European Respiratory Society


Ischaemia–reperfusion injury of the lung is a major cause of morbidity and mortality, particularly following lung transplantation, the mainstay treatment for patients with end-stage pulmonary disease. Effective measures to prevent this complication are lacking. Thrombomodulin (TM) is an endothelial cell receptor and cofactor for thrombin-mediated generation of the anticoagulant and anti-inflammatory activated protein C (APC). The N-terminal lectin-like domain (LLD) of TM has no direct effects on coagulation, but has distinct anti-inflammatory properties, interfering with leukocyte adhesion, complement activation and cytokine generation, all of which are hallmarks of ischaemia–reperfusion injury. Using a murine model of lung ischaemia–reperfusion injury (90 min ischaemia, 4 h reperfusion), the present study shows that mice lacking the LLD of TM respond with increased extravasation of neutrophils and macrophages into the lung parenchyma and bronchoalveolar fluid (BALF), with augmented BALF levels of cytokines interleukin (IL)-1β and granulocyte-monocytic colony-stimulating factor (GM-CSF). Pre-treatment of wild-type mice with recombinant LLD, as compared with controls, significantly suppresses protein leakage and accumulation of leukocytes in the BALF.

These novel findings support further evaluation of recombinant lectin-like domain of thrombomodulin to protect the lung against tissue-damaging pro-inflammatory responses following ischaemia-reperfusion.

Lung transplantation has become the standard treatment for patients with most end-stage parenchymal and vascular pulmonary diseases. In spite of major surgical advances, better organ preservation techniques, and the development of safer and more effective immunosuppressive approaches, lung transplantation continues to be complicated by ischaemia–reperfusion injury, which significantly impacts on short- and long-term outcomes 1, 2.

The cellular and molecular mechanisms underlying ischaemia–reperfusion injury are complex and remain incompletely understood. Nonetheless, major insights have recently been gained, and there are several well-defined features. These include: an influx of activated inflammatory cells; the release of pro-inflammatory cytokines 35; local and systemic accumulation of chemokines and reactive oxygen species 6; activation of complement 7; increases in thromboxanes; cellular apoptosis; and reductions in nitric oxide (reviewed in 1). Thus, therapeutic strategies have been devised to limit the severity of ischaemia–reperfusion injury with, for example: 1) complement inhibition; 2) neutralisation of interleukin (IL)-8 8; 3) use of antioxidants 8, 9; 4) use of vasodilators, such as nitric oxide; 5) inhibition of macrophage activation 10; 6) interference with leukocyte adhesion 11, 12; 7) blocking generation of inflammatory arachidonic acid metabolites 13; 8) suppression of the nuclear factor (NF)-κB pro-inflammatory signalling pathway 14; and 9) promoting cell survival 15, 16. Although these innovative therapeutic manoeuvres are showing promise in animal models, few have entered the clinic. Thus, ischaemia–reperfusion injury remains a highly relevant but largely unsolved problem, and novel therapeutic and/or preventative measures are urgently needed.

Thrombomodulin (TM) is a pan-endothelial cell, transmembrane glycoprotein receptor that, via its structural domains, provides vasculoprotection through several distinct molecular pathways (reviewed in 17). Structurally, TM is comprised of five domains: a C-terminal cytoplasmic tail, a single membrane-spanning domain, an extracellular juxtamembranous serine-threonine rich region, six epidermal growth factor (EGF)-like repeats, and finally an N-terminal C-type lectin-like domain (LLD) joined to the first EGF-like repeat via a short hydrophobic module. EGF-like repeats three to six provide critical cofactor activity for thrombin-mediated activation of the physiologically relevant anticoagulant and anti-inflammatory protein C (PC). Recent studies have revealed that the LLD of TM has distinct anti-inflammatory properties, interfering with neutrophil and monocyte adhesion, suppressing activation of mitogen-activated protein (MAP) kinase and NF-κB pathways, promoting cell survival, interfering with complement activation 18, 19, and blocking the activity of the pro-inflammatory cytokine, high mobility group box 1 (HMGB1) 20. Mice that lack the LLD of TM (TMLeD/LeD mice) are more sensitive to pro-inflammatory stimuli and myocardial ischaemia-reperfusion injury, whereas administration of recombinant LLD of TM protects against endotoxin-induced sepsis and inflammation 19, 20.

