Correspondence Hans Van Vlierberghe, MD, PhD, Department of Hepatology and Gastroenterology, Ghent University Hospital, 1K12 IE, De Pintelaan 185, 9000 Ghent, Belgium Tel: +32 9 3322370 Fax: +32 9 3324984 e-mail: firstname.lastname@example.org
Nowadays, liver cancer, cirrhosis and other liver-related diseases are the fifth most common cause of mortality in the UK. Furthermore, chronic liver diseases (CLDs) are one of the major causes of death, which are still increasing year-on-year. Therefore, knowledge about the pathophysiology of CLDs and its complications is of uttermost importance. The goal of this review is to clarify the role of angiogenesis in the disease progression of various liver diseases. Looking closer at the pathophysiology of portal hypertension (PH), fibrosis, cirrhosis, non-alcoholic steatohepatitis (NASH) and hepatocellular carcinoma (HCC), we find that angiogenesis is a recurring factor in the disease progression. In PH, several factors involved in its pathogenesis, such as hypoxia, oxidative stress, inflammation and shear stress are potential mediators for the angiogenic response. The progression from fibrosis to cirrhosis, the end-point of CLDs, is distinguished by a prolonged inflammatory and fibrogenic process that leads to an abnormal angioarchitecture distinctive for cirrhosis. In several stages of NASH, a link might be made between the disease progression and hepatic microvasculature changes. HCC is one of the most vascular solid tumours in which angiogenesis plays an important role in its development, progression and metastasis. The close relationship between the progression of CLDs and angiogenesis emphasises the need for anti-angiogenic therapy as a tool for blocking or slowing down the disease progression. The fact that angiogenesis plays a pivotal role in CLDs gives rise to new opportunities for treating CLDs and its complications.
The formation of new blood vessels is a key mechanism in the pathogenesis of chronic liver diseases (CLDs), irrespective of their underlying aetiology. CLDs cannot be approached as a single disease, given the fact that they usually progress from hepatocyte damage to inflammation, fibrosis, cirrhosis and in some cases to hepatocellular carcinoma (HCC). During the pathogenesis of CLDs, neovascularisation and establishment of an abnormal angioarchitecture are related to its pathological progression.
Cirrhosis is currently ranked as the tenth leading cause of death in the Western world (1) and an established cirrhosis has a 10-year mortality of 34–66%. The evolution of cirrhosis is characterised by intrahepatic vascular remodelling, with capillarisation of sinusoids, fibrogenesis and the development of intrahepatic shunts. One of the most common complications in patients with CLDs is the occurrence of portal hypertension (PH) (2), which results out of an increased resistance to blood flow caused by cirrhosis. Its pathophysiology is linked to an increased hepatic and splanchnic neovascularisation (3, 4).
Owing to the epidemic burden of obesity, diabetes and metabolic diseases, non-alcoholic fatty liver disease (NAFLD) and non-alcoholic steatohepatitis (NASH) will probably become the most common hepatic disease worldwide (5). NASH, the most severe form of NAFLD, is one of the major consequences of the current obesity epidemic (6). NAFLD is a progressive disease ranging from simple steatosis, to steatohepatitis, fibrosis, cirrhosis and HCC. Given the wide range of clinical features, it is not surprising that angiogenesis could play a role in its pathophysiology.
Every year, nearly half a million patients are diagnosed with HCC, making it the fifth most common cancer worldwide. HCC is a malignant liver tumour that generally develops in a background of CLDs. Because HCC is a well vascularised, solid tumour that relies on neo-angiogenesis to meet with its increasing needs for oxygen and nutrients, angiogenesis is a pivotal factor in its pathophysiology.
The goal of this review is to give a clear view on the role of angiogenesis in CLD and its complications, and in particular in NASH and HCC.
Molecular factors in angiogenesis
The formation of new blood vessels is initiated by hypoxic or ischaemic conditions. Oxygen levels are detected by prolyl hydroxylase domain-containing proteins that bind oxygen and hydroxylate-specific proline residues on hypoxia-inducible factors (HIFs) that are rapidly degraded after hydroxylation (7–9). Low physiological oxygen levels lead to an accumulation of HIFs that stimulate the expression of angiogenic growth factors (10).
