Hepatic Cell Cycle Activity in Initiation and Progression of Liver Fibrosis

Initiation of liver fibrosis is associated with cell cycle activity of several hepatic cell entities including hepatocytes, hepatic stellate cells, endothelial cells and others. This is plausible, considering that liver fibrosis is based on chronic liver damage and liver cell death. Dying hepatocytes are replaced through compensatory hepatocyte proliferation. In addition, hepatocyte cell death and tissue injury triggers activation and proliferation of hepatic stellate cells (HSCs) as well as other hepatic cell types. Activated HSCs transdifferentiate into myofibroblasts, which continue to proliferate and produce extracellular matrix (Figure 1), which is the basis for liver scarring with fibrogenesis and the risk of disease progression toward cirrhosis as well as toward hepatocellular carcinoma (HCC). Thus, hepatic proliferation is a key event during liver fibrosis initiation and targeting cell cycle activity in the liver could be a promising anti-fibrotic treatment approach.

One key for preventing liver fibrogenesis by targeting the cell cycle machinery will be the identification of cell cycle factors that are specifically important for driving liver fibrosis without affecting general liver homeostasis or liver regeneration. Basically, the mammalian cell cycle is regulated by specific combinations of cyclin-dependent kinases (CDKs) and their regulatory subunits referred to as cyclins. Importantly, many of the known cyclins and Cdks are dispensable for overall cell cycle activity presumably due to high functional redundancy between these proteins. For instance, we have shown that single genetic loss of Cyclin E1, Cyclin E2, or Cdk2 in mice does not prevent proper liver regeneration after partial hepatectomy, and even the concomitant genetic deletion of all three genes still allows sufficient liver mass restitution to prevent mortality of hepatectomized mice (2, 3). However, this does not exclude the possibility that any of these factors might perform essential, cell type specific functions in the context of disease development. In addition to cyclins and Cdks, an abundance of other mitogens (i.e., growth factors, transcription factors) is capable of modulating and/or controlling the cell cycle activity. The proto-oncogene c-myc is a transcription factor that activates expression of many pro-proliferative genes including E-type cyclins and Cdk2 (4). In a recent study, it has been observed that patients with advanced liver fibrosis or liver cirrhosis showed significant up-regulation of hepatic c-myc gene expression (5). These findings were mirrored in experimental mouse models. Briefly, mice with over-expression of c-myc in hepatocytes (alb-myc^tg) revealed increased basal liver collagen disposition and spontaneous HSC activation. Importantly, primary HSC derived from alb-myc^tg mice showed enhanced proliferation and accelerated transdifferentiation into myofibroblasts in vitro. Accordingly, fibrosis initiation in vivo after chronic carbon tetrachloride (CCl₄) treatment was accelerated in alb-myc^tg mice compared to controls. Mechanistically, some data pointing to a paracrine crosstalk of c-myc over-expressing hepatocytes and HSC as livers of alb-myc^tg mice revealed increased levels of PDGF-B. As a consequence, over-expression of c-myc in hepatocytes triggered accelerated experimental liver fibrogenesis.

The sequels of elevated hepatic c-myc expression were also evaluated in the clinical relevant setting of liver fibrogenesis due to alcoholic liver disease (ALD). It has been demonstrated that c-myc was induced in human and murine (i.e., experimental) ALD (6). Moreover, patient-derived data clearly showed a significant correlation of hepatic c-myc expression with the strength of ALD progression. Overall, expression of c-myc and alcohol-uptake synergistically accelerated the progression of ALD, and additional data strongly indicated that this was most likely due to a loss of p53-dependent protection mechanisms (6). Thus, c-myc can be considered a new potential marker for the early detection of ALD and for the identification of risk patients. Unfortunately, due to its complex functions in almost all cellular processes, c-myc is still considered an undruggable target. However, several approaches are currently under development which could help to modulate c-myc dependent effects in the future (7).

Another aspect of cell cycle regulation during liver fibrosis initiation has become evident in the last years by focussing on the pro-fibrotic role of E-type cyclins. Of note, Cyclin E/Cdk2 kinase was shown to be regulated by c-myc in several independent studies (8–10). In good agreement with the published investigations on the role of c-myc, increased expression of Cyclin E1—but not of Cyclin E2 has been shown in human and murine liver fibrosis (11). In humans, Cyclin E1 mRNA expression was significantly up-regulated in patients with advanced hepatic fibrosis and liver cirrhosis, compared to healthy control livers or patients with mild fibrosis. In contrast, Cyclin E2 was not aberrantly expressed in liver fibrosis at any stage. In mice, Cyclin E1, but not Cyclin E2 was induced in the liver after repeated treatment with the fibrosis-inducing toxin CCl₄, and correlated with the strength of fibrosis progression. Genetic experiments revealed that the constitutive inactivation of Cyclin E1 prevents the initiation at least of experimental liver fibrosis in mice. Further investigations showed that Cyclin E1 is essential for cell cycle progression, transdifferentiation and survival of HSCs suggesting that HSCs are a key effector cell of the pro-fibrogenic function of Cyclin E1. As Cyclin E1 is not required for liver regeneration (2), these findings suggest that its therapeutic inhibition could reduce liver fibrogenesis without affecting the regenerative capacity of the liver.

Advanced studies then addressed whether these key findings could be translated into a pre-clinical therapeutic approach. So far, pharmacological small-molecule inhibitors directly targeting Cyclin E1 have not been developed; however, there are several groups of inhibitors available targeting the kinase activity of Cdk2. Yet, it has to be mentioned that these inhibitors are not completely specific to Cdk2 but also mediate off-target effects on other Cdks such as Cdk1, Cdk5, 7, 9 and others (12). In order to treat experimental liver fibrosis by targeting Cyclin E/Cdk2, basically two different strategies have been used. For targeting Cyclin E1 in vivo, liposome-based delivery of Cyclin E1- specific small interfering RNA (siRNA) was applied, whereas for targeting Cdk2 in HSCs, the efficacy of the second-generation pan-Cdk inhibitor CR8 (13) was tested in a series of in vitro experiments. Importantly, both approaches turned out to be highly promising and will be briefly reviewed below.

In wild type (WT) mice, systemic delivery of stabilized siRNA, using a liposome-based carrier, targeted ~95% of HSCs, 70% of hepatocytes, and 40% of CD45-positive leukocytes after single injection, and the accumulation of siRNA after 24 h was mainly limited to the liver (14). Importantly, ubiquitous delivery of Cyclin E1 siRNA to all liver cells turned out to be strongly beneficial during acute toxic liver injury and in the course of hepatic fibrogenesis. Major side effects regarding toxicity or survival were not observed using this strategy. In more detail, it could be demonstrated that preventive systemic delivery of stabilized CcnE1-siRNA in vivo only once a week was sufficient to inhibit Cyclin E1 induction in settings of both, acute and chronic liver injury, and this was sufficient to prevent the initiation of experimental liver fibrosis. For the underlying mechanisms it has been suggested that the CcnE1-siRNA primarily prevents proliferation and survival of activated HSCs during chronic fibrosis progression thereby reducing the formation of extracellular matrix. In addition, it has been shown that CcnE1-siRNA reduces the proliferation and infiltration of pro-fibrotic leukocytes in the challenged liver and thereby may attenuate the overall inflammatory response after a pro-fibrotic stimulation.

However, in a real clinical situation, a patient would only be treated after the development of liver fibrosis. It is therefore important to note that Cyclin E1 siRNA was shown to still act anti-fibrotic, if it is administered after induction of liver fibrosis. In conclusion, this approach could basically be a promising treatment option for patients with liver fibrosis in the future (14). In this context it is very promising to note that currently two independent siRNA-based therapeutics have been approved by the U. S. Food and Drug Administration (FDA) (15, 16), which suggests that RNA-interference is a suitable technology for the treatment of diseases in humans.