In view of the diverse and potent anti-inflammatory properties of the LLD of TM, loss-of-function and gain-of-function approaches in mouse models 21 were used to examine the role of this structure in protecting against ischaemia–reperfusion injury of the lung.



Female Swiss outbred mice were obtained from Janvier (Savigny/Orges, France). Mice lacking the N-terminal LLD of TM (TMLeD/LeD mice) were previously generated 18. TMLeD/LeD mice and their wild-type counterparts (TMwt/wt mice), were obtained from intercrosses of TMLeD/wt mice. Both lines were maintained in the same genetic background at the same generation, thereby allowing direct comparisons. TMLeD/LeD mice express normal antigenic levels of TM, and activation of PC is intact and unaffected by deletion of the LLD 18. Experiments were performed on mice aged 8–10 weeks. Animals were housed in cages under specific pathogen-free conditions and fed ad libitum with a standard diet. Studies were approved by the University of Leuven Animal Ethics Committee (Leuven, Belgium).

Hydrodynamic gene delivery

In vivo gene delivery of the LLD in mice has been previously reported 19. The cDNA encoding the first 155 amino acids of the LLD of murine TM was subcloned into pSecTag2HygroA-CMV (Invitrogen, Merelbeke, Belgium; referred to as pSec), resulting in pSecTag2HygroA-CMV-mLec155 (herein referred to as pSec-mLec). For gene delivery in vivo, plasmid DNA 2 μg per gram body weight in 100 μL of saline per gram body weight, was injected over 5–8 s via the tail vein.

Murine model of in situ lung ischaemia–reperfusion injury

Lung ischaemia–reperfusion injury was induced in mice as previously described 21. Briefly, animals were anaesthetised with intraperitoneal pentobarbital 60 mg·kg−1 (Nembutal; Sanofi, Brussels, Belgium) and kept at 37°C. Animals were ventilated mechanically via an endotracheal tube (120 strokes·min−1; 320 μL stroke volume) with 50% oxygen (TMLeD/LeD and TMwt/wt mice) or room air (Swiss mice). The hilum of the left lung was occluded for 90 min via a left thoracotomy, followed by a reperfusion period of 4 h. Mice were then sacrificed by ventilator switch-off.

Bronchoalveolar lavage

After sacrifice, bronchoalveolar lavage (BAL) was performed with the right lung clamped to obtain BAL fluid (BALF) from the left lung 21. Four 0.5-mL aliquots of sterile saline were gently instilled into the trachea. The returned fractions were centrifuged at 500×g for 10 min at 4°C. The supernatants of the first lavage were frozen and stored at -80°C for subsequent analyses. Pellets of the four lavages were pooled in 250 μL of saline for total and differential cell counts.

Cell counts and protein assays

Total cell counts in BALF were performed in triplicate using a Bürker chamber. Differential cell counts using cytospin preparations, on ≥300 cells per sample, were determined as previously reported 21. BALF IL-1β levels were measured in undiluted supernatants by ELISA (Biosource International, Nivelles, Belgium). Interferon (IFN)-γ, tumour necrosis factor (TNF)-α, IL-2, granulocyte-monocyte colony-stimulating factor (GM-CSF), keratinocyte cytokine (KC), monocyte chemotactic protein (MCP)-1, and IL-6 were measured with Multiplex Cytometric Bead Array Flex kits (BD Biosciences, Erembodegem, Belgium). Total protein was quantified with the Bradford assay kit (Sigma, Bornem, Belgium). Plasma levels of thrombin–antithrombin (TAT) complexes and D-dimers were quantified by ELISA (Dade Behring, Brussels, Belgium).

