Journal of Thrombosis and Thrombolysis

, Volume 35, Issue 3, pp 342–345

Bench to bedside: new developments in our understanding of the pathophysiology of thrombosis


    • Departments of Medicine and Pathology and Laboratory MedicineUniversity of North Carolina

DOI: 10.1007/s11239-013-0898-8

Cite this article as:
Key, N.S. J Thromb Thrombolysis (2013) 35: 342. doi:10.1007/s11239-013-0898-8


There has been a resurgent interest in the molecular and cellular mechanisms of thrombogenesis in venous thromboembolism (VTE). Improved animal models of VTE, combined with high-resolution microscopy techniques, have helped to uncover novel roles for blood cells including platelets and innate immune cells, particularly neutrophils. These insights are likely to result in novel disease biomarkers and perhaps even adjunctive anti-thrombotic therapies.


Venous thrombosisNeutrophil extracellular trapsVenous valve sinusPolyphosphates


In 1856, the German pathologist and statesman Rudolf Virchow proposed that thrombosis is a multifactorial event caused by some combination of changes in (1) blood flow; (2) the state of the vessel wall; and (3) the composition of blood. More than 150 years later, Virchow’s triad is still believed to encompass the major alterations that underlie arterial or venous thrombosis. While perturbations in the normal high shear stress laminar flow in arteries and structural abnormalities in the vessel wall (particularly related to atherosclerosis) dominate the patho-mechanism of arterial thrombosis, abnormalities in the plasma clotting factors have received the most attention in venous thromboembolism (VTE). Through the power of large case–control studies, the major familial and acquired risk factors for VTE have been defined [1]. One appeal of these ‘potent’ inherited thrombophilias to the clinician is that with even a basic understanding of coagulation biochemistry, it can be appreciated how each defect (i.e. deficiencies of antithrombin, protein C or protein S, factor V Leiden and prothrombin G20210A mutations) contributes to excess thrombin generation, and thereby to an enhanced thrombotic risk.

More recently, large-scale genetic studies have begun to identify thrombosis-associated single nucleotide polymorphisms (SNPs) within the human genome. Although in isolation these SNPs are usually very weak predictors of VTE, in combinations (of five SNPs or more), they may have clinically meaningful predictive value for VTE [2]. Frequently, the pathophysiologic mechanism whereby individual SNPs contribute to the risk of thrombosis is unknown. While intellectually less satisfactory, this limitation does not necessarily detract from their value as risk biomarkers.

In contrast to the remarkable progress just cited, it can be argued that sparse attention has been paid to the study of the interactions between cellular components of blood and the vessel wall in the pathogenesis of VTE. Until recently, the National Institutes of Health (NIH) definition of ‘translational research’ emphasized a uni-directional transfer of knowledge from the bench to the bedside [3]. However, as applied to VTE, this paradigm is inherently limited by the validity of animal models of venous thrombosis. In particular, in murine models of VTE, thrombosis is most frequently initiated by some chemical or physical injury to the vessel wall (e.g. application of ferric chloride or laser), complete ligation of the inferior vena cava, or in the case of pulmonary embolism, systemic injection of a procoagulant such as tissue factor or collagen [4, 5]. In all cases, the relevance of these human models to human VTE can be questioned, since venous thrombi generally occur on intact and grossly normal-appearing endothelial cells.

Venous thromboembolism instead lends itself to the revised bi-directional concept of translational research that is now espoused by the NIH [3]. In this paradigm, poorly understood clinical observations are used to frame the relevant research questions in a ‘bedside to bench’ fashion. For example, with respect to VTE, one might ask the following questions: [a] why do some venous thrombi present initially with deep vein thrombosis (DVT), while others present with pulmonary embolism?; [b] why is venous thrombosis, as typified by superficial thrombophlebitis, associated with the classic signs of acute inflammation (pain, heat, redness, swelling), and why are these symptoms often promptly relieved by heparin?; [c] why are venous thrombi ‘red clots’…..are erythrocytes merely trapped in the clot as passive bystanders?; and [d] if platelets have no role to play in VTE, why is there increasing evidence that aspirin may be beneficial in VTE prevention [68]?

In this brief review, I would like to highlight some recent basic research that is beginning to address some of these intriguing conundrums.

Initiation of lower extremity DVT

The genesis of thrombi in the venous valvular sinuses has classically been attributed to stasis and vortices of blood in this location. It has been assumed that stasis would prevent the efflux and clearance of activated clotting factors from the sinus, which would be sufficient to initiate thrombosis. While this may be part of the story, stasis is also associated with exaggerated hypoxemia, a fact that was pointed out more than 30 years ago by direct measurement of pO2 in the blood within the venous valves in the hind limbs of immobilized dogs and humans [9]. It has been demonstrated that mice exposed to ambient hypoxia (6 % oxygen) develop thrombosis in the vasculature of the lungs [10]. Thrombosis triggered by this systemic hypoxemia is mediated by tissue factor induction in macrophages following activation of the transcription factor early-growth-response (Egr-1) gene [11]. Egr-1, along with HIF-1 (hypoxia inducible factor), is also an important transcription factor mediating adaptive responses to hypoxia in endothelium [12]; in vitro, these responses include proinflammatory and pro-adhesive changes [13]. However, until recently, little attention was paid to the possibility that localized hypoxemia within valve sinuses could trigger local activation of endothelial cells to initiate venous thrombosis [14]. Evidence in support of this hypothesis includes a recent morphologic study of the pro-/anti-coagulant phenotype of endothelial cells from various locations within and around venous valve sinuses in explanted human veins. Endothelial cells within the valve sinus—which were exposed to the most profound hypoxemia in vivo—demonstrated a thromboresistant phenotype compared to endothelium further removed from the sinus [15]. This observation suggests that vascular endothelial cells adapt to local hypoxemia in a manner that helps to maintain blood in a fluid, anticoagulated state during health [16]. However, the molecular and cellular events by which impaired venous blood flow evolves into thrombogenesis still require elucidation.