Vascular endothelial growth factors and receptors
The vascular endothelial growth factor (VEGF) family consists out of five homologues, VEGF-A, -B, -C, -D and placental growth factor (PlGF). All VEGFs bind with different affinities to their receptors, VEGF receptor (VEGFR)-1, VEGFR-2 and VEGFR-3, of which only the first two are responsible for angiogenic signal transduction. VEGF-A is the key factor in the induction of angiogenesis and vasculogenesis, by binding to VEGFR-2 and increasing vascular permeability through a nitric oxide (NO)-dependent pathway (11). It causes vasodilatation by induction of endothelial NO synthase and increasing NO production (12). VEGF-A promotes endothelial cell (EC) survival by inducing the expression of anti-apoptotic proteins through the activation of phosphatidylinositol-3-kinase pathways. Binding of VEGF-A to VEGFR-1 does not induce signalling transduction; it merely serves as a decoy receptor.
The VEGF-B is a ligand for VEGFR-1, but can also form heterodimers with VEGF-A (13). It is probably involved in the formation of coronary collaterals (14) and inflammatory angiogenesis (15); nevertheless, its precise function remains unknown.
The VEGF-C has a high affinity for VEGFR-2 and VEGFR-3. It is primarily a lymphangiogenic growth factor that induces mitogenesis, migration and survival of ECs through activation of VEGFR-3. The angiogenic action of VEGF-C lies in its potential to bind VEGFR-2 (16–18).
The VEGF-D is responsible for the proliferation of ECs (through VEGFR-2) and also shows lymphangiogenic potential (through VEGFR-3) (19, 20).
Placental growth factor was originally discovered in the human placenta, but is expressed by a variety of cell types, such as hepatic stellate cells (HSC) and hepatocytes. It binds to VEGFR-1 and induces autophosphorylation, thereby activating the VEGFR-1 signalling cascade. Nevertheless, only weak phosphorylation occurs when binding to VEGFR-1; therefore, the major angiogenic action might lie in its potential to bind to neuropilin (NRP)-1 and NRP-2 (21). In addition, by binding to VEGFR-1, PlGF prevents the attachment of VEGF to this decoy receptor, and therefore causes an indirect stimulation of the VEGFR-2-dependent angiogenic pathway. In contrast to the other members of the VEGF family, PlGF is not essential for physiological angiogenesis during development and reproduction. Together with VEGFR-1, it is the key mediator in regulating the angiogenic switch in pathological conditions (3, 22, 23).
Fibroblast growth factors
Members of the fibroblast growth factor (FGF) family are also inducers for angiogenesis. Cellular responses to FGFs are mediated via specific binding to FGF receptors, which consist of an intracellular domain with high-affinity tyrosine kinase activity (24, 25). Receptor dimerisation by FGF is facilitated by heparin and induces a signal transduction cascade that stimulates cell migration, proliferation, differentiation and matrix dissolution (26). While VEGF members mostly regulate capillary formation, the FGF family is primarily involved in arteriogenesis.
Two types of angiopoetins (Ang) have been described, Ang-1 and Ang-2, both able to bind Tie-1 and Tie-2. Ang-1 stimulates tyrosine phosphorylation of Tie-2 in ECs, inhibiting endothelial apoptosis, stimulating EC sprout formation and stabilising vessels (27). Ang-2 is an antagonist of Ang-1, causing vessel destabilisation by shifting the ECs from a stable, growth-arrested state to a proliferative phenotype.
Platelet-derived growth factor
Although the angiogenic effect of platelet-derived growth factors (PDGFs) is not as strong, as seen in VEGFs and FGFs, in vivo studies have shown that PDGF can stimulate blood vessel formation (28, 29) and that PDGF receptors are expressed on capillary ECs (30, 31). In addition, PDGF is involved in the regulation of the tonus of blood vessels, inducing vasoconstriction (32), dilatation (33) and it plays a role in blood vessel maturation (34).
The cell–matrix interactions mediated by integrins play a pivotal role in vascular remodelling, EC attachment and EC migration. While it has long been considered that αvβ3 and αvβ5 integrins are positive regulators of the angiogenic switch (35), recent studies suggest that vascular integrins might inhibit angiogenesis (36, 37).