Lung histology and immunohistochemistry

After BAL, the left lung was excised, fixed in 4% paraformaldehyde in PBS, and embedded in paraffin 21. Transverse 10-μm sections were cut and prepared for staining with haematoxylin and eosin (HE), or for immunostaining with antibodies against HMGB1 (Abcam, Cambridge, UK) or CD45 (Becton Dickinson, Erembodegem, Belgium). Quantification of cells per 400× high-power field (hpf) was performed manually on three nonoverlapping sections by an investigator blinded to the identity of the samples. Averages of a minimum of three measures for each lung were obtained, and the means of the results from six mice were obtained.

Statistical analyses

Results are presented as mean±sem. Unpaired t-tests were used to identify differences between groups. When required, either a t-test with Welch correction, or the Mann–Whitney test was used. Correlation analyses were performed using a Pearson correlation test. A p-value <0.05 was considered significant.


Lack of the LLD of TM increases BALF leukocyte and cytokine levels after ischaemia–reperfusion injury

The role of the LLD of TM in lung ischaemia–reperfusion injury was examined by using a previously developed in situ murine model 21, and comparing the response of TMLeD/LeD mice with their wild-type counterpart TMwt/wt mice. TMLeD/LeD mice lack the entire N-terminal LLD, including the C-type lectin-like module and the adjacent hydrophobic module. The regimen of 90 min ischaemia followed by 4 h of reperfusion was previously demonstrated in the current authors' laboratory to induce BALF accumulation of leukocytes and cytokines in a temporal pattern comparable to that seen in other animal models 21, 22.

Under nonstress baseline conditions, analyses of BALF from TMLeD/LeD or TMwt/wt established that myeloperoxidase activities were not significantly different between TMwt/wt and TMLeD/LeD mice 21, consistent with a lack of spontaneous leukocyte infiltration into the alveolar space, and consistent with the absence of clinical disease in both.

In pilot studies, the present authors determined that it was not possible to perform the lung ischaemia–reperfusion injury protocol in the TMLeD/LeD mice room air due to a high operative mortality rate. Thus, both the TMLeD/LeD and the TMwt/wt mice were ventilated with 50% oxygen during the entire ischaemia–reperfusion injury procedure, in which case no deaths were encountered.

Under these conditions, it was necessary to confirm that any changes that might occur in BALF accumulation of cytokines or leukocytes in response to ischaemia–reperfusion were not due to ventilation alone. Therefore, the mice were exposed to 5.5 h of ventilation with 50% oxygen (n = 3 per group). BALF was collected from the left lung and analysed. Both lungs were then harvested for histological analysis. Total leukocyte counts in the BALF of TMwt/wt and TMLeD/LeD mice were 2.9±0.1×104 cells·mL−1 and 3.6±0.6×104 cells·mL−1, respectively (p = 0.363); total protein levels were 168±36 μg·mL−1 and 192±7 μg·mL−1, respectively (p>0.5); and cytokine levels (IL-1β, GM-CSF, TNF-α) were below the limits of detection of the assays in both groups. These results are no different from BALF studies without ventilation (data not shown). Thus, ventilation alone did not significantly alter the BALF results.

