The endocannabinoid system in advanced liver cirrhosis: pathophysiological implication and future perspectives


  • Maurizio Baldassarre,

    1. Department of Medical and Surgical Sciences, Center for Applied Biomedical Research (C.R.B.A.), Alma Mater Studiorum University of Bologna, Bologna, Italy
    Search for more papers by this author
  • Ferdinando A. Giannone,

    1. Department of Medical and Surgical Sciences, Center for Applied Biomedical Research (C.R.B.A.), Alma Mater Studiorum University of Bologna, Bologna, Italy
    Search for more papers by this author
  • Lucia Napoli,

    1. Department of Medical and Surgical Sciences, Center for Applied Biomedical Research (C.R.B.A.), Alma Mater Studiorum University of Bologna, Bologna, Italy
    Search for more papers by this author
  • Alessandra Tovoli,

    1. Department of Medical and Surgical Sciences, Center for Applied Biomedical Research (C.R.B.A.), Alma Mater Studiorum University of Bologna, Bologna, Italy
    Search for more papers by this author
  • Carmen S. Ricci,

    1. Department of Medical and Surgical Sciences, Center for Applied Biomedical Research (C.R.B.A.), Alma Mater Studiorum University of Bologna, Bologna, Italy
    Search for more papers by this author
  • Manuel Tufoni,

    1. Department of Medical and Surgical Sciences, Center for Applied Biomedical Research (C.R.B.A.), Alma Mater Studiorum University of Bologna, Bologna, Italy
    Search for more papers by this author
  • Paolo Caraceni

    Corresponding author
    • Department of Medical and Surgical Sciences, Center for Applied Biomedical Research (C.R.B.A.), Alma Mater Studiorum University of Bologna, Bologna, Italy
    Search for more papers by this author


Paolo Caraceni, MD, Department of Medical and Surgical Sciences, Alma Mater Studiorum Universitv of Bologna, Via Massarenti 9, 40138 Bologna, Italy

Tel: +39 051 6362919

Fax: +39 051 6362930



Endogenous cannabinoids (EC) are ubiquitous lipid signalling molecules providing different central and peripheral effects that are mediated mostly by the specific receptors CB1 and CB2. The EC system is highly upregulated during chronic liver disease and consistent experimental and clinical findings indicate that it plays a role in the pathogenesis of liver fibrosis and fatty liver disease associated with obesity, alcohol abuse and hepatitis C. Furthermore, a considerable number of studies have shown that EC and their receptors contribute to the pathogenesis of the cardio-circulatory disturbances occurring in advanced cirrhosis, such as portal hypertension, hyperdynamic circulatory syndrome and cirrhotic cardiomyopathy. More recently, the EC system has been implicated in the development of ascites, hepatic encephalopathy and the inflammatory response related to bacterial infection. Rimonabant, a selective CB1 antagonist, was the first drug acting on the EC system approved for the treatment of obesity. Unfortunately, it has been withdrawn from the market because of its neuropsychiatric side effects. Compounds able to target selectively the peripheral CB1 receptors are under evaluation. In addition, molecules stimulating CB2 receptor or modulating the activity of enzymes implicated in EC metabolism are promising areas of pharmacological research. Liver cirrhosis and the related complications represent an important target for the clinical application of these compounds.

The endocannabinoid system

Although the antiemetic, analgesic and appetite-stimulating properties of marijuana or Cannabis sativa have been described many centuries ago by the ancient Oriental pharmacopoeia, in Western societies marijuana has been known mostly for its ability to alter sensory perception and cause euphoria. For this reason, until a few decades ago, when the existence of endogenous cannabinoids (EC) and their specific receptors was finally demonstrated, the research on biological properties of marijuana was considered an esoteric field [1].

