Abstract
Acute respiratory distress syndrome (ARDS) is defined as a syndrome of acute onset, with bilateral opacities on chest imaging and respiratory failure not caused by cardiac failure, leading to mild, moderate, or severe oxygenation impairment. The syndrome is most commonly a manifestation of sepsis-induced organ dysfunction, characterized by disruption of endothelial barrier integrity and diffuse lung damage. Imbalance between coagulation and inflammation is a predominant characteristic of ARDS, leading to extreme inflammatory response and diffuse fibrin deposition in vascular capillary bed and alveoli. Activated platelets, neutrophils, endothelial cells, neutrophil extracellular traps, microparticles, and coagulation proteases, participate in the complex process of immunothrombosis, which is a key event in ARDS pathophysiology. The present review is focused on the elucidation of immunothrombosis in ARDS and the potential therapeutic implications.
Introduction
According to the recent Berlin definition, acute respiratory distress syndrome (ARDS) is defined as a syndrome of acute onset, with bilateral opacities on chest imaging and noncardiogenic respiratory failure, leading to mild, moderate, or severe oxygenation impairment [1].The syndrome might be the result of a direct lung insult (pneumonia, aspiration) or the expression of a systemic event as sepsis, massive transfusion, or trauma. New insights into the pathophysiology of ARDS have not lead to equal therapeutic implications; therefore, the treatment remains supportive, mainly based on lung-protective mechanical ventilation (MV) as a buying-time strategy, neuromuscular blockade, prone position, and conservative fluid administration [2,3,4]. The mortality rate for the syndrome is estimated to be 30-50%, underlying the need for new effective therapeutic strategies.
ARDS is characterized by diffuse lung inflammation which results in endothelial and epithelial damage and subsequent increased vascular permeability [5,6]. The syndrome is most commonly a manifestation of sepsis-induced widespread organ dysfunction, with predominant features being the disruption of endothelial barrier integrity and diffuse lung damage. In the pathophysiology of ARDS, many cell types are involved, such as platelets, neutrophils, alveolar macrophages, and monocytes, as well as endothelial and alveolar epithelial cells. Following the invasion of a pathogen, neutrophils and other phagocytizing cells are recruited in the microvasculature, interact with platelets and contribute to the defense machinery [7]. Platelets aggregate to any site of tissue lesion, forming a platelet clot, followed by the activation of coagulation cascade. There is ample evidence that sepsis and ARDS are characterized by a procoagulant state, leading to a massive production of thrombin. The endothelial barrier is directly affected by thrombin, the predominant coagulation protein, which is converting fibrinogen to fibrin. The diffuse alveolar and interstitial fibrin deposition induces the formation of microthrombi [8,9]. Moreover, thrombin is the main activator of platelets, which further accumulate to the site of endothelial lesion, interacting with innate immune cells. In addition, the activation of protease-activated receptors (PARs) on immune cells by coagulation proteases induces pro- and anti-inflammatory reactions [10]. The recruitment and interaction of platelets and neutrophils at the site of endothelial injury, regulated by coagulation and inflammatory mediators, are therefore considered as a humoral regulatory process, defined as immunothrombosis (Fig. 1) [11]. Immunothrombosis results in the formation of an intravascular scaffold, enhancing the recognition and destruction of pathogens and supporting the endothelial integrity [12]. However, uncontrolled immunothrombosis might induce collateral tissue damage, and contribute to organ dysfunction. The aim of the present review is to summarize current knowledge on the role of immunothrombosis in ARDS pathophysiology, and to examine potential therapeutic implications.
Immunothrombosis in ARDS Pathophysiology
Inflammation and Coagulation
Activation of coagulation and excessive inflammatory response are essential characteristics of ARDS pathophysiology. Innate inflammatory response is initiated when specific pathogen-associated molecular patterns (PAMPs) are recognized by pattern recognition receptors (PRRs) expressed by immune cells [13]. Among them, Toll-like receptors (TLRs) have been identified in endosome, while nucleotide-binding oligomerization domain-like receptors (NLRs) are cytosolic PRRs. Both receptors are also recognizing nonendogenous PAMPs and endogenous damage-associated molecular patterns (DAMPs) produced by injured cells. The prominent TLR ligands are lipopolysaccharides (LPS), double-stranded RNA and lipoproteins. Interestingly, DAMPs are also released during sterile cell injury produced by trauma or other noninfectious causes of ARDS and multiple organ failure [7]. The recognition of PAMPs and DAMPs by immune cell receptors triggers a proinflammatory response. A pyrin domain-containing member of the NLR family (NLRP) is a component of a multiprotein complex, known as inflammasome. The activation of this complex by hypoxic cellular injury is implicated in the inflammation process during ARDS [13]. In a murine ARDS model, NLRP3 inflammasome is substantially contributing to the development of hypoxemia, by interacting with histones, while the subsequent production of cytokines is related to worse prognosis in humans suffering from ARDS [14,15].
