formalin-fixed, paraffin-embedded tissue specimens were cut into 4-μm-thick serial sections. Each serial section was mounted on a silane-coated glass slide, deparaffinized in xylene, rehydrated in decreasing concentrations of ethanol, and rinsed in PBS. Slides were treated in Ag retrieval solution for 15 min at 120°C using Dako REAL Target Retrieval Solution (Dako, Carpinteria, CA). Endogenous peroxidase was quenched by incubation at room temperature in 0.3% hydrogen peroxide, followed by rinsing with PBS. After washing three times with PBS, the sections were blocked in PBS containing 0.1% Triton X-100 (PBS-T) with 5% horse serum for 2 h at room temperature. After blocking, the sections were incubated with rabbit anti-fibrinogen Ab (1:500; Bioss Antibodies, Woburn, MA) overnight at 4°C. The next day, the sections were rinsed in PBS three times and incubated in PBS-T containing 5% horse serum and Biotinylated PAN-Specific Secondary Antibody (1:400; Vector Laboratories, Burlingame, CA) for 2 h followed by Reagent A and Reagent B (1:1; Vectastain ABC Kit; Vector Laboratories) for 2 h at room temperature. Immunohistochemical staining was performed using 3,3′-diaminobenzidine and 0.3% hydrogen peroxide.

Stained slides were recorded using digital microscopy (BZX700; Keyence, Osaka, Japan), applying Z-stack technology to improve the quality of images. The density of stained cells (cytosol) in three different areas per section was quantified and standardized using the measurement module BZ-H3C (Hybrid Cell Count Ver.1.1; Keyence).

The hepatic Mφs (Kupffer cells [KCs]) were isolated by collagenase digestion and differential centrifugation using Nycodenz (Nycomed Pharma AS, Oslo, Norway) (26), as described in our previous work with modifications (27).

The AV Mφs were collected, and the lavage of the lungs was repeated 20 times until only occasional cells were present in the lavage fluid. Thereafter, the lung tissue was sliced into small pieces. The IT lung Mφs were isolated according to the method presented by Holt et al. (17, 28) and modified by Sjöstrand et al. (17, 29). For isolation of the peritoneal Mφs, the contents of the peritoneum were lavaged by injecting 50 ml of PBS. Peripheral blood was also collected from the aorta, and peripheral mononuclear cells were isolated by differential centrifugation using NycoPrep 1.077A (Nycomed Pharma AS) (21).

Isolated cells were seeded onto culture dishes and incubated for 1 h in DMEM (GIBCO Laboratories Life Technologies, Grand Island, NY) supplemented with 10% FCS and antibiotics. After incubation, adhesive cells were collected using a scraper and washed three times with PBS for further experiments.

Isolated tissue Mφs were seeded onto 24-well plates (1 × 10⁵/well) and cultured in DMEM supplemented with 10% FBS and antibiotics at 37°C in the presence of 5% CO₂ for 24 h. Cells were coincubated with LPSs (10 μg/ml in media) in the presence or absence of rhTM (5 μg/ml in media). Medium was collected and kept at −80°C until assay.

Determination of the TNF-α, HMGB-1, and IL-10 production by isolated Mφs was performed using ELISA kits (TNF-α and IL-10 from R&D Systems [Minneapolis, MN] and HMGN-1 from SHINO-TEST [Tokyo, Japan]).

Data are expressed as the mean ± SEM. Continuous variables were compared between groups using the Student t test. Noncontinuous variables were compared between groups using the Mann–Whitney U test. Comparisons between groups were analyzed by repeated measures ANOVA, adjusted for the baseline values as a covariate, and using the Bonferroni test. A p value <0.05 was considered significant.

In the sham operation groups, all animals survived after surgery, and the treatment with rhTM did not affect mortality in normal animals (Fig. 1). Mortality was 80% at 168 h after CLP in the rhTM(−) group (Fig. 2). In contrast, mortality was 50% in the rhTM(+) group and prevented by rhTM treatment. There was a significant difference between the two groups (p < 0.05).

In the sham operation group, there were no pathological changes as expected (Fig. 3A). CLP caused significant pathological changes in the lungs, including inflammatory cell infiltration into the AV and IT spaces, septal thickness, and hemorrhage, in the rhTM(−) group (Fig. 3B, 3D). These pathological changes were markedly blunted in the rhTM(+) group (Fig. 3C).

No microthrombosis was observed in the sham operation groups (data not shown). Severe microthrombosis was observed in the rhTM(−) group after CLP (Fig. 3E), whereas it was significantly blunted by rhTM treatment, although lung IT edema was detected (Fig. 3F).

By the image analysis, the fibrinogen-positive area was significantly blunted in the rhTM(+) group compared with the rhTM(−) group (Fig. 4).

After CLP, plasma endotoxin levels increased. Furthermore, bacterial loads also increased (Fig. 5). Treatment of rhTM did not affect both plasma endotoxin levels and bacterial loads. There were no significant differences in plasma endotoxin levels and bacterial load between the groups (Fig. 5).

