Role of the blood coagulation cascade in hepatic fibrosis

Abstract

Liver is the primary source of numerous proteins that are critical for normal function of the blood coagulation cascade. Because of this, diseases of the liver, particularly when affiliated with severe complications like cirrhosis, are associated with abnormalities of blood clotting. Although conventional interpretation has inferred cirrhosis as a disorder of uniform bleeding risk, it is now increasingly appreciated as a disease wherein the coagulation cascade is precariously rebalanced. Moreover, prothrombotic risk factors are also associated with a more rapid progression of fibrosis in humans, suggesting that coagulation proteases participate in disease pathogenesis. Indeed, strong evidence drawn from experimental animal studies indicates that components of the coagulation cascade, particularly coagulation factor Xa and thrombin, drive profibrogenic events, leading to hepatic fibrosis. Here, we concisely review the evidence supporting a pathologic role for coagulation in the development of liver fibrosis and the potential mechanisms involved. Further, we highlight how studies in experimental animals may shed light on emerging clinical evidence, suggesting that beneficial effects of anticoagulation could extend beyond preventing thrombotic complications to include reducing pathologies like fibrosis.

HEMOSTATIC SYSTEM FUNCTION IN PATIENTS WITH LIVER DISEASE

Viral hepatitis, alcoholic and nonalcoholic fatty liver disease, and various autoimmune conditions remain major causes of liver-related morbidity and mortality (21). If liver damage driven by these conditions is not met by adequate liver repair, chronic liver injury leads to activation of hepatic stellate cells (and portal fibroblasts), resulting in a replacement of healthy hepatic parenchyma with excess extracellular matrix, including collagen (21). Excessive deposition of extracellular matrix is termed liver fibrosis, which can advance to a severe form termed cirrhosis. Cirrhosis is associated with a loss of hepatic function and can lead to liver failure and liver cancer. Treatment and prevention of cirrhosis in patients with liver disease are essential, because once cirrhosis develops, liver transplantation is often the only life-saving option.

Among its various functions, the liver is tasked with synthesis of numerous proteins involved in the coagulation cascade, the plasma pathway responsible for blood clot formation. The coagulation system is a complex series of serine proteases activated by tissue factor, a transmembrane receptor for coagulation factor VIIa (51, 62). The pinnacle coagulation protease is thrombin (factor IIa), which proteolytically enables conversion of plasma fibrinogen to insoluble fibrin clots. Fibrin clot formation is counterbalanced by a process called fibrinolysis. The tissue- and urokinase- plasminogen activators convert plasminogen to plasmin, the latter capable of degrading fibrin (52). Because the liver is so central in this process, patients with liver fibrosis and cirrhosis display profound changes in the components of the coagulation cascade. Levels of procoagulant proteins, such as factors II, V, VII, X, and others are decreased in patients with hepatic fibrosis and cirrhosis (47, 76). This decrease in procoagulant factors in patients with cirrhosis is counterbalanced by a concomitant decrease in anticoagulant proteins, such as protein C, protein S, and antithrombin. Similarly, thrombocytopenia (i.e., low platelet count) and thrombocytopathy (i.e., platelet dysfunction) in cirrhotic patients are rebalanced by compensatory elevation in the levels of the platelet-adhesive protein von Willebrand factor (VWF) and a decrease in the VWF-cleaving protease, ADAMTS13 (45, 47). Likewise, a decrease in antifibrinolytic factors is rebalanced by the concomitant decrease in profibrinolytic factors (49). The functional consequence of these changes is a rebalanced hemostatic system with normal to increased thrombin generation (49, 76, 77). In patients with compensated cirrhosis, wherein remaining parenchymal cells effectively manage hepatic functions, resulting in largely asymptomatic cirrhosis with few symptoms, the hemostatic system usually remains in this rebalanced state. However, in a decompensated state, wherein concurrent disease challenges cause remaining hepatic function to fail, the hemostatic rebalance can tip easily toward either bleeding or thrombosis (49, 76). Therefore, although patients may suffer from spontaneous or procedural bleeding episodes (58), cirrhosis is also associated with hypercoagulation, and patients are at a higher risk for venous thromboembolism, portal vein thrombosis (PVT), and pulmonary embolism (61, 70, 76).