The main psychoactive constituent of marijuana, Δ9-tetrahydrocannabinol (THC), was characterized in 1964 [2], followed by the isolation of many other bioactive molecules derived from the plant [1, 3-5]. Afterwards, synthetic radio-labelled cannabinoid analogues allowed the identification of an orphan G protein-coupled receptor, named CB1, as one of the binding sites for cannabinoids [6]. The CB1 receptor was initially found to be widely expressed in specific areas of the brain, often on presynaptic neurones, where it participates in the regulation of neuronal excitability [7-9]. However, it is also expressed in many other organs and systems [10, 12]. Few years later, another G protein-coupled cannabinoid receptor, named CB2 and located primarily on immune and haematopoietic cells such as macrophages and lymphocytes, was isolated [13, 14] (Table 1).

Table 1. Expression of CB1 and CB2 receptors in central and peripheral normal tissue
Central nervous system
Olfactory bulbCB1, CB2
Cortical regionsCB1, CB2
Basal gangliaCB1
Thalamic and hypothalamic nucleiCB1, CB2
CerebellumCB1, CB2
Brainstem nucleiCB1, CB2
Endocrine system
Adrenal glandCB1
Endocrine pancreasCB1, CB2
Respiratory system
LungCB1, CB2
Immune and haematopoietic system
SpleenCB1, CB2
ThymusCB1, CB2
TonsilsCB1, CB2
Bone marrowCB2
Immune cellsCB1, CB2
Reproductive system
OviductCB1, CB2
TestisCB1, CB2
SpermCB1, CB2
ProstateCB1, CB2
Gastrointestinal tract
StomachCB1, CB2
Urinary tract
KidneyCB1, CB2
BladderCB1, CB2
Cardiovascular system
HeartCB1, CB2
Other tissue
Skeletal muscleCB1

The discovery of receptors for plant-derived cannabinoids in mammalian cells paved the way for the identification of their endogenous ligands, the EC, lipid signalling molecules that mimic the activity of THC. The lipid arachidonoyl ethanolamide, named anandamide (AEA), was the first EC isolated from porcine brain [15] followed by the characterization of a second EC, 2-arachidonoylglycerol (2-AG) [16, 17].

Endocannabinoids are synthesized on demand from membrane phospholipids and immediately released from cells; once released, they activate their specific receptors to trigger a biological response to be rapidly inactivated, thus acting in a paracrine or autocrine fashion [1].

Arachidonoyl ethanolamide is synthesized from the phospholipid precursor N-arachidonoyl-phosphatidyl ethanolamine and acts as a partial or full agonist of CB1 receptors, while it shows a lower affinity for CB2 receptors. It is removed from the site of action by the cellular reuptake mechanism, involving the putative AEA membrane transporter [18], and is metabolized in intracellular membranes by the enzyme fatty acid amide hydrolase (FAAH). Furthermore, AEA is able to interact also with non-CB1 and non-CB2 receptors, including the transient receptor potential vanilloid type-1 (TRPV1). Finally, it seems that other binding sites exist even though they still need further investigation [19-21].

On the other hand, 2-AG is generated from a diacylglycerol precursor and shows a similar affinity for both CB1 and CB2 receptors. After triggering its biological response, it is degraded in the cytosol and at intracellular membranes level by monoacylglycerol lipases (MAGLs). As for AEA, 2-AG also seems to bind to other receptors not precisely identified as yet [19]. A schematic representation of AEA and 2-AG metabolism is shown in Fig. 1.

Figure 1.

Schematic representation of anandamide (AEA) and 2-arachidonoilglycerol (2-AG) metabolisms and chemical compounds modulating the endogenous cannabinoid system. AEA synthesis occurs at the cellular membrane level as a result of a specific phospholipase-D (NAPE-PLD) activity on a phospholipid-derived precursor. Once in the cytosol, a specific endocannabinoid membrane transporter (EMT) mediates the AEA trafficking across the cellular membrane. AEA interacts mainly with the CB receptors, being less affine for the CB2, but also with the intracellular domain of the TRVP-1 receptor, while more controversial is the interaction with GPR55. AEA is rapidly removed from the extracellular space and degraded at the cytosol level by the fatty acid amide hydrolase (FAAH) releasing arachidonic acid (AA) and ethanolamine (EA). 2-AG is also synthesized at the cellular membrane level through the activity of a specific diacil-glycerol lipase (DAGL) from a phosphatidylinositol precursor. 2-AG could interact with CB receptors, with the same affinity for both CB1 and CB2, and other non-CB receptors. 2-AG inactivation occurs through its reuptake in the cellular compartment and the subsequent degradation by monoacylglycerol lipase (MAGL), producing AA and glycerol (GLY). Chemical compounds such as the first-generation CB1 receptor inverse agonist (rimonabant, taranabant) and the novel peripherally restricted neutral antagonist (AM6545) or inverse agonist (JD5037) were successfully used in the setting of experimental chronic liver disease. On the other hand, potentiating CB2 receptor signalling through the administration of selective agonists (HU-308, JWH-133) has proved to be effective in reducing experimental liver injury. Finally, inhibiting MAGL through JZL184 administration, thus potentiating 2-AG signalling, may represent an effective and alternative pharmacological approach.

