The pathogenesis of edema is regulated by trans-vascular differences in hydrostatic and osmotic pressure and by vascular permeability. Edema disturbs organ function and results in poor outcome in severe diseases. In particular, lung edema is critical in all cases: this condition disturbs gas exchange in the lung and causes hypoxia, leading to morbidity and mortality. Infection stimulates the secretion of inflammatory cytokines, yielding increased permeability. Severe infectious diseases induce massive tissue destruction, thereby accelerating coagulation processes. Some coagulation products in turn lead to increased endothelial permeability [1,2]. Lung edema by this mechanism is a major cause of hypoxia and death in sepsis, DIC (disseminated intravascular coagulopathy), and ARDS (acute respiratory distress syndrome). The regulation of such critical edema has been investigated as a therapeutic target. However, insights into hyper-permeability have not been translated into improved therapy and current treatments are limited to supportive modalities [3, 4].
Some hemostasis factors are closely related to inflammation and vascular permeability. For example, thrombin is the proto-typical protein that has strong effects on the endothelial barrier function and is involved in critical edema, as noted above [5,6]. Some endogenous molecules are known to regulate hyper-permeability; these factors have been proposed as candidate drugs for treatment of this condition. Recently, recombinant activated protein C was withdrawn from development as a therapeutic for sepsis . Furthermore, anti-coagulation therapy appears to have been abandoned as a candidate therapeutic . Thus, re-appraisal of known and novel endogenous molecules is needed to identify further candidate therapeutic agents. Here, we discuss the hemostatic factors that suppress endothelial permeability in the face of events that lead to increased permeability, and consider the potential of these hemostatic factors as therapeutic treatments.
The increase of endothelial permeability is a physiological reaction that occurs during inflammation. Following tissue injury by wound or infection, the body needs to increase endothelial permeability to permit survival and tissue repair [9,10]. In normal tissues, the concentrations of plasma proteins are low because of filtration by the vessel wall. However, high vascular permeability allows extravasation of plasma proteins of relatively large size, including coagulation factors. "Leakage" of these proteins causes extravascular coagulation, thereby yielding a tissue environment rich in serum proteins. In vitro exposure of mesenchymal cells to serum has been shown to result in changes in gene expression patterns. Additionally, fibrin generated by coagulation provides a matrix that facilitates the migration of cells involved in wound healing. Additionally, decreased intercellular adhesion of endothelial cells probably contributes to extravasation of leukocytes for local immune protection, while also facilitating migration of cells that participate in angiogenesis.
A variety of molecules participate in the deterioration of the endothelial barrier function and increased permeability, including VEGF (vascular endothelial growth factor), thrombin, histamine, serotonin, PAF (platelet activating factor), TNF-alpha, substrate P, endothelin-1, LPA (lysophosphatidic acid), nitric oxide, and bradykinin [11-13]. In addition to those endogenous molecules, a number of pathogen-generated toxins (such as LPS (lipopolysaccharide)) also cause increased endothelial permeability. Most of these factors (endogenous and exogenous) are related to injury and infection, in other words, to hemostasis and inflammation. Hemostasis and inflammation are closely related and the two processes generally provide mutual positive feedback. For example, thrombin-activated inflammation is stimulated by inflammatory cytokines such as TNF-alpha, interleukin-1-beta, and interleukin-6 . These inflammatory cytokines in turn increase the expression of several thrombophilic factors . This positive feedback system often activates numerous molecules for inflammation and coagulation; the combination of these processes can run out of control, resulting in the serious diseases noted above.
Permeability-decreasing molecules are outnumbered by permeability-increasing molecules. This distinction could explain why pathological hyper-permeability often results in critical diseases. Based on promising results from multiple animal studies, some permeability-decreasing molecules were expected to be effective against hyper-permeability in sepsis, DIC, and ARDS. Unfortunately, efforts to translate these laboratory studies into clinical applications have not generally been successful. However, several novel molecules have been proposed as candidate targets for the treatment of hyper-permeability. Protein C is a plasma protein generated by the liver; like several other factors in the coagulation pathway, protein C activity is vitamin K-dependent. The activation of protein C zymogen is accelerated via the thrombin-thrombomodulin-endothelial protein C receptor (EPCR) complex, which is located on the surface of endothelial cells . In addition to its anticoagulant effects, activated protein C (APC) has cytoprotective effects on endothelial cells . APC reduces cytokine responses, suppresses cell migration, and protects endothelial barrier function; APC also enhances the generation of sphingosine-1-phosphate (S1P; see below) by sphingosine kinase. In early limited trials, treatment of septic patients with APC revealed significant efficacy . In larger trials, the increases in adverse bleeding were reported (7). However, small improvements in outcomes were found in subgroups that were administered early in sepsis or who presented with DIC . Related work was performed with thrombomodulin, a cell surface protein that is displayed on endothelial cells and plays important roles in the protein C activation process. Recombinant soluble thrombomodulin has been reported to suppress vascular permeability [20-22]. In a large randomized trial of patients with sepsis, recombinant thrombomodulin provided minimal but statistically significant improvements [23, 24].
