2. From Microvascular Thrombosis to Hepatic Fibrosis
Studies on the consequences of acute murine hepatitis virus infection have provided the first evidence that coagulation is an important factor in the pathogenesis of liver disease. Indeed, numerous sinusoidal microthrombi directly related to hepatic parenchymal necrosis were described in these animal models, and microvascular thrombosis was associated with more severe hepatitis
[4,5,6][4][5][6].
In 1995, examining histological specimens from the livers of patients with chronic heart failure, Wanless et al. found a substantial correlation between sinusoidal fibrosis and the occurrence of local thrombotic events, indicating that liver fibrosis may be the result of microvascular thrombosis. More specifically, it has been proposed that disruption of blood flow by sinusoidal microthrombi and sinusoidal fibrosis cause reactive hyperemia and congestion, which activate fibroblasts and increase collagen deposition, worsening blood flow, inducing extension of thrombosis, neoangiogenesis, and parenchymal necrosis. The end result is the loss of hepatocytes and the formation of fibrous septa that completely alter the architecture of the liver
[7]. The replacement of liver parenchyma with fibrotic tissue as a result of microvascular disruption has been called parenchymal extinction
[8]. Inflammation is considered to be one of the main mechanisms inducing hepatic fibrogenesis in response to parenchymal damage
[9]; however, a preclinical model has shown that mild inflammation at both histological and serological levels is associated with the development of liver fibrosis due to chronic venous congestion
[10]. According to this study, sinusoidal thrombosis appears to be crucially involved in the direct potentiation of fibrogenesis in congestive liver disease.
On the other hand, hemostasis has also been firmly linked to the inflammatory response, and recently the concepts of immune-coagulation and thrombo-inflammation have been proposed
[11,12][11][12]. Several conditions, including ischemia-reperfusion syndrome and infections, can lead to the formation of microvascular thrombosis, which triggers an inflammatory response
[13]. As part of inflammatory processes, numerous cell types, including immune cells and fibroblasts, participate in the coagulation process, enhancing inflammation and increasing ECM deposition in different organs
[14]. In fact, hemostatic activation could be both the cause and the consequence of the inflammatory process.
However, there are few examples of clinical models of hepatic microvascular thrombosis in the literature, probably because microvascular thrombosis is rapidly replaced by fibrous tissue and is rarely detected on biopsy in patients with chronic liver disorders
[8]. Cases of systemic microvascular thrombotic disease in humans suddenly increased following the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic, prompting multiple studies, including some on liver pathology. SARS-CoV-2 causes COVID-19, which is a systemic infectious disease in which endothelial cell dysfunction and microvascular thrombosis are likely to play a major role in the development of multi-organ complications
[15]. Although the airways and lungs are the main organs affected by SARS-CoV-2, liver injury is a frequent condition in patients with the most severe clinical forms
[16,17][16][17]. Sinusoidal microthrombi were frequent in liver biopsy specimens taken from COVID-19 patients and were related to more severe liver damage
[18,19][18][19]. Mild portal and lobular inflammation, confluent parenchymal necrosis, and fibrosis are other histological findings
[18]. Because there is no evidence of direct cytopathic virus damage, the latter appears to be associated with the marked inflammatory activation produced by the infection
[20]; in fact, molecular analyses revealed that during severe COVID-19, genes frequently linked to hepatic stellate cells (HSCs) activation and liver fibrosis, such as interleukin 6 (IL6), interleukin 1 (IL1), tumor necrosis factor α (TNF-α), interleukin 10 (IL10), and interferon α (IFN-α), as well as vascular endothelial growth factor (VEGF) and monocyte chemoattractant protein 1 (MCP-1), are overexpressed
[21]. In addition, a higher noninvasive fibrosis score appears to be correlated with a higher risk of developing severe COVID-19
[22].
The fascinating link between sinusoidal thrombosis and the activation of molecular pathways of liver fibrosis in humans is supported by the beneficial effects of anticoagulants on the development of fibrosis
[23,24][23][24]. In line with the direction set out by Wanless et al. in 1995, developments in molecular medicine and the identification of new intra- and intercellular networks have strengthened this connection.
