SEARCH

SEARCH BY CITATION

Keywords:

  • angiogenesis;
  • nitric oxide;
  • portosystemic collateral circulation;
  • vasodilation

Abstract

  1. Top of page
  2. Abstract
  3. Intrahepatic circulation
  4. Splanchnic and systemic circulations
  5. Future direction
  6. Acknowledgment
  7. References

Portal hypertension is caused by an increased intrahepatic resistance, a major consequence of cirrhosis. Endothelial dysfunction in liver sinusoidal endothelial cells (LSECs) decreases the production of vasodilators, such as nitric oxide, and favours vasoconstriction. This contributes to an increased vascular resistance in the intrahepatic/sinusoidal microcirculation and develops portal hypertension. Portal hypertension, in turn, causes endothelial dysfunction in the extrahepatic, i.e. splanchnic and systemic, circulation. Unlike dysfunction in LSECs, endothelial dysfunction in the splanchnic and systemic circulation causes overproduction of vasodilator molecules, leading to arterial vasodilation. In addition, portal hypertension leads to the formation of portosystemic collateral vessels. Both arterial vasodilation and portosystemic collateral vessel formation exacerbate portal hypertension by increasing the blood flow through the portal vein. Pathological consequences, such as oesophageal varices and ascites, result. While the sequence of pathological vascular events in cirrhosis and portal hypertension has been elucidated, the underlying cellular and molecular mechanisms causing endothelial dysfunctions are not yet fully understood. This review article summarizes the current cellular and molecular studies on endothelial dysfunctions found during the development of cirrhosis and portal hypertension with a focus on the intra- and extrahepatic circulations. The article ends by discussing the future directions of the study for endothelial dysfunction.

Abbreviations
α-SMA,

alpha-smooth muscle actin;

BH4,

tetrahydrobiopterin;

CO,

carbon monoxide;

EMT,

epithelial–mesenchymal transition;

EndoMT,

endothelial–mesenchymal transition or endothelial–myofibroblast transition;

eNOS,

endothelial nitric oxide synthase;

EPC,

endothelial progenitor cell;

ET-1,

endothelin-1;

ETR,

endothelin receptor;

FSP-1,

fibroblast specific protein-1;

GTP-cyclohydrolase,

guanosine triphosphate-cyclohydrolase;

H2S,

hydrogen sulfide;

HSC,

hepatic stellate cell;

KLF,

Kruppel-like factor;

MDA,

malondialdehyde;

miRNA,

microRNA;

MyD88,

myeloid differentiation protein;

NO,

nitric oxide;

O2,

superoxide radical;

ONOO,

peroxynitrite;

PIGF,

placental growth factor;

SOD,

superoxide dismutase;

TLR,

toll-like receptor;

VEGF,

vascular endothelial growth factor.

Portal hypertension is a detrimental complication resulting from cirrhosis (1, 2). Intra- and extrahepatic endothelial dysfunction is a key factor that causes and worsens portal hypertension (3) (Fig. 1). In the intrahepatic microcirculation, hypoactive endothelial cells contribute to an increased intrahepatic resistance mainly by decreasing nitric oxide (NO) production, which in turn initiates portal hypertension. Portal hypertension, once it develops, affects extrahepatic vascular beds in the splanchnic and systemic circulation, leading to arterial vasodilation and collateral vessel formation, resulting in greater blood flow into the portal vein. This increase in portal blood flow further exacerbates portal hypertension (1, 2). In contrast to hypoactive endothelial cells in the intrahepatic microcirculation, endothelial cells in the splanchnic and systemic circulation are hyperactive and increase NO production. Thus, endothelial cells with opposing phenotypes are found in the intra- vs. extrahepatic circulation. Both contribute to the development and exacerbation of portal hypertension (3).

image

Figure 1.  Overview of the development and consequences of portal hypertension in cirrhosis. Endothelial dysfunction plays important roles in the pathophysiology of portal hypertension. LSEC; liver sinusoidal endothelial cell.

Download figure to PowerPoint

With knowledge of vascular biology, our understanding of the pathogenesis of portal hypertension has significantly advanced, revealing how vascular abnormalities both in and outside the liver contribute to portal hypertension, i.e. how these endothelial dysfunctions relate to those vascular abnormalities. However, how these endothelial dysfunctions occur in cirrhosis and portal hypertension still remains to be elucidated, particularly at the cellular and molecular levels.

This review article thus summarizes current cellular and molecular studies on endothelial dysfunctions in cirrhosis and portal hypertension, first in the area of the intrahepatic/sinusoidal microcirculation and second in the extrahepatic, i.e. splanchnic and systemic, circulation. This article concludes with a discussion of the future directions that the study of endothelial dysfunctions and the vascular abnormalities associated with them will take in relation to cirrhosis and portal hypertension. This review article focuses on endothelial cells. However, other cell types, such as hepatic stellate cells (HSCs), smooth muscle cells and Kupffer cells, are equally important to endothelial dysfunctions and the subsequent vascular changes found in cirrhosis and portal hypertension.

Intrahepatic circulation

  1. Top of page
  2. Abstract
  3. Intrahepatic circulation
  4. Splanchnic and systemic circulations
  5. Future direction
  6. Acknowledgment
  7. References

Overview – hypoactive endothelial cells

Endothelial cells are the first line of defense protecting the liver from injury (3). Thus, endothelial cells are vulnerable to various types of liver injury. Endothelial dysfunction seen in the intrahepatic/sinusoidal microcirculation during the development of cirrhosis and portal hypertension is largely because of hypoactive endothelial cells. Endothelial dysfunction by hypoactive endothelial cells, together with increased intrahepatic contractility because of activated HSCs (4) and fibrosis, contributes to an increased intrahepatic vascular resistance and consequently leads to portal hypertension (1–3). These activated contractile HSCs exhibit a decreased response to vasodilators, such as NO (4). Furthermore, it is suggested that an increased recruitment of these activated HSCs around newly formed sinusoidal vessels contributes to an increased intrahepatic vascular resistance in cirrhosis (5, 6). Liver sinusoidal endothelial cells (LSECs) facilitate recruitment of activated HSCs through paracrine signal (7).