Antithrombin (AT) is a plasma protein generated by the liver . AT forms a complex with thrombin and inhibits thrombotic function as well as the functions of coagulation factors VII, IX, X, XI, and XII. Notably, AT suppresses endothelial permeability via inhibition of thrombin activity. A large-scale trial of AT treatment in septic patients did not reveal any significant improvement . However, AT recently has been reported to provide a slight but significant decrease in overall mortality in septic patients with DIC . S1P is not a hemostatic factor, but this molecule is closely related to the coagulation system . Specifically, S1P is a bioactive lipid mediator that regulates cell proliferation, migration, and apoptosis. S1P is a key molecule for inflammatory responses and endothelial permeability . Sphingosine is released from ceramides and phosphorylated by sphingosine kinase (SphK1 and SphK2). S1P is present at high concentrations in blood and at low concentrations in tissue, reflecting the presence of an S1P lyase activity in tissues . S1P binds to a series of G-protein-coupled receptors designated S1P receptors (S1PRs) 1 to 5 . Effects of S1P vary considerably depending on the type of receptor. This lipid has been reported to enhance endothelial barrier function by activation of a small GTPase, Rac, upon binding to S1PR1. In contrast, signaling via S1PR2 or S1PR3 seems to mediate signaling in the opposing direction, resulting in an increase of permeability. Thrombin activates RhoA/ROCK (Rho-associated protein kinase) signaling via PAR1 (proteaseactivated receptor 1), increasing endothelial permeability. S1P signaling counteracts the thrombin effect and protects vasculature from hyper-permeability. A number of animal experiments have shown the efficacy of S1P or S1PR agonists in preventing hyper-permeability in models of acute lung injury, cecal ligation and puncture, sepsis (via LPS injection), experimental ischemic stroke, or hemorrhagic shock . Additionally, patients rendered septic by various causes have been noted to have decreased S1P blood concentrations. Therefore, S1P is a potential target for preventing hyper-permeability in cases with sepsis, DIC, or acute lung injury. S1P exhibits a short physiological half-life; to address this challenge, synthetic agonists of S1PRs (FTY720, FTY720 S phosphate, and ONO-4641) have been developed. Coagulation factor IX (FIX) is essential for normal coagulation; FIX deficiency results in hemophilia B, evidenced as a tendency to bleed or bruise . The FIX protein consists of an N-terminal light chain, a C-terminal heavy chain, and an intervening connecting region . During the coagulation process, FIX is cleaved at two Arg residues by activated factor XI or by the factor VIIa/TF (tissue factor) complex. This cleavage releases the connecting region, also known as the activation peptide, resulting in FIX activation. Activated FIX, which consists of a light chain and a heavy chain linked by disulfide bonds, in turn cleaves factor X. Recently, Mamiya et al. have reported that FIX suppresses the permeability of the endothelial cell layer . This function of FIX was localized to the FIX activation peptide (FIX-AP) . Native FIX and FIX-AP (generated as part of the coagulation process) have functions in regulating endothelial permeability, such that native FIX circulating in blood vessels maintains vascular permeability. A 23-residue synthetic FIX-AP peptide induces cell spreading and the closing of intercellular gaps between endothelial cells in vitro and reduces endothelial permeability in a mouse model of lung sepsis induced by LPS injection. FIX-AP also has been shown to decrease levels of eNOS (endothelial NO synthase) protein and intercellular cGMP (cyclic guanosine monophosphate) in vitro. Because NO potently increases vascular permeability, suppression of NO production could be a mechanism of FIX-AP function .
Factor XIII (FXIII) is a transglutaminase that cross-links glutamine and lysine residues . FXIII, which works at the final stage of coagulation process, is activated by thrombin. Activated FXIII cross-links fibrin monomers, stabilizing and strengthening clots. There are two types of FXIII, plasma FXIII and intracellular FXIII. Plasma FXIII is a heterotetramer that consists of two catalytic A subunits and two B subunits. Cellular FXIII a homodimer that consists of two A subunits. Noll et al. reported that thrombinactivated FXIII enhances endothelial barrier function, and demonstrated that this function localizes to subunit A2 .
Recombinant subunit A2 suppressed endothelial permeability in models of heart ischemia-reperfusion, hemorrhagic shock, and gut ischemic reperfusion injury [40,41]. Treatment with the recombinant FXIII subunit A2 decreased cytokine responses and limited polymorphonuclear leukocyte activation. Recombinant FXIII subunit A2 also has been reported to minimize the incidence of myocardial edema during extracorporeal circulation in surgery for congenital heart disease . Given the 20-year history of FXIII concentrate use, administration of exogenous FXIII is considered safe enough for clinical use . The mechanism of FXIII's effect on endothelial permeability is still not fully understood. Further studies will be needed to clarify FXIII's potential utility as a therapeutic agent for hyper-permeability.
The treatment of sepsis with selected anti-coagulation factors has provided promising results in animal studies testing regulation of hyper-permeability in lung and decreased mortality. Notably, these factors yielded significant effects in small-scale clinical trials, but not in large-scale trials. Analysis of subgroups within the large-scale studies yielded significant improvements.
These results suggest that some anti-coagulation factors may indeed block some pathways in septic patients, but that these effects are not sufficient to provide a cure. Life-threatening endothelial hyper-permeability arises as a result of hemostasis and inflammation. These two processes clearly occur in concert, creating a solid network, particularly in severe conditions such as sepsis. It seems unlikely that the administration of a single molecule will be sufficient to simultaneously ameliorate the combined effects of both processes. A number of anti-coagulation, anti-platelet, or antiinflammation drugs have already been characterized. The use of combinations of drugs should be considered for the treatment of hyper-permeability in sepsis, DIC, and ARDS.
Davis et al. suggested that future therapies for sepsis will need to differentiate between pathogenic coagulation and requisite hemostasis . From this perspective, molecules that circulate in their active form, such as S1P and FIX, are of interest. Hemostasis and inflammation are ordinarily local events that should be strictly restricted to the site of injury. It is possible that S1P and FIX normally protect the healthy endothelial barrier (in tissues with normal circulation) from the effects of local infection and injury. Further study of hemostatic molecules will be needed.