COAGULATION FACTORS CONTRIBUTE TO FIBROSIS

Although changes in coagulation parameters are often observed as a clinical manifestation of liver cirrhosis, studies suggest that coagulation factors themselves, play important roles in the progression of hepatic fibrosis and/or cirrhosis (8). Progression of fibrosis is faster in hepatitis patients who have elevated plasma prothrombin levels, resulting from a single nucleotide polymorphism in the prothrombin gene, and patients who have increased production of thrombin resulting from a single nucleotide polymorphism in the factor V Leiden (53, 63, 80). Similarly, hepatitis C infection-related liver disease and fibrosis is less severe in patients with hemophilia, and elevated plasma VWF levels are associated with poor disease prognosis in patients with hepatic cirrhosis (6, 18, 27, 54). These studies suggest that components of the coagulation cascade, especially the prothrombotic factors, are associated with advanced progression of fibrosis in patients with chronic liver diseases. One of the suggested mechanisms behind this is the formation of intrahepatic microthrombi, which could cause ischemia and disrupt localized blood flow, ultimately leading to replacement of hepatic parenchymal cells with fibrous tissue. This and other mechanisms have been the subject of intense study in experimental settings of liver fibrosis. Thus, while liver disease, particularly cirrhosis, drives marked changes in the coagulation system, there is also evidence that coagulation activity can also worsen the progression of liver fibrosis. In the following section, we discuss examples of how experimental settings of fibrosis, primarily rodent models, have revealed the mechanisms whereby coagulation proteases contribute to liver fibrosis (Fig. 1).

Fig. 1.

Mechanisms of coagulation-directed liver fibrosis discovered in experimental settings of chronic liver injury. Chronic liver injury elicited by various insults is associated with activation of the blood coagulation cascade in experimental settings of liver fibrosis. This leads to generation of the coagulation proteases, coagulation factor Xa, and thrombin. Anticoagulants, including direct inhibitors of factor Xa and thrombin, have been shown to reduce liver fibrosis in animal models. Mice harboring prothrombotic risk factors comparable to those in humans (e.g., factor V Leiden) also develop more severe fibrosis. Studies with mice lacking protease-activated receptors (PARs) strongly suggest that PAR-1 (activated by thrombin) and PAR-2 (activated by factor Xa) contribute to hepatic fibrosis. Therefore, inhibition of PAR-1 and PAR-2 could be a novel strategy to reduce hepatic fibrosis. Fibrin(ogen) and fibrin clots may also contribute to hepatic fibrosis, but published evidence supporting this pathway is very limited. Rather, in specific experimental contexts, such as biliary injury, published studies suggest fibrin(ogen) can inhibit hepatic fibrosis.

ANTICOAGULATION AND ITS EFFECT ON EXPERIMENTAL LIVER FIBROSIS

Liver fibrosis can be induced in rodent models by inducing chronic liver injury through multiple approaches. Hepatotoxic xenobiotics, including carbon tetrachloride (CCl₄), thioacetamide, α-naphthylisothiocyanate (ANIT), dimethyl- or diethyl-nitrosamine, and tetrachlorodibenzo-p-dioxin have been used as the source of this chronic insult (35, 44, 55, 59, 60). Additionally, obstructive cholestasis, modeled in animals by bile duct ligation (BDL), also causes liver fibrosis (13, 75). Intrahepatic activation of the coagulation cascade is evident in these experimental settings (41). Moreover, numerous studies have documented a pathologic role for coagulation proteases in the pathogenesis of liver fibrosis. For example, treatment with warfarin, an anticoagulant that blocks the production of coagulation factors by the liver, reduced liver fibrosis in CCl₄-challenged mice, as well as in a mouse model of liver injury induced by chronic hepatic congestion (5, 68). Administration of anticoagulant low-molecular weight heparins (LMWHs) reduced hepatic fibrosis in rats induced by chronic CCl₄ administration or by BDL (1, 2). Similarly, the thrombin inhibitor SSR182289 reduced liver fibrosis in rats following chronic CCl₄ exposure (16). In addition, the direct oral anticoagulants (DOACs) rivaroxaban and dabigatran, novel small-molecule inhibitors of either coagulation factor Xa or thrombin, respectively, have been shown to reduce experimental liver fibrosis in rats (26, 78). Consistent with protective effects of reducing coagulation, mice harboring the procoagulant factor V Leiden mutation, which are genetically prone to increased coagulation activity, developed more fibrosis than wild-type counterparts, mirroring observations in humans (5, 80). Collectively, these studies indicate an important contribution of coagulation protease activity, particularly factor Xa and thrombin, to the pathogenesis of liver fibrosis.