After lung ischaemia–reperfusion injury, the number of leukocytes in the BALF of the TMLeD/LeD and the TMwt/wt mice increased (fig. 1a). However, leukocyte counts in the BALF of TMLeD/LeD mice were significantly higher than in the TMwt/wt mice (n = 6; p<0.01). Differential cell counts revealed that the higher BALF leukocyte counts from the reperfused lungs of the TMLeD/LeD mice as compared with the TMwt/wt mice were due to significant increases (p<0.05) in both macrophages and neutrophils, without a notable difference in the number of lymphocytes (fig. 1). These cellular changes in the BALF of TMLeD/LeD mice were associated with selective increases in cytokine levels (fig. 2). Thus, there were higher levels of IL-1β (p<0.001) and GM-CSF (p<0.05) in the BALF of TMLeD/LeD mice versus the TMwt/wt mice. The number of macrophages correlated directly with IL-1β levels (r = 0.60; p<0.05), while the number of neutrophils directly correlated with IL-1β (r = 0.79; p<0.01) and GM-CSF (r = 0.59; p<0.05). Protein leakage was not notably affected by deletion of the LLD, since BALF total protein counts were similar (569±70 μg·mL−1 versus 622±90 μg·mL−1 in TMwt/wt and TMLeD/LeD mice, respectively). There were also no significant effects of deleting the LLD of TM on BALF levels of IL-6, KC, TNF-α or MCP-1 (fig. 2).

Fig. 1—

a–d) Lack of the lectin-like domain (LLD) of thrombomodulin (TM) augments bronchoalveolar lavage fluid (BALF) leukocytes after lung ischaemia–reperfusion injury. White cell counts in BALF from mice lacking the LLD of TM (TMLeD/TMLeD) and their wild-type counteparts (TMwt/wt) were quantified after lung ischaemia (90 min) and reperfusion (4 h). n = 6 mice per group. Lack of the LLD of TM resulted in significant increases in BALF neutrophils and monocytes. *: p<0.05; **: p<0.01; ***: p<0.001.

Fig. 2—

a–f) Bronchoalveolar lavage fluid (BALF) cytokine levels after lung ischaemia–reperfusion. Cytokines in BALF from mice lacking the lectin-like domain of thrombomodulin (TMLeD/LeD mice) and their wild-type counterparts (TMwt/wt mice) were quantified after ischaemia–reperfusion. n = 6 mice per group. Granulocyte-monocytic colony-stimulating factor (GM-CSF) and interleukin (IL)-1β were significantly elevated in BALF from TMLeD/LeD mice. KC: keratinocyte cytokine; MCP: monocytic chemotactic protein; TNF: tumour necrosis factor. *: p<0.05; **: p<0.01; ***: p<0.001.

Microscopic examination of lung sections from the mice after ischaemia–reperfusion injury indicated notably more cellular infiltration into the lungs of TMLeD/LeD mice as compared with their wild-type counterparts (fig. 3). Few cells were observed in the alveolar space, but this was attributed to bronchial lavage being performed prior to tissue preparation. CD45 staining confirmed the significantly increased lung infiltration with leukocytes (12.3±1 cells·hpf−2 and 24.2±3 cells·hpf−1 in TMwt/wt and TMLeD/LeD lungs, respectively; n = 6 per group, p<0.005; fig. 3). The histological changes could not be attributed either to the ventilation or to the lavage procedures. TMLeD/LeD and TMwt/wt mice were ventilated for 5.5 h as described, and, in addition to BALF from the left lung, both lungs were examined histologically. There were no discernible differences in the structure of the lungs under these otherwise baseline conditions, with only minimal evidence of neutrophil accumulation in the lung parenchyma of TMLeD/LeD mice (fig. 3). Overall, the findings suggest that the LLD of TM confers protection against accumulation and extravasation of neutrophils and monocytes into the alveolar space in the setting of ischaemia–reperfusion injury.