In the last decade, several EC-like molecules, such as N-palmitoyl-ethanolamine (PEA) and N-oleoyl-ethanolamine (OEA), have been also identified [22]. Even if these compounds have a biosynthetic and degradative process similar to that of AEA, they do not interact with CB receptors, but with other target receptors, including the peroxisome proliferator-activated receptor-alpha, the GPR55 receptor and, in the case of OEA, the TRPV1 receptor [4, 8, 19].

The understanding of the EC effects in several physiological and pathological processes has been made possible by the synthesis of effective and highly selective CB1 and CB2 receptor pharmacological agonists and antagonists. As a result, it has become evident that EC provide several central and peripheral actions implicated in many biological processes such as energy balance, food intake, mood, analgesia and motor function, immune and inflammatory responses, cell proliferation, fertility, cardiovascular and gastrointestinal function [1, 3, 4].

The endocannabinoid system and the liver

Under physiological conditions, both hepatocytes and non-parenchymal cells (Kupffer and endothelial cells) are able to produce EC, but the basal expression of CB1 and CB2 in the adult healthy liver is low or even absent [23, 24].

On the other hand, EC synthesis and hepatic expression of CB1 and CB2 receptors are upregulated in several conditions characterized by chronic liver damage, such as alcoholic and non-alcoholic fatty liver, viral hepatitis and fibrosis [25]. A dysregulation of the EC system has been also shown during acute liver injury, such as that caused by ischaemia-reperfusion, contributing to the development of the inflammatory response and cell death [26-28].

The EC system appears to be strongly activated both in human and experimental cirrhosis. Batkai et al. showed that AEA levels are increased in monocytes isolated from both cirrhotic rats and humans as compared to the non-cirrhotic counterparts and contribute to the development of the hemodynamic alterations [29]. The circulating plasma level of AEA, but not that of 2-AG, was also found to be elevated in patients with cirrhosis [30, 31]. The AEA levels directly correlated with biochemical parameters of liver function, like international normalized ratio (INR) and bilirubin, but not with hemodyamic variables [31]. Interestingly, the plasma EC-like molecules, OEA and PEA, which have no activity on CB receptors, were markedly more elevated than AEA in cirrhotic patients [31], but the biological and clinical meaning of this finding needs to be clarified.

Here, we reviewed the experimental and clinical evidences showing the involvement of the EC system in the pathogenesis of cardiovascular abnormalities, which characterize advanced cirrhosis and favour the development of severe clinical complications, such as portal hypertension, hyperdynamic circulatory syndrome, sodium retention and ascites, cirrhotic cardiomyopathy, hepatic encephalopathy (HE) and susceptibility to bacterial infections.

The endocannabinoid system and portal hypertension

Advanced cirrhosis leads almost invariably to portal hypertension that contributes to the onset of gastro-oesophageal varices and eventually bleeding, ascites, HE, hepatorenal and hepatopulmonary syndromes. Two major factors contribute to the development of portal hypertension: the increased intrahepatic resistance and the increased portal blood flow, secondary to splanchnic arterial vasodilatation. The intrahepatic resistance is the result of a structural component, related to the distortion of the liver architecture because of the deposition of fibrotic tissue and the formation of regenerative nodules, and a dynamic component, related to the imbalance of local vasoactive substances, thus favouring vasoconstriction [32].