Cadherins compromise a large family of Ca2+-binding transmembrane molecules that promote homotypic cell–cell interactions (38, 39). ECs possess two cadherins: VE-cadherin and N-cadherin; only the first plays a role in cell-to-cell contact during neovascularisation (40). VE-cadherin mediates contact between ECs and regulates the passage of molecules across the endothelium (41).
Thrombospondin-1 (TSP-1) is an anti-angiogenic protein that directly affects EC migration and apoptosis (42), as well the bioavailability of VEGF. By inhibiting the activation of matrix metalloproteinases (MMPs), TSP-1 suppresses the release of VEGF from the extracellular matrix (ECM) (43). In addition, TSP-1 binds directly to VEGF, and recent data indicate that TSP-1 can mediate the uptake and clearance of VEGF (44). Similarly, TSP-2 binds to MMPs and mediates their uptake and clearance, thus affecting VEGF-levels (45).
Endostatin is a broad-spectrum angiogenesis inhibitor and interferes with growth factors such as FGF and VEGF. Endostatin also induces the activation of caspase 3, an intracellular protease that initiates apoptosis and cellular breakdown (46). Endostatin does not affect non-ECs and therefore only inhibits the proliferation of ECs.
Angiostatin is an angiogenic inhibitor that induces apoptosis of ECs (47). In addition, it inhibits the FGF and VEGF activation of extracellular signal-regulated kinase (ERK)-1, ERK-2 by temporary dephosphorylation in human microvascular ECs (48), but not smooth muscle cells (SMC) or fibroblast cells.
Mechanisms of angiogenesis
Vasodilatation is the first step in the formation of blood vessels. The influence of Ang-2 and VEGF gives rise to the occurrence of fenestrations, which increases the vascular permeability allowing extravasation of plasma proteins (Fig. 1, 1–2) (11, 49). These plasma proteins will serve as a scaffold for migrating ECs. Integrins are crucial in the migration of ECs to communicate between the scaffold proteins and the ECs, thus providing information about the location of the angiogenic site. Next, the basement and the ECM are degraded by activated MMPs, to allow subsequent EC migration and proliferation (Fig. 1, 2). TSP-1, endostatin and angiostatin can inhibit angiogenesis through interference with VEGF, FGF and integrins (Fig. 1). Several angiogenic factors (VEGF, FGF and EGF) induce EC proliferation and migration through the matrix (Fig. 1, 3). VE-cadherin and integrins coordinate the EC binding while tumour necrosis factor-alpha (TNF-α), FGF and PDGF induce tube formation (Fig. 1, 4–5). Endothelial progenitors differentiate in to ECs and form a primitive vessel (Fig. 1, 5–6). The surrounding vessel layers, composed of pericytes [or SMC (SMC in large vessels], need to be recruited (a PDGF- and VEGF-mediated process) (Fig. 1, 6). Under the influence of transforming growth factor (TGF)-β ECs will tighten up, pericytes are recruited and a new basement membrane and ECM are generated (Fig. 1, 7). Ang-1 is responsible for vessel stabilisation leading to durable and mature blood vessels (Fig. 1, 8).
Angiogenesis and liver inflammation
Inflammation of the liver is a recurrent factor in the pathophysiology of CLDs and its complications, like fibrosis, cirrhosis, steatosis, NASH and HCC. It is a biological response of the liver, in an attempt to initiate the tissue's healing process after hepatocyte damage. Inflammatory cells are already present in the liver, and HSC will undergo activation and release inflammatory mediators. Pro-inflammatory mediators can elicit an angiogenic response through the induction of HIF-1α and HIF-1-dependent transcriptional activity. Kupffer cells, mast cells and activated HSC are able to initiate angiogenesis through different pathways (Fig. 2).