Fig. 3—

Histological sections of lungs from mice lacking the lectin-like domain of thrombomodulin (TMLeD/LeD; b, d, f and h) and their wild-type counterparts (TMwt/wt; a, c, e and g) mice before and after ischaemia–reperfusion. Representative histological sections of lungs from mice before (a–d; n = 3 per group) and after ischaemia–reperfusion (e–g; n = 6 per group) are shown, stained with haematoxylin and eosin (a, c, e, g) or for CD45 (b, d, f, h) to identify leukocytes. Histological examination of sections from the left lungs of mice after 5.5 h of ventilation (no ischaemia) and collection of bronchoalveolar (BALF; a–d) was indistinguishable from the right lungs of corresponding mice (ventilation but no BALF collected), and also from baseline (no ventilation or BALF) (not shown). Only a few CD45-positive leukocytes (arrows in d) were detected in the parenchyma of lungs from TMLeD/LeD mice. e–h) After ischaemia–reperfusion, lung sections from TMLeD/LeD mice (g, h), as compared with those from TMwt/wt mice (e, f), exhibited notably increased cellularity, and a significant increase in CD45-positive leukocytes (g, h). Scale bars = 50 μm.

HMGB1 is a pro-inflammatory cytokine, the action of which is, in part, regulated by the LLD of TM 20. Recent studies suggest that HMGB1 may not only mediate late inflammatory events, but also may be involved in early events, i.e. in acute injury 23. Moreover, HMGB1 has recently been shown to have pro-thrombotic effects by stimulating tissue factor and by interfering with PC activation 24. The expression of HMGB1 in the lung sections of the TMwt/wt and TMLeD/LeD mice was examined in the present study before and after ischaemia–reperfusion and, in that time frame, HMGB1 remained mostly localised to the nuclei, with minimal accumulation in the cytoplasm. HMGB1 was not detected in the BALF by ELISA (not shown). Plasma levels of the haemostatic markers, TAT and D-dimers were measured from TMwt/wt and TMLeD/LeD mice after lung ischaemia–reperfusion. Under baseline conditions or after ventilation only, there were no differences between the two mouse lines. Following ischaemia–reperfusion, D-dimers rose significantly over baseline (for TMwt/wt mice; 15.5±7 ng·mL−1 after reperfusion versus 1.6±0.4 ng·mL−1 ventilation only; for TMLeD/LeD mice; 20.9±7 ng·mL−1 after reperfusion versus 2.6±0.9 ng·mL−1 ventilation only), but the increases in the two mouse lines were not significantly different (p>0.5). In both the TMwt/wt mice or the TMLeD/LeD mice, plasma accumulation of TAT complexes remained unchanged from baseline in response to ischaemia–reperfusion (data not shown; p>0.2). Overall, when comparing the response to lung ischaemia–reperfusion injury in the TMwt/wt and TMLeD/LeD mice, differences in the haemostatic system activity or a differential role for HMGB1 could not be identified.

Treatment with recombinant LLD of TM suppresses ischaemia–reperfusion injury

Based on the heightened inflammatory response of the TMLeD/LeD mice, the current authors predicted that administration of recombinant LLD of TM might protect against ischaemia–reperfusion injury. The present authors have previously shown that hydrodynamic gene delivery of a cDNA vector (pSec-mLec) encoding the LLD of TM (the C-type lectin-like module) 19 protects against inflammatory arthritis in vivo, and that serum levels of the recombinant protein could be detected for up to 72 h, peaking at ∼600 ng·mL−1 at 6–24 h. Therefore lung ischaemia–reperfusion injury studies were performed on wild-type Swiss mice, pretreating them 24 h prior to clamping with pSec-mLec or with the vector-alone control (pSec; n = 6 per group). Swiss mice were selected for these studies, because they tolerated room air (rather than requiring 50% oxygen) during the procedure.

Treatment with pSec-mLec resulted in a significant decrease in the total number of BALF leukocytes following 90 min of ischaemia and 4 h of reperfusion (p<0.01; fig. 4), and this was due to a significant decrease in the number of macrophages (p<0.01), lymphocytes (p<0.05) and neutrophils (p<0.001).