The field emphasizes two primary mechanisms, whereby coagulation proteases could contribute to liver fibrosis. Intrahepatic fibrin clot deposition in the form of microthrombi could block blood flow, leading to hypoxia and cellular injury. Coagulation proteases could also promote liver fibrosis by direct activation of hepatic stellate cells, or indirectly by promoting local inflammatory cell activity. We briefly review the evidence supporting each hypothesis below, primarily from experimental studies.

FIBRIN(OGEN) AND HEPATIC FIBROSIS

Hepatic fibrin(ogen) deposits are a conspicuous feature of models of liver fibrosis (41). In principle, fibrin(ogen) deposits could occlude the hepatic sinusoids, disrupting blood flow, serve as a provisional matrix driving fibrosis, or act as a proinflammatory trigger advancing liver disease. The suggestion that fibrin(ogen) deposits drive hepatic fibrosis is anchored in the observed reduction in hepatic fibrin(ogen) upon administration of anticoagulants, which also reduce fibrosis. However, at present, there is a lack of definitive experimental evidence indicating that fibrin(ogen) contributes to liver fibrosis. In fact, the role of fibrin(ogen) in experimental fibrosis may depend heavily on the context of insult driving the chronic liver injury. For example, complete genetic deficiency in fibrin(ogen) did not affect gross liver histopathology in mice chronically exposed to CCl₄ (7). Interestingly, in the setting of ANIT-induced biliary injury and fibrosis, several studies have shown that fibrin(ogen) exerts protective effects through activation of integrin receptors on platelets and leukocytes (34–36, 50). Thus, at present, there is no definitive evidence illustrating a unified pathologic role for fibrin(ogen) in liver fibrosis. Fundamental differences in the cause or severity of disease in experimental and clinical settings may ultimately dictate fibrin(ogen) function. Much remains to be learned about the role of fibrin(ogen) in experimental and clinical settings of liver injury, repair, and fibrosis.

PROTEASE-ACTIVATED RECEPTORS SIGNALING AND HEPATIC FIBROSIS

Coagulation proteases trigger cell activation through cleavage of the G protein-coupled protease-activated receptors (PARs), which consist of four family members, PAR-1, PAR-2, PAR-3, and PAR-4 (3, 11, 12, 25, 31). Proteolytic cleavage of an extracellular NH₂-terminus of PARs exposes unique tethered ligands, which stimulate distinct intracellular signaling pathways (71, 83). PARs are implicated in a large number of processes and diseases, including hemostasis, thrombosis, inflammation, and cancer progression (3, 25, 83). PAR-1, the primary receptor for thrombin, and PAR-2, a receptor for numerous proteases, including factor Xa, are expressed by hepatic stellate cells (22). Thrombin and factor Xa, as well as selective peptide agonists for these receptors, have been shown to promote a profibrogenic phenotype in cultured stellate cells (19, 22). Liver fibrosis was reduced in PAR-1^−/− mice in multiple experimental settings, and administration of a PAR-1 antagonist reduced fibrosis in BDL rats (19, 60, 66, 73). These studies indicate that PAR-1 contributes to experimental liver fibrosis. Although direct activation of PAR-1 or PAR-2 on stellate cells is a plausible mechanism, this has not been unequivocally demonstrated in vivo, and it is possible that PAR-1 expressed by inflammatory cells also contributes to fibrosis by exacerbating inflammation (38). In addition to PAR-1, several studies indicate that PAR-2 also contributes to experimental liver fibrosis (40, 67, 74). Collectively, there is strong experimental evidence to indicate that coagulation proteases can drive hepatic fibrosis through activation of PARs. Although not the primary focus of this review, it is important to note that PARs can be activated by diverse proteases, such as cathepsins and neutrophil elastase (17, 57, 65, 82), widening the pool of potential effectors of fibrosis through PAR activation. The signaling pathways activated by PARs are also heavily dependent on the activating protease, illustrating a need to determine the precise intracellular pathways driven by coagulation protease activation of PARs, particularly as new antagonists are developed (24, 83).