The characterization of the pleiotropic Cdk inhibitor CR8 for its anti-fibrotic properties was so far restricted to in vitro analyses on established immortalized HSC lines, primary HSCs and primary hepatocytes (17). CR8 treatment of HSCs resulted in cell cycle arrest, apoptosis and down-regulation of pro-fibrotic genes. In contrast, it has been shown that hepatocytes tolerate substantially higher doses of CR8 than HSCs without prominent cytotoxic effects. Therefore, this proof-of-concept study demonstrated that pharmacological Cdk-inhibition restricts the pro-fibrotic properties of HSCs, while preserving the regeneration capacity of hepatocytes under the same conditions. Thus, CR8 and related drugs might be promising therapeutic agents for the treatment of liver fibrosis. However, it needs to be addressed in the future if other non-parenchymal liver cells such as Kupffer cells, infiltrating immune cells or biliary epithelial cells will also tolerate CR8 to the same extend as hepatocytes. In addition, this concept will need intensive in vivo validations in the future before pan-Cdk inhibitors can be considered for anti-fibrotic therapies.

The findings on the role of cell cycle activation in liver fibrogenesis are summarized in Figure 2: Initiation of liver fibrogenesis is associated with increased cell cycle activity involving the proto-oncogene c-myc and Cyclin E1. The latter one act in complex with its kinase subunit Cdk2. Overexpression of c-myc is observed in patients with liver fibrosis of several etiologies including ALD; yet it is considered undruggable at present due to its numerous functions. Cyclin E1 is indispensable for initiation of liver fibrosis and also essential for proliferation, differentiation and survival of HSCs. Direct targeting of Cyclin E1 at present is only feasible by using RNA interference, but was shown to act highly anti-fibrotic in mice without any notable side effects. It needs to be evaluated in future if this can be developed into a human siRNA therapeutic. Finally, anti-fibrotic effects in HSCs can also be triggered by the use of pharmacological inhibition of Cdk-activity without major effects on hepatocyte viability at least in vitro. However, before Cdk inhibitors can be considered for anti-fibrotic therapies in humans, an abundance of in vivo trials need to be performed. It is concluded that the modulation of the hepatic cell cycle activity could be general approach for the treatment of patients with moderate liver fibrosis.

Open in a separate window

Figure 2

Role of cell cycle mediators for liver fibrosis initiation. Initiation of liver fibrogenesis is associated with increased hepatic cell cycle activity in HSCs. Important cell cycle mediators in these processes are c-myc and the Cyclin-dependent kinase 2 (Cdk2) in complex with its regulatory subunit Cyclin E1. In general, c-myc is involved in activation of Cyclin E1/Cdk2. Moreover, over-expression of c-myc in hepatocytes triggers activation of HSCs and is observed in patients and mice with liver fibrosis. Cyclin E1 is indispensable for initiation of liver fibrosis and also essential for proliferation, differentiation and survival of HSCs. Direct targeting of Cyclin E1 is currently feasible by using RNA interference (siRNA) in vivo. Anti-fibrotic effects in HSCs can also be triggered by the use of pharmacological inhibition of Cdk-activity.

Programmed Cell Death in Non-alcoholic Steatohepatitis and Fibrosis

Non-alcoholic fatty liver disease (NAFLD), now the most prevalent liver disease in Western countries, represents an enormous threat to our health care systems. NASH, the chronic progressive manifestation of NAFLD, is the second leading risk factor for hepatocellular carcinoma (HCC), the most frequent form of liver cancer (18). Currently, there are no existing pharmacological treatment options for NASH, therefore the current treatment is primarily aimed at lifestyle change with more exercise and modification of food intake.

The occurrence of cell death—e.g., caused by lipotoxicity—represents a critical event in NASH. It triggers the recruitment of immune cells and consequently drives disease progression toward fibrosis and liver cancer development (19). How cell death and inflammation are associated with the malignant transformation of hepatocytes is currently poorly understood. In the past, apoptosis, for decades a synonym for programmed cell death, was considered the main driver for the development of NASH (20). However, recent work revealed that key molecules of necroptosis—the programmed form of necrosis—mediated by receptor interacting protein kinase (RIPK) 1, RIPK3 and mixed lineage kinase domain-like pseudokinase (MLKL)—are major contributors in the pathophysiology of NASH, fibrosis and liver cancer development (21–23). Necroptosis is a type of cell death whose features are similar to apoptosis and necrosis. However, it is initiated by ligand binding to the tumor necrosis factor receptor 1 (TNFR1), thereby forming a specific complex with capacity to initiate different downstream cascades in which the cells rupture and leak their content into the intercellular space (24). Necroptosis plays a vital role in different stages of liver disease including cholestatic liver disease, alcoholic liver disease, NASH, viral hepatitis, and liver cancer (19, 25).

As such, it was shown, that in livers of NASH patients, apoptosis is only negligibly activated while the necroptosis mediator RIPK3 is strongly overexpressed (26, 27). Of note, genetic inhibition of RIPK3 resulted in a very effective inhibition of liver fibrosis in a murine model of NASH [methionine/choline-deficient (MCD) diet], whereas apoptosis inhibition even increased disease progression (26, 28).

In contrast to the MCD model, Ripk3 deficient mice fed with a high fat diet (HFD), exhibited more steatosis, fibrosis, and inflammation compared to HFD-treated WT mice (29). Moreover, Ripk3 deficient mice showed an aggravation of systemic insulin resistance and increased compensatory adipocyte apoptosis in the choline deficient-high fat diet (CD-HFD) model (27). Interestingly, aggravated insulin resistance was reverted to WT conditions upon additional inhibition of apoptosis, probably in adipocytes (27). Thus, RIPK3 may have a protective, counterbalancing function in adipose tissue. This could explain the observed increased expression of RIPK3 in adipose tissue of obese patients (27). Consistent with RIPK3, increased expression of RIPK1 was found in adipose tissue of obese patients (30). Thus, genetic knock out of Ripk1 using antisense oligonucleotides in HFD treated wild type mice resulted in reduced diet-induced obesity and adipose tissue inflammation and improved insulin resistance. These protective effects were independent of a kinase activity, as RIPK1^K45A kinase dead knock in mice did not improve insulin resistance or reduce obesity on HFD feeding (30). However, Tao et al. observed less liver injury, less steatosis and decreased inflammation in HFD- and MCD-fed RIPK1^K45A mice (31). Further investigations revealed that RIPK1′ kinase activity in hematopoietic-derived macrophages contributed mostly to the disease progression in NASH (31).

Finally, the terminal executor of necroptosis, MLKL was found activated in livers of NASH patients (32). In contrast to RIPK3, recent studies showed protective effects in the previous mentioned high fat models upon genetic deletion of Mlkl. As such, Saeed et al. reported reduced hepatic steatosis and inflammation in Mlkl-deficient HFD-fed mice (33), as well as enhanced hepatic insulin sensitivity (34). A further study using the Western diet (FFC diet, high in fat, fructose and cholesterol), also showed reduced liver injury and hepatic steatosis upon genetic ablation of Mlkl, probably through a necroptosis-independent but autophagy-dependent mechanism (35).

Together, these studies suggest that the necroptosis-associated proteins are overexpressed in human liver and/or adipose tissue and that they are involved in the transition from NAFLD to NASH and NASH-fibrosis in mice (Figure 3). Moreover, there is evidence that the respective involvement of RIPK1, RIPK3, and MLKL may be independent of their primary function in necroptosis activation. It is so far completely unclear, whether necroptosis is executed in vivo at all and under which pathological circumstances. The metabolic studies carried out with the different knockout mice represent a helpful tool to approach this issue, but despite the fact that necroptosis-independent functions have been identified for all three proteins, no final conclusion can be given whether necroptosis is relevant in NASH disease. Here, an in vivo visualization in real time would be required, showing the execution of necroptosis with the typical associated morphological criteria. Moreover, since constitutive knockout mice were used in most murine studies, the specific function in the individual organs is still not evident. Therefore, it would be important to repeat the studies with cell type-specific conditional knockout mice in order to more precisely describe the specific function of RIPK1, RIPK3, and MLKL in NASH-fibrosis.