3. Hepatic Stellate Cells and Protease Activated Receptor
In response to repeated injury, HSCs can differentiate into myofibroblasts, which proliferate and produce ECM
[25]. In addition, HSCs stimulate other cell types that participate in the inflammatory response and fibrogenesis through the production of growth factors and chemokines, playing a critical role in driving the initiation and progression of the fibrogenic process
[26]. As expected, microvascular sinusoidal thrombosis, which leads to ischemic injury and inflammatory response, causes overexpression of VEGF, platelet-derived growth factor (PDGF) and transforming growth factor beta (TGFβ) by hepatocytes and HSCs, as well as increased synthesis of type I and type IV collagen by activated HSCs
[27]. However, a specific molecular mechanism has been proposed as responsible for the apparent association between microvascular thrombosis and fibrogenesis. In 1998, Marra et al. hypothesized that the protease-activated receptor (PAR) signaling pathway was a major molecular pathway involved in HSCs activation and hepatic scarring
[28]. PARs are G-protein-coupled receptors with proteolytic activity that stimulate cellular responses by interacting with coagulation factor Xa (FXa) and neutrophil elastase (PAR 1, 2), thrombin (PAR 1, 3, 4), coagulation factor VIIa (PAR 1), and tryptase (PAR 2, 4), which cleave the N-terminal of the receptor
[28]. PARs are expressed by different types of cells involved in the fine regulation of vascular homeostasis, and their signaling pathways are complex because they can be linked to G proteins with different functions. Consequently, they interact with a myriad of signaling transducers (e.g., extracellular signal-regulated kinase [ERK] 1/2, Rho/Rho-kinase, c-Jun N-terminal kinase, inositol 1,4,5-trisphosphate [IP3], phosphoinositide 3-kinases [PI3K], and Janus kinase/signal transducers and activators of transcription [JAK-STAT]), resulting in pleiotropic effects
[29,30,31][29][30][31]. PARs have been linked to the progression of fibrosis in several organs, and because they have a high affinity for factors in the coagulation cascade, they have been proposed as the main link between hemostasis and pulmonary fibrosis
[32,33][32][33] or renal fibrosis
[2]. In addition, PAR 1 induces cardiac fibroblast activation in response to thrombin or FXa and, by modulating the ERK1/2 pathway, leads to cardiac remodeling and fibrosis
[29,34][29][34].
PARs are found in liver cells and are abundantly overexpressed in chronic liver disease, as in the case of PAR 1 in the myofibroblast group of cirrhotic individuals
[35]. A preclinical rat liver stellate cell model revealed a progressive increase in PAR 1 and PAR 2 expression during transformation to a myofibroblastic phenotype
[36]. Increasing concentrations of thrombin transform HSCs into myofibroblasts increases the production of α-smooth muscle actin (α-SMA), pro-collagen, TGFβ-1, matrix metalloproteinase 2 (MMP-2), and other cellular signals essential for wound healing
[37,38][37][38]. Furthermore, when HSCs are treated with a combination of FXa and thrombin, there is an increase in α-SMA, procollagen, TGFβ-1, and significantly improved cell contraction compared with FXa or thrombin alone
[39].
Preclinical studies have confirmed the importance of PARs in fibrogenesis, showing how inhibition or lack of PAR 1 and PAR 2 reduces the evolution of liver fibrosis
[38,40][38][40].
However, this evidence about coagulation factor-mediated HSCs activation by PAR has been based mainly on in vitro research. Poole et al. recently studied the effect on the liver of chronic exposure to carbon tetrachloride (CCl4) in vivo, using a mouse model with PAR 1 deletion specific for HSCs. PAR 1 deletion was linked to decreased activation of HSCs and collagen deposition but was not protective against acute liver damage after CCl4 exposure
[41]. Thus, PAR 1 appears to play a role in the “healing process” that occurs after liver injury rather than in its acute phase. Indeed, enoxaparin treatment significantly reduced portal hypertension, hepatic fibrosis, HSCs activation, and desmin expression in mice with CCl4-induced cirrhosis without having any effect on the acute injury. In addition, molecular analysis revealed decreased hepatic fibrin deposition in enoxaparin-treated rats, implying the role of intrahepatic microthrombosis as a primary mechanism of PAR activation
[23].
Regarding the etiology of liver fibrosis, thrombin-mediated HSC activation appears to be closely related to the progression of nonalcoholic fatty liver disease (NAFLD)
[42]. In addition, administration of a direct thrombin inhibitor to mice with NAFLD reduces HSCs activation, α-SMA expression, and hepatic collagen type 1 mRNA levels
[42]. While thrombin has been shown to play a crucial role in the progression of NAFLD, a preclinical investigation suggested that PAR 1, its receptor, is essential for the development of hepatic steatosis in mice fed a Western diet
[43]. Nault et al. studied the involvement of PAR 1 signaling in liver damage caused by the contaminant 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). C57BL/6 mice exposed to TCDD acquire NAFLD-like features, such as steatosis, liver damage, inflammation, and fibrosis. Subchronic exposure to TCDD also causes increased intrahepatic coagulation, as reflected by increased thrombin production and deposition of fibrin and fibrinogen in the liver. As measured by serum alanine aminotransferase activity, PAR 1 deficiency had no effect on TCDD-induced hepatocellular damage and hepatic lipid accumulation but nevertheless was linked to a significant reduction in liver fibrosis and histologic evidence of inflammation
[44].