Kupffer cells, the resident liver macrophages, patrol the hepatic sinusoids to phagocytose foreign particles. Activation of Kupffer cells leads to the release of cytokines, reactive oxygen species (ROS) and platelet-activating factor (PAF) (50, 51). All these factors can induce angiogenesis. (i) TNF-α is a cytokine involved in systemic inflammation. It is mainly produced by macrophages, but can also be expressed by a broad variety of cell types, including lymphoid cells, mast cells, ECs, adipose tissue fibroblasts and neuronal tissue. TNF-α can activate the mitogen-activated protein kinase (MAPK)/ERK pathway, a signal transduction pathway implicated in cell migration, proteinase-induction, regulation of apoptosis and angiogenesis (52, 53). (ii) The increase of ROS and nitrogen species in the liver can stimulate angiogenesis through an increased expression of TNF-α, NO, HIF-1 and VEGF (54). (iii) PAF enhances angiogenesis by inducing nuclear factor (NF)-κB activation, which in turn promotes the production of angiogenic factors such as VEGF (55, 56).
Mast cells are involved in the regulation of physiological and pathological vasculogenesis, by producing mediators, such as heparin, histamine, tryptase, TGF-β1, TNF-α, interleukins and cytokines (such as VEGF), and they increase the number of different types of ECs in vitro, including ECs making up liver sinusoids (57, 58).
Activated hepatic stellate cells
The recruitment of inflammatory cells is also stimulated by factors produced by activated HSCs and sinusoidal ECs. Activated HSCs play a direct role in the inflammatory process by expressing different chemokines (59), such as hepatic growth factor, PDGF, PlGF, VEGF and NO, which are able to induce angiogenesis (60, 61, 117). Chemokine expression by HSC is regulated by soluble mediators, in particular pro-inflammatory cytokines, as well as growth factors, proteases and products of oxidative stress. They may also play a role in modifying hepatic microvascular dynamics through the regulation of sinusoidal calibre and hepatic microvascular blood flow (62, 63) (Fig. 2).
The inflammatory response is not only triggered through the activation of cells present in the liver. Leucocytes, which normally reside in the blood stream, will move to the liver tissue via extravasation. The influx of leucocytes is regulated through the expression of chemokines, which can be induced by almost all types of liver injury. These chemokines modulate a number of critical biological actions, including angiogenesis (64). Leucocytes affect many angiogenic processes given the fact that they can produce a myriad of angiogenic factors, such as VEGF, PlGF, PDGF, FGF, Ang-2, TGF-β1, epidermal growth factor and various interleukins (49). The angiogenic and inflammatory pathways are tightly cross linked together (Fig. 3). In the event of hypoxia, HIF-1α will not only induce an upregulation of angiogenic factors, it also stimulates the NF-κβ pathway, thus inducing an inflammatory action. Furthermore, both events are capable of sustaining each other (65) (Fig. 2).
Angiogenesis in chronic liver disease
The progression from fibrosis to cirrhosis, the end-point of CLDs, is distinguished by a prolonged inflammatory and fibrogenic process that leads to an abnormal angioarchitecture distinctive for cirrhosis. Several mechanisms are responsible for the angiogenic switch during the pathogenesis of CLDs (Fig. 4). First, CLDs are characterised by chronic inflammation (66). Increased intrahepatic vascular resistance is primarily caused by anatomical changes, such as fibrotic scar tissue compressing portal and central venules. In addition, the formation of fibrotic septa, as well as sinusoidal capillarisation, can result in an increased resistance to blood flow and oxygen delivery. This results in hypoxia and the transcription of hypoxia-sensitive pro-angiogenic genes, usually modulated trough HIF. Also, an increased contribution of the hepatic artery to the sinusoidal blood flow leads to sensitisation of the hepatocytes to abnormal high-oxygen conditions (67).