Cellular injury during sepsis, trauma and other ARDS subtypes can cause the release of mitochondrial DAMPs into the circulation. These cellular components further activate polymorphonuclear neutrophils, propagating the inflammatory response [16,17]. The presence of mitochondrial DAMPs (DNA in particular) in plasma has been associated with the increased endothelial permeability observed in ARDS and might contribute to higher mortality in ARDS patients [18]. The above-mentioned mechanisms induce a proinflammatory response, aiming at the elimination of pathogens. However, the exaggerating proinflammatory procedure might result in collateral tissue damage and subsequent endothelial dysfunction.
The innate host response to endothelial damage is associated with the activation of coagulation, which in turn regulates and is regulated by the inflammatory process. The major pathological role of coagulation in the innate host response has recently been elucidated and defined as immunothrombosis [11,12], a humoral regulated process that might equally contribute to the protection of endothelial integrity and to the propagation of the inflammatory process. The immunothrombotic process is initiated by the formation of microthrombi in the microvasculature, as a result of endothelial injury (Fig. 1) [19,20]. Diffuse endothelial damage contributes to the exposure of subendothelial collagen, and to the expression of tissue factor (TF) and von-Willebrand factor on endothelial cells.
TF plays a constitutional role in the activation of coagulation, which is tightly correlated with the inflammatory process [21]. It is a membrane protein expressed in the fibroblasts of the adventitia of vessels of several organs like the brain, lung, and kidney. It is also distributed by epithelial cells, endothelial cells, platelets, and microparticles (MPs), particularly under inflammatory conditions. Normally, TF is encrypted, thus the normal host is protected against activation of coagulation [22]. Endothelial damage leads to the release of TF into the bloodstream and to subsequent interaction with the proteases of the coagulation cascade [23]. Therefore, TF binds to fVII, while the TF:fVIIa complex consequently activates fX. fXa and fVa constitute the prothrombinase complex in the presence of calcium and phospholipid surfaces provided by activated platelets. The prothrombinase complex activates prothrombin to thrombin. Thrombin production leads to the ample formation of fibrin and microthrombi. Microthrombi create a barrier against the invasion of pathogens, participating in the initial defense procedure. Furthermore, they contribute to the formation of a distinct complex with antimicrobiotic properties: on the active surface of microthrombi, innate immune cells are recruited and activated, generating the inflammatory response and enhancing further TF expression [24].
The roles of coagulation and inflammation in ARDS have already been extensively investigated prior to the introduction of the immunothrombotic concept. In this respect, pulmonary vascular lesions in ARDS patients have been described in pathological specimen but also using a specific angiographic technique, the balloon occlusive pulmonary angiography (BOPA), performed at the bedside [25,26]. Widespread presence of thromboemboli of various morphological types was detected, depending on the stage of ARDS. Interestingly, the diagnosis of diffuse intravascular coagulation (DIC) was not necessarily associated with the presence of capillary microthrombi. Although Greene et al. [27] reported that filling defects in ARDS patients could prognosticate mortality, the exact role of thrombosis as a triggering or propagating factor of lung injury was not clarified. Interestingly, Greene et al. [28] observed that streptokinase infusion in 5 ARDS patients with pulmonary vascular thrombosis, led to thrombi lysis and improved hemodynamic parameters. Moreover, Vesconi et al. [26] did not confirm a prognostic significance of filling defects in ARDS; they reported instead a correlation of the microthrombi presence with ARDS etiology (posttraumatic syndrome) rather than with ARDS severity. However, immunothrombosis cannot easily be depicted by imaging studies since it evolves in microvessels that are not completely occluded, and the size of thrombi can be as small as <10 μm [29,30].
Procoagulant state has long been recognized as an essential part of ARDS pathophysiology [31]. Saldeen [31] showed that thrombin infusion in animal models caused the “microembolism syndrome” characterized by diffuse microemboli in the pulmonary circulation. Indeed, the presence of fibrin microthrombi was confirmed in autopsies obtained from patients with ARDS [32]. The fibrin production is the result of diffuse thrombin generation, which regulates the 3 hemostatic domains: coagulation, anticoagulation, and fibrinolysis [33]. The ample thrombin production results in consumption of clotting factors. Moreover, thrombin interacts with thrombomodulin (TM), thus mediating protein C activation, which has been bound on the endothelial cell protein C receptor (EPCR) [34,35]. Activated protein C (APC) is an endogenous anticoagulant as it inhibits fVa and fVIIIa, while it exerts anti-inflammatory effects as well, by suppressing cytokine production. In addition, APC activates the fibrinolytic process, as it inhibits the plasminogen activating inhibitor-1 (PAI-1) [36,37]. Ware et al. [38] showed that in ARDS patients, the plasma levels of APC were decreased, while the PAI-1 levels were increased, compared to controls with cardiogenic pulmonary edema. These investigators additionally reported that the low APC and high PAI-1 levels were associated with increased mortality in ARDS patients, underlying that the procoagulant and antifibrinolytic state might be an attractive therapeutic target [38]. However, the concept of immunothrombosis reveals a different aspect of the previously described, possibly detrimental widespread thrombosis: the organized recruitment of innate cells and platelets at the site of endothelial injury leads to the release of molecular mediators contributing to the intravascular immune mechanism [30].