Although the function of microvascular thrombosis in the evolution of a chronic disease such as NAFLD seems well established, in contrast, the importance of PARs in the progression of chronic viral hepatitis to fibrosis has been less studied. One study examined the PAR 1 genotype in people with chronic HCV infection and found that a specific PAR 1 polymorphism (1426 C/T) was linked to histological evidence of increased liver fibrosis
[45]. However, the role of PAR 1 in the progression of viral hepatitis-related liver fibrosis is mainly unknown. The PAR signaling pathway of HSCs is currently believed to be the main element explaining the direct association between microvascular thrombosis and fibrogenesis in the liver parenchyma, although there may be variations depending on the etiology of liver disease. It is also possible to hypothesize a different role of PAR 1 expressed by different cell types; in fact, a number of non-parenchymal liver cells and other cells that infiltrate the damaged liver, including sinusoidal endothelial cells, inflammatory cells (monocytes, neutrophils, and lymphocytes), resident hepatic macrophages (Kupffer cells), and bile duct epithelial cells may express PAR 1
[46,47,48][46][47][48]. Therefore, since microvascular thrombosis may be a manifestation of a harmful hepatic stimulation, PAR 1 activation of HSCs may be the primary cause of the abnormal response leading to fibrosis and its progression. However, considering the wide variety of interactions that occur during hepatic fibrogenesis, it may not be the only one and may indeed be part of a more complex network of molecular and cellular pathways yet to be defined.
4. The Role of Liver Sinusoidal Endothelial Cells and Neutrophil Extracellular Traps
Another important cell type involved in the relationship between microvascular thrombosis and fibrosis progression is hepatic sinusoidal endothelial cells (LSECs). It is commonly recognized that LSECs interact with other cells, such as neutrophils, lymphocytes, HSCs, hepatocytes, and Kupffer cells, which is critical in the progression of nonalcoholic steatohepatitis (NASH) to fibrosis
[49]. Healthy endothelial cells, on the other hand, express molecules that inhibit platelet activation, coagulation, and thrombosis
[50]. Expression of pro- and anti-thrombotic elements changes when LSECs lose their antithrombotic phenotype due to endothelial dysfunction
[51,52][51][52]. The expression of thrombomodulin, nitric oxide, or prostaglandin I2 is attenuated in the presence of dysfunctional LSECs, which also expose Von Willebrand factor (VWF), integrins, and other receptors that interact with activated platelets and cause clot formation
[53,54][53][54]. It has also been shown that hepatitis viruses, including hepatitis B virus (HBV) and murine acute hepatitis virus 3, a member of the
Coronaviridae, can induce LSECs to overexpress Fgl2/fibroleukin prothrombinase, which is crucial for the initiation and progression of fibrin deposition
[55,56,57][55][56][57]. Recent histological and molecular studies in a mouse model of congestive hepatopathy demonstrated that LSECs exposed to mechanical stretch upregulate Notch-dependent transcription factors through an integrin-dependent pathway and interaction with the mechanosensitive piezo-calcium channel. As a result, LSECs increase the production of neutrophil chemoattractant C-X-C motif ligand (CXCL) 1, attracting platelets and neutrophils into hepatic sinusoids and leading to the formation of extracellular neutrophil traps (NETs). NETs are complexes consisting of a backbone of extracellular DNA fibers bound to histones and granular proteins, such as myeloperoxidase and neutrophil elastase, which are strongly linked to sinusoidal thrombosis
[58]. Interestingly, the progression of NASH has been linked to parenchymal neutrophil infiltration and NET development in mice
[59]. In addition, in biopsy samples of the lungs, liver, and kidneys from patients with severe COVID-19, NETs have been linked to microvascular thrombosis
[60,61][60][61]. As previously shown, when LSECs are dysfunctional, they lose their antithrombotic activity, promoting local thrombotic events and, consequently, the progression of PAR 1-induced fibrosis. In addition, LSECs themselves overexpress PAR 1, which is activated by the coagulation cascade in response to injury, as was demonstrated in a preclinical model of ischemia in mouse liver. In LSECs, PAR 1 activation blocks the ERK1/2 pathway and promotes apoptotic signaling, exacerbating liver damage and causing inflammation and fibrosis
[62]. Kruppel-like factor 2 (KFL2), a transcription factor, has recently been recognized as a crucial regulator of endothelium homeostasis in response to inflammatory stimuli, coagulation factors, and hemodynamic stresses such as laminar shear stress
[63,64,65][63][64][65]. In LSECs of cirrhotic mice, transcriptome analysis revealed downregulation of Kruppel-like factor (KLF) 2 and 4. In addition, Marrone et al. showed that overexpression of KLF2 in LSECs and HSCs derived from cirrhotic rats reduces HSC activation and enhances paracrine cross-talk between LSECs
[66,67][66][67]. This is consistent with the decrease in fibrosis and portal pressure associated with KFL2 overexpression in animal studies
[68]. Interestingly, KLF2 is primarily activated by the extracellular signal-regulated kinase 5 (MEK5)-extracellular signal-regulated kinase 5 (ERK5) pathway
[69], the overexpression of which has recently been linked to suppression of PAR 1 signaling in cell types such as pneumocytes and alveolar barrier endothelial cells
[70]. Further research revealed the function of the long noncoding RNA Airn in controlling KLF2. In fact, Airn interacts with subunit 2 of the Polycomb Enhancer Of Zeste 2 repressive complex to maintain the differentiation of LSECs through the KLF2 pathway, preventing the capillarization of LSECs, maintaining the quiescence of HSCs and attenuating the progression of fibrosis. Airn is highly expressed in the liver and serum of patients with fibrosis and in mouse fibrotic livers
[71].