Endogenous cannabinoids and liver fibrosis

Liver fibrosis is the common response to liver injury promoted by a series of conditions, with viral hepatitis, alcohol abuse and non-alcoholic fatty liver being the most frequent causes. The fibrogenic process requires the activation of hepatic stellate cells (HSC), which under the stimuli of profibrogenic cytokines and growth factors switch their phenotype into myofibroblast-like cells responsible for extracellular matrix deposition [33].

The initial clinical observation that patients with hepatitis C consuming marijuana present a faster progression of fibrosis has been confirmed by a series of experimental data indicating the involvement of the EC system in the fibrogenic signalling pathways [34].

Both genetic and pharmacological modulation of the EC receptors has an impact on the fibrogenic process [35]. In different models of chronic liver damage, CB1-deficient animals, as well as those treated with the CB1 antagonist rimonabant, present a lower amount of fibrotic tissue. The reduced fibrosis was accompanied by a reduced expression of profibrogenic cytokines, such as transforming growth factor-β, and a marked decrease in activated HSCs as a consequence of reduced cell growth and enhanced apoptosis [23, 36].

In contrast, CB2-knockout mice exposed to chronic CCl4 administration show a greater deposition of fibrotic tissue compared to their wild-type counterparts, suggesting a protective antifibrogenic effect mediated by the CB2 receptor [37, 38]. Consistently, pharmacological chronic stimulation of CB2 receptors prevents fibrosis progression during CCl4 treatment by lowering the hepatic gene expression of PDGFRβ and modulating TIMP/MMP balance [39].

Recently, Rossi et al. confirmed the role of the CB2 receptor signalling also in humans by showing that the presence of a functional variant (Q63R) of the CB2 receptor, which impairs its immunomodulating function, correlates with the severity of liver inflammation in obese children with non-alcoholic fatty liver disease [40].

Different experimental studies suggest that the pharmacological modulation of the EC system is also able to achieve extracellular matrix remodelling and fibrosis regression even when a stage of full-blown cirrhosis was already reached. CCl4-induced cirrhotic rats with ascites treated with the CB2 receptor agonist JWH-133, or the CB1 receptor antagonist rimonabant, showed an improved liver architecture because of the partial regression of fibrosis as compared with cirrhotic rats treated with the vehicle [36, 41, 42].

In other words, these data clearly indicate that CB1 and CB2 receptor stimulation exerts an opposing effect in the liver: the activation of CB1 promotes fibrogenesis, while CB2 agonism triggers antifibrogenic responses. However, during chronic liver injury, the profibrogenic CB1 signalling seems to prevail on the antifibrogenic action of CB2 (Fig. 2).

Figure 2.

Endogenous cannabinoid (EC) contribution to hemodynamic alteration and ascites formation in chronic liver disease. Gut-derived lipopolisaccharide (LPS) triggers AEA synthesis from both macrophages and lymphocyte (WBC). Once released, AEA activates CB1 and TRPV1 receptors in the splanchnic compartment, leading to splanchnic vasodilation and effective hypovolaemia that, in turn, evoke compensatory system (RAA, ADH, catecholamines) to promote renal water and sodium retention. Hepatic EC activation also contributes to faster fibrosis progression and increased intrahepatic resistances by enhancing CB1 receptor signalling, thus contributing to ascites formation. Finally, increased circulating AEA, together with the enhanced intrinsic production by cardiac muscle cells, contributes to the impairment of cardiac contractility by activating the CB1 receptor.

Endogenous cannabinoids and intrahepatic vascular resistance

The EC system is also able to affect portal pressure by affecting the dynamic, reversible component of the intrahepatic resistance. AEA causes a dose-dependent increase in eicosanoid production, thus inducing vasoconstriction in the isolated perfused rat liver model, this effect being more evident in cirrhotic than in healthy animals [43]. Furthermore, in rats with biliary cirrhosis, the pharmacological antagonism of the CB1 receptor with AM251 is able to downregulate the expression levels of vasoconstrictors, thus achieving a fall in intrahepatic vascular resistance [44].