Regarding the different pathogenic factors associated with ARDS immunothrombosis, sepsis is the prevailing cause. There are several reports on the role of activated platelets and their interaction with leucocytes during infection [2,19]. Gando and Otomo [39] have recently described the role of immunothrombosis in acute coagulopathy of trauma shock; they concluded that in trauma patients the mechanism of immunothrombosis and hemostasis is dysregulated and the endothelial injury leads to fibrinogenolysis and subsequent diffuse oozing. Moreover, immunothrombosis and the particular role of extracellular histones have been extensively studied in transfusion-related acute lung injury (TRALI) models [40,41]. However, it seems that, irrespectively of the different pathogenic factors causing ARDS, the inflammatory process is the common pathway triggering the mechanism of microvascular thrombosis.
The influence of MV and positive end-expiratory pressure (PEEP) on microvessel thrombosis has not been investigated in large studies. There are several reports concerning the kinetics of coagulation and fibrinolysis components in bronchoalveolar lavage fluid (BALF) during MV. Schultz et al. [42] showed that levels of thrombin-antithrombin complexes (TAT), TF, and fVIIa in BALF were increased in patients with ventilator-associated pneumonia. Moreover, MV with higher tidal volume and PEEP was associated with higher levels of soluble TM, TF, TAT, fVIIa, and lower levels of APC in BALF, reflecting a procoagulant activity [43]. Haitsma et al. [44] have shown, in an animal model of Streptococcus pneumonia, that the implementation of high tidal volume MV was associated with local activation of coagulation and attenuation of fibrinolysis in the lung, whereas the levels of IL-6 in BALF were increased. The abovementioned phenomenon was described as “ventilator-induced coagulopathy.” Moreover, the bronchoalveolar activation of coagulation and inhibition of fibrinolysis have been associated with ventilation-associated lung injury [45]. Fewer studies investigated potential associations of MV and PEEP with plasma indices of coagulation. In this respect, data from the ARDS network study showed that activation of coagulation and depression of fibrinolysis, as expressed by increased plasma TAT and PAI-1 levels, respectively, were associated with higher mortality rates in ARDS patients [46]; interestingly, application of protective lung ventilation in these patients did not have any effect on PAI-1 plasma levels.
Links between Inflammation and Coagulation
The molecular link between inflammation and coagulation is mainly the PARs on immune cells, platelets, and endothelial cells. PARs constitute a subfamily of related G protein-coupled receptors, consisting of 4 members. PAR-1, 3, and 4 are activated by thrombin, while PAR-2 and 1 are activated by the complex of TF:fVIIa, and by fXa [47]. Activation of PAR-1 by thrombin signals the upregulation of inflammatory genes in the lung [48]. Additionally, activation of PAR-1 by APC or low levels of thrombin exhibits cytoprotective function, whereas the respective activation by high levels of thrombin might disrupt endothelial barrier function [7]. Interaction of thrombin with PAR-1, 3, and 4 on platelet's membrane leads to further platelet activation and recruitment. Moreover, the activation of PAR-2 by the TF:fVIIa complex results in the expression of adhesion molecules, enhancing the inflammatory process [49]. Inversely, proinflammatory cytokines, such as tumor necrosis factor-α, interleukin-1, interferon-γ, and LPS of Gram-negative bacteria, regulate the expression of TF [50]. In human studies, the intravenous administration of LPS leads to an increase in MPs containing TF into the circulation, resulting in thrombin production [51]. Moreover, in patients with ARDS, TF is increased in BALF, while its blockage by a special antibody results in decreased procoagulant activity [52]. TF pathway inhibitor (TFPI), the endogenous inhibitor of TF, is also increased in BALF of ARDS patients [53]. However, it appears to remain in an inactive form; thus, the procoagulant activity of TF is not inhibited. The consequent impaired balance between TF/TFPI contributes to the deleterious procoagulant and proinflammatory effects of TF on ARDS pathophysiology, via the diffuse intraalveolar and intravascular fibrin deposition [52,54].
Platelets are another essential component of ARDS pathogenesis, as they participate in neutrophil recruitment and in the subsequent development of immunothrombosis in capillary vascular beds. Thus, immunothrombosis is a regulated immune process evolving on the active surface of the fibrinogen-platelet plug, and supported by immune cells, platelets and coagulation-related molecules activating PAR receptors (Fig. 1) [12,19]. This biological process might enhance the elimination of pathogens and the preservation of endothelial integrity. Nevertheless, the exaggerated activation of immunothrombosis might promote the inflammatory process and play an important role in ARDS development [30].