Hepatic stellate cell motility and migration promote coverage of HSC around sinusoids, causing sinusoidal constriction and contributing to the hepatic resistance in cirrhosis. The characteristic fenestrated phenotype of the sinusoidal ECs is lost and an organised basement membrane is established, which leads to an impairment of oxygen diffusion even though the increased arterial flow provides a high supply of oxygenated blood (68). The deposition of collagen in the space of Disse accentuates the narrowing and distortion of the sinusoidal lumen, further restricting microvascular blood flow. This is aggravated by leucocytes, either mechanically trapped in the narrowed sinusoids or adhering to the endothelium, as a result of activation of a hepatic microvascular inflammatory response. The hypoxic liver tissue causes an up-regulation of VEGF, Ang and their receptors in HSCs, not only enhancing the hypoxia-induced angiogenesis but also stimulating activation and migration of HSC (69, 70). As stated previously, activated HSCs induce an inflammatory response and enhance angiogenesis (71–73). New vessels themselves can significantly contribute to the perpetuation of the inflammatory response by expressing chemokines and adhesion molecules promoting the recruitment of inflammatory cells. Furthermore, angiogenesis, early in the course of a CLD, may contribute to the transition from acute to chronic inflammation (74).
The close relationship between the progression of CLDS and angiogenesis brings about two potential clinical goals: (i) detection of selected pro-angiogenic molecules that may serve as a non-invasive way to monitor both disease progression, as well as therapeutic response; (ii) anti-angiogenic therapy may be an effective tool for blocking or slowing down fibrogenic progression of CLDs. Studies in several animal models have shown the potential benefits of anti-angiogenic therapy in CLDs (75–78). A study performed on chronic hepatitis C virus (HCV)-infected patients has tried to correlate circulating levels of molecules related to angiogenesis, disease progression and efficacy of standard pegylated interferon α-2b plus ribavirin therapy (79). Thirty-six HCV-infected patients with all possible grades of inflammation, stages of fibrosis (except established cirrhosis) and viral genotypes were included in the study. Serum levels of VEGF, PlGF, Ang-2 and soluble Tie-2 were determined, before and after therapy. A sustained virological response to antiviral combination therapy was observed in 18 CHC patients, defined by the absence of detectable HCV RNA in the serum at 24 weeks after the end of treatment. The increased VEGF and Ang-2 levels were significantly decreased after therapy in these patients.
Angiogenesis in portal hypertension
Portal hypertension is one of the more common and severe complications that develop in patients with CLDs. Development of a hyperdynamic splanchnic circulatory state is the foremost component of the portal hypertensive syndrome. The increased blood flow in splanchnic organs draining into the portal vein, and subsequent increase in portal venous inflow, aggravates the portal hypertensive syndrome, especially when portal–systemic collaterals are widespread (80, 81). The development of hyperdynamic splanchnic circulation is an active modulated angiogenic process induced by VEGF (4, 82), PDGF (83) and PlGF (3). In addition, VEGF- and PlGF-dependent angiogenesis also plays an important role in the formation of portal–systemic collateral vessels (3–4). Inhibition of the mTOR pathway with rapamycin prevents mesenteric neo-angiogenesis and reduces the splanchnic blood flow in portal hypertensive mice (78). The precise mechanisms by which the angiogenesis-associated response in PH is modulated remain to be defined. Several factors involved in the pathogenesis of PH, such as hypoxia, oxidative stress (84), inflammation (85) and shear stress are potential mediators for the angiogenic response (3) (Fig. 4).
Angiogenesis in non-alcoholic steatohepatitis
Pathogenesis of non-alcoholic fatty liver disease and non-alcoholic steatohepatitis
The liver is an important organ that has several pivotal functions, such as detoxification, synthesis of bile, proteins, lipids and carbohydrates. When an imbalance between production and utilisation of triglycerides in the liver metabolism occurs, lipids can accumulate in the cytoplasm of hepatocytes and result in hepatic steatosis (86). Historically, steatosis was considered as a non-progressive benign disease. Nowadays, NAFLD is considered to be a chronic inflammatory disease. The most severe form of NAFLD is NASH given its progressive disease course and considering that NASH can evolve from steatohepatitis to fibrosis, cirrhosis and eventually can lead to HCC (6, 87) (Fig. 4). The two-hit theory, proposed in 1998 by Day and James, is the first theory that gave a plausible explanation for the pathogenesis of NASH (88). The first hit is hepatic steatosis, which is an accumulation of intrahepatic triglycerides, which can lead to lipotoxicity. The second hit consists of lipid peroxidation and the production of cytokines and ROS mediating liver injury (88).