The lung injury score was low in the normal rats (Fig. 6A), but the score significantly increased in the rhTM(−) group after CLP (p < 0.01). This increase was significantly blunted in the rhTM(+) group (p < 0.05).

Consistent with the pathophysiological changes of the lungs, the increased wet/dry weight ratio was significantly blunted in the rhTM(+) group compared with that in the rhTM(−) group after CLP (p < 0.05) (Fig. 6B).

Lung EBD concentration was low in the normal rats (Fig. 6C), but the concentration significantly increased in the rhTM(−) group after CLP (p < 0.01), and this increase was significantly blunted in the rhTM(+) group (p < 0.05).

Plasma TNF-α levels were low in the sham operation group. In contrast, the values increased and peaked at 9 h after CLP and gradually decreased thereafter in the rhTM(−) group (Fig. 7A). In the rhTM(+) group, they also peaked at 9 h after CLP; however, they were significantly lower than those in the rhTM(−) group at 6, 9, and 12 h after CLP.

Plasma HMGB-1 levels were low in the sham operation group. In contrast, the values increased gradually up to 24 h after CLP in the rhTM(−) group (Fig. 7B). In the rhTM(+) group, they also increased gradually up to 24 h after CLP; however, they were significantly lower compared with those in the rhTM(−) group at 12, 18, and 24 h after CLP.

Plasma IL-10 levels were low in the sham operation group. In contrast, the values gradually increased and peaked at 12 h after CLP and gradually decreased thereafter in the rhTM(−) group (Fig. 7C). In the rhTM(+) group, the values also peaked at 12 h after CLP; however, they were significantly greater compared with those in the rhTM(−) group at 12 and 18 h after CLP.

TNF-α production was detected in isolated tissue Mφs stimulated with LPS in vitro and was the greatest in the IT Mφs (Fig. 8A). However, the production was significantly blunted in the AV and IT Mφs incubated with rhTM.

HMGB-1 production was detected in isolated tissue Mφs stimulated with LPS in vitro and was the greatest in the KC Mφs (Fig. 8B). Alternatively, the production was significantly blunted in the KC, AV, and IT Mφs incubated with rhTM.

IL-10 production was detected in isolated tissue Mφs stimulated with LPS in vitro and was the greatest in the isolated KCs (Fig. 8C). Furthermore, this production was not markedly changed in those cells incubated with rhTM.

In this study, rhTM prevented mortality caused by sepsis induced by a perforated peritonitis model in rats. The critical mechanism of this effect was the inhibition of pulmonary microvascular thrombosis formation and expression of inflammatory cytokines, which enables the prevention of ALI.

Coagulation activation induces pulmonary microvascular thrombosis, leading to multiple organ dysfunction in sepsis, and the severity of coagulopathy is directly correlated with mortality (30). In this study, pulmonary microvascular thrombosis was induced in the sepsis model (Fig. 3). In septic patients, the levels of physiological anticoagulants are decreased; therefore, restoring anticoagulant activity is a therapeutic strategy. As recommended in the latest Clinical Practice Guidelines for the Management of Sepsis and Septic Shock from Japan (31), anticoagulation is considered a therapeutic target in patients with DIC. rhTM has been constructed and has been clinically available for the treatment of DIC in Japan since 2008. rhTM administration reduced the consumption of endogenous anticoagulants, including antithrombin and protein C by DIC (32). In this study, pulmonary microvascular thrombosis was prevented by rhTM administration (Figs. 3, 4). This effect may lead to the reduction of microvascular permeability, the wet/dry weight ratio, and ALI (Fig. 6). Furthermore, mortality was significantly prevented in the rhTM(+) group compared with the rhTM(−) group (Fig. 2). In clinical treatment, ∼13% reduction in the risk for mortality was observed in the rhTM group, although a difference was not significant (relative risk, 0.87; 95% confidence interval, 0.74–1.03; p = 0.10) (33). Furthermore, rhTM treatment has additive effects in septic patients undergoing continuous hemodiafiltration because of intestinal perforation (11). In that study, the risk for serious complications, including postoperative bleeding and liver dysfunction, did not increase with rhTM administration in the early period after surgery. Thus, rhTM is effective for anticoagulation therapy in sepsis.

Because of the continuous gaseous exchange, the lungs serve as a very easy target organ for airborne pathogens, allergens, and other toxicants to cause infections or inflammation (34). For example, sepsis is a leading cause of ALI (6–42%) (35). Acute microbial infections responsible for pneumonia or sepsis cause severe inflammatory damage to the lungs, leading to the development of ALI or acute respiratory distress syndrome in critically ill patients. Pulmonary dysfunction is extremely frequent in patients with sepsis and is associated with increased mortality rates, particularly when acute respiratory distress syndrome is diagnosed. Furthermore, ALI causes severe tissue damage, and irreversible pulmonary damage may lead to death in severe cases. In this study, ALI was induced and mortality was severe after CLP (Figs. 2, 3, 4, 6). Thus, the lungs are the critical target organ in sepsis.