The major hepatic endothelial cells are LSECs, comprising ∼98% of the total number of endothelial cells found in the non-parenchymal cell population of the liver. Lymphatic and non-sinusoidal endothelial cells comprise the remainder (8). Thus, the endothelial dysfunction observed in the intrahepatic circulation is apparently because of LSECs. LSECs have far-reaching effects on liver functions such as blood clearance, vascular tone, immunity, hepatocyte growth (8) and angiogenesis/sinusoidal remodelling (6, 9). LSEC dysfunction thus results in a pathology contributing to impaired vasomotor control (primarily vasoconstrictive), inflammation, fibrosis and impaired liver regeneration (1–3, 10). These physiological impairments facilitate the development of cirrhosis and portal hypertension. Recent findings regarding the factors that cause LSEC dysfunction and its consequences are discussed below (Fig. 2).

image

Figure 2.  Overview of liver sinusoidal endothelial cell (LSEC) dysfunction in the intrahepatic/sinusoidal circulation leading to increased intrahepatic resistance and consequent portal hypertension. SOD; superoxide dismutase, O2; superoxide radical, MDA; malondialdehyde, TLR4; toll-like receptor 4, MyD88; myeloid differentiation protein, NO; nitric oxide, TXA2; thromboxine A2, KLF2; Kruppel-like factor 2.

Download figure to PowerPoint

Liver sinusoidal endothelial cells dysfunctions

Oxidativex stress attenuates nitric oxide bioavailability and endothelial nitric oxide synthase (eNOS) activity in liver sinusoidal endothelial cells

Oxidative stress causes endothelial dysfunction. In fact, patients with cirrhosis have significantly elevated levels of circulating malonic dialdehyde (MDA), an indicator of oxidative stress (11). It was demonstrated that administration of the anti-oxidant vitamin C to cirrhotic patients markedly attenuates post-prandial increases in hepatic venous pressure, i.e. portal pressure, and significantly decreases MDA levels, suggesting that increased oxidative stress in cirrhotic patients contributes to portal hypertension. In the LSECs of cirrhotic livers, there is an increased production of the superoxide radical (O2), which reacts with NO to form peroxynitrite (ONOO). Consequently, NO bioavailability in the intrahepatic circulation is decreased (12) favouring vasoconstriction. Xanthine oxidase and cyclooxygenase (COX), but not nicotinamide adenine dinucleotide phosphate (NADPH) oxide, have been shown to be the primary causes of these elevated O2 levels. Furthermore, decreased activity of superoxide dismutase (SOD), an O2 scavenging enzyme, also contributes to elevated O2 levels in cirrhotic livers (12). Adenoviral delivery of SOD to cirrhotic livers attenuates O2 levels and can ameliorate portal hypertension (13). All these observations indicate that oxidative stress in LSECs decreases NO bioavailability and that anti-oxidants such as vitamin C and SOD help to reduce portal hypertension.

It is also known that oxidative stress impairs eNOS activity in at least three ways(14): (i) by increasing association of eNOS with caveolin-1, an inhibitor of eNOS activity; (ii) by inhibiting endothelin-1 (ET-1)-induced eNOS phosphorylation; (iii) by increasing dissociation of eNOS from endothelin receptor B (ETRB), presumably the ETRB1 component. Thus, oxidative stress, by increasing O2 levels and decreasing eNOS activity, leads to decreased NO production, a major characteristic of endothelial dysfunction (14, 15).

Toll-like receptor signalling in liver sinusoidal endothelial cells enhances angiogenesis

Liver sinusoidal endothelial cells express many types of Toll-like receptors (TLRs) (9, 16). Particularly, TLR4, a receptor of endotoxin, regulates angiogenic responses (9). Using TLR4-mutant mice, which express a spontaneous mutation conferring loss of TLR4 function, Jagavelu et al. (9) showed that LSECs isolated from TLR4-mutant mice diminish tubulogenesis, an indicator of angiogenic capacity of vascular cells, in response to the bacterial endotoxin, lipopolysaccharide (LPS). TLR4 conveys downstream signals through an adapter molecule, myeloid differentiation protein 88 (MyD88) and a MyD88-independent pathway (17). Overexpression of MyD88 in human LSECs results in enhanced tubulogenesis. The peptide IMG-2005-1 blocks MyD88 function by inhibiting homodimerization of MyD88 and suppresses tubulogenesis. Given that angiogenesis is postulated to contribute to portal hypertension by enhancing fibrogenesis (6, 18), these observations suggest that LSECs, being the first cells exposed to portal venous LPS, mediate angiogenesis through the TLR4/MyD88 signaling pathway and result in fibrosis and portal hypertension.

Mechanical stimulus such as shear stress changes gene expression in liver sinusoidal endothelial cells

Sinusoidal distortion resulting from fibrosis/cirrhosis alters the flow pattern in the intrahepatic microcirculation. LSECs sense these haemodynamic changes and influence cellular functions by changing their gene expression (19). For example, Kruppel-like factors (KLFs) are transcription factors that regulate cellular growth and tissue development. One of the KLFs, KLF2, is highly expressed in vascular endothelial cells and protects the endothelium by upregulating the expression of a wide variety of vasoprotector genes (20, 21), including that for eNOS (22).

Shear-stress is the most potent known inducer of KLF2 expression (19, 23, 24). A study by Gracia-Sancho et al. (19) showed that LSECs express KLF2 in response to shear stress and that KLF2 expression is increased in cirrhotic livers. This increased KLF2 level then changes gene expression patterns in LSECs by increasing expression of its target vasoactive agents, such as eNOS, thrombomodulin (a blood coagulation inhibitor) and c-type natriuretic peptide (CNP). Furthermore, as there is no increase in eNOS and CNP protein levels in cirrhotic livers, changes in the stability of eNOS mRNA and CNP mRNA may also play a role in their final levels. In contrast, thrombomodulin protein levels remain significantly higher in cirrhotic livers compared with normal ones (19). Collectively, these observations indicate that changes in flow-mediated mechanical forces may influence not only the gene expression of vasoactive molecules via modulating KLF2 levels, but also the stability of mRNA in cirrhotic livers.