The pathogenesis of NASH is still not fully understood, the recognised mechanisms as stated above do not fully explain the range of symptoms and physiological processes found in the disease progression. Nonetheless, the pathophysiology of NASH should probably be approached as a multifactorial process. In several stages of NASH, there can perhaps be made a link between the disease progression and hepatic microvasculature changes.
Effect of non-alcoholic steatohepatitis on the liver microvasculature
To develop NASH, a switch from simple steatosis to steatohepatitis should occur. Fat accumulation damages the hepatocytes leading to deregulation of the microvascular blood flow and lipotoxicity. Reduction in sinusoidal perfusion initially arises from the effects of hepatocytes swollen with accumulated lipids. This results in reduction of the intrasinusoidal volume, as well as altering the sinusoidal-architecture.
Fat accumulation in the cell might continue to induce lipotoxicity, either directly or through sensitisation of other agents (89). Lipotoxicity activates cytokines, which will subsequently induce the recruitment of inflammatory cells and platelets. Inflammation triggers vascular permeability by recruiting monocytes, macrophages, platelets, mast cells and other leucocytes. As stated previously, these cells are able to initiate angiogenesis through different pathways. In this manner, inflammation can contribute to the formation of new vascular structures in the liver (90) (Fig. 3).
In more severe fatty liver disease, simple steatosis may progress over steatohepatitis to fibrosing steatohepatitis. This stage is characterised by adhesion of leucocytes to the sinusoidal endothelium, followed by leucocyte infiltration into the hepatic parenchyma to form inflammatory foci (91). Enlarged, fat-laden and inflamed hepatocytes together with perivascular fibrosis, narrow the sinusoidal lumens, making vessels more tortuous and impairing sinusoidal perfusion (91). Chronic liver damage might initiate physiological hepatic angiogenesis leading to the formation of new functional sinusoids (90).
Angiogenesis in non-alcoholic steatohepatitis
Nowadays, some studies suggest that angiogenesis might play a role in the progression of NASH. This was first brought to light by a study of Kitade et al. (92), which suggested that leptin-mediated neovascularisation, coordinated by VEGF, plays an important role in the development of liver fibrosis and HCC in a rat model for NASH. A macroarray gene expression analysis on the liver of obese patients with severe NASH, showed that VEGF, TGF-β1, connective tissue growth factor and FGF were over-expressed, compared with control patients (93). More recently, Kitade et al. described a significant up-regulation of CD34 expression, which is widely used as a marker of neovascularisation, in liver biopsies of patients with NASH. The increase in neovascularisation was almost parallel to the development of liver fibrosis, whereas almost no neovascularisation could be observed in patients with simple steatosis or a healthy liver (94). These new findings could give a new perspective to investigate the pathophysiology of NASH.
Angiogenesis in hepatocellular carcinoma
Hepatocellular carcinoma is a primary malignancy that mostly emerges on a background of CLDs (Fig. 4). It is one of the most vascular solid tumours in which angiogenesis plays an important role in its development, progression and metastasis (95). The hypervasculature found in CLDs facilitates the progression from small dysplastic nodules, through neoplastic lesions to large HCC-tumours. Dysplastic nodules in the premalignant (cirrhotic) environment are responsible for the angiogenic switch and angiogenesis carries on throughout the process of tumour progression (96) (Fig. 4). The phenotypic switch to angiogenesis is linked to a disturbance in the equilibrium between positive and negative angiogenic regulators (97). Tumours secrete a number of angiogenic growth factors, such as VEGF, PDGF, PlGF and TGF-β1. Furthermore, the expression of endogenous inhibitors, for instance TSP-1, endostatin and angiostatin, are downregulated (98). This activates ECs and basement membranes to remodel existing vessels, and stimulates the release of endothelial progenitor stem cells from the bone marrow to form new vessels. HCC-cells display rapid growth and are consequently in need for a high oxygen and nutrient supply. Hence, tumour cells induce the formation of new blood vessels to fulfil this aspiration. HCC lesions are characterised by arterial hypervascularity to provide the tumour with highly oxygenated blood (99). However, these new vessels are marked by a disorganised vasculature, consisting of leaky, haemorrhagic and torturous vessels. The irregularity of tumour vessels results in chaotic blood flow with poorly oxygenated blood. Furthermore, these leaky blood vessels provide a passage for tumour cells into the blood circulation, and hence, facilitate metastasis.