Open in a separate window

Figure 3

Summary of the involvement of the necroptosis-associated proteins RIPK1, RIPK3 and MLKL in obese/NASH patients and in murine NASH models. The necroptosis-associated proteins RIPK1, RIPK3, and MLKL are overexpressed in the liver and/or adipose tissue of NASH patients. Murine NASH studies with distinct knock-in and knock-out mouse models confirmed an essential function of RIPK1, RIPK3, and MLKL in the transition from NAFLD to NASH and NASH-fibrosis, including both protective and inducing mechanisms. While RIPK3 deletion led to a decrease in fibrosis in the MCD model, NASH progression was observed in the HFD and CD-HFD models. Deletion of RIPK1 and MLKL was predominantly protective in the MCD and HFD models.

Defining HGF/c-Met Dependent Diagnostic Markers and Novel Therapeutic Targets in NASH-Development

c-Met acts as the cellular receptor for hepatocyte growth factor (HGF), thereby regulating cell growth, motility and morphogenesis. This growth factor is mainly produced by cells of mesenchymal origin and it acts on epithelial and endothelial cells. Originally, c-Met has been identified as an oncogene. It serves as an important regulator for stem cell growth during embryonic development and in particular also for the growth of hepatocytes. In this regard, it is of interest that HGF knockout mice as well as mice deficient for its receptor c-Met display severe developmental defects and both die between day 13 and 16 of embryonic development. The lack of both genes results in impaired placental and liver development (36, 37). Binding of HGF to c-Met induces dimerization and phosphorylation of the receptor and lead to recruitment of intracellular adapter proteins that bind to c-Met and lead to activation of specific intracellular signaling cascades. Those activate PI3K, Ras and ERK-dependent signaling pathways and control HGF-induced pro-mitogenic and anti-apoptotic events. The latter are of particular importance, as interference with apoptosis regulating pathways can have deleterious consequences for the whole organism. Usually, the apoptosis inducing death-receptor Fas is sequestered by c-Met. This prevents an uncontrolled Fas-ligand induced death-receptor activation and apoptosis is inhibited (38). This control mechanism is disturbed during NASH pathogenesis. Fas-Ligand is produced in excess and the physiological inhibition through c-Met is hampered (39). As a consequence, apoptosis is induced and liver damage occurs (40). Thus, HGF/c-Met has direct implications for the pathogenesis of NASH and seems to be hepatoprotective. Further light into the manifold c-Met controlled activities that regulate liver physiology became evident through the introduction of conditional c-Met knockout mice by Borowiak et al. (41). Here, mice carrying the Mx-cre-induced c-Met deletion displayed a reduced liver regeneration. Analysis of hepatocellular cell cycle progression in conditional Met knockout mice indicated a defective exit from quiescence and diminished entry into S-phase. This was accompanied by a reduced activation of Erk1/2 kinase, while Akt-phosphorylation was still intact—highlighting the importance of cytokine induced signaling cross talks.

Studying the role of c-MET in murine NASH models revealed that hepatocyte specific c-Met knockout mice (c-Met^Δhepa) exhibited increased steatosis and liver infiltration upon a methionine-choline deficient (MCD) diet compared to wild type controls (c-Met^loxP/loxP). This was accompanied by an outstanding upregulation of genes involved in fatty acid metabolism, generation of reactive oxygen species and enhanced fibrosis. Interestingly, c-Met^Δhepa mice also display more apoptosis in the liver which could be reverted in c-Met/Caspase-8 double knockout animals, emphasizing the prominent role of c-Met in the regulation of cell survival (42). Of note, a later publication could indicate that the protective role of c-Met is not restricted to hepatocytes since genetic deletion of c-Met in Kupffer cells/macrophages, myofibroblasts and CK19⁺ cells also attenuated steatohepatitis and fibrosis in the MCD-model, highlighting a more global role of c-Met beyond liver parenchymal cells (43). Noteworthy, the mineralocorticoid receptor (MR), which has been implicated in the pathogenesis of insulin resistance and type 2 diabetes, on myeloid cells (i.e., macrophages), might contain protective HGF/c-Met signaling in hepatocytes whereby these two cells could be functionally linked in the formation of steatohepatitis and related fibrosis (44). In line with hepatoprotective consequences of c-Met signaling in experimental NASH models, Pioglitazone, a ligand of peroxisome proliferator-activated receptor gamma (PPARγ), which was shown to alleviate human steatohepatitis in clinical trials, acts through activation of c-Met (45).

Accordingly, it has also been demonstrated that the c-Met ligand HGF counteracts steatohepatitis since mice with transgenic HGF overexpression were largely protected from the deleterious effects of MCD-diet by amelioration of fibrosis and inflammation and protection from oxidative stress (46). In a subsequent study the hepatoprotective effect of HGF in NASH could be attributed to the JAK2-STAT3 pathway thereby limiting inflammation (47). Furthermore, HGF exerts anti-oxidative properties by inducing glutathione and related enzymes (48) and dampens insulin resistance and liver triglyceride and cholesterol content which might involve nuclear receptors such as the Farnesoid X receptor (FXR) (49). Blocking c-Met mitigated the beneficial effects of HGF indicating the importance of HGF/c-Met downstream signaling. Furthermore, an earlier pivotal study indicated that the HGF/c-MET axis suppresses hepatic glucose uptake promoting insulin responsiveness by direct engagement of the insulin receptor (INSR) in a INSR/c-Met complex (50). The potential role of recombinant HGF in the treatment of NASH was also demonstrated by Yang et al., who could show that mice which were fed a choline-deficient amino acid defined diet (CDAA) display improved inflammation, steatosis and lipid profile when treated with recombinant feline HGF (51).

As already mentioned, oxidative stress plays a critical role during the progression from simple steatosis toward manifest steatohepatitis. The transcription factor termed nuclear factor-E2-related factor-2 (NRF2) serves as a cellular sensor for oxidative stress. NRF2 itself is sequestered in the cytosol by Kelch-like ECH-associated protein (Keap1). During oxidative challenge, modification of Keap1 sulfhydryl groups results in the stabilization and nuclear translocation of NRF2 (52). NRF2 is crucial for antioxidant, responsive element/electrophile-responsive element (ARE/EpRE)-mediated induction of detoxifying enzymes, anti-oxidative stress genes and other target genes involved in cellular protection. Activation of these target genes serves to decrease the oxidative burden of the cells (53). This seems to be pathogenetically related to the development of NASH, as Nrf2^−/− mice are more prone to develop fatty liver degeneration (54). More refined investigations of related molecular mechanisms and potential cross talks to other signaling pathways involved in NASH pathogenesis are still scant and need to be better explored. Given this role as a master-regulator, pharmacologic or genetic inhibition of Keap1 function results in constitutive activation of Nrf2 signaling that has been shown to exert hepato-protection against chemical-induced cytotoxicity, ischemia/reperfusion injury and alcohol-induced liver steatosis (55, 56). However, in spite of multiple published reports, results supporting an involvement of this transcription factor in the modulation of hepatic lipid accumulation and steatohepatitis remain deeply controversial. Indeed, whereas genetic Nrf2 deletion has been demonstrated to accelerate the transition from simple steatosis to NASH (57), genetic activation of Nrf2 resulted in different and sometimes contradictory effects. These results seem to depend on the site of Keap1 deletion, the experimental NASH model and the experimental conditions (58). Indeed, Nrf2 over-activation in hepatocyte-specific Keap1 knockout mice results in amelioration of steatosis but does not affect liver inflammation and fibrosis (59). A recent study highlighted the potential synergistic hepatoprotective role of c-Met and NRF2 in NASH by investigating the consequences of combined c-Met/Keap1 gene deletion in hepatocytes in rodent models of steatohepatitis. It was clearly demonstrated that double-knockout mice exhibit less inflammation, steatosis and oxidate stress in comparison to c-Met^Δhepa animals indicating that the anti-oxidative factor NRF2 can override the detrimental effects of c-Met deficiency by restoring the cellular redox homeostasis (60). This finding emphasizes the crucial role of reactive oxygen species for the development and progression of steatohepatitis.