Liver sinusoidal endothelial cell dysfunction impairs regulation of vascular tone

Liver sinusoidal endothelial cells play an important role in the regulation of hepatic vascular tone by releasing vasoactive molecules (1, 2). However, LSECs lose their vasodilatory properties in cirrhosis and portal hypertension. They decrease the production/bioavailability of vasodilator molecules (such as NO) and increase the levels of vasoconstrictor molecules (such as thromboxane A2), thereby increasing intrahepatic resistance.

Nitric oxide (NO) is probably the most important vasodilator regulating vascular tone in the intrahepatic microcirculation. It is produced primarily by eNOS (25, 26) in LSECs. In cirrhotic livers, NO levels are decreased, which causes an increased intrahepatic vascular resistance. As mentioned earlier, NO levels are decreased at least in two manners. One is a decreased bioavailability of NO because of increased superoxide radicals (O2) (12) which quench NO. The other is because of a decreased activity of eNOS itself (3, 27, 28). In general, the level of eNOS does not differ between normal and cirrhotic livers (19, 26). In cirrhotic livers, however, complex post-translational modifications of eNOS decrease its enzymatic activity (3, 27, 28). For example, an increased association of eNOS with caveolin-1, a negative regulator of eNOS, inhibits NO production (26). Another important regulatory pathway of eNOS is via ET-1. ET-1 has dual vasoactive effects (Fig. 3). It promotes vasoconstriction by binding to two cell surface receptors on hepatic stellate cells: endothelin A receptor (ETRA) and endothelin B2 receptor (ETRB2) (29, 30). ET-1 also promotes vasodilation by binding to the endothelin B1 receptor (ETRB1) on LSECs and thus activating eNOS (31). However, endotoxin levels, elevated in cirrhotic livers, increase caveolin-1 levels, which facilitate caveolin-1's binding to eNOS and consequently block ET-1-induced eNOS activation in LSECs (32, 33). Similarly, the influence of decreased high-density lipoprotein (HDL) levels has also been pointed out. Thabut et al. (34) report that decreased HDL levels in cirrhotic rats lead to an increase in serum endotoxin (i.e. LPS) levels. As a result, the increased endotoxin level facilitates caveolin-1's association with eNOS and decreases phosphorylation of Akt (an activator of eNOS), thereby decreasing eNOS activity in cirrhotic livers (34).

image

Figure 3.  Endothelin-1 (ET-1) mediates both vasodilation and vasoconstriction through ET-1 receptors in the intrahepatic/sinusoidal microcirculation. Endothelin type-A receptors (ETRAs), residing on hepatic stellate cells (HSCs) and vascular smooth muscle cells, mediate vasoconstriction. ET type-B receptors (ETRBs) induce both vasoconstriction and vasodilation depending on their cellular location. ETB1 receptors, which reside on endothelial cells, cause vasodilation through nitric oxide (NO) production, whereas ETB2 receptors on HSCs and smooth muscle cells induce vasoconstriction, similar to ETA receptors (29–31).

Download figure to PowerPoint

Thromboxane A2 (TXA2), a vasoconstrictor molecule, is generated by the action of COX-1 in LSECs. In cirrhotic livers, increased TXA2 production contributes to an increased intrahepatic resistance. It was shown that blocking TXA2 (by a prostaglandin H2/TXA2 receptor blocker, SQ-29548) or inhibiting COX-1 activity (by a COX-1 inhibitor, SC-560) can be beneficial for attenuating an increased intrahepatic resistance in fibrotic/cirrhotic livers (35, 36).

Liver sinusoidal endothelial cells dysfunction facilitates hepatic inflammation

Liver sinusoidal endothelial cells are primary mediators of hepatic immune tolerance. Injury from hepatotoxins causes LSECs to get defenestrated (37) thus promoting inflammation through the secretion of an array of cytokines and chemokines (38). Accordingly, these series of changes in LSECs enhance the capacity for antigen capture and induce T-cell proliferation. Furthermore, in fibrotic livers, as a result of these changes in LSECs, LSECs do not reject dendritic cell priming of T-cells and this subsequently enhances immunogenicity. Consequently, inflammation in the intrahepatic microcirculation is increased in cirrhotic livers (39).

One recently proposed mechanism of regulation of immunogenicity involves a molecule expressed in LSECs, called LSEC-specific lectin (LSECtin) (40). LSECtin recognizes activated T-cells and inhibits their immune response. Mice lacking LSECtin gene exhibit accelerated T-cell-mediated immune response, which enhances liver injury. Conversely, exogenous administration of recombinant LSECtin protein or plasmid ameliorates liver injury by suppressing the T-cell immune response. It is not known whether LSEC dysfunction decreases LSECtin levels in LSECs in cirrhosis. However, this study indicates that molecules such as LSECtin expressed in LSECs ameliorate inflammation and could provide a novel approach for the treatment of cirrhosis and portal hypertension.

Liver sinusoidal endothelial cells defenestration promotes hepatic dysfunction

Pathophysiological alterations of LSEC structure result in a wide range of adverse effects on the liver and on metabolism (41). LSECs are perforated with fenestrations, pores that facilitate the transfer of lipoproteins and macromolecules between the blood and hepatocytes. Loss of this pore structure, termed defenestration, is caused by exposure to various pathological endogenous and exogenous agents and significantly impairs hepatic function.

Defenestration entails endothelial thickening and the deposition of excessive extracellular matrix in the subendothelial space of Disse. Altogether, these changes are referred to as cirrhotic capillarization and impede the transfer of many substrates from the sinusoidal lumen to hepatocytes through the space of Disse. This capillarization is also thought to contribute to an increase in intrahepatic resistance (1). The review article by Cheluvappa et al. (41) details LSECs and fenestration.