Vascular endothelial growth factor is one of the uttermost important factors involved in the angiogenic switch of HCC. Already in an early phase of CLDs an upregulation of VEGF is observed (100, 101), an increase that is tremendously enhanced during HCC progression (102–104). The degree of VEGF expression increases with tumour stage (96) and is associated with poor prognosis (105–108). Furthermore, VEGF levels have been correlated to vascular invasion (109, 110), metastasis (109, 111, 112), recurrence (113, 114), vascular density (109), differentiation (115) and tumour aggressiveness (109, 116). Recently, the importance of PlGF in the pathogenesis of CLD has been established. Both in human (117) and mouse (118), PlGF levels are strongly upregulated in HCC. Although PDGF is not a major regulator of angiogenesis in HCC, it has been shown that PDGF expression is associated with venous invasion and is linked to an increased metastatic potential (119).
Inhibition of angiogenesis as systemic therapy for hepatocellular carcinoma
Hepatocellular carcinoma is potentially curable by surgical resection and liver transplantation. However, the majority of patients present with advanced-stage disease, which is most commonly accompanied by severe CLDs. Therefore, surgery is only feasible for a small fraction of patients. Moreover, systemic chemotherapy has very limited impact on advanced HCC, partly because HCC is a chemotherapy-resistant tumour (120, 121). Furthermore, the underlying cirrhosis in most patients may lead to PH with hypersplenism, varices and gastrointestinal bleeding, hypoalbuminemia, hepatic encephalopathy, altered drug binding, distribution and pharmacokinetics, limiting the selection and optimal dosing of most cytotoxic agents (122). Until recently, systemic therapy of advanced HCC provided marginal benefit. Since the discovery of anti-angiogenic agents as potential inhibitors for tumour growth (123), anti-angiogenic therapy has fundamentally landmarked the clinical new era of anticancer therapy (Table 1).
Table 1. Overview of current trials with anti-angiogenic agents as potential therapies against hepatocellular carcinoma
Mechanism of action
Terminated due to adverse events related to sunitinib.
Significant progress on the treatment of advanced HCC has been made by the introduction of sorafenib. Sorafenib is a small molecular inhibitor that targets several tyrosine protein kinases in the Raf/MEK/ERK-pathway (antiproliferative effect); and PDGF, VEGFR-1 and VEGFR-2 (anti-angiogenic effect). By inhibiting angiogenesis, the tumour is withheld from its blood supply and therefore its growth is inhibited (124). In vitro, sorafenib has been shown to stop tumour growth and induced apoptosis of HCC cell lines, and thus inhibited the growth of HCC xenografts (125, 126). A large phase III clinical trial (SHARP) was conducted in 602 patients with advanced HCC resulting in a significant increase of survival compared with placebo (127). Sorafenib has become the standard of care for patients with advanced HCC and also for those progressing after loco-regional therapies.
Limitations of anti-angiogenic therapy
The use of VEGF pathway inhibitors to impair angiogenesis now represent a clinically validated treatment for a variety of solid tumours. Targeting of a single angiogenic protein, such as VEGF, might initially be effective, but ultimately leads to the failure because of escape mechanisms. Escape from anti-angiogenesis therapy is likely to involve multiple mechanisms (171).
The benefits of VEGF-targeted agents in the treatment of late-stage cancers can be temporary, resulting in eventual drug resistance, tumour growth and/or regrowth and rapid vascular recovery when therapy is stopped. Ebos et al. and Pàez-Ribes et al. (172, 173) demonstrated that angiogenesis inhibition in mice can lead to opposing effects on tumour growth and metastasis depending on tumour stage and treatment duration. Anti-angiogenic therapy disturbs the tumour vasculature and in some cases (e.g. sunitinib) also disrupts pericyte coverage, which destabilises vessels, makes them more leaky and immature, and hence facilitates intravasation of tumour cells and metastasis (174).
The hypoxic conditions created by anti-angiogenic therapy may select for more invasive tumour variants better adapted to survive and proliferate under reduced oxygen tension, leading to increased intravasation and metastatic dissemination (175–178).