Fig. 4—

a–d) Bronchoalveolar lavage fluid (BALF) accumulation of leukocytes after treatment with recombinant lectin-like domain (LLD) of thrombomodulin (TM). White cell counts in BALF were quantified after lung ischaemia–reperfusion in mice that were pre-treated with recombinant LLD of TM via hydrodynamic gene delivery (LLD) or with the control vector-alone. n = 6 mice per group. Treatment with recombinant LLD of TM significantly reduced all leukocytes. *: p<0.05; **: p<0.01; ***: p<0.001.

Gene delivery of recombinant LLD of TM caused a significant reduction in MCP-1 levels in the BALF as compared with the control pSec-treated mice (139±13 pg·mL−1 with LLD versus 77±6 pg·mL−1 with vector alone; p = 0.0079; n = 6 mice per group), while not affecting other measured cytokines. BALF protein levels decreased significantly with administration of the LLD (from 1165±140 μg·mL−1 to 626±140 μg·mL−1; p = 0.0364), suggesting protection against alveolar leakage. These responses, as compared with leukocyte and cytokine changes observed in the TMLeD/LeD and TMwt/wt mice, likely reflect strain differences, as has frequently been reported 2528, or the effects of different local oxygen concentrations. Importantly, histological analyses of the lungs after ischaemia–reperfusion injury were also consistent with a beneficial effect of pretreatment with recombinant LLD (fig. 5). HE-stained lung sections from vector-alone treated mice revealed evidence of increased leukocyte infiltration that was diminished in the lungs from mice pretreated with pSec-mLec. CD45 staining of the lung sections (fig. 5) confirmed that administration of recombinant LLD of TM caused a significant reduction in the numbers of leukocytes in the lung parenchyma (30.7±4 cells·hpf−1 and 12.7±2 cells·hpf−1 with administration of pSec and pSec-mLec, respectively; n = 6 per group; p<0.005).

Fig. 5—

Histological changes in lungs in response to recombinant lectin-like domain (LLD) of thrombomodulin (TM). Representative histological sections of lungs from mice after ischaemia–reperfusion are shown (n = 6 per group), stained with haematoxylin and eosin (a and b) or for CD45 (c and d) to identify leukocytes. Mice pretreated with recombinant LLD of TM as compared with controls (a and c) exhibited a significant reduction in infiltration with leukocytes (arrows). Scale bars = 25 μm.


In seeking innovative therapeutic approaches to interfere with lung injury associated with ischaemia–reperfusion, the current authors have drawn from advances in the understanding of the coagulation system, which have revealed the molecular basis underlying the links with the inflammatory system. The clinical relevance of this is best appreciated by the successful use of activated PC (APC), originally recognised for its anticoagulant properties, in decreasing sepsis-induced mortality in patients 29, and protecting against ischaemia–reperfusion injury in preclinical models 30. The goal of the present study was to assess whether a structural domain of TM might also protect against ischaemia–reperfusion lung injury.

Full-length TM has multiple mechanisms by which it protects the vascular endothelium and underlying tissue, i.e. by modulating coagulation, fibrinolysis, cell proliferation, inflammation and innate immunity 17. These are accomplished via direct and indirect interactions of its structural domains with several protein partners, including thrombin, PC, thrombin-activatable fibrinolysis inhibitor, the endothelial cell PC receptor (EPCR), platelet factor 4, HMGB1 and the complement system 31, 32. Recombinant forms of TM that include the active anticoagulant EGF-like repeats have been used in vivo to interfere with sepsis, disseminated intravascular coagulation and lung injury, activities that are attributed to its cofactor activity for thrombin-mediated generation of APC 33, 34. As is the case for APC, the anticoagulant property of the EGF-like domain of TM, with its inherent risk of bleeding, raises concerns about its therapeutic use for ischaemia–reperfusion injury. Since the initiating events in ischaemia–reperfusion injury are primarily inflammatory in nature, it is reasonable to design therapeutic agents that target these steps, while avoiding haemorrhagic side-effects.