The cannabinoid system and hyperdynamic circulatory syndrome

Patients with advanced cirrhosis present features typical of the hyperdynamic circulatory syndrome (arterial hypotension, tachycardia and hyperkinetic peripheral pulse), resulting from a reduction in peripheral vascular resistances, mainly located in the splanchnic area. This leads to effective hypovolaemia that, in turn, evokes a compensatory increase in the cardiac output through the activation of sodium-retaining and vasoconstrictor systems [45]. The molecular mechanisms responsible for arterial vasodilatation are yet not fully clarified, but it is widely accepted that an imbalance of vasoactive substances, such as nitric oxide (NO), carbon monoxide and prostacyclin, with a prevalence of vasodilators occurs [46].

Experimental data suggest that the EC system contributes to the vasodilatation and arterial hypotension in advanced cirrhosis. The administration of the CB1 receptor antagonist rimonabant was able to reverse the arterial hypotension and to increase the splanchnic vascular resistances in cirrhotic rats, leading to a concomitant decrease in the mesenteric arterial blood flow and portal pressure [29, 41, 47]. Moreover, monocytes isolated from cirrhotic rats producing high levels of AEA induced a long-lasting hypotensive effect after being infused in normal recipient animals; this effect was lost when monocytes isolated from healthy rats were used [47]. Thus, as lipopolysaccharide (LPS) represents the major trigger for EC generation from platelets, monocytes and lymphocytes [48-50], it has been hypothesized that these cells are stimulated to produce large amounts of EC, particularly AEA, by the increased endotoxemia occurring in advanced cirrhosis as a result of bacterial translocation [51]. This would lead to vasodilatation and long-lasting arterial hypotension by directly activating the CB1 receptor located in the vascular and myocardial endothelium and in the perivascular nerves [29, 47].

The cellular mechanisms that mediate the vasodilating effect of AEA in cirrhosis have been further elucidated by Domenicali et al. in a model of mesenteric resistance arteries isolated from rats with CCl4-induced cirrhosis and ascites [52]. The authors showed that AEA causes a dose-dependent relaxation, which is greater in cirrhotic than in control arteries. The same effect was not seen in femoral arteries from the same rats, prompting evidence of a selective AEA action in the splanchnic circulation. Additionally, the hypotensive effect was not influenced by pretreatment with the NO synthase inhibitor l-NAME or endothelial denudation, suggesting that the endothelium-derived NO does not play a major role in this response. Furthermore, pretreatment with capsaicin, which blocks the response of primary sensory nerves, fully abolished the AEA-induced relaxation, indicating that EC act on vessel adventitia, where sensory nerves are located, rather than in the endothelial layer [52].

However, CB1 is not the only receptor implicated in the EC-mediated vasorelaxation. Indeed, AEA can also interact with the TRPV1 receptor, which is expressed in perivascular nerves, as demonstrated by the finding that the concomitant use of rimonabant and capsazepine, which blocks the TRPV1, almost fully abolish the AEA-induced vasodilatation, demonstrating that both CB1 and TRPV1 receptors mediate the AEA hypotensive effect in cirrhotic rats [52-55].

The sum of the above data indicates that splanchnic vasodilatation is mediated, at least in part, by AEA and can be potentially reversed by the pharmacological antagonism of CB1 and TRPV1 receptors (Fig. 2).

The endocannabinoid system and sodium retention/ascites

Ascites is one of the most common complication of cirrhosis. It represents the combined result of hemodynamic and hepatic alterations that occur during the advanced stages of the disease, and its appearance carries a poor prognostic meaning in terms of both expectancy and quality of life. The effective hypovolaemia caused by splanchnic vasodilatation, activating sodium-retaining systems, such as the renin-angiotensin-aldosterone axis, catecholamines and vasopressin, promotes renal retention of sodium and water, while portal hypertension leads the fluid to accumulate in the peritoneal cavity [56]. Thus, if the CB1 receptor antagonism is able to interfere with the hemodynamic alterations of cirrhosis, it also appears reasonable to hypothesize that it can also counteract ascites formation.