Ebos et al. (172) show that pretreatment of healthy mice with VEGF inhibitors (before tumour cell implantation) promotes metastasis. A possible hypothesis is the creation of a ‘premetastatic niche’. Administration of VEGF inhibitors prunes quiescent vessels in healthy tissues, which will show rapid rebound growth after withdrawal of therapy (179) and such a well-vascularised niche may promote metastasis. Furthermore, VEGF inhibitors are known to induce a state of chronic inflammation, which leads to the upregulation of several inflammatory cytokines that recruit angiogenic bone marrow-derived progenitors (180) and might promote the formation of a premetastatic niche (181).
The close relationship between the progression of CLDs and angiogenesis emphasises the need for anti-angiogenic therapy as a tool for blocking or slowing down the fibrogenic progression. The fact that angiogenesis plays a pivotal role in CLDs gives rise to new opportunities for treating CLDs and its complications. Studies in experimental models for PH and cirrhosis showed that anti-angiogenic treatment with a number of receptor tyrosine kinase inhibitors, including imatinib, sunitinib, sorafenib and soluble anti-PlGF antibodies improved the disease progression (77). Sorafenib became a breakthrough in the field of liver cancer; the improvement in median survival (almost 3 months) represented a reduction of >30% of the probability of death along the follow-up time. Sorafenib has now become the standard-of-care for patients with advanced HCC. The positive data regarding sorafenib have confirmed the importance of molecular targeted therapies in HCC and opened the quest for new developments that should exceed the current benefits (182). Other agents that are under clinical investigation may exceed the activity of sorafenib and are currently being tested as second-line agents or in combination with sorafenib. Therefore, anti-angiogenic therapies are considered as a promising tool in managing CLDs and its complications.
When using anti-angiogenic treatments in patients with CLDs, some aspects should be taken into consideration. Firstly, the side effects of long-term administration of anti-angiogenic agents remain unknown. Secondly, the interactions between anti-angiogenic therapies and common medications would also require further investigation. Thirdly, the tolerable and safe dose needed to properly modulate fibrosis and hepatic hemodynamics should be closely regulated. Although, often described in the literature as ‘non-toxic’ compounds, angiogenic inhibitors exhibit dose-related adverse effects, such as fatigue, fever, sedation, hypertension, thrombocytopaenia, leucopenia and liver-function abnormalities. Most anti-angiogenic therapies have been tested extensively – and often almost exclusively – in cancer-related malignancies. Thus, it should be taken in consideration that the tolerance to side effects in patients without cancer, could be different than in patients with an advanced tumour (183).
Currently, liver biopsies are the golden standard for disease monitoring, although this invasive technique has several disadvantages, for instance abdominal bleeding. Consequently, there is a vast need for non-invasive markers to identify the exact stage of disease progression in CLDs. The use of angiogenic molecules as markers for disease progression and response to antiviral therapies in hepatitis C patients, has recently been suggested (79). VEGF, PlGF, Ang-2 and PDGF expression in HCC was found to correlate with the tumour progression and metastasis. These findings confirm that the detection of angiogenic molecules might serve as potential non-invasive markers to monitor disease progression and to select for potential responders to therapy, in several CLDs and their complications.
In conclusion, the close relationship between angiogenesis in the progression of CLDs, offers multiple clinical applications. Firstly, anti-angiogenic therapies can be used to manage CLD progression, ranging from fibrosis, cirrhosis, NASH to HCC. Secondly, these anti-angiogenic drugs could be used as a preventive strategy against complications of cirrhosis. Thirdly, angiogenic molecules might be of great importance to detect CLD progression and potential responders to therapeutic agents.
The authors wish to thank Tim Taelman for his contribution to the illustrations, and the University Hospital Ghent, Department of Gastroenterology and Hepatology for the material support. S. C. received a scholarship (BOF 09/24J/012) from the University Ghent Research Fund (BOF). F. H. received a scholarship (FWO09/ASP/161) from the Fund for Scientific Research (FWO-Flanders).
Financial disclosures: All the authors have no financial disclosures to a company related to the submitted review or with a company making a competing product.