Despite these promising findings, hepato-cancerogenic effects of HGF/c-Met and Keap1/Nrf2 activity currently limit the therapeutic exploitation of the potential beneficial properties of HGF/c-Met and NRF2 in human NASH. For example, aberrant NRF2 activation has been identified to accelerate the development of hepatocellular carcinoma (HCC) from pre-cancerous lesions (61) and the HGF/c-Met axis is fundamentally involved in HCC pathogenesis via canonical and non-canonical pathways (62). More research is therefore warranted to unravel potential applications of c-Met and Nrf2 agonists in fatty liver disease.

Targeting MIF Family Proteins in Liver Fibrosis

Macrophage migration inhibitory factor (MIF) is an evolutionarily conserved defense protein and upstream regulator of innate immunity that was more recently re-defined as a broadly expressed and pleiotropic inflammatory cytokine. The 3D architecture of the MIF trimer is similar to tautomerase enzymes found in various kingdoms, but the MIF monomer also exhibits similarities to dimeric CXC chemokines such as CXCL8. In fact, MIF engages in high-affinity non-cognate interactions with the CXC chemokine receptors CXCR2 and CXCR4 and accordingly, has been recognized to be a prototypical member of the emerging family of atypical chemokine (ACK) (77–79). MIF promotes monocyte and neutrophil recruitment through CXCR2, while enhancing T-cell, B-cell, and progenitor cell recruitment, as well as cancer cell metastasis through CXCR4. Combined activation of the MIF/CXCR2 and MIF/CXCR4 axes drives monocyte/macrophage atherogenic recruitment and is a major driver of atherosclerotic lesion development (78, 80). In addition to the CXC chemokine receptors, MIF interacts with CD74, the membrane form of invariant chain (Ii). While Ii functions as an MHC class II chaperone with a critical role in class II trafficking and control of peptide loading in the endolysosomal compartment, the membrane form of CD74 acts as cytokine receptor for MIF (81). Under inflammatory and neoplastic conditions, CD74 is upregulated also in the absence of class II (82). Receptor activities of CD74 encompass proliferation and survival responses through NF-κB/TAp63 and ERK-MAPK, while blocking p53-mediated apoptosis (82). Signaling through MIF/CD74 requires recruitment of signaling-competent co-receptors, which can either be the accessory protein CD44 mediating Src tyrosine kinase and PI3K/AKT activation, MIF associated chemokine receptors, which can form hetero-oligomeric complexes with CD74 (78), or nuclear translocation of the intracellular domain (ICD) of CD74. The more recently described MIF homolog D-dopachrome tautomerase (D-DT/MIF-2) shares an inflammatory activity spectrum with MIF in sepsis, but has opposing properties in adipose inflammation, and also engages the MIF cognate receptor CD74 (83, 84).

MIF is a critical mediator in the pathogenesis of various inflammatory and immune diseases such as septic shock, rheumatoid arthritis, and atherosclerosis, as well as several cancers (80, 85). MIF levels correlate with the inflammatory status and disease stage and the predominant MIF gene promoter polymorphisms, a −173 G/C SNP and −794 (CATT)₅₋₈ microsatellite repeat, are associated with the severity of asthma, atherosclerosis, kidney injury, and rheumatoid arthritis (86, 87). With regard to fibrotic disorders, MIF has been firmly linked to cystic fibrosis, fibrosis in the bladder, and myocardial interstitial fibrosis (88, 89). Here, we discuss current knowledge on MIF's role in liver disease and fibrogenesis.

In the liver, MIF is not only produced by immune, endothelial cells, and cancer cells, but also by hepatocytes. In a study on alcoholic liver injury, MIF serum levels were significantly augmented in patients with alcoholic steatohepatitis (ASH) and alcoholic cirrhosis as compared to controls and MIF expression in ASH could be pinpointed to (ballooned) hepatocytes and infiltrating inflammatory cells i.e., neutrophils (90). This data is supported by a recent study identifying the liver—and more specifically hepatocytes—as significant source of circulating MIF levels in ASH, which correlated with disease severity and mortality (91). On the other hand, one study analyzing MIF expression and the MIF gene −173 G/C polymorphism in NASH failed to detect a correlation between MIF expression in hepatocytes and fibrosis stage, but observed that MIF expression of mononuclear cells in liver tissue significantly increased according to fibrosis stage.

Besides metabolic liver injury, MIF has also been functionally implicated in the course of human chronic liver diseases of various etiologies such as autoimmune disorders or chronic viral infections (92–94). In HCV-induced liver fibrosis, microsatellite polymorphism −794CATT₅₋₈ and the −173G/C SNP were predictive of fibrosis severity and cirrhosis-associated complications such as impaired liver function and prevalence of hepatocellular carcinoma (94). Moreover, MIF promoter polymorphisms and serum levels in patients correlated with the presence of autoimmune hepatitis and primary biliary cholangitis (92, 93).

Functional studies in two independent mouse models of hepatotoxin-driven chronic liver injury revealed a protective effect of MIF, which could be attributed to inhibitory effects on hepatic stellate cell activation and proliferation—key events in liver fibrogenesis—via engaging the CD74/AMPK pathway (95). Moreover, MIF/CD74/AMPK signaling in hepatocytes prevented fatty degeneration in a model of NASH (96). The activation of these protective pathways in the liver is reminiscent of reported cardioprotective activities of MIF mediated through the CD74/AMPK axis (80). In experimental models of ethanol-induced liver injury, however, MIF's impact seems to be context-dependent. While MIF exacerbates liver injury during chronic ethanol feeding by modulating chemokine production and immune cell infiltration (91, 97), it mediates protective effects following chronic-binge ethanol feeding by regulating the unfolded protein response in hepatocytes (98). A similar effect might also be seen during the course of NASH. While MIF conveys anti-steatotic effects on hepatocytes, it contributes to liver fibrogenesis in a murine model of NASH (MCD diet feeding) by skewing the intrahepatic immune milieu toward a pro-fibrotic polarization of innate lymphocytes (99).

In summary, these results underline a pivotal, but complex role of MIF during chronic liver diseases, with even dichotomic effects in response to the same insults. Figure 6 summarizes the regulatory role of the MIF/receptor network in chronic liver diseases. Considering the complexity of this network is of specific importance when designing MIF-directed therapeutic strategies in the setting of chronic liver disease. Of note, several efforts to therapeutically target the MIF system are currently assessed in early clinical trials concerning various disease settings such as rheumatic, cardiovascular disease, or neoplastic disorders, including small molecule-, antibody- and peptide-based approaches targeting either MIF or specific MIF receptor-mediated pathways (80). Based on the mechanistic insight from the experimental data, one might speculate that a key determinant of MIF-mediated outcome could be the predominance of MIF's impact on immune cell recruitment/polarization over its direct effects on liver resident cells. This predominance might be guided by specific intrahepatic MIF receptor expression patterns. Whether these patterns are a valid and stable surrogate to predict MIF-mediated effects warrants further investigations, which also should consider potential distinct activities of D-DT/MIF-2. A further comprehensive understanding of MIF's role in the different settings and even at distinct disease stage is a prerequisite to guide MIF-directed therapeutic interventions and predict the outcome of these therapeutic strategies in clinical practice during the course of chronic liver injury.