The experimental proof of this hypothesis was provided by Domenicali et al., who showed that a 2-week treatment with rimonabant, started in compensated rats with CCl4-induced cirrhosis, was able to prevent ascites formation in up to 90% of the cases [41]. The dose-dependent increase in sodium urinary excretion was associated with an improvement in the hyperdynamic circulation documented by the rise of peripheral vascular resistances and mean arterial pressure, eventually leading to a greater renal arterial blood flow [41].

Thus, this experimental study clearly showed that CB1 receptor antagonism improves sodium balance and delays decompensation in preascitic experimental cirrhosis likely through a dose-dependent improvement in systemic and renal hemodynamics, providing the rationale for a potential novel application of the pharmacological modulation of the EC system.

The endocannabinoid system and cirrhotic cardiomyopathy

During advanced cirrhosis, patients often develop a series of cardiac functional alterations, which are identified by the term cirrhotic cardiomyopathy, even if it still awaits a precise and universally accepted definition. Although its pathogenesis is still unclear, it has been showed that both central neuronal dysregulation and humoral factors are involved [57, 58].

Irrespective of the aetiology of cirrhosis, cardiac impairment is characterized by a diastolic and systolic dysfunction that remains usually subclinical, likely being masked by the reduced left ventricular afterload because of peripheral vasodilatation [59-61]. However, under stressful conditions, such as sepsis, transjugular intrahepatic porto-systemic shunting and liver transplantation, cardiac impairment may become clinically manifest and contribute to the development of severe complications, such as hepatorenal syndrome [59, 62]. Besides altered contractility, a cardiac chamber hypertrophy and electrophysiological abnormalities, including electromechanical uncoupling, chronotropic incompetence and the electrocardiographic QT interval prolongation, have been also identified [59, 62].

The participation of the EC system in the pathogenesis of cirrhotic cardiomyopathy in cirrhosis was first investigated by Gaskari et al. who showed that AEA reduces the contractile response to isoprotenerol of cardiac papillary muscles isolated from rats with BDL-induced cirrhosis, which instead is restored by the CB1 receptor antagonist AM251 [63]. This effect was mediated through the activation of CB1 receptors [63], which is known to affect L-type calcium channels [64] and the myocardial cAMP content [65]. Moreover, the administration of VDM11 and AM404, two AEA reuptake inhibitors, produces a relevant relaxation in the cardiac papillary muscles isolated from cirrhotic but not in those from control animals [63]. Additionally, in contrast to what appears to occur in the splanchnic area, the vasorelaxant effect of AEA in the heart seems to be mediated only by the CB1 receptors and not by the TRPV channels [63]. These results were then confirmed in an in vivo model, as the increased myocardial AEA level has been related to a lower responsiveness to the β-adrenergic stimulation in BDL-induced cirrhotic rats [66]. Despite the above observation, the cardiac expression of the CB1 receptor seems to be not affected by cirrhosis [63, 66], leading to the hypothesis that the EC system lowers cardiac contractility only by increasing the ligand release [63, 66]. Finally, a recent paper suggests that inflammation represents the major trigger for AEA synthesis, as in vitro experiments in cultured cardiomyocytes from BDL cirrhotic rats produce large amounts of AEA under stimulation with tumour necrosis factor alpha [67].

Taken together, these experimental observations indicate that AEA can contribute to the pathogenesis of cirrhotic cardiomyopathy activating the CB1 receptor (Fig. 2). Thus, besides the positive effects on splanchnic and systemic hemodynamics, the inotropic activity of CB1 receptor antagonism may turn out to be useful in stressful conditions characterized by a depressed cardiac contractility, such as hepatorenal syndrome and sepsis.

The endocannabinoid and hepatic encephalopathy

Hepatic encephalopathy is characterized by central neurological symptoms ranging from lower attention to stupor and coma [68]. Although it was considered for long time as a consequence of ammonia neurotoxicity, it has been shown that the pathophysiological process leading to HE involves alterations in several neurotransmission systems, such as the GABA-ergic, monoaminergic and opioidergic [69, 70].