5. ADAMTS 13—Von Willebrand Factor, a Bridge between Coagulation and Fibrosis
Liver cirrhosis has long been considered an acquired bleeding condition because of altered coagulation parameters. Indeed, there is a reduction in liver-related coagulation factors, such as Factors II, V, VII, IX, and XI, but levels of vitamin K-dependent anticoagulant proteins, protein C (PC), protein S, and antithrombin III are also reduced
[94,95][72][73]. However, it is now recognized that the risk of bleeding is not a consequence of coagulation imbalance but is related to portal hypertension, and that cirrhotic patients maintain a perfect hemostatic balance in most cases
[96][74]. In fact, studies by Tripodi et al. proved that the thrombin generation assay remains preserved in these patients despite these alterations
[97,98][75][76]. Increased levels of Factor VIII (FVIII) and VWF have also been observed in cirrhotic patients with portal hypertension, suggesting that a procoagulant milieu may prevail
[99,100][77][78].
These factors are produced and stored in the endothelial cells of portal and hepatic veins but not in hepatic sinusoids
[101][79]. However, in advanced stages of liver fibrosis, endothelial cells of hepatic sinusoids acquire the phenotype of vascular endothelium due to chronic inflammation and endotoxemia, which are the main contributors to portal hypertension
[102,103[80][81][82][83][84],
104,105,106], and begin to produce FVIII and VWF. VWF is a multimeric glycoprotein synthetized and released by vascular endothelial cells in the bloodstream in the form of multimers, with the ability to bind platelets proportionally to their size, demonstrating a pivotal role in hemostatic balance
[107][85]. It is stored in Weibel–Palade bodies (WPBs) and excreted upon stimulation. Some studies have shown a direct correlation between circulating FVIII levels and the severity of portal hypertension as measured by hepatic venous pressure gradient (HVPG) or the presence of ascites and the risk of variceal hemorrhage in patients with advanced liver disease
[102,103,104,105,106][80][81][82][83][84]. In addition, the association between FVIII alterations and liver fibrosis is well-known: in particular, FVIII and VWF have been observed in capillaries and pericellular regions alongside necrotic sites in the liver parenchyma, and inflammatory injury may promote the deposition of these factors along with fibrosis
[108][86]. A Disintegrin and Metalloproteinase with a Thrombospondin Type 1 motif, member 13 (ADAMTS-13), is another factor influenced by chronic inflammation and advanced liver disease. It is a metalloproteinase that cuts the multimeric VWF between Tyr1605 and Met1606 in its A2 domain. ADAMTS-13 is mainly produced in the liver by HSCs
[109[87][88][89][90][91],
110,111,112,113], and finely regulates hemostatic balance by controlling the size of VWF multimeters, thus their ability to aggregate platelets by forming microthrombi. In chronic liver disease, HSCs that acquire a myofibroblastic phenotype lose the ability to produce and store ADAMTS-13
[111,114,115,116][89][92][93][94].
Moreover, ADAMTS-13 is inversely correlated with the severity of liver dysfunction in terms of antigen production and activity, whereas VWF antigen and activity increase, as previously discussed, resulting in the release of high molecular weight VWF (HMWVWF) multimers
[117][95].
Therefore, ADAMTS-13 and VWF imbalance may have a pivotal role in the paradigm of parenchymal extinction and liver fibrosis progression in chronic liver diseases
[118,119][96][97]; indeed, the upregulation of VWF levels in the presence of chronic inflammation, vascular damage, or endotoxemia
[102,103][80][81] is not counterbalanced due to the deficiency of ADAMTS-13, promoting the formation of platelet microthrombi and fibrin deposition in hepatic sinusoids, with the loss of liver parenchyma and fibrogenesis
[120,121][98][99].
This condition is similar to thrombotic thrombocytopenic purpura, a primary or acquired clinical disorder induced by the loss of ADAMTS-13 activity or its deficiency that lead to microthrombi formation in small vessels such as glomeruli, cerebral vessels, and cutaneous capillaries
[115,121,122,123][93][99][100][101].