Summary – intrahepatic microcirculation

Liver sinusoidal endothelial cells regulate a wide range of liver functions. In response to various agents, such as bacterial endotoxin, viruses, drugs and ethanol, LSECs experience oxidative stress (11–13). This results in the activation of inflammatory pathways, such as TLR4/MyD88 signalling (9, 42), which causes the LSECs to become dysfunctional (Fig. 2). Furthermore, mechanical stimuli, such as shear stress, change the gene expression pattern via a transcription factor KLF2 in cirrhotic livers and worsen LSEC dysfunction (19). Dysfunction of LSECs in the intrahepatic circulation is characterized particularly by vasoconstriction brought about by decreased levels of vasodilators (e.g. NO) (2, 3, 25–27) and/or by increased levels of vasoconstrictors (e.g. TXA2) (35, 36). Furthermore, LSEC dysfunction results in increased inflammation because of impaired immune tolerance and defenestration (37–39, 41). LSEC dysfunction thus facilitates cirrhotic fibrosis. This results in angiogenesis and pathological sinusoidal remodelling, which further worsen LSEC dysfunction and increase intrahepatic resistance, leading to portal hypertension. Although the discussion above is limited to the role LSECs dysfunction plays in increased intrahepatic resistance in cirrhosis and portal hypertension, other cells such as HSCs and Kupffer cells are also important in this regard.

Splanchnic and systemic circulations

  1. Top of page
  2. Abstract
  3. Intrahepatic circulation
  4. Splanchnic and systemic circulations
  5. Future direction
  6. Acknowledgment
  7. References

Overview – hyperactive endothelial cells

An increase in portal pressure triggers mechanical signals that induce diverse vascular changes in the splanchnic and systemic circulation (Fig. 4). There are two major vascular changes that develop in response to an increase in portal pressure. One is arterial vasodilation, and the other is formation of portosystemic collateral vessels. These vascular changes increase the blood flow to the portal vein, thereby exacerbating portal hypertension, as reviewed elsewhere (1, 2, 43, 44). In contrast to hypoactive LSECs in the intrahepatic microcirculation, hyperactive endothelial cells characterized by increased NO production play a critical role in the vascular changes in the splanchnic and systemic circulation (3, 45). Molecular and cellular mechanisms underlying hyperactive endothelial cells and subsequent vascular changes, which in turn worsen portal hypertension, are discussed in this section.

image

Figure 4.  Overview of endothelial dysfunction/hyperactivation in the arterial splanchnic and systemic circulation, leading to an increased blood flow through the portal vein and a consequent worsening of portal hypertension. VEGF; vascular endothelial growth factor, PIGF; placental growth factor, NO; nitric oxide, PGI2; prostacyclin, CO; carbon monoxide, EDHF; endothelium-derived hyperpolarizing factor.

Download figure to PowerPoint

Portal pressure triggers endothelial dysfunction/hyperactivation in the splanchnic and systemic circulation

A study by Abraldes et al. (46) demonstrated that portal pressure is sensed at different vascular beds depending on the severity of portal hypertension (Fig. 5). A small increase in portal pressure is first sensed by the intestinal microcirculation, followed by the arterial splanchnic circulation (e.g. the mesenteric arteries) and then the arterial systemic circulation (e.g. the aorta). Thus, the intestinal microcirculation serves as a ‘sensing organ’ to portal pressure (47). These observations were demonstrated using a surgical technique termed partial portal vein ligation (PVL) (2), which enables varying degrees of portal hypertension to be created in a rat model. The portal vein is ligated along with a needle. When the needle is removed at the end of surgery, a luminal opening corresponding to the needle gauge is created. Therefore, using different needle gauges, different stages of portal hypertension can be induced in proportion to the degree of the luminal diameter. This model allows us to determine how specifically changes in portal pressure affect local vascular beds/endothelial cells without influencing liver function (2).

image

Figure 5.  Portal pressure is sensed at different vascular locations depending on the severity of portal hypertension. Local nitric oxide production is affected by the severity of portal hypertension. The intestinal microcirculation may be the most sensitive to changes in portal pressure, followed by the arterial splanchnic circulation (e.g. the mesenteric arteries) and then arterial systemic circulation (e.g. the aorta). The diagram adapted from Iwakiri (47).

Download figure to PowerPoint

When mild portal hypertension is induced using PVL, the change in portal pressure is too small to develop splanchnic arterial vasodilation. However, there is a significant increase in the production of intestinal vascular endothelial growth factor (VEGF) with a subsequent increase in eNOS levels in the intestinal microcirculation (46). This model of mild portal hypertension may most likely represent the portal pressure changes observed in early-stage cirrhosis during which portal hypertension generally progresses slowly. When portal pressure is further increased and reaches a certain level, vasodilation in the arterial splanchnic circulation develops. It is postulated that mechanical forces generated as a result of an increased portal pressure, presumably cyclic strains and shear stress generated by an increased blood flow, activate eNOS and thus lead to NO production (46–50). Once vasodilation is established in the intestinal microcirculation and arterial splanchnic circulation, systemic circulatory abnormalities seem to follow (47).

Hyperactive endothelial cells overproduce vasodilators and cause arterial vasodilation and subsequent arterial wall thinning

Arterial vasodilation in the splanchnic and systemic circulation is mediated by an increased production of vasodilator molecules in endothelial cells. Subsequently, chronic overproduction of NO, for example, contributes to thinning of arterial walls, which may help to sustain arterial vasodilation and to attenuate a response to vasoconstrictors (Fig. 6). This decreased response to vasoconstrictors is also mediated by impaired signalling through smooth muscle cells [see a review article by Hennenberg et al. (51)]. Splanchnic arterial vasodilation is a prerequisite for the development of the hyperdynamic circulatory syndrome observed in patients with cirrhosis and portal hypertension. This syndrome is characterized by decreased mean arterial pressure, increased cardiac output and decreased peripheral resistance. These changes ultimately lead to life-threatening complications such as oesophageal varices and ascites (52, 53).

image

Figure 6.  Arterial vasodilation in the splanchnic and systemic circulation in portal hypertension. In response to an increased portal pressure, raised levels of vasodilators, such as nitric oxide (NO), cause vasodilation. Chronic NO overproduction causes arterial wall thinning, which may also contribute to sustained vasodilation. SMA, superior mesenteric artery; SMV, superior mesenteric vein; eNOS, endothelial nitric oxide synthase.