Open in a separate window

Figure 6

Macrophage migration inhibitory factor in chronic liver injury and fibrogenesis. In the liver, the atypical chemokine Macrophage migration inhibitory factor (MIF) is mostly expressed by infiltrating immune cells as well as hepatocytes especially under stress condition e.g., fatty degeneration or inflammation during NASH, ASH and viral hepatitis. In addition, it is most likely that MIF is involved in the pathogenesis of hepatic autoimmune disorders because this cytokine was previously shown to be associated with many autoimmune disorders including rheumatoid arthritis, inflammatory arthritis, inflammatory bowel disease, multiple sclerosis, autoimmune uveitis., autoimmune glomerulonephritis, systemic lupus erythematosus, sarcoidosis and many other autoimmune inflammatory disorder [for review see (85, 100)]. MIF engages high-affinity non-cognate interactions with three different surface receptors expressed by liver resident cells as well as infiltrating cells, the CXC chemokine receptors CXCR2 and CXCR4 as well as with CD74, the membrane form of invariant chain (Ii). While MIF engagement with the CD74 receptor on hepatocytes and hepatic stellate cells exert protective effect via AMPK signaling during chronic liver injury, MIF also mediates pro-inflammatory effects orchestrating the intrahepatic recruitment and activation of inflammatory immune cells via engagement of CXCR2 and CXCR4. The balance between these opposing, regulatory roles of the MIF/receptor network is crucial for the overall impact of MIF on severity and progression of chronic liver disease in distinct settings and has to be considered when designing MIF-directed therapeutic strategies. Blue arrow indicates MIF release. Red arrow marks pro-inflammatory properties and green arrow protective features of MIF on respective pathways.

Monocyte and Macrophages in Liver Fibrosis

Macrophages comprise the most abundant immune cell type of the liver, playing an essential part in the maintenance of liver homeostasis and in reparative or propagating mechanisms following acute or chronic liver injury. Two major subtypes of hepatic macrophages can be distinguished; the liver resident Kupffer cells (KCs), which are considered to be a self-sustaining and often tolerogenic phagocyte population, and the immunogenic monocyte-derived infiltrating macrophages (MoMΦs) (101, 102). Interestingly, upon liver disease, dynamic changes in these macrophage subsets can be observed, such as a rapid loss of KCs after injury, and an increased recruitment and subsequent accumulation of inflammatory MoMΦs that are characterized by Ly6C^high (Gr-1) expression in mice. Moreover, such infiltrated Ly6C^high MoMΦs, of which the recruitment was shown to be critically controlled by the chemokine CCL2/CCR2 signaling pathway, represent the dominant macrophage population during the early phases of liver injury, promoting the progression of liver fibrosis by activating hepatic stellate cells (HSCs) and other myofibroblasts (103). Furthermore, the pharmacological inhibition of CCL2 with the Spiegelmer-based antagonist mNOX-E36 efficiently inhibits the infiltration of Ly6C^high MoMΦs into chronic CCl₄- or methionine-choline-deficient (MCD)-diet-induced injured liver in mice, ameliorating hepatic inflammation and steatosis (104). Administration of mNOX-E36 during the regression phase of these murine models of liver disease, however, significantly accelerated fibrosis resolution, by causing a shift in hepatic dominance from the pro-inflammatory, tumor necrosis factor (TNF)-secreting, Ly6C^high MoMΦs toward “restorative” Ly6C^low MoMΦs (105). Also cenicriviroc (CVC), an orally available dual CCR2/CCR5-inhibitor has been shown to efficiently block monocyte recruitment, and to promote amelioration of insulin resistance as well as anti-inflammatory-, and anti-fibrotic effects in murine NASH models (106). This drug has been evaluated in a large (n = 289 patients) biopsy-controlled phase 2b clinical trial in NASH patients, in which CVC promoted accelerated fibrosis regression after 1 year of therapy (107). However, the anti-fibrotic efficacy appears neither increased nor sustained over 2 years of therapy (108). CVC had entered phase 3 clinical development with about 2,000 patients (109), but the trial was prematurely terminated since the primary endpoint of fibrosis improvement after 1 year was not reached upon pre-specified interim analysis.

Besides the CCL2/CCR2-pathway, also the CCL1/CCR8-signaling cascade represents an important chemoattraction axis. Indeed, Ccr8^−/− mice undergoing chronic CCl₄-injections or bile duct ligation (BDL) display reduced infiltration of inflammatory MoMΦs, neutrophils and natural killer (NK) cells, whereas the number of hepatic CD4+ T cells was increased. An overall protection from liver fibrosis was observed in these models in Ccr8^−/− mouse, when compared to their respective controls (110). The importance of chemokine-axes can be further demonstrated by the regulation of hepatic inflammation through the CXCR6/CXCL16-axis during liver disease. CXCL16 was found to be secreted by hepatic endothelium and macrophages. It can control the recruitment and functionality of CXCR6-expressing pro-inflammatory NK T-cells, especially during the early response in experimental liver damage. Moreover, reduced NK T-cell accumulation in Cxcr6^−/− mice led to the diminished presence of the pro-inflammatory cytokines IFN-γ, TNF-α, and IL-4 in the hepatic microenvironment, and overall reduced fibrogenesis (111).

While all above-mentioned research suggests the interception of chemokine-axes as potential therapeutic strategy, it should be noted that the inflammatory system has a dual function in liver disease, and that some chemoattractants may thus be necessary for disease amelioration and tissue repair (Figure 7). For example, CX₃CL1 (also known as fractalkine) is shed by HSCs and hepatocytes upon liver injury, creating a chemo-attractive gradient for the CX₃CR1-expressing leukocytes. Surprisingly, the abrogation of this axis in CCl₄- or BDL-challenged mice causes an increased presence of pro-inflammatory TNF/iNOS- producing MoMΦs and subsequent extent of fibrosis (112). The CCR6/CCL20-axis has a similar –protective—functionality. Upon liver damage, the secretion of CCL20 by parenchymal cells is elevated, causing increased recruitment of CCR6-expressing T-helper (Th)17, regulatory, and gamma-delta (γδ) T-cells. Elimination of this signaling cascade during liver disease caused by CCl₄ and MCD-diet, using mice with a Ccr6^−/− phenotype, identified reduced accumulation of interleukin (IL)-17 and −22 expressing γδ T-cells, and causing more inflammation and fibrosis as compared to wild type mice. These effects would be partly derived through the potential of γδ T-cells to induced HSC apoptosis in a cell-cell contact dependent manner involving Fas-ligand (CD95L) (113).

Open in a separate window

Figure 7

The dual role of inflammatory cells in liver fibrosis. While the chemokine-axes CCL2/CCR2, CCL1/CCR8, and CXCL16/CXCR6 have been shown to play a role in disease propagation through induction of a pro-inflammatory and pro-fibrogenic environment, the axes CX₃CL1/CX₃CR1, and CCL20/CCR6 were identified to be essential to obtain amelioration of liver function after acute or chronic damage. Besides its effects on inflammatory cell recruitment, the liver environment also influences the function of immune cells, such as through secretion of histidine-rich glycoprotein (HRG), promoting the polarization of macrophages toward an inflammatory phenotype. While therapeutic agents targeting the inflammatory system may be a promising strategy, their potential off-target effects limit their future use. Cell-specific nano-scale delivery systems such as liposomes, polymers and microbubbles may therefore aid in the development of such inflammatory-specific therapeutic tools.