Platelet activation during endothelial damage leads to a change of their shape and to the release of their granules' content, consisting of immunomodulatory mediators and several cell membrane proteins. One such protein is P-selectin, which in turn causes neutrophil migration and activation at any site of injured endothelium [2]. Moreover, the platelet-induced expression of the intracellular adhesion molecule-1 (ICAM-1) on endothelial cells further facilitates neutrophil adhesion. The produced platelet-neutrophil complexes participate in the inflammatory process via expression of proinflammatory cytokines, and in the defense mechanism by increasing phagocytosis capacity [55,56,57]. The sequestration of these complexes at target organs, including the lung, is implicated in the pathophysiology of multiple organ failure, as a result of immunothrombosis and of the subsequent development of vaso-occlusive thrombi in pulmonary capillary vessels.
Immunothrombosis is further promoted by neutrophil extracellular traps (NETs) and MPs (Fig. 1). Activated neutrophils release the chromatin of their nucleus and the proteins of their granular contents, forming the NETs [11]. NETosis, a procedure of cell death, is an important mechanism of innate immune response. Although NETs show antimicrobial properties by trapping microorganisms in blood vessels, they might cause collateral tissue damage. Interestingly, NET formation has been implicated in lung injury after LPS and bacterial challenge [57].
MPs are involved in another mechanism regulating platelets and innate cell interplay. MPs are small cell-derived vesicles released from activated cells, playing an important role in intracellular communication. More specifically, MPs are generated by platelets and endothelial cells, as an effect of various triggering factors, including mechanical injury and inflammation [58]. They contain several enzymes and proteins, which serve as biological signals, promoting information exchange. MPs participate in the inflammatory response during various diseases, as diabetes, hypertension, and atherosclerosis. During lung injury, they interact with platelets, endothelial and innate cells, stimulating the generation of proinflammatory cytokines and exerting a procoagulant and immunomodulatory role [59]. The unique role of the aforementioned cells and cell mediators in the development of immunothrombosis in ARDS will be further analyzed.
Platelets and Innate Immune Response
Platelets are circulating anucleate cells produced by megakaryocytes. During endothelial damage, the exposure to subendothelial collagen leads to platelet activation, consisting of a change of their shape and subsequent release of cell membrane proteins and granular contents, including chemokines, cytokines, coagulation proteases, adhesive molecules, growth factors, and mediators of angiogenesis [60,61]. Platelet-derived mediators lead to further platelet aggregation and activation, and promote hemostatic mechanisms and clot formation. However, many of these agents, apart from their considerable role in the coagulation cascade, participate in the regulation of inflammatory response. Surface proteins CD40 and CD154 interact with antigen-presenting cells, such as dendritic cells or macrophages. The interaction results in the expression of ICAM-1 and vascular cell adhesion molecule-1 (VCAM-1), and leads to neutrophil recruitment at the site of tissue damage [62]. Additionally, activated platelets express TLRs, recognizing several PAMPs. This activation enhances the production of various proinflammatory cytokines, such as interleukin-1 and tumor necrosis factor-α.
Platelets additionally participate in the immune response through direct bactericidal activity. This occurs by releasing antimicrobial peptides (AMPs) also called defensins, which directly destroy bacterial membranes (Fig. 1) [63]. AMPs are stored in α-granules, with the exception of human β-defensin (hBD-1), which is stored in submembrane cytoplasmic sites. It is noteworthy that hBD-1 is released during the terminal lysis of platelets, thus protecting the host from an inappropriate release and a subsequent cytotoxicity [64].
Interestingly, platelets seem to have different characteristics during health and disease [65]. During sepsis, the hemostatic properties of platelets are decreased; meanwhile, the secretion of adhesion molecules and growth factors is preserved [66]. In an animal model of sepsis, produced by cecal ligation, circulating platelets have shown increased expression of acute-phase proteins [67]. Moreover, acidosis produced by tissue hypoxia, enhances neutrophil-platelet interactions during the inflammatory process, albeit decreasing hemostatic properties [68].
In ARDS, the accumulation and activation of platelets at sites of lung injury is linked to the production of vaso-occlusive thrombi in lung capillaries and to the subsequent development of immunothrombosis. In postmortem studies of ARDS patients, considerable deposition of platelets has been found in the pulmonary capillary vascular bed [69]. Moreover, the severity of lung injury has been tightly correlated with the presence of platelet-derived α-granule mediators in BALF [70]. Taken together, these results highlight the considerable role of platelets in immunothrombosis in terms of ARDS pathophysiology. The contribution of platelets in immune defense exceeds their traditional role in the thrombotic process, and is an issue of ongoing research with possible therapeutic implications, regarding the use of antiplatelet treatment in ARDS prevention [71,72].