Download figure to PowerPoint

Increased levels of vasodilators

Endothelial nitric oxide synthase-derived NO has been considered the most important vasodilator molecule leading to excessive arterial vasodilation in the splanchnic and systemic circulation in portal hypertension. Other important vasodilator molecules involved in endothelium-derived arterial vasodilation include carbon monoxide (CO), prostacyclin (PGI2), endocannabinoids such as anandamide, and endothelium-derived hyperpolarizing factor (EDHF) (2, 3, 27). Hydrogen sulphide (H2S) is a vasodilator in the aorta (54, 55) and mesenteric arteries (56). In the context of portal hypertension, the role of H2S in arterial vasodilation in the splanchnic and systemic circulation has not yet been explored. Considering its role in the regulation of vascular tone and arterial pressure (57), it is highly possible that this molecule may have an important role in both splanchnic and systemic vasodilation in portal hypertension.

Other mediators associated with arterial hypocontractility/vasodilation

There are several important mediators that cause hypocontractility/vasodilation of the arterial splanchnic and systemic circulation in portal hypertension by not directly influencing the endothelium, but the smooth muscle cells and neurons. An increase in LPS levels during bacterial translocation mediates the arterial hypocontractility observed in cirrhosis and portal hypertension (58). In addition, endocannabinoids (59, 60), neuropeptide Y (61), urotensin II (62, 63), angiotensin (64) and bradykinin (65, 66) are also mediators that cause hypocontractility/vasodilation of the arterial splanchnic and systemic circulation in portal hypertension.

Endocanabinoids are lipid-like substances that act on two inhibitory G protein-coupled receptors, CB1 and CB2. CB1 is up-regulated in the endothelium of cirrhotic rats. Activation of CB1 receptor by endogenously produced endocannabinoids causes marked vasodilation in cirrhotic rats (59). Administration of a CB1 receptor antagonist, AM251, decreases the blood flow in the superior mesenteric artery (i.e. the arterial splanchnic circulation) in cirrhotic rats, suggesting that an increased flow in the superior mesenteric artery, most likely because of excessive vasodilation, is at least in part mediated by CB1 receptor (60).

Neuropeptide Y is a sympathetic neurotransmitter known to facilitate α-adrenergical vasoconstriction (67, 68). RhoA/Rho-kinase regulates various cellular functions such as cell contractility through phosphorylation of myosin light chain (69, 70). An impaired RhoA/Rho-kinase signalling pathway has been suggested to contribute to vasodilation and vascular hypocontractility in BDL-induced cirrhosis (71). However, acute administration of neuropeptide Y increases arterial contractility in the mesenteric arteries of cirrhotic rats by restoring impaired RhoA/Rho-kinase signalling (61). This observation indicates that neuropeptide Y can be used for the treatment of hypocontractility/vasodilation of the arterial splanchnic circulation in cirrhosis.

Urotensin II, a cyclic peptide with a structural similarity to somatostatin, can cause both vasoconstriction and vasodilation depending on vascular beds (72). In the systemic vessels such as the aorta and coronary artery, urotensin II works as the most potent vasoconstrictor known (73, 74). In contrast, urotensin II induces vasodilation in rat mesenteric arteries (75). In cirrhosis, plasma urotensin II levels are up-regulated. An urotensin II receptor antagonist, palosuran, ameliorates hypocontractility/vasodilation in the mesenteric arteries of biliary cirrhotic rats by increasing RhoA/Rho-kinase expression and Rho-kinase activity (thereby more contraction) and decreases nitrite/nitrate levels (62). These observations suggest that elevated urotensin II contributes to hypocontractility/vasodilation in the mesenteric arteries of cirrhotic rats. Thus, blocking urotensin II-mediated signalling pathway is beneficial for attenuation of hypocontractility/vasodilation of the arterial splanchnic circulation in cirrhotic rats.

Bradykinin is a nine amino acid peptide known to cause vasodilation (76). Bradykinin decreases sensitivity to glypressin (a long lasting vasopressin analogue) in portal hypertensive and cirrhotic rats (65, 66), thereby increasing vasodilation.

Arterial wall thinning

Endothelial-derived NO plays a central role in regulating the structure of the vessel wall (77). Using cirrhotic rats with ascites, studies showed intensive arterial thinning, as indicated by a decreased thickness of the vascular walls of the thoracic aorta, abdominal aorta, mesenteric arteries and renal artery (78, 79). Treatment with a nitric oxide synthase (NOS) inhibitor significantly improved wall thickness and attenuated the degree of the hyperdynamic circulatory syndrome by increasing arterial pressure and peripheral resistance (78). As the predominant source of NO in these arteries is endothelial cells, these observations suggest that increased eNOS-derived NO, at least in part, contributes to arterial wall thinning. Understanding the mechanisms of arterial wall thinning will be important for the development of useful treatments for patients with portal hypertension.

Portosystemic collateral vessel formation

In addition to arterial vasodilation in the splanchnic and systemic circulation, the development of portosystemic collateral vessels is also thought to worsen portal hypertension (1, 2). The development of collateral blood vessels is probably an adaptive response to increased portal pressure, which may help to delay development of the pathology. However, the formation of these vessels can also cause detrimental complications. As the vessels are fragile, they rupture easily, causing oesophageal and gastric variceal bleeding. Furthermore, as these vessels bypass the liver, the portal blood flow carrying toxic substances, such as drugs, bacterial toxins and toxic metabolites, returns to the systemic circulation and can result in portal-systemic encephalopathy and sepsis (1, 2).