The injured liver environment not only tightly regulates the inflammatory response through secretion of chemoattractants, and thus the recruitment of inflammatory cells, but also by shaping the inflammatory cell polarization and functionality. One example includes histidine-rich glycoprotein (HRG), a liver-derived plasma protein, which promotes the polarization of macrophages toward the proinflammatory (“M1-like”) phenotype, thus stimulating the propagation of liver injury and subsequent fibrogenesis (114). Disease-induced effects on immune cell function and polarization are even suggested to be obtained in the precursory bone marrow monocytes, and remain conserved in their functional derivates, the hepatic myeloid cells. For example, in murine models of NAFLD, a prominent down-regulation of calprotectin coding genes S100a8 and S100a9 is observed in the myeloid cells of both the liver and bone marrow (115). Peroxisome proliferator-activated receptors (PPARs) are nuclear receptors, subdivided into 3 isoforms (PPARα, PPARγ, and PPARδ/β), with PPARγ and PPARδ being involved in macrophage polarization. Such effects where proven through administration of lanifibranor, a pan-PPAR agonist, to palmitic acid-stimulated murine macrophages and patient-derived circulating monocytes, leading to a reduction in the expression of proinflammatory genes. When administered to choline-deficient amino acid-defined high fat (CDAA)- and Western-diet (WD) mice models, lanifibranor showed significant beneficial effects on the extent of steatosis and hepatitis, thus suggesting its potential use as therapeutic agent for NAFLD/NASH (116).

While modulators of immune cell recruitment, functionality and polarization may represent interesting therapeutic options in the battle against liver disease, their systemic administration may cause unwanted side-effects. Nanomedicine-based approaches, targeting specific cell types, may eliminate this drawback. Different nano- and micrometer-sized drug delivery systems have been developed and were shown to target distinct cellular players of the inflammatory process (117). Indeed, after systemic administration, PEGylated liposomes are rapidly taken up by the dendritic cells of the liver, lung, and kidney, poly(N-(2-hydroxypropyl) methacrylamide) polymers by the endothelial cells in the liver, neutrophils and alveolar macrophages, and poly n-butylcyanoacrylate (PBCA) microbubbles by KC and splenic red pulp macrophages (118). The absence of any effects of these drug carrier systems with regard to hepatotoxicity or inflammation, regardless of its efficient uptake by myeloid immune cell in the liver, is promising concerning future clinical application (119). It should be noted that the fibrotic liver environment significantly affects the targeting efficiency of the different mentioned carrier systems, therefore suggesting the need for an adapted nanomedicine-based approach for interfering with early events in the pathogenesis of chronic liver disease (120).

Innate Lymphoid Cells and Liver Fibrosis

Innate lymphoid cells (ILCs) represent a heterogeneous family of innate immune cells with lymphoid phenotypes, but lack rearranged antigen receptors (121). Traditionally, ILCs have been divided into 3 groups based on the expression of specific transcription factors, cell-surface markers, and signature cytokines. Members of group 1 are IFN-γ producing and T-bet dependent ILCs (ILC1s). Group 2 ILCs (ILC2s) are a population of cells that preferentially produce type 2 cytokines, including IL-5 and IL-13, and require GATA3. Group 3 ILCs (ILC3s) can produce IL-17 and/or IL-22, and are dependent on RORγt.

This classification was recently revised. Two new members were added to the ILC family: conventional NK cells (cNK) and lymphoid tissue-inducer cells. However, the unambiguous definition of ILC1 and cNK and their differentiation from each other is still challenging and is likely to differ between mice and men (122). For instance, transcriptional profiling of hepatic ILCs demonstrated mouse liver ILC1s to display unique expression patterns for several chemokine receptors and adhesion molecules, including CXCR6, CD103, CD49a, CD69 (123, 124), markers that have been proposed to characterize liver resident NK (lrNK) cells in humans (125, 126). Within each group, ILCs are heterogeneous in terms of phenotype and cytokine profiles within tissue and between different organs, both in mice and humans. Moreover, there is increasing data suggesting that localization of ILC subsets in specific compartments relates to their roles in immune and inflammatory responses.

cNK cells are critically involved in the immune-pathogenesis of liver disease and both murine and human cNK cells have been shown to exhibit anti-fibrotic activity by the induction of apoptosis and/or killing of activated star cells (127, 128). This anti-fibrotic function of cNK cells is linked to the surface expression of activating NK cell receptors including NKG2D, NKp46, and NKp30, that recognize specific molecules expressed on activated HSC. In addition to NK cell receptors, members of the TNF superfamily are also involved in the anti-fibrotic activity of NK cells. Activation of HSCs leads to an increased expression of TRAIL receptors resulting in an increased susceptibility to apoptosis induction by TRAIL-expressing NK cells (128).

Of note, most of these data were obtained in studies with peripheral cNK cells, so it is unclear to what extent these results can be extrapolated to lrNK cells. In addition, further work is needed to clarify which subsets of lrNK cells are involved in modulating liver fibrosis. Moreover, it is important that the function of NK cells is influenced by the surrounding microenvironment, such as local cytokine concentration or interactions with other immunocompetent cells. For instance, we demonstrated that CD4⁺ T cells effectively trigger anti-fibrotic cNK cell activity in an IL-2 dependent fashion (129) whereas regulatory T cells can exert inhibitory effects on NK cell anti-fibrotic activity, thus promoting fibrogenesis (130). In addition, chronic liver cell damage leads to increased levels of Transforming growth factor-β (TGF-β), which inhibits the anti-fibrotic function of NK cells by downregulating NKG2D and 2B4. With regard to the other members of the ILC family, data on their possible role in hepatic fibrogenesis are rather sparse. McHedlidze et al. demonstrated that in mice interleukin-33 (IL-33), secreted by injured hepatocytes, activates ILC2 to produce IL-13, which then induces activation of hepatic stellate cells, thereby promoting hepatic fibrogenesis (131). This observation provided first evidence that ILCs other than cNK cells also modulate hepatic fibrosis.

In the context of human liver disease, Forkel and co-workers observed a correlation between the severity of fibrosis and the proportion of intrahepatic ILC2, which may produce IL-13 and mediate pro-fibrotic activity. The increased production of IL-33 and thymic stromal lymphopoeitin by hepatocytes, HSCs and Kupffer cells is discussed as a possible mechanism of ILC2 activation (132). However, given the low abundance of ILC2 in the human liver the exact role of this subset still has still to be defined. With respect to ILC3, Wang et al. observed an intrahepatic accumulation of IL-17 and IL-22 producing ILC3, which displayed pro-fibrotic activity in CCl₄-induced liver fibrosis (133). Whether human group 3 ILCs also play a role in hepatic fibrogenesis is unclear at the moment. However, given their important role in maintaining intestinal health by promoting immunity to pathogens, limiting inappropriate inflammatory responses to commensal bacteria or dietary antigens, or mediating repair following tissue damage ILCs may also indirectly modulate liver fibrosis via affecting the so called “gut-liver axis.” Both, derangement of the gut microflora (the microbiome) as well as defects of the intestinal barrier integrity have been shown to promote microbial translocation (MT) and there is accumulating evidence, mainly obtained in mouse models, indicating that ILCs critically affect both parameters. Exposure of hepatic immune cells to such gut-derived microbial products is considered to trigger hepatic inflammation in an inflammasome-dependent manner and to modify immune responses of intrahepatic immune cells, thereby accelerating hepatic fibrogenesis. Thus, it is tempting to speculate that alterations of the intestinal ILC population in liver disease may result in increased microbial translocation, thereby promoting hepatic inflammation and fibrogenesis.

Despite relevant progress, our understanding of hepatic ILCs and their role in fibrogenesis is still incomplete. Important questions concerning the role of specific subsets, the influence of the hepatic microenvironment and the gut-liver axis still need to be answered in detail to improve our understanding of the immunopathogenesis of liver cirrhosis.