Interactions among Platelets, Neutrophils, and Endothelial Cells in ARDS
ARDS is characterized by deranged endothelial barrier function and increased vascular permeability. Lung alveoli are occupied by fibrin-rich exudates caused by activation of coagulation, resulting in excessive thrombin formation [23]. While platelets emerge at the site of endothelial damage in order to support endothelial barrier integrity, thrombin causes direct barrier dysfunction, by affecting endothelial cytoskeleton via its receptor PAR-1 [5]. As already mentioned, activation of PAR-1 by APC or low levels of thrombin has cytoprotective function through the regulation of the balance between sphingosine 1 phosphate receptor 1 (SIP1) and SIP3 [73]. During lung injury, platelets contribute to endothelial cell stimulation by causing the expression of adhesion molecules, such as ICAM-1 and VCAM-1, and promoting the respective migration of neutrophils [74]. “Neutrophil rolling” is facilitated by platelet-derived P-selectin and thromboxane-A2 (Fig. 1). P-selectin creates a phospholipid framework that enhances neutrophil adhesion to platelets and attenuates neutrophil circulation in the capillaries, thereby accelerating their migration through the endothelium [75]. Interestingly, the presence of neutrophils at the site of lung injury is essential for the respective recruitment of more platelets; in animal models, the depletion of neutrophils prior to the induction of endotoxemia reduces further migration of platelets to the lung [76].
The platelet-neutrophil complex shows different properties than platelets or neutrophils alone in terms of increased adhesion molecule expression, greater phagocytosic activity, and production of toxic oxygen radicals [77]. Triggering receptors expressed on myeloid cells-1 (TREM-1) are expressed by neutrophils and recognize specific ligands on platelets. TREM-1 activation is linked to the release of numerous proinflammatory cytokines and chemokines, which amplifies the inflammatory response and generates the immunothrombotic procedure (Fig. 1). As excessive inflammatory reaction might lead to collateral tissue damage, blocking TREM-1 might be an attractive therapeutic target aiming to depress inflammation [78]. The release of leukocytes' granule proteins during ARDS, including azurocidin, elastase, and several matrix metalloproteinases, promotes the activation of coagulation, apoptosis of epithelial cells, and degradation of surfactant proteins [79]. Neutrophil elastase (NE), an important proteinase of neutrophil granules, is implicated in lung injury by direct damage of endothelial or epithelial cells, i.e. by degradation of capillary-alveolar membrane [80]. Interaction of platelets with neutrophils also leads to the release of platelet cytokines, such as PF4-RANTES and thromboxane A2, which cause upregulation of adhesion molecules and contraction of endothelial cells, thus contributing to lung injury [81].
There is ample evidence derived from animal and human studies, showing the important role of platelet-neutrophil interaction in ARDS immunothrombosis. In a murine model of acid-induced acute lung injury, the inhibition of platelet-neutrophil complex accumulation improved gas exchange and led to prolonged animal survival [74]. In a different model of LPS-induced lung injury, platelet depletion resulted in decreased neutrophil accumulation in the lung. In addition, antibodies against platelet-derived chemokines reduced lung capillary permeability and subsequent lung edema, supporting the significant role of chemokines in lung injury development [79]. In a small human study, platelet activation was greater in ARDS patients than in healthy controls [82]. In other studies investigating the genetic predisposition to ARDS, the presence of genetic variants of genes implicated in inflammation, coagulation, and endothelial barrier integrity has been associated with increased ARDS risk [83,84].
Neutrophil Extracellular Traps
NET formation is playing an important role in platelet-neutrophil interaction during immunothrombosis. NETs are released from neutrophils on contact with bacteria, under certain circumstances such as the interaction with activated platelets or the presence of inflammatory stimuli [85]. NETosis is an additional mechanism used by neutrophils in order to extinguish pathogens. Activated neutrophils mobilize their chromatin, which decondensates and expands outside the cell together with granular antimicrobial factors. NETs share antimicrobial and procoagulant properties. They are trapping Gram-negative and Gram-positive bacteria, which are inactivated by NET-associated proteases [86]. They additionally have procoagulant properties, since they cause platelet activation and aggregation, and the activation of the coagulation pathway [87]. This procedure is mediated by TF expressed in neutrophils and delivered to the extracellular space via NET release [88]. Indeed, it has recently been shown that in septic patients, an increased amount of TF enclosed in NETs is released, thus contributing to thrombin generation [89,90]. Moreover, NETs bind fXII, thus contributing to its activation and to the intrinsic coagulation cascade. This process has been associated with deep vein thrombosis (DVT) development [88]. In an animal model of DVT, the administration of DNAse led to NET cleavage and to a subsequent inhibition of the thrombotic process [91]. NETs also participate to the procoagulation process by trapping and activating platelets, the central orchestrator of immunothrombosis. Activated platelets further enhance NET formation [92] leading to the propagation of thrombi.