These collateral vessels are formed through the enlargement of pre-existing vessels as well as through angiogenesis (44, 45). The process of angiogenesis is modulated by growth factors exhibiting vasodilatory activity, such as VEGF. These collateral vessels can contribute to an increase in the blood flow through the portal vein, exacerbating portal hypertension (44). How are these angiogenic growth factors up-regulated? One mechanism would be triggered by an increase in portal pressure. Studies using portal hypertensive rats indicated that a sudden increase in portal pressure is sensed at the intestinal microcirculation and induces VEGF expression as described previously (46, 49). This sudden increase in portal pressure may generate local mechanical forces, which can induce VEGF expression.

Anti-angiogenic agents, such as blockers of VEGF receptor-2 (SU5416, anti-VEGFR2 monoclonal antibody) (49, 80) and inhibitors of receptor tyrosine kinases (Sorafenib and Sunitinib) (81, 82), have been shown to decrease portosystemic collateral vessel formation and reduce portal pressure. Besides VEGF, placental growth factor (PIGF), another member of the VEGF family, has also been shown to be up-regulated in the intestinal microcirculation of portal hypertensive mice (83). Portal hypertensive mice lacking PIGF or given an anti-PIGF monoclonal antibody show decreases in portal pressure and portosystemic collateral vessel formation. Collectively, these studies suggest that blocking angiogenic activity, thereby reducing collateral blood vessel formation, is beneficial for the treatment of portal hypertension.

Circulating endothelial cell levels are elevated in patients with cirrhosis and portal hypertension

It is not certain whether levels of circulating endothelial cells are related to endothelial dysfunction. Furthermore, circulating endothelial cells may originate from not only the splanchnic and systemic circulation, but also from the intrahepatic circulation. Nonetheless, circulating endothelial cells are an interesting topic that has received attention in recent years, and are thus discussed here briefly. Elevated levels of circulating endothelial cells have been observed in a variety of disease conditions associated with vascular injury (84–86), and are considered to reflect the severity of vascular injury (87). For example, a recent study by Abdelmoneim et al. (86) reported that circulating endothelial cell levels are significantly higher in patients with cirrhosis compared with healthy subjects. Cirrhosis and portal hypertension are characterized by prominent changes in the vascular endothelium of both the intra- and extrahepatic circulations (1). Furthermore, changes in the levels of inflammatory cytokines and vasoactive molecules, such as tumour necrosis factor (TNF) and NO respectively may also contribute to changes in the endothelium and increased levels of peripheral circulating endothelial cells (88–90).

The origin of circulating endothelial cells in patients with cirrhosis and portal hypertension is not clear. Originally, it was thought that circulating endothelial cells may derive from mature endothelial cells shed from the vessel walls in response to vascular injury, or from endothelial progenitor cells (EPCs), which are bone marrow-derived and are assumed to contribute to vascular repair (91–93). However, the study by Abdelmoneim et al. (86) showed that circulating endothelial cells observed in patients with cirrhosis and portal hypertension were mature endothelial cells, not EPCs. This was concluded by the circulating endothelial cells being positive for CD146 and CD105 markers, but negative for CD34, a marker for EPCs (86). This finding may indicate that circulating endothelial cells arise from endothelial cells released from the damaged vasculature of patients with cirrhosis and portal hypertension. The mechanism of their release, whether or not they are still functional, and their eventual fate require further research.

Summary – splanchnic and systemic circulations

In cirrhosis and portal hypertension, endothelial cells in the splanchnic and systemic circulation are hyperactive, producing increased levels of NO derived from eNOS (1–3). This causes arterial vasodilation and thinning of arterial walls in the splanchnic and systemic circulation (1–3, 78, 79). Besides NO, other vasodilator molecules, such as CO, PGI2, endocannabinoids and EDHF, also contribute to arterial vasodilation in the splanchnic and systemic circulation. Hyperactive endothelial cells also contribute to angiogenic collateral vessel formation (46, 49).

An increase in portal pressure is a key factor that causes hyperactive endothelial cells in the splanchnic and systemic circulation (46, 47). Studies using surgically induced portal hypertension in rats demonstrated that changes in portal pressure are first sensed by the vascular beds of the intestinal microcirculation, followed by the arteries in the splanchnic and then systemic circulation (46). An increase in portal pressure induces production of VEGF and PIGF, potent angiogenic growth factors that facilitate collateral vessel formation (80–83). Blocking these angiogenic growth factors attenuates collateral vessel formation and ameliorates portal hypertension, suggesting potential therapeutic targets for the treatment of portal hypertension.

Circulating endothelial cells are increased in patients with cirrhosis and portal hypertension (86). The cause of increased levels of circulating endothelial cells in these patients, whether they are still functional, and their fate require further research. Although the discussion above is largely confined to endothelial cells, smooth muscle cells also play an important role in increased vasodilation and thinning of arterial walls in the splanchnic and systemic circulation observed in cirrhosis and portal hypertension.

Future direction

  1. Top of page
  2. Abstract
  3. Intrahepatic circulation
  4. Splanchnic and systemic circulations
  5. Future direction
  6. Acknowledgment
  7. References

The basic pathology and mechanisms of portal hypertension can be closely related to those of cardiovascular disease (1, 2). Thus, a thorough familiarity with cardiovascular disease in particular and vascular biology in general will be of great advantage in the study of portal hypertension and our understanding of the cellular and molecular mechanisms of vascular changes underlying it. It has become increasingly clear in the study of vascular biology that mediating factors, such as microRNAs (miRNAs) and endothelial cell progenitor cells, are involved in the regulation of vascular changes, such as angiogenesis and vascular remodelling in vascular homeostasis and pathophysiological conditions. The roles of inflammation and endothelial–mesenchymal transition (EndoMT) in endothelial dysfunction are another significant areas in the study of portal hypertension. This section will briefly discuss potential roles for these mediators/factors in the vascular changes seen in cirrhosis and portal hypertension.