Cellular mechanisms of venous thrombosis

Recent discoveries in the biology of innate immune cells have provided important insights into a novel role for these cells in the pathogenesis of venous thrombosis. The involvement of these inflammatory cells is compatible with the observation that systemic inflammatory disorders are associated with VTE, as well as the fact that acute venous thrombosis is frequently associated with the clinical symptoms and signs of acute inflammation.

Neutrophil extracellular traps (NETs), first described in 2004 [17], are comprised of DNA, histones, and antimicrobial proteins (such as elastase), and are released as a defensive mechanism by activated neutrophils in response to pathogens. This process, known as NETosis, complements other defensive mechanisms such as phagocytosis, and the production of oxidants and cytotoxic granule proteins and peptides [18]. In this situation, NETs serve to trap microbes in their adhesive web-like scaffold, where they are killed by locally-concentrated antimicrobial effector molecules.

It is generally believed that tissue factor expressed by circulating monocytes and monocyte-derived microparticles can trigger coagulation during pathological thrombosis based on experimental (mouse) models [19, 20]. However, using the currently available methods for measuring microparticle tissue factor (TF) in human blood, it remains an open question whether microparticle TF plays an important role in VTE [21]. More recently, NETs have been shown to be a second mechanism whereby innate inflammatory cells can activate coagulation in the pathogenesis of venous thrombosis. In experimental murine models of DVT induced by partial occlusion/stenosis of the inferior vena cava, the release of endothelial von Willebrand factor leads to the recruitment of platelets [22]. P-selectin expression on activated endothelium and platelets then mediates the recruitment and binding of neutrophils [23, 24]. NETs released by these cells as a result of platelet-induced NETosis promote further platelet as well as red cell binding. NETs also promote thrombin generation via activation of factor XII, thereby implicating the ‘intrinsic’ coagulation pathway in thrombus propagation [25]. Depletion of platelets or neutrophils—or the genetic absence of P-selectin—is protective against the development of thrombosis in this model. Similarly, systemic administration of DNase I, which degrades the chromatin backbone of NETs, prevents DVT formation in these mice [25]. Activated platelets are well known to contribute to thrombin generation through the exposure of negatively charged phospholipids that are essential for assembly of coagulation enzymatic complexes. More recently, it has been demonstrated that activated platelets also release polyphosphates that provide a co-factor for activation of factor XI by thrombin [26, 27]. And finally, factor XII deficient mice appear to be protected from both arterial and venous thrombosis, although they do not have any hemorrhagic phenotype [28].

In aggregate, these studies point to a novel role for innate immune cells in the development of venous thrombosis. These functions of neutrophils and monocytes in the development of venous thrombosis are distinct from their participation in the later resolution phase, the success of which may determine whether post thrombotic syndrome develops or not [29].

Evidence for role of NETs in VTE in humans

Nucleosomes—consisting of a DNA segment coiled around four histone protein cores—are derived from released NETs, and may be detected in the circulation. A recent case–control study demonstrated that plasma levels of nucleosomes and complexes of elastase-α1-antitrypsin were associated with a three fold increased risk of DVT, with an apparent dose–response relationship [30]. If confirmed by other groups, this study supports an association between neutrophil activation and venous thrombosis in humans, even in the absence of a systemic inflammatory disorder.


Basic research into the molecular and cellular mechanisms of venous thrombosis has enjoyed significant advances in recent years. It remains to be proven whether some of the questions posed at the outset of this brief review can be answered by appropriately designed experimental systems. For example, is the possible efficacy of aspirin in VTE prevention due to the inhibition of platelet trapping within NETs? Furthermore, it will be of interest to determine whether erythrocytes are anchored to venous thrombi exclusively through attachment to NETs, or possibly through additional mechanisms. Contrary to popular belief that red cells are merely trapped in the clot, recent data have shown that they may contribute to thrombin generation through the exposure of cell surface phosphatidylserine, analogous to activated platelets [31]. Finally, the dilemma of why some venous thrombi embolize might be a function of nucleosome-mediated stabilization of the thrombus through factor XII activation. A greater understanding of these mechanisms could lead to the development of novel diagnostic and/or prognostic biomarker(s) for VTE, as well as targeted non-anticoagulant based therapeutic adjunctive strategies, such as deoxyribonucleases [25], or inhibitors of factors XI, XII [32] or platelet polyphosphate [33], that may be associated with fewer bleeding risks compared to existing anticoagulant therapies.


The author acknowledges support from NIH grant HL095096.

Copyright information

© Springer Science+Business Media New York 2013