NET formation may lead to an excessive inflammatory response, in case of sepsis and ARDS. NET release is favored by the high concentration of neutrophil activating factors in alveolar space. Moreover, reduced surfactant proteins result in impaired clearance of histones [57]. NET components, including histones, myeloperoxidase (MPO), NE, and cathepsin G may exert cytotoxic activity against lung epithelial and endothelial cells [93]. In particular, NE degrades endothelial cytoskeleton and causes the loss of function of E-cadherin and VE-cadherin, which affect the integrity of alveolar capillary barrier. NE also participates in the release of proinflammatory cytokines and is associated with apoptosis of epithelial cells. NE contributes to the proteolytic cleavage of TFPI as well, while it enhances the activation of fX, thus regulating the procoagulant process [94,95]. In addition, MPO is leading to the production of reactive oxygen species, contributing to epithelial cells apoptosis [96]. In a mouse model of sepsis, blockage of histones led to decreased mortality, while these DNA structures were associated with diffuse alveolar damage and hemorrhage [97]. Lung injury has been related to NET release after LPS, bacterial, or fungal stimuli [98]. It is noteworthy that positively charged extracellular histones could be a target for negatively charged molecules such as heparin or albumin [99,100]. These molecules might share a promising role in the treatment of sepsis and ARDS, due to their potential antihistone capacity. Indeed, there are several studies regarding the administration of nebulized heparin in animal models of ARDS, showing an attenuation of pulmonary inflammation and coagulopathy, without affecting fibrinolysis [101,102]. However, there are few related clinical studies, and further evaluation is required [103,104].
In ARDS, platelets participate in NET activation, through TLR4 engagement. In a human study consisting of patients with severe sepsis caused by blood stream infection, platelets' TLR4 recognized TLR4 ligands in neutrophils. Consequently, neutrophils were activated, with subsequent release of NETs, especially in pulmonary capillaries, resulting in severe lung injury [2]. NETs have also been implicated in the pathophysiology of TRALI, which is one of the most severe complications of transfusion, defined as lung injury developing within 6 h of transfusion of 1 or more units [41]. Animal and human studies concerning TRALI have shown that NETs are found in abundance in the lungs and circulation [40]. During immunothrombosis, extracellular histones along with neutrophil proteins might create a tangle web in which platelets are trapped in lung microcirculation (Fig. 1). Moreover, extracellular histones are directly toxic to endothelial cells, contributing to increased permeability of endothelial monolayer [76]. The increased endothelial permeability, combined with microcirculation obstruction and immunothrombosis, promotes lung injury [105]. In an animal model of TRALI, pretreatment with a platelet inhibitor, aspirin or an inhibitor of IIb/IIIa glucoprotein, reduced platelet sequestration at the lung and decreased lung permeability and edema [106]. Additionally, pretreatment of mice with a DNAse, which dismantles the web of extracellular chromatin, decreased NET formation and also reduced lung edema [76]. Consequently, the platelet-induced NET production could be a worthy target for TRALI and lung injury treatment. Taken together, the main activity of NETs is the extracellular expression of various intracellular components essential to the prothrombotic and innate immune response, thus contributing to immunothrombosis.
Microparticles
MPs are small vehicles modulating intracellular communication that participate in immunothrombosis met in ARDS. MPs are generated from many cell types, including platelets, endothelial cells, polymorphonuclear neutrophils and lymphocytes. However, in humans, platelet- and megakaryocyte-derived MPs predominate (PMPs) [107]. MPs are shed from precursor cells after several triggering factors, like mechanical injury and inflammation (Fig. 1). They preserve the lipid bilayers of parent cells, while they contain proteins, ribosomal RNA, messenger RNA and microRNA [59,108]. These mediators are secreted by MPs towards various target cells, generating intercellular information exchange. Moreover, MPs have been identified as biomarkers and important regulators of cell interactions, in disproportionate inflammatory responses such as sepsis and multiorgan dysfunction syndrome, including ARDS [109]. Interestingly, in ARDS MPs are implicated in the inflammatory and coagulation response, conveying beneficial and detrimental effects [59].
PMPs, which are released from platelets during inflammation, mainly participate in polymorphonuclear neutrophil activation and in the subsequent degranulation and leukocyte accumulation. Moreover, LPS-stimulated platelets contribute to the activation of human endothelial cells through the release of PMPs which induce VCAM-1, cytokine and inflammatory mediator production [110]. PMPs are also activated by phospholipase A2, leading to arachidonic acid synthesis and to the subsequent production of thromboxane A2 [111]. The latter is an important contributor to the increased vascular permeability, which characterizes ARDS. Taken together, PMPs interplay with platelets, leukocytes, and endothelial cells, in terms of generating proinflammatory and immunomodulatory effects, thus participating in immunothrombosis. These effects are more pronounced during TRALI, since storage of blood products further enhances MP shedding [112].
