The study in this issue by Brill et al.  describing the role of activated neutrophil-derived DNA and associated nuclear proteins (termed neutrophil extracellular traps [NETs]) in an experimental model of deep vein thrombosis (DVT) is intriguing both mechanistically and in terms of implications for human disease prevention or treatment. Since the discovery of NETs as a rapid response to microbial stimulation , their impact on the hemostatic system has been studied in increasing detail . Nevertheless, new studies of the functional effect of NETs in vivo continue to expand upon extracellular roles for DNA, and in doing so renew interest in the hemostatic role of the leukocyte.
Postmitotic neutrophils retain their DNA in a condensed form, and, in an innate immune system response to infection/inflammation, neutrophils and platelets are induced to eject their chromatin, producing NETs consisting of web-like DNA–histone structures decorated with granule proteins, including proteases and cytotoxic mediators, as well as procoagulant tissue factor [2,3]. These highly electrostatically charged adhesive structures minimize the dispersion of bacteria and other invaders, trigger intrinsic coagulation, and destroy the pathogens. Using a murine model of DVT, Brill et al.  demonstrate that the venous thrombi formed in these mice are rich in citrullinated histone H3, and that intravenous administration of either histone preparations or DNase I (which degrades DNA), respectively, exacerbates or protects mice from the onset of DVT. Although the biochemical mechanisms underpinning these observations of early neutrophil activity still need to be established in detail, this work raises the intriguing possibility of targeting NET release in the clinical setting as a novel strategy to prevent venous thrombosis.
Venous thromboembolism (VTE), comprising DVT and pulmonary embolism, is a major clinical problem for hospitalized patients and those undergoing surgery. Transient risk factors include immobility, venous stasis, obesity, inflammation, and active malignancy, often enhanced by inherited prothrombotic states. Current treatment for VTE is based on preventing further activation of coagulation factors through the use of heparin-derived anticoagulants and vitamin K antagonists (warfarin). Such therapies do not address the contributions of endothelial and leukocyte activation to VTE formation. If neutrophil activation and associated NET formation are important in the early stages of clinical VTE development, then these events could potentially be targeted in at-risk patients. It is recognized that venous and arterial thrombi form differently, with a more prominent role for platelets in the arterial system, forming a platelet-rich ‘white’ thrombus, than in the venous system, where, although platelets contribute, erythrocytes and leukocytes dominate, forming ‘red’ thrombus. This relative contribution of platelets in arterial vs. venous thrombosis is relevant to the use of antiplatelet drugs, for example aspirin, which is indicated for arterial rather than venous thrombosis. In the study of Brill et al., histones and DNA were limited to the ‘red’ or erythrocyte-rich regions of the venous thrombus.
Under arterial conditions, pathologic shear stress (e.g. resulting from arterial stenosis) or atherosclerotic plaque rupture exposing extravascular matrix can lead to rapid platelet adhesion, activation, secretion and platelet-mediated coagulation (exposure of platelet-surface phospholipids, or secretion of procoagulant factors that activate intrinsic/extrinsic clotting under arterial shear conditions). The platelet von Willebrand factor (VWF) receptor, glycoprotein (GP)Ibα (of the GPIb–IX–V complex), is critical for thrombus formation under conditions of high physiologic or pathologic shear in flowing blood. GPIbα, which contains a highly anionic sulfated tyrosine sequence within the ligand-binding domain, interacts with large VWF multimers on activated endothelium, in plasma or subendothelial matrix, as well as with thrombin and other ligands, including leukocyte integrin αMβ2 and endothelial P-selectin. In venous thrombus formation, stasis or disturbance of blood flow and activation of endothelial cells is thought to initiate tissue factor-dependent thrombin formation, leading to a fibrin clot. In sepsis or infectious disease, pathogens are also strongly linked to the activation of coagulation factors and platelets, and thrombosis can also result from off-target effects of some drugs, including heparin (where misguided host defense against microorganisms can also play a key role [4,5]). Much clinical interest is centered on improving strategies for prophylactic therapy in individuals at risk of VTE, with a particular focus on abrogating the early initiating events of VTE. Studies like these of Brill et al. , pointing to leukocytes as early activators, could help to clarify the sequence and timing of steps leading to thrombus formation (Fig. 1), enabling the identification of new diagnostic or treatment options.
The mainstay of initial anticoagulation in VTE comprises heparin, a strongly negatively charged polysaccharide, and its low molecular weight derivatives. Unfractionated heparin is known to have many other activities besides its anticoagulant effects on thrombin and factor Xa; these may include binding to cationic histones and extracellular DNA following NET formation . Newer, highly specific anticoagulants have been developed that inhibit either thrombin or FXa directly and data on bleeding events and the recurrence of VTE in patients being treated with these new generation anticoagulants indicate that these drugs are viable treatment options . Although these agents appear to be effective in the treatment and prevention of VTE, they do not share the negative charge properties of the heparinoids, and may have differing activity against NET-mediated events in the vasculature. What is needed, however, is a far better understanding of the biochemistry of the key interactions underpinning these links between leukocyte activation and NET formation and coagulation/thrombus formation. In this regard, the direct interaction between histone and the GPIbα-binding domain of human VWF was initially thought ‘unlikely to be physiological’ . Later studies reported prothrombotic effects of exogenous histone, and histone colocalization with VWF in murine models of DVT [1,9]; whether blocking VWF–histone is protective is unclear.
Finally, neutrophils make up more than 70% of the circulating population of leukocytes, and during an inflammatory response the number of neutrophils released may reach 1012 per day. Studies from several laboratories have now indicated that neutrophil activation and NET formation are early events in DVT and, along with neutrophil-derived tissue factor, vascular endothelial growth factor, interferons, and tissue necrosis factor-α, NETs are likely to contribute to the onset of thrombosis [1–3]. Indeed, even after adjustment for underlying clinical covariates, including inflammation and thrombocytosis, leukocytosis is strongly associated with an increased risk of VTE and mortality in cancer patients receiving systemic chemotherapy . Leukocytosis also correlated with increased thrombotic risk and decreased survival in at-risk patients with essential thrombocythemia .
In conclusion, although heparin is unlikely to be excluded from the front line of therapeutic anticoagulants any time soon, the anti-NET benefits could negate some of the contributory effects of NETs to DVT in patients, and could be worthwhile evaluating for new-generation direct thrombin/FXa inhibitors.