MicroRNAs

MicroRNAs are short non-coding RNAs that generally function as negative regulators of expression for their target gene (94). A growing body of evidence suggests that miRNAs are crucial regulators of gene expression associated with many biological processes, such as angiogenesis, vessel remodelling, fibrosis and apoptosis (95). Because of their involvement in various biological processes, dysregulation of miRNA expression leads to a number of human diseases (96).

While an increasing number of studies has demonstrated important regulatory roles of miRNAs in the gene expression of HSCs in many pathological conditions, little is known about miRNA regulation of endothelial cells, particularly LSEC gene expression in cirrhosis and portal hypertension. To date, one study by Yeligar et al. (97) has reported that miRNAs control gene expression in LSECs. Using an experimental model of alcohol-induced liver injury, they showed that elevated concentrations of ethanol decrease miR-199 levels in LSECs. MiR-199 is known to bind to an untranslated region of ET-1 mRNA and decrease its level. Thus, overexpression of miR-199 reduces ET-1 mRNA expression in LSECs. Conversely, blocking miR-199 increases ET-1 mRNA expression. Therefore, it can be considered that ethanol increases intrahepatic resistance by suppressing miR-199, which increases expression of miR-199-target genes such as ET-1, an important regulator of vascular tone [ET-1 can cause both vasodilation and vasoconstriction depending on which receptors it binds to (Fig. 3)]. As ethanol increases portal pressure (98), it is possible that ethanol changes the expression of miRNAs, such as miR-199, and causes the LSEC dysfunction that facilitates an increase in intrahepatic resistance. These miRNAs can be novel targets for the attenuation of an increased intrahepatic resistance in cirrhosis, thereby ameliorating portal hypertension.

Identification of miRNAs and miRNA-target genes related to the vascular changes in cirrhosis and portal hypertension will provide a tremendous opportunity to elucidate the mechanisms of hepatic angiogenesis, collateral vessel formation and arterial wall thinning. As many miRNAs exhibit patterns of gene regulation strikingly specific to cell types (99–101), identification of relevant miRNAs would be more successful by performing in a cell-specific manner, rather than in a procedure involving the entire vasculature.

Endothelial progenitor cells

Circulating EPCs enhance angiogenesis and vascular repair. In recent years, however, the definition of EPCs has become controversial (102). It has become increasingly clear that so-called EPCs are not genuine endothelial progenitors, but are predominantly of monocytic lineage (103). Accordingly, the name EPC may not be appropriate. Nevertheless, these so-called EPCs exhibit an angiogenic activity and play a role in vascular repair (102). Therefore, it is possible that EPCs may contribute to the formation of new intrahepatic vessels and portosystemic collateral vessels that develop during the progression of cirrhosis and portal hypertension. Furthermore, circulating endothelial cells, which do not appear to be EPCs and are found to be elevated in patients with cirrhosis and portal hypertension as mentioned previously (86), may play a role in angiogenesis seen in cirrhosis and portal hypertension.

Bone marrow-derived cells, including EPCs, can repopulate LSECs (104–106). Efficiency of the repopulation depends upon the injury models used. In a transplanted liver after 18 h of cold ischaemic storage, an engraftment of bone marrow-derived cells into functioning LSECs is only between 1% and 5% (105). In contrast, in livers injured by exposure to monocrotaline, bone marrow-derived cells replace more than one quarter of LSECs and could eliminate the histological evidence of centrilobular haemorrhagical necrosis observed in early injury (104). These engrafted bone marrow-derived LSECs exhibit the presence of fenestrae, a typical feature of LSECs, suggesting that these LSECs could be functional. Future studies with the goal of increasing efficiency of engraftment by bone marrow-derived cells and the subsequent restoration of LSEC function in cirrhotic livers may have a potential to ameliorate increased intrahepatic vascular resistance and portal hypertension.

Endothelial to mesenchymal/myofibroblast transition

Endothelial to mesenchymal transition has been implicated in the pathogenesis of fibrosis in various organs. Furthermore, EndoMT involves a process similar to that found in epithelial–mesenchymal transition. Thus, it is possible that the loss of fenestrae in LSECs during the development of cirrhosis could be related to EndoMT. Furthermore, as its name implies, during EndoMT, endothelial cells exhibit gene expression patterns similar to those of myofibroblasts, including gene expression of α-smooth muscle actin, fibroblast-specific protein-1 and collagen type I (107). Thus, EndoMT of LSECs, by gaining gene expression patterns similar to those of myofibroblasts, may contribute to an increased intrahepatic resistance found in cirrhosis and portal hypertension.

Inflammation

Inflammation is coupled with angiogenesis in many pathological conditions, such as atherosclerosis and diabetes. One of the consequences of inflammation in the vasculature is an increase in endothelial cell permeability, frequently induced by VEGF, NO or other mediators. Increased permeability permits plasma components and immune cells to enter the subendothelial space from the bloodstream, initiating and sustaining an inflammatory response (108).

In cirrhotic rats, it was reported that endothelial cell permeability significantly increases in the liver and mesenteric vascular beds compared with normal rats (109). In this study, endothelial cell permeability was assessed by Evan's Blue dye extravasation and colloidal carbon deposition at the basal lamina of endothelial cells. This increase in endothelial cell permeability was accompanied with elevated levels of VEGF-A and angiopoietin-2 and was attenuated by the administration of VEGFR-2 blockers. These observations suggest that increased levels of VEGF-A and possibly angiopoietin-2, generated in response to inflammation occurring at the endothelium, enhance endothelial cell permeability and destabilize the endothelial cell junction in the liver and mesenteric vascular beds of cirrhotic rats.

Furthermore, it is possible that this increase in endothelial cell permeability facilitates the formation of hepatic angiogenesis and portosystemic collateral vessels during the development of cirrhosis and portal hypertension, as increased endothelial permeability is, in general, an initial and necessary event for VEGF-mediated angiogenesis. The disruption of intercellular junctions between endothelial cells allows for increased passage of macromolecules through the endothelium which favours the generation of the proangiogenic microenvironment (110, 111).