PMPs are considered to contribute to the hypercoagulable and hypofibrinolytic state of the lung, which is tightly associated with ARDS pathophysiology. TF pivotally contributes to this prothrombotic state, through activation of PAR1 and PAR2, thus leading to ample generation of fibrin [113]. TF activity is counterbalanced by the antithrombotic activity of TFPI. During the inflammatory process, monocyte-derived MPs reduce TFPI expression, while they abundantly express TF, contributing to the progression of coagulation procedure [114]. Bastarache et al. [115] demonstrated that the pulmonary edema fluid from ARDS patients contained a higher amount of MPs expressing TF compared to the fluid obtained from control patients. Interestingly, these procoagulant MPs conferred a trend towards higher mortality. In addition, PMPs' membranes contain large quantities of phosphatidyloserine, providing a suitable environment for the activation of the coagulation cascade [116]. In line with this observation, PMPs show 100-fold higher prothrombotic effects than their precursor platelets [117].
Apart from these proinflammatory and procoagulant detrimental effects of MPs on ARDS pathogenesis, MPs may show various beneficial effects constituting several anti-inflammatory and anticoagulant properties. In sepsis and ARDS, it has been shown that an increased number of lymphocyte-derived MPs or endothelial-derived MPs is associated with a higher survival rate [118]. Even more, 12-lipoxygenase, an eicosanoid stored in PMPs, leads to the release of the anti-inflammatory leukotriene lipoxin A4. In animal models, lipoxin A4 analogues attenuated acute lung injury [119]. According to a recent analysis in humans, lymphocyte-derived MPs were higher in ARDS patients who survived, compared to mechanically ventilated or spontaneously breathing controls [118]. In addition, MPs might show anticoagulant activity by expressing natural anticoagulants, such as TM, TFPI, and EPCR [120]. Moreover, MPs might protect the integrity of the endothelial barrier, since they limit reactive oxygen species production, enhance nitric oxide (NO) production and contribute to endothelial repair [121].
Taken together, MPs might confer beneficial and detrimental effects in ARDS pathophysiology, interfering in platelet, immune cell, and endothelial cell interactions, which regulate the immunothrombosis process (Fig. 1).
Immunothrombosis in the Resolution of ARDS and Possible Therapeutic Implications
Immunothrombosis as a Therapeutic Target
The abovementioned reports in animal and human studies, underscore the important role of immunothrombosis as a part of intravascular immunity, contributing to ARDS pathogenesis. The diffuse fibrin production, a predominant characteristic of the procedure, provides a scaffold that captures pathogens, while the accumulated coagulation factors regulate the inflammatory response via PARs [122]. The procoagulant state met in the ARDS lung has been the target of several therapeutic approaches. Despite the initial enthusiasm of recombinant human (rh)APC administration in septic ARDS patients, the randomized clinical ADDRESS trial more recently showed that rhAPC does not alter ARDS outcome [123]. In a similar respect, antithrombin administration in critically ill patients did not decrease the mortality rate, while it increased the risk of bleeding [124]. Moreover, human recombinant TFPI in a phase III trial did not show any benefit on the prognosis of patients with ARDS and severe sepsis [125]. A promising drug is the soluble recombinant human TM (rhTM), which has been approved in Japan for the treatment of patients with DIC. A more recent trial showed that the combination therapy with the NE inhibitor sivelestat and rhTM in ARDS patients with DIC was associated with improved survival [126]. However, larger studies are required in order to elucidate the usefulness of the above regimen.
In recent years, there has been increasing interest in the use of inhaled anticoagulants in ARDS patients. There are several reports concerning the administration of inhaled heparin, APC, and antithrombin in animal ARDS models, showing effects on pulmonary coagulopathy [101,102,127]. A recent study of Rehberg et al. [128] on the effect of intravenous antithrombin along with nebulized heparin and nebulized tissue plasminogen activator in a sheep ARDS model showed that, despite the observed improvement in gas exchange, the anti-inflammatory effects of antithrombin were abolished; the authors concluded that there may be possible interactions between anticoagulants. Given the paucity of large trials on effective treatments, the association of endothelial injury with endovascular immunity and coagulation, as expressed by immunothrombosis, might provide an effective research model for further therapeutic implications and a promising alternative target.