On the basis of ample evidence linking PAR activation by the coagulation cascade to stellate cell activation and fibrosis, there may be therapeutic utility of PAR antagonists in the treatment of liver fibrosis. For example, the U.S. Food and Drug Administration-approved PAR-1 antagonist vorapaxar could be considered, but the risk of bleeding may outweigh potential benefit (10, 20). Unlike in mice, PAR-1 is the primary receptor for thrombin on human platelets (37, 72). However, it is worth noting that beneficial effects of other antiplatelet drugs have been documented in both experimental and clinical settings of liver fibrosis (33, 48, 69, 81). New strategies to target PAR-1, including pepducins and parmodulins, may also offer novel opportunities to treat liver fibrosis (4, 23). On the basis of mechanistic studies, PAR-2 is also an attractive target, particularly because factor Xa activation of PAR-2 could be blocked without affecting hemostasis. Emerging strategies to inhibit PAR-2, including pepducins and small-molecule antagonists are logical candidates to explore as antifibrotic agents (32, 67, 74).

ANTICOAGULATION OF CIRRHOTIC PATIENTS: BENEFITS BEYOND PREVENTION OF THROMBOSIS?

Coagulation disturbances featured in cirrhotic patients, and the potential risk of PVT, have prompted several studies evaluating anticoagulants as a preventative strategy to reduce PVT. This is perhaps the best highlight of how clinical and laboratory evidence has changed the perception of cirrhosis imposing an inherent bleeding risk. Interestingly, there is emerging evidence to suggest that anticoagulation might confer benefits beyond prevention of PVT in cirrhotic patients. For example, anticoagulation with enoxaparin, a LMWH, not only prevented development of PVT in cirrhotic patients, but also delayed hepatic decompensation and improved survival (79). Further, reduction in posttransplant fibrosis was observed in patients who were anticoagulated with warfarin (15). Particularly when viewed in the context of the strong connection between coagulation and fibrosis development in experimental animals, these studies suggest that anticoagulant treatment may, in fact, be a novel approach to reduce the progression of fibrosis or improve prognosis of patients with hepatic cirrhosis. However, much more research is required. Indeed, several clinical trials are addressing this possibility (NCT02643212, NCT02271295).

Emerging questions include which patients should receive anticoagulation, and with the availability of DOACs, which anticoagulant drug is best suited for this patient population (14, 46). Studies on warfarin and LMWHs, like dalteparin and nadroparin, have shown that these drugs have very few severe side effects and low hemorrhagic risk in cirrhotic patients (9, 39, 43). A newer class of DOACs, such as dabigatran and direct factor Xa inhibitors (apixaban, rivaroxaban, and edoxaban), which have been shown to be safe in cirrhotic patients, could be attractive candidates to study the effects of anticoagulation in fibrosis progression in patients with chronic liver diseases (29, 30, 56). We might also predict a greater efficacy of factor Xa inhibitors in fibrosis, as these drugs would more readily inhibit signaling elicited by PAR-2, as well as a secondary reduction in thrombin activity and downstream PAR-1 activation. However, using these clinically approved drugs in cirrhotic patients is complicated because impaired hepatic function in patients could potentially alter the anticoagulation potency, as well as pharmacokinetic/pharmacodynamic characteristics for these drugs (42, 64). Additionally, cirrhotic patients are frequently excluded from clinical trials investigating anticoagulants because of their precariously rebalanced coagulation status and bleeding risks (28). Therefore, additional studies are required to inform on type, dose, and treatment regimens for the existing anticoagulants in patients with cirrhosis.

CONCLUSIONS

Coagulation abnormalities detected by laboratory tests in patients with cirrhosis do not uniformly predispose them to a bleeding risk. In fact, a normal to hypercoagulable coagulative state is evident in cirrhotic patients. Interestingly, strong experimental evidence from animal models suggests that the coagulation cascade can also promote fibrosis, most likely through activation of PAR signaling and stellate cell activation. These experimental observations are mirrored by observations in humans that prothrombotic risk factors are associated with a more rapid progression of fibrosis, and consistent with emerging observations from clinical studies delineating the effect of anticoagulants on liver disease pathogenesis. Several new options for anticoagulation have emerged, concurrent to a deeper understanding of the complex hemostatic changes observed in cirrhotic patients. This combination of new knowledge and therapeutic options makes this an exciting time to consider the possibility that anticoagulant drugs, or newer agents targeting downstream pathways (e.g., PAR antagonists) discovered in experimental settings, could represent novel strategies to reduce liver fibrosis in patients.

GRANTS

This work was supported by the National Institutes of Health Grants R01-ES-017537 and R01-DK-105099.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.P., A.K.K., and J.P.L. prepared figures; A.P., A.K.K., and J.P.L. drafted manuscript; A.P., A.K.K., and J.P.L. edited and revised manuscript; A.P., A.K.K., and J.P.L. approved final version of manuscript.