The LLD of TM does not directly affect the haemostatic system, but rather, it interferes with key pro-inflammatory intracellular signalling pathways, suppressing the expression of leukocyte adhesion molecules, tissue accumulation of neutrophils and monocytes 18, and the activation of complement pathways 19. The wide array of protective effects of the LLD of TM in the inflammatory cascade prompted evaluation of its role in lung ischaemia–reperfusion injury. This was accomplished by using an established murine model of warm ischaemia–reperfusion injury, in which changes in leukoctye trafficking and cytokine release, caused by genetic or pharmacological manipulation, could be quantified 21. The present results showed that, as compared with matched wild-type controls, mice lacking the LLD of TM exhibited significantly higher BALF levels of macrophages and neutrophils, and these changes correlated with significantly elevated BALF concentrations of the cytokines IL-1β and GM-CSF. This enhanced inflammatory response in the TMLeD/LeD mice cannot be attributed to alterations in thrombin–TM mediated activation of PC, since the EGF-like repeats of TM that are critical for PC activation are intact in the TMLeD/LeD mice, and the antigenic levels of TM are the same as in TMwt/wt mice. The current authors have previously confirmed that PC activation in vivo in the TMLeD/LeD mice is unaltered by loss of the LLD of TM 18. Finally, the present study showed that the formation of TAT complexes and D-dimers in the blood was not altered after ischaemia–reperfusion by lack of the LLD of TM. Thus, the conclusion can be made that it is the lack of the LLD of TM, rather than diminished PC activation, that causes enhanced infiltration of leukocytes and augmented release of cytokines into the BALF of TMLeD/LeD mice in this model of lung ischaemia–reperfusion injury.

These findings provided a strong rationale for testing the hypothesis that treatment with recombinant LLD might protect against lung ischaemia–reperfusion injury. Indeed, when the LLD of TM was administered via gene delivery prior to the ischaemia–reperfusion protocol, BALF accumulation of macrophages, lymphocytes and neutrophils, as well as total protein and MCP-1 levels, were significantly suppressed, as was neutrophil lung infiltration. Peripheral circulating blood counts were not measured, and thus it could not be determined whether the anti-inflammatory effects were localised or systemic. Nonetheless, the data provide evidence of the potential clinical utility of the LLD of TM in protecting against lung ischaemia–reperfusion injury.

The current lung ischaemia–reperfusion injury studies were restricted to a single reperfusion interval, as well as a single “dose”. Thus, further study will be required to assess the short- and long-term effects of this intervention, as well as the physiological significance of the LLD-induced changes in cytokine release. IL-1β, GM-CSF and MCP-1 are all inflammatory cytokines, each of which has been linked to the pathogenesis of tissue injury associated with ischaemia–reperfusion. MCP-1 is important for inflammation-associated monocyte recruitment and activation, as well as for neutrophil accumulation 35. In a murine model of ischaemia–reperfusion injury similar to that used in the present study, Zhao et al. 36 showed that MCP-1 expression is dependent on the presence of alveolar macrophages, and that BALF levels of MCP-1 and TNF-α correlate with the extent of lung injury. GM-CSF is best characterised by its ability to stimulate the mobilisation and proliferation of neutrophil and monocyte precursor cells. However, in the setting of ischaemia–reperfusion, it may also have pro-apoptotic effects, and thus contribute to tissue damage 37. IL-1β plays a critical role in mediating the acute phase of the inflammatory response after ischaemia–reperfusion, and in a pig model of ischaemia–reperfusion injury, BALF levels of IL-1β correlate with graft function, suggesting that it may be a marker of graft viability 5. Although the LLD of TM interferes with the activity of the pro-inflammatory HMGB1 20, the present study did not detect differences in the subcellular pattern of expression of HMGB1 in the lung tissue sections of TMwt/wt versus TMLeD/LeD mice or treated versus sham-treated mice. Indeed, even after the ischaemia–reperfusion lung injury, HMGB1 remained mainly localised to the nuclei, where it is not believed to exert a pro-inflammatory effect. Likewise, no differences in coagulation were detected between LLD-treated and sham-treated mice that might be differentially affected by HMGB1 release 24. Until recently, HMGB1 was considered to be a late mediator of inflammation. Thus, it was not expected to play a role in the present model of acute ischaemia–reperfusion injury. However, HMGB1 is now also recognised to mediate acute inflammation, including, for example, the response to severe haemorrhagic lung injury 23. The inability of the current authors to demonstrate a contribution of HMGB1 may be explained by several factors, including, for example, the type and duration of the stress, its relative severity and the strain of mice. Nonetheless, the findings do not exclude the possibility that the LLD of TM might still confer protection via HMGB1 after ischaemia–reperfusion injury at time periods beyond those evaluated. Overall, the current authors are confident from these studies that the LLD of TM does modulate cytokine release, and that this is reflected by therapy-induced attenuation in the number of leukocytes accumulating in the lung and entering the alveoli after reperfusion. More extensive analyses will provide insights into when the LLD of TM acts, and on which cell populations it is most active.