In the CLP model, the cause of death is usually hypoxia (21). In this study, the pulmonary microthrombosis was induced, and the lung injury score, the microvascular permeability, and the wet/dry lung weight ratio increased after CLP (Figs. 3, 4, 6). These pathophysiological changes induce hypoxia, leading to the death of the host. However, the pulmonary pathophysiological changes and mortality were significantly reduced by rhTM treatment in this study (Figs. 2, 3, 4, 6). Thus, treatment with rhTM may be an adjuvant strategy to attenuate distant lung injury in sepsis.

IL-10 is an anti-inflammatory cytokine (36–38). Previously, it was reported that tissue hepatic Mφs are the major source of IL-10 (Fig. 8) (21). Exposure of mononuclear phagocytes or dendritic cells to IL-10 inhibits the synthesis of proinflammatory cytokines and the release of reactive oxygen and nitrogen intermediates, in addition to inhibiting the Ag-presenting capacity of these cells (39). Furthermore, IL-10 protects mice against endotoxin shock by preventing excessive production of proinflammatory cytokines (40). Moreover, in this study, administration of rhTM maintained the blood IL-10 levels (Fig. 7) and prevented mortality after CLP (Fig. 2). Thus, rhTM does not affect the IL-10 value, and this effect also turns off the signaling of the inflammatory cytokine cascade and inhibits the development of multiple organ injuries such as ALI.

ALI occurs as a result of the immunological recognition of the pathogen responsible for inducing a proinflammatory immune response. In this study, plasma inflammatory cytokines increased after CLP (Fig. 7), but rhTM inhibited the production of TNF-α by isolated tissue Mφs (Fig. 8). Plasma TNF-α levels were inhibited in the rhTM(+) group compared with the rhTM(−) group (Fig. 7). The inhibition of TNF-α production of tissue Mφs by rhTM treatment most likely leads to inhibition of plasma TNF-α levels, reducing ALI and mortality after CLP (Figs. 2, 3, 4, 6, 7). Thus, rhTM inhibits the initiation of the inflammatory cytokine cascade (41).

Alternatively, rhTM increased the plasma IL-10 levels (Fig. 7C). IL-10 acts in conjunction with the inhibition of the proinflammatory cytokine IL-6 in vivo. In this study, plasma IL-6 levels were also blunted in the animals treated with rhTM (data not shown). Based on these results, the mortality and ALI after CLP were prevented by rhTM treatment (Figs. 2, 3, 4, 6).

The N-terminal lectin-like domain (D1) of rhTM has unique anti-inflammatory properties. Using the D1, TM binds to HMGB-1 DNA-binding protein, a factor closely associated with necrotic cell damage after its release from the nucleus, thereby preventing in vitro leukocyte activation, in vivo UV irradiation–induced cutaneous inflammation, and in vivo LPS-induced lethality (14, 42). Furthermore, rhTM not only binds to HMGB-1 but also aids in the proteolytic cleavage of HMGB-1 by thrombin. These findings highlight the novel anti-inflammatory role of TM, in which thrombin–TM complexes degrade HMGB-1 to a less proinflammatory form (43). In this study, plasma HMGB-1 levels were significantly lower in the rhTM(+) group compared with the rhTM(−) group (Fig. 7). Furthermore, HMGB-1 production by the tissue Mφs was blunted in the presence of rhTM in vitro (Fig. 8). Thus, rhTM can reduce the expression of HMGB-1, as reported previously (43).

The pulmonary innate immune system plays a pivotal role in ALI pathogenesis (37). In this study, in isolated tissue Mφs, the production of inflammatory cytokines, including TNF-α and HMGB-1, increased in the AV and IT pulmonary Mφs stimulated with endotoxin in vitro. These productions were greater in the IT pulmonary Mφs than in the AV Mφs. Furthermore, rhTM treatment in vitro inhibited the production of those cytokines (Fig. 8). The production of anti-inflammatory cytokine IL-10 was the greatest in the hepatic Mφs, and the production was not changed by rhTM treatment in vitro (Fig. 8). Alternatively, there were no significant differences in plasma endotoxin levels and bacterial burden, suggesting that rhTM has no effects on the bacterial burden and peripheral endotoxin levels (Fig. 5). Therefore, the mechanism for the effects of rhTM on ALI was partially involved in inhibiting activation of the tissue Mφs, leading to inhibition of the inflammatory cytokine levels. Thus, the mechanism of rhTM in sepsis is partially involved in inhibiting the activation of tissue Mφs, leading to the suppression of the proinflammatory cytokines.

This study investigated the effects of rhTM on sepsis. The mechanism of preventing mortality by rhTM treatment involved the inhibition of the inflammatory immune reaction and pulmonary microvascular thrombosis, leading to the prevention of ALI. Thus, treatment with rhTM may be an adjuvant strategy to attenuate distant lung injury in sepsis.

The authors have no financial conflicts of interest.