Strategies to reduce intrahepatic vascular resistance

Strategies to increase intrahepatic NO availability are useful for ameliorating portal hypertension. As mentioned previously, while NO bioavailability is decreased in the intrahepatic circulation, NO overproduction in the splanchnic and systemic circulation worsens portal hypertension. Therefore, liver specific NO-delivery or NO-generating systems are essential for attenuating portal hypertension.

Liver-specific nitric oxide donors

NCX-1000 is a liver-specific NO-releasing agent conjugated to ursodeoxycholic acid. It releases NO from parenchymal and non-parenchymal hepatic cells into the liver microcirculation with no detectable effect on the extrahepatic circulation. Regarding its effect on portal hypertension, however, two studies using different experimental models of cirrhosis report different results. A study by Fiorucci et al. (112) used a biliary cirrhotic model induced by bile duct ligation, while Loureiro-Silva et al. (113) employed a cirrhotic model generated by CCl4 inhalation. Both studies used the same dose of NCX-1000 (28 mg/kg) and showed a significant reduction in intrahepatic perfusion pressure in response to α-adrenoceptor agonists as a result of NCX-1000 treatment. However, the study by Fiorucci et al. (112) reported a marked reduction of baseline portal pressure, whereas NCX-1000 did not reduce portal pressure in the study by Loureiro-Silva et al. (113). This difference in the outcomes may reflect differences in the experimental models they used. This drug was also tested in 11 patients with cirrhosis and portal hypertension (114). While it is safe to administer this drug to patients, there was no reduction in their portal pressure.

Adenoviral gene transfers

Nitric oxide synthase genes, including the endothelial (115) and neuronal (116) isoforms of NOS, have been introduced to cirrhotic livers generated by CCl4 in gavage (115, 116), resulting in increased NO synthesis and a marked decrease of portal pressure. However, in another BDL study, eNOS gene transfer did not reduce the hepatic vascular tone probably because of a marked increase in caveolin-1 levels in BDL livers (117).

Akt, also known as protein kinase B, phosphorylates and activates eNOS (118, 119). Introduction of constitutively active Akt also increased eNOS activity and NO production, leading to a marked attenuation of portal pressure in CCl4-induced cirrhotic rats (120). Overall, these studies provide the proof-of-concept that increased NO availability in cirrhotic livers is a useful strategy to attenuate portal hypertension.

Statins

Simvastatin is a family of statins, cholesterol-lowering drugs known to increase Akt-dependent eNOS phosphorylation and NO synthesis (121). It was reported that cirrhotic patients given Simvastatin (20 mg/day for 14 days and 40 mg/day for 17 days) showed a significantly increased hepatic NO production (122) and attenuation of portal hypertension as indicated by a hepatic venous pressure gradient (123). In cirrhotic rats with ascites, administration of Simvastatin increased eNOS expression, Akt-dependent eNOS phosphorylation and cGMP content, compared with those rats given a placebo (124). Collectively, these observations suggest the beneficial effects of Simvastatin in the treatment of cirrhosis and portal hypertension.

Tetrahydrobiopterin (BH4)

A pteridine metabolite, BH4 is an essential cofactor for the activity of all three NOS isoforms (125). In the absence of BH4, NOS is unable to catalyse the oxidation of a substrate, l-arginine. As a result, NOS generates superoxide anion (O2), instead of NO by transferring electrons from NADPH (a reduced form of nicotinamide adenine dinucleotide phosphate) to O2 forming O2 (126). The process, generating a superoxide radical by NOS, is known as NOS uncoupling. In cirrhotic rats, levels of BH4 and guanosine triphosphate (GTP)-cyclohydrolase (a rate-limiting enzyme of BH4 synthesis) were significantly diminished (127, 128). Supplementation of BH4 to those rats resulted in increased hepatic BH4 levels, enhanced eNOS activity, a subsequent increase in endothelium-dependent vasodilation of cirrhotic rat livers in response to acetylcholine, and attenuated portal hypertension (127, 128). These findings suggest that BH4 supplementation decreases intrahepatic resistance and portal pressure by increasing NO generation instead of O2 in cirrhotic rats without influencing NO production in the arterial splanchnic and systemic circulation. However, it is also important to note that the bacterial translocation observed in cirrhosis enhances GTP-cyclohydrolase activity and BH4 production in the mesenteric arterial beds leading to an increase in NO production and vasodilation in the mesenteric arterial beds of cirrhotic rats (129, 130).

Conclusion

Although portal hypertension leads to the most lethal complications of liver disease such as gastro-oesophageal varices and ascites, there are very limited options for its treatment. Facing this serious clinical situation, there is a strong need for further studies of the vascular changes associated with cirrhosis and portal hypertension (131). Studying the roles of miRNAs, EPCs, EndoMT, and inflammation in the vascular changes associated with cirrhosis and portal hypertension advances our understanding of the molecular and cellular mechanisms underlying these changes. The practical benefit of these studies enables the development of novel target molecules for the treatment of portal hypertension. Further, a cell-specific approach in vivo and in vitro (e.g. cell/tissue specific gene manipulation, drug delivery, etc.) will be important in understanding these mechanisms in detail. Finally, although this review article addresses mostly endothelial cells, it is also important to understand the contribution of other cell types, such as HSCs, smooth muscle cells and immune cells, to endothelial dysfunctions and the subsequent vascular changes found in cirrhosis and portal hypertension.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Intrahepatic circulation
  4. Splanchnic and systemic circulations
  5. Future direction
  6. Acknowledgment
  7. References

I thank Mr Jay Prendergast for editing and Dr Teruo Utsumi for useful discussion.

Financial support: This work was supported by grants from the NIH/NIDDK (R01DK082600, K01DK067933) and the Yale Liver Center Pilot Grant (P30-DK034987).

References

  1. Top of page
  2. Abstract
  3. Intrahepatic circulation
  4. Splanchnic and systemic circulations
  5. Future direction
  6. Acknowledgment
  7. References