Pulmonary edema and lesions are frequently observed in inhalation anthrax [29, 47, 48]. Elevated plasma D-dimer indicates coagulopathy in anthrax patients [29, 30]. These manifestations are reflected in the results of in the rat experiments described in this study. Compared to the rat model, several pathological features of coagulant regulation are distinct in mice. LT induces thrombocytopenia in mice [5, 8], but not in rats . The results of our D-dimer analyses suggest that coagulopathy is involved in LT-mediated pathogenesis in rats (Figure 5B), which is consistent with anthrax in human [29, 30], but not in mice . The protective role of aPC in LT-mediated mortality in rats (Figure 4B), further suggests a role of coagulopathy. In contrast, the lack aPC-mediated amelioration of LT-induced pathogenesis in the C57Bl/6J mice in our lethal dose experiments indicates that coagulopathy does not contribute significantly to mortality in mice (data not shown). Despite these distinct differences in coagulopathy, LT-induced cardiopathy is similar in the rat and mouse models [3, 4, 7, 12–14]. The significant differences in lung pathology indicate a mechanism other than heart dysfunction. Thus, coagulopathy was a consideration. Although no obvious systemic changes in clotting time or plasma anticoagulants were observed, fibrin deposition was observed in the lung sections, suggesting that the lung is the primarily organ affected by LT-induced coagulopathy in rats.
In the absence of systemic coagulopathy, localized activation of coagulation in the lung may cause significant pathology. LT-induced pathogenesis in rats shares common features with acute lung injury and acute respiratory distress syndrome in humans, including multiple organ dysfunction, intra-alveolar coagulation with fibrin deposition on the hyaline membrane [49, 50], and decreased mortality with aPC treatment . Coagulant activation and the presence of fibrin is known to be associated with the pathophysiology of lung injury [19, 20, 51, 52], and LT has been shown to increase the paracellular permeability of endothelial cells . Our data indicate that this increased permeability results in the extravasation of fibrinogen and fibrin into the alveolar lumen (Figure 1B, 2L).
Fibrinogen, fibrin and related degradation products have been shown to suppress surfactant functions in the alveoli [51, 52]. The activation of coagulation pathways can induce local leukocyte traffic and endothelial cell permeability  that may exacerbate lung injury [20, 54]. The elevation of D-dimer following LT treatment indicates the initiation of the coagulation-fibrinolysis cascade (Figure 1C; Figure 5B). Unlike endothelial cells, the cells comprising the alveolar surface do not express thrombomodulin. The LT-induced release of endothelial thrombomodulin into the plasma that supposedly will suppress vascular anticoagulant protein C pathway. Because the LT-induced coagulopathy is primarily restricted to the lung, the circulating plasma level of aPC may not increase in the same manner as occurs in LPS-treated animals (Figure 5F, LT vs. LPS). Since aPC can inactivate the coagulation cascade and block fibrin formation , this suggests the ameliorative role of aPC in LT-treated animals (Figure 4f, 5). The anticoagulant and anti-inflammatory activities of aPC play protective roles in lung injury and asthma . Because LT-mediated pathogenesis does not induce significant secretion of proinflammatory cytokines [8, 11], anticoagulant activity of aPC is likely the major effect that contributes to the amelioration of LT-induced pathogenesis.
Recombinant aPC (Xigris) was previously approved by United States Food and Drug Administration (FDA) for the reduction of mortality in adult patients with severe sepsis . Xigris was, however, withdrawn from the market by the manufacturer after the failure of its world wide trail to treat severe sepsis [57, 58]. LT-mediated pathogenesis does not induce a prominent inflammatory response [8, 11], and the changes in aPTT, PT, protein C, and antithrombin III are distinct from those of endotoxin-induced sepsis (Figure 5D and 5G). Thus, LT-induced pathogenesis is significantly different from sepsis. As a result, Xigris may, nonetheless, ameliorate LT-mediated pathogenesis by re-establishing the regulation of coagulation. Further investigations of the ameliorative effects of Xigris on LT-mediated pathogenesis are warranted.
Consistent with the results of previous studies [12, 13], the mean blood pressure of the LT-treated rats significantly decreased at approximately 4 h following LT treatment (Additional file 1: Figure S2A vs. Figure S2B). To ascertain whether the elicitation of lung edema occurred prior to the circulatory collapse, we examined the extent of lung injury by measuring the blood pressure and the lung wet-to-dry weight ratio of LT-treated rats. We found that both pulmonary and circulatory abnormalities were induced over a similar time course (Additional file 1: Figure S2A-B; 4-hour, LT + vehicle vs. vehicle groups). The changes in levels of plasma D-dimer and thrombomodulin were also associated with a similar time course (Figure 5B-C, 4-hour groups). Thus, the LT treatments elicited both coagulation-mediated lung injury (Figure 5 and Additional file 1: Figure S2C) and heart dysfunction (Additional file 1: Figure S2B) in rats. Therefore, we propose a hypothetical model for LT-mediated pathogenesis that is based on the hemodynamic changes and the altered coagulation occurring simultaneously (Additional file 1: Figure S3).
Severe heart failure leads to the redistribution and accumulation of body fluid in the lung [15, 16], which would likely exacerbate the increased vascular permeability and the coagulation-activation in the lungs of LT-treated rats. Coagulopathy may also negatively influence cardiac function, resulting in further hemodynamic changes [17, 18]. Dysregulation of protein C activation and thrombosis are also caused by secondary pulmonary hypertension, a known effect of severe heart failure . As a result, these two pathogenic events may have potential for mutual exacerbation (Additional file 1: Figure S3). Thus, the interruption of this exacerbating feedback by aPC treatment may contribute to the improved survival rate (Additional file 1: Figure S3). The cross-talks between these two types of pathogenic regulations are interesting issues and worthy to be further investigated.
In addition to LT, B. anthracis releases other toxins that perturb coagulation and vascular processes during infection. Bacteria-derived metalloproteases have been shown to degrade von Willebrand factor and ADAMTS13, which contribute to the recruitment of platelets to the injured vessel wall . In addition, bacterial protease InhA has been shown to inhibit fibrinolysis by activating plasminogen activator inhibitor-1 [61–63], and anthrolysin O, a cholesterol-dependent cytolysin and a Toll-like receptor 4 agonist, has been shown to disrupt endothelial and epithelial barriers [64–66]. Combined treatments of anthrolysin O and edema toxin significantly induced thrombin activity in human umbilical vein endothelial cells . These studies indicate that B. anthracis-derived secretory factors play important roles in anthrax consumptive coagulopathy. Further investigations are needed to reveal the combined effects of these virulence factors on circulatory homeostasis during infection.