NLRP3 Inflammasome Activation in Liver Disease Progression

Inflammasomes are intracellular multi-protein complexes expressed in parenchymal as well as non-parenchymal cells in the liver. They act as key regulators of inflammation and cell fate in various diseases (134). In humans, Nlrp3 gain-of-function mutations are causing a spectrum of rare auto-inflammatory disorders known as cryopyrin associated periodic syndromes (CAPS). The nucleotide-binding oligomerization domain-like receptor (NLR) pyrin domain containing 3 (NLRP3) inflammasome consists of NLRP3, apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC), and pro-caspase-1. The assembly of these three main components is triggered by different molecular and pathogenic structures such as bacterial LPS, adenosine triphosphate (ATP), uric acid, reactive oxygen species (ROS), fatty acids, bile acids (134). The activation of the functional inflammasome requires two signals (or two hits), one for inducing gene expression and assembly and the second for activating the effector component caspase 1. Active caspase-1 mediates the cleavage and release of the pro-inflammatory cytokines IL-1β and IL-18. Alongside Nlrp3-associated programmed cell death termed pyroptosis is initiated ultimately causing cell swelling and disruption of plasma membrane due to the assembly of the pore forming gasdermin D. This form of cell death is triggered by pro-inflammatory signals and associated with inflammation in which processes are triggered that enhance or initiate attraction and activation of immune cells but also perpetuate an abnormal wound-healing response (135, 136). This is also reflected in the term pyroptosis composed of “pyro” and “ptosis.” While “pyro” comes from the Greek word fire indicating the properties of an inflammatory reaction, the Greek word term “ptosis” standing for falling indicate the processes that are associated with the process of cell death (136).

Methionine- and choline-deficient diet (MCDD) or prolonged high fat diet (HFD) induced murine NASH characterized by steatosis and immune cell infiltrates, display increased hepatic mRNA expression of IL-1β, NLRP3, caspase 1, and ASC alongside elevated caspase-1 activity (137, 138). NLRP3 knockout mice fed a choline-deficient amino acid-defined [CDAA] diet were protected from similar inflammation and fibrosis development (139). Hepatic stellate cell (HSC) specific NLRP3-inflamasome-activation induced transdifferentiation to pro-fibrotic myofibroblasts, with livers of 24 weeks old mice showing increased expression of fibrotic α-smooth muscle actin (α-SMA) and collagen independent of inflammation (140). Interestingly, severe inflammatory changes associated with universal overactive Nlrp3 were almost completely rescued by TNF knockout alongside decreased IL-1β levels, while IL-17 deletions had only minor influence on the induced phenotype (141). Specific blocking of TNF by eternacept also rescued the gain of function phenotype, while reducing serum IL-1ß and IL-18 levels significantly in vitro and in vivo (142). Targeting the inflammasome with the selective small-molecule inhibitor MCC950 suppressed infiltration with immune cells in NASH caused by overnutrition in atherogenic diet-fed foz/foz mouse model alongside decreased IL-1β, IL-6, and MCP-1 levels (143). These results warrant targeting Nlrp3 as a possible therapeutic approach. In addition to its effect on the liver, global Nlrp3 knockout mice fed with MCDD for 3 weeks showed hepatic steatosis alongside alterations in the gut microbiota. Co-housing resulted in exacerbated fatty liver phenotype in wild type animals (144). In the MDR2 knockout model of primary sclerosing cholangitis, intestinal dybiosis was associated with pronounced Nlrp3 inflammasome activation in the gut-liver axis. Microbiome transfer from healthy mice markedly reduced liver injury in the recipient, so did the pan caspase inhibitor IDN-7314 hinting toward the importance of caspase activation to promote liver injury (145). Nlrp3-deficient mice that underwent bile duct ligation (BDL) as a model for primary sclerosing cholangitis (PSC) had significantly less inflammation in acute (2 days) and chronic (28 days) injury in liver and kidney, hinting toward an important role of the inflammasome as well. This finding was confirmed with the use of the specific Nlrp3 inhibitor MCC950 that cut down disease progression in wild type mice in the BDL model (146).

Recently, a lot of effort was focused on the evaluation of gut microbiota, which was found to be important in the constant interplay between gut and liver. NASH condition induced by MCD diet was markedly improved using the depletion of gut microbiota and consequent repopulation. Commensal microbiota appears to be hepatoprotective in that regard (147). Gastrointestinal dysbiosis associates with increased production of PAMPs and DAMPs, which then enter portal circulation and promote NLRP3 inflammasome activation and specifically inflammation in the liver most importantly by interacting with TLR4 (Figure 8) (148). In obesity, increased intestinal permeability attenuated the expression of tight junctions, which facilitated the effect (149). Depletion of the G-Protein coupled receptor CX3CR1 was associated with significantly altered intestinal microbiota composition due to an impaired intestinal barrier. Endotoxin levels in portal serum and consequently inflammatory macrophages in liver were increased in CX3CR1 deficient mice, indicating an increased inflammatory response (150). As persistent inflammation of gut and fibrosis are preconditions for the development of hepatocellular carcinoma (HCC), effort focused on connecting inflammation in the liver with tumor biology. Signaling of Toll like receptors through the universal adaptor Myd88 and consequent IRAK resulting in NF-κB activation might be important in this process. Functional evaluation of the CC-chemokine ligand 5 in murine model of HCC using hepatocyte specific knockout of NF-κB essential modulator (NEMO) reduced TNF induced apoptosis in those mice, alongside reduced immune cell infiltration of granulocytes and pro inflammatory monocytes (151). While hepatocyte-specific deletion of MyD88/NEMO promotes the tumor progression in mice, this development was ablated in global Myd88/hepatocyte-specific knockouts of NEMO providing further evidence that inflammatory signaling through TLR in non- parenchymal cells of the liver is a driver of disease progression and might be a relevant target for disease treatment or prevention (152).

Open in a separate window

Figure 8

Multitudes of metabolic products like alcohol and free fatty acids can lead to enhanced intestinal permeability by disrupting tight junctions of intestinal epithelial cells. PAMPs and DAMPs that enter the liver initiate gene transcription of pro-IL-1β, pro-IL-18, and NLRP3 itself by binding to an appropriate receptor (here TLR4; signal 1). Injured and dying hepatocytes release DAMPs including endogenous ATP or uric acid that promote the assembly of the three main effectors NLRP3, ASC and procaspase 1 to form the inflammasome. The active caspase 1 cleaves pro-IL-1β and pro-IL-18 as well as Gasdermin (GSDMD) into their mature forms. The N-terminal GSDMD fragments form a membrane pore to enable the release of IL-1β and IL-18 in order to attract further immune cells (Figure was partly created by Biorender).

In summary (see Figure 8), the Nlrp3 inflammasome was shown to be important in fibrosis progression in various models of murine liver disease. Especially continuous inflammatory conditions are often causing for advanced stages of liver fibrosis including HCC. Hence, targeting inflammation and especially NLRP3 seems a viable choice in decreasing severity and slowing fibrosis progression in the future.