The Role of Platelets and Immune Cells in the Resolution of ARDS - Preventive and Therapeutic Implications
The apoptosis of neutrophils and the progressive inhibition of further neutrophil migration are promoting the resolution of lung injury. During this process, in which macrophages and lymphocytes play an important role, reformed epithelial and endothelial cell junctions contribute to the restitution of endothelial barrier integrity [71]. The resolution process is regulated, among others, by lipid mediators such as resolvins and lipoxins produced from the eicosapentaenoic acid. Lipoxins share anti-inflammatory and immunomodulatory activity. Platelets and neutrophils participate in the production of lipoxins, through the activation of the arachidonic acid pathway [2]. Lipoxins are increased during the resolution of the inflammatory process, in the course of which they induce macrophage activity and subsequent phagocytosis of immune cells [72]. Lipoxin production is promoted by platelet inhibitors, such as aspirin. Aspirin induces the acetylation of cyclooxygenase, which in turn converts arachidonic acid to a precursor molecule, through which neutrophils produce lipoxins that are activated by platelets [129]. Therefore, aspirin triggers the formation of metabolically active lipoxins that promote the resolution of inflammation during ARDS and sepsis.
Antiplatelet therapy has been widely studied in sepsis models. In particular, the administration of glycoprotein IIb/IIIa blockers, such as abciximab and eptifibatide in rabbits with Escherichia coli endotoxin-induced shock, has led to decreased mortality [130]. Several studies have shown beneficial effects of antiplatelet therapy in ARDS: In animal models, antibodies against platelet derived chemokines decreased the development of lung injury [79]. Moreover, in a mouse model of TRALI, the administration of aspirin led to increased survival, while it reduced thromboxane B2 levels and the amount of extravascular lung water [106]. There are several very recent human studies concerning the antiplatelet therapy in ARDS. Regarding the prevention of ARDS, the administration of antiplatelet treatment in patients with at least one risk factor led to a reduced incidence of the syndrome [131]. However, these results were not confirmed in a larger trial studying the prehospital administration of aspirin, which showed only a trend towards reduced occurrence of the syndrome [132]. Moreover, another prospective trial including patients with ARDS showed that prehospital administration of aspirin led to reduced intensive care unit mortality [133]. Nevertheless, there are several studies showing no significant effects of prehospital aspirin therapy on the incidence of ARDS [134,135]. In order to elucidate the effectiveness of aspirin on the prevention of ARDS, a multicenter randomized control trial was performed, which did not confirm a preventive role of aspirin in the development of ARDS in patients at risk [136].
A promising approach to ARDS treatment is conferred by mesenchymal stromal cells (MSCs) [137]. These are precursor cells isolated predominantly from bone marrow and from other tissues like fat, umbilical cord, and peripheral blood. They have the potential to differentiate to chondrogenic, osteogenic, and adipogenic directions [138]. Several clinical trials on MSCs are ongoing in neurological, autoimmune, and cardiovascular diseases [139]. There is also one completed trial in COPD patients, confirming the safety of MSCs administration, albeit no efficacy was proved [140]. There are 2 clinical studies currently ongoing in ARDS patients in the US and China concerning the administration of allogeneic MSCs of bone marrow and adipose tissue origin, respectively (NCT01775774 and NCT01902082). MSC activity produces a modulation of the immune response via paracrine mechanisms through the release of several modulators [138]. In preclinical ARDS models, mesenchymal cells decrease neutrophil accumulation and participate in the resolution process by modifying the alveolar macrophage phenotype [141]. The paracrine effects of MSCs are mediated, at least in part, by the secretion of MPs, containing pre-miRNA. These effects have been shown in rat models, in which the intracellular transfer of mRNA and miRNA reduced the apoptosis of kidney epithelial cells [142]. Moreover, MSCs might cause mitochondrial transfer to lung parenchymal cells through the secretion of MPs in cases of acute lung injury [143]. Taken together, MSCs constitute a promising therapeutic strategy in ARDS. However, further investigation is needed in order to clarify their exact mechanism of action and the subsequent efficacy in lung injury.
Conclusions
This review has highlighted the role of immunothrombosis as an innate host response during ARDS. Immunothrombosis is evolving on the surface of the fibrinogen-platelet plug and is supported by immune cells, platelets, and coagulation-related molecules activating PAR receptors. Endothelial injury triggers the initial host response, leading to recruitment of platelets and immune cells, and the subsequent development of platelet-neutrophil complexes. The initially produced clot is providing an active field for further activation of immune cells and coagulation pathways, leading to ample production of thrombin and fibrin in the pulmonary capillary bed. Coagulation proteases interact with immune cells inducing further pro- and anti-inflammatory reactions. Moreover, platelet and neutrophil-derived MPs and NETs regulate the intercellular communication, playing an important role in immune response, while causing collateral tissue damage. The consequent derailment of inflammatory and coagulation response is a key event in ARDS development. Platelet inhibitors and MSCs might be promising treatments for ARDS by modulating the immune response. Further understanding of the complex role of immunothrombosis in ARDS, as well as prospective randomized studies, will promote new therapeutic strategies focused on this devastating syndrome.
Acknowledgments
F.F. would like to thank Prof. Oreanthi Travloufor her mentoring in the field of coagulation pathophysiology. The authors thank Mr. Antonis Makriyannis for his excellent artwork (Fig. 1).