After systemic gene delivery of the LLD and 4 h of lung ischaemia–reperfusion injury, LLD could not be detected in the BALF. Likewise, increased deposition of the LLD in the lung parenchyma could not be confirmed 19, either by Western immunoblotting of lung lysates or by immunohistochemical staining of tissue sections (not shown). However, even in the setting of inflammation, where TM levels are suppressed 38, the endogenous expression of TM in the lung vasculature is relatively high, and recombinant LLD of TM could not be distinguished from endogenous sources. Thus, further study is required to identify the precise mechanisms by which the LLD of TM protects the lungs against ischaemia–reperfusion injury. In addition to interfering with leukocyte adhesion to endothelial cells and suppressing complement activation, the LLD does have direct anti-apoptotic effects on cultured endothelial cells 18, and one or more of these mechanisms may account for its protective role. Increased C1q or C3 deposition could not be detected in the lungs of the TMLeD/LeD mice (not shown). In addition, the effects of the LLD on pulmonary epithelium have not been examined, although alveolar type-II cells may express TM and release proteolytic fragments of TM 39. It is notable that APC, which has anti-inflammatory and anti-apoptotic effects on several cells, including lung epithelial cells, shares several intracellular signalling pathways with the LLD of TM, including suppression of NF-κB translocation and MAP kinase activation 40. APC, however, acts primarily via activation of protease-activated receptor-1 and EPCR 41, while it is not yet fully understood how the LLD of TM regulates intracellular signalling pathways. Nonetheless, it is likely that APC and the LLD of TM act cooperatively, and delineating how this occurs at a molecular level will provide important therapeutic insights.

In view of the overwhelming success of lung transplantation in saving lives of patients with end-stage lung disease, there is an urgent need to develop multiple approaches that may be used singly or in combination to prevent ischaemia–reperfusion injury, thereby substantially decreasing morbidity and mortality. The present authors have, for the first time, demonstrated that the lectin-like domain of thrombomodulin, a molecular fragment with potent anti-inflammatory properties but lacking anticoagulant effects, may be a useful adjunct in preventing ischaemia–reperfusion injury of the lung, and future studies will determine its utility in models of lung transplantation.

Support statement

This work is supported by grants from the KU Leuven (OT/03/55) and the Flanders Fund for Scientific Research (G.0282.07, G3C04.99), where D.E.M. Van Raemdonck is a senior research fellow. E.M. Conway was supported by National Institutes of Health (Bethesda, MD, USA) Grant R01-HL067846.

Statement of interest

None declared.


This study was presented in part at the 27th Annual Meeting and Scientific Sessions of the International Society for Heart and Lung Transplantation, San Francisco, CA, USA, April 25–28, 2007.

  • Received November 22, 2007.
  • Accepted May 9, 2008.


View Abstract