Targeting CCN1/CYR61 as a New Treatment Modality in Hepatic Fibrosis

The protein family of Cellular Communication network (CCN) factors contains six matricellular proteins (CCN1-CCN6). A CCN protein is composed of a signal sequence and four distinct structural modules including an insulin-like growth factor binding domain, a von Willebrand factor type C motif, a thrombospondin type I module, and a carboxyl-terminal cystine knot motif containing two disulfide bridges (163–165). These proteins are critically involved in the control of development, cell fate, angiogenesis, tumorigenesis, osteogenesis, cell adhesion, mitogenesis, migration, chemotaxis, cell survival, and extracellular matrix production. CCN proteins are further capable to bind pro-fibrogenic cytokines such as TGF-β and the production of reactive oxygen species (ROS) that are both involved in the production of liver damage and initiation of hepatic fibrogenesis. Most strikingly, individual CCN members can mutually inhibit each other's expression and drive opposite effects of biological processes (166, 167). In this dualistic Yin and Yang activity concept, individual CCN proteins are independently acting but interconnected in a regulatory dynamic network that decides about the outcome of both, physiological and pathological processes. In regard to liver fibrosis, the interplay of CCN actions is even more complex and far beyond the complementary nature of Yin and Yang. This was exemplarily documented in a study showing that the adenoviral overexpression of CCN2 suppressed CCN3 expression, while the overexpression of CCN3 as well as the suppression of CCN3 by targeted siRNAs both resulted in enhanced CCN2 expression (168). In vivo, the expression of CCN2 and CCN3 are both increased in models of ongoing fibrogenesis. However, the cellular subsets expressing CCN2 or CCN3 are different and are strongly dependent on the model. In the BDL model, CCN3 expression in damaged liver is majorly found in mesenchymal and proliferating bile duct epithelia cells along the fibrotic septa, while CCN3 is predominant in persinusoidal areas peripheral to centrilobular hepatic necrosteatosis in the CCl₄ model (168). In contrast, CCN2 expression is markedly increased in damaged and cultured hepatocytes, which however do not express CCN3 (168). Artificial overexpression of CCN3 reduced expression of CCN2 in cultured hepatocytes, but failed to reduce liver fibrogenesis in the BDL model (169). Interestingly, adenoviral overexpression of either CCN2 or CCN3 in cultured hepatocytes induced reactive oxygen species formation and activated p38 and JNK pathways, thereby triggering hepatocyte apoptosis (169).

In the liver, CCN1 has several important activities (Figure 10). It acts as a senescence inducer that might be particularly important in later stages of wound healing by avoiding progressive fibrosis and by initiating resolution of fibrotic scar tissue (170). When massively overexpressed, CCN1 is directed to the endoplasmic reticulum (ER), resulting in ER stress and unfolded protein response (UPR). The resulting apoptosis of HSC is correlated with reduced collagen expression confirming that CCN1 has the capacity to attenuate liver fibrogenesis by modulating the three phases of endoplasmic reticulum stress, namely adaptation, alarm, and apoptosis (170). Moreover, CCN1 was shown to attenuate TGF-β signaling by scavenging TGF-β, thereby mitigating the overall fibrogenic response in vitro and in vivo during phases of liver insult (171). In portal myofibroblasts, CCN1 induced ROS formation, p38 phosphorylation, and upregulation of Fas, suggesting that resulting apoptosis requires excessive formation of free radicals and modulation of associated downstream cellular signaling pathways (171). Similarly, the overexpression of CCN2, CCN3, or CCN4 effectively induced ER stress and UPR in HSC, hepatocytes, and portal myofibroblasts suggesting that CCN proteins are generally associated with processes involved in hepatic tissue repair following liver injury (172, 173). Although UPR-mediated hepatocyte apoptosis might hinder hepatic tissue repair, it was suggested that CCN expression might be therapeutically attractive to mitigate liver fibrosis, especially when gene expression is specifically directed to HSC, MFB and portal myofibroblasts (172, 173). Hepatocytes express large quantities of CCN2. This alone does not cause hepatic injury or fibrosis per se, but renders the liver more susceptible to injurious actions of other fibrotic stimuli including TGF-β (174, 175). In addition, the reverse (i.e., the blockade of CCN2) has the potential to be an effective treatment for liver fibrosis. This again highlights the complexity in CCN protein biology and demonstrates that targeting of a specific member of this family might result in an unexpected outcome. Therefore, it was proposed that CCN proteins are part of an interconnected team that forms a scaffolding system in which the different CCN proteins act as connectors to permit a balanced series of biological effects (165).

Open in a separate window

Figure 10

CCN1 in liver homeostasis and disease. CCN1 consists of a secretory signal (SP), an insulin-like growth factor-binding protein domain (IGFBP), a von Willebrand type C domain (VWC), a thrombospondin-1 domain (TSP-1), and a cysteine knot (CT). The biological activities of CCN proteins manifest during liver injury. They stimulate the activation and transdifferentiation of hepatic stellate cells (HSC) to matrix-producing myofibroblasts (MFB), modulate cytokine activity, regulate apoptosis/necrosis, and fine-tune mechanisms involved in control of cell-cell contacts, cell renewal, epithelial-to-mesenchymal transition (EMT), and (neo-)angiogenesis. Large quantities induce endoplasmic reticulum stress and unfolded protein response. Moreover, they can bind cytokines such as TGF-β, thereby modulating their activities and pathways.

Angiotensin Stimulated Hepatic Fibrogenesis and Portal Hypertension: Receptor Regulation and Intracellular Signaling

Role of Renin-Angiotensin-System in Hepatic Fibrosis and Portal Hypertension

In patients and animals with liver cirrhosis, RAS is activated (Figure 11), leading to increased levels of circulating angiotensin II (179). In human liver samples of cirrhotic patients, the components of the classical renin-angiotensin-system (angiotensinogen, renin, angiotensin-converting-enzyme) are upregulated, and Angiotensin-II-type-1 receptor (AT1R) stimulation is increased (180, 181). The link between AT1R stimulation and development of fibrosis with portal hypertension has been well-established in animal models. The continuous injection of angiotensin II induced fibrosis in rats as shown by elevated hepatic hydroxyproline-content. The absence of AT1R in mice resulted in reduced fibrosis upon liver injury, which is in agreement with previous data and supports our working hypothesis that the AT1-receptor is important for the development of fibrosis and portal hypertension. This was underlined in TG(mREN2)27 rats, which overexpress Renin, especially in the liver, and develop spontaneous liver fibrosis and portal hypertension (182) without additional experimental hepatic injury.

Open in a separate window

Figure 11

Renin-angiotensin-system (RAS) in liver fibrosis. Angiotensinogen is cleaved by renin into Angiotensin I (Ang I), which is further converted to Angiotensin II (Ang II) by Angiontensin-Coverting-Enzyme (ACE). Ang II is the agonist of AT1R, which signals G-protein dependent via Janus-kinase 2 (JAK2), Argef1/RhoA/Rho-kinase. This constitutes the classical RAS, which is known to lead to activation of hepatic stellate cells (HSC) and thereby to fibrosis and their contraction. The G-protein coupled pathway is terminated by beta-arrestin-2 binding to AT1R, which may terminate the contraction, but still may induce fibrosis via ERK-activation. Ang II may be further metabolized to Ang1-7 by ACE2, which represents the alternative RAS-pathway. The alternative RAS-pathway may block contraction via mas-receptor (masR) stimulation. The role of masR and beta-arrestin-2 are still under investigation and be crucial to elucidate the mechanisms in HSC, but also may offer therapeutic options for liver fibrosis and portal hypertension.

On the other side, the alternative RAS, including angiotensin-converting-enzyme-2 and mas-Receptor (masR), is also increased in cirrhotic livers (179, 183). This mediates vasodilation by formation of NO. Thus, the stimulation of the masR using angiotensin (1–7) and/or the non-peptidic orally active agonist AVE 0991 could decrease hepatic resistance and portal pressure in cirrhotic rats (183, 184). Since both receptors are upregulated in liver fibrosis with portal hypertension, it should be important to distinguish interaction of pathways induced by stimulation of the respective receptors (Figure 11).

The principal investigators of the SFB/TRR57 are grateful to all members of the participating institutions and clinics that helped to establish a close translational research network on the pathogenesis of hepatic disease in Aachen and Bonn. We are grateful to our secretaries Mrs. Heike Schrinner, Lucie Hrvat-Delforge, and Jana Lippe, who performed excellent work in coordinating the administrative work of our consortium. Finally, we would like to thank the German Research Foundation that sponsored our work over 13 years and the reviewers who endorsed refunding decisions during the interim assessments.