Regarding the EC system, increased 2-AG and CB1 and CB2 receptors have been recently measured in the brains of mice with thioacetamide-induced acute liver failure [69-71]. Furthermore, the inhibition of CB1 receptors by rimonabant or activation of CB2 receptors with HU-308 improves the cognitive function and the neurological score in mice with HE because of acute liver failure [69]. The direct involvement of CB2 receptors was confirmed by the development of neurological manifestations similar to HE in CB2 receptor knockout mice, although the contribution to the neuroprotective mechanism of other receptorial molecules, such as the TRVP1, was also suggested [72]. The involvement of non-CB receptors has been also demonstrated by the finding that the administration of cannabidiol (CBD), a non-psychoactive constituent of marijuana with no activity on CB1 or CB2 receptors, ameliorates the neuroinflammation and the neurological score during HE, suggesting instead the implication of the hippocampal A2A adenosine and 5-HT1A receptors [73, 74]. Thus, a combined administration of CB2 agonists and CBD could represent a novel potential pharmacological approach to treat HE [75].

The endocannabinoid system and bacterial infection

Cirrhosis is associated with a higher susceptibility to bacterial infections mainly resulting from intestinal bacterial translocation, caused by an excessive gut permeability and bacterial overgrowth, and from the impairment of the immune defence, caused by both genetic predisposition and acquired alterations of several components of the immune system [76]. Bacterial infections are also associated with a higher morbidity and mortality risk in cirrhotic patients. Once the infection is established, it frequently activates an abnormal inflammatory response that leads eventually to sepsis, septic shock and multiorgan failure. Indeed, increased levels of circulating LPS trigger the release of large amounts of proinflammatory cytokines and vasoactive mediators, which exacerbate the vasodilatation state and depress cardiac contractility, thus worsening the effective hypovolaemia of advanced cirrhosis.

Endogenous cannabinoids appear to act as an important mediator between bacterial infection and the hemodynamic alterations. Maccarrone et al. demonstrated that LPS increases AEA levels in human peripheral lymphocytes through the inhibition of FAAH, the main catabolic enzyme for AEA [49, 77]. Similarly, Varga et al. showed that rat platelets and macrophages are stimulated to produce 2-AG and AEA, respectively, when exposed to LPS [48]. More recently, the administration of ciprofloxacin was associated with a fall in the hepatic AEA and 2-AG levels which was accompanied by the amelioration of hepatic microcirculation and portal pressure [78]. Interestingly, in these experiments, the expression of the CB1 receptor in the liver declined after antibiotic treatment, while that of the CB2 receptor was conversely increased [78]. Preliminary results from our group showed that rimonabant administration improves survival in cirrhotic animals exposed to a single lethal dose of LPS and reduced the expression of both proinflammatory and vasodilating substances [79].

Besides the effect on systemic hemodynamics, recent evidence suggests that the EC system impairs the chemotactic response of macrophages in cirrhosis [80]. Even if AEA and 2-AG have been found to be powerful inducers of chemotaxis in macrophages and monocytes, including U937 cells [81], the mRNA expression of CB2 receptor was found to be reduced in circulating monocytes and peritoneal macrophages isolated from cirrhotic patients with spontaneous bacterial peritonitis [80]. From in vitro experiments in U937 cells, it appears that the downregulation of the CB2 receptor associated with endotoxemia is responsible for the blunted chemotactic response even in the presence of elevated EC levels as treatment with SR144528, a specific antagonist of the CB2 receptor, fully prevents the chemotaxis triggered by EC [80].

Based on the above experimental observations, it can be postulated that bacterial infections stimulate the synthesis of EC, which can precipitate the hemodynamics mainly through a CB1 receptor mechanism, and alter the immunological response mainly through a downregulation of the CB2 receptor signalling. Thus, the acute modulation of the EC system may represent a potential therapeutical target in the treatment approach against bacterial infection and related complications in advanced cirrhosis.

Research and therapeutic perspectives

Alterations in EC receptor expression have been documented to be implicated in the pathogenesis of several diseases, such as multiple sclerosis, atherosclerosis, rheumatic diseases, insulin resistance, obesity, allergic asthma, gut and liver diseases [82, 83].

In the last decade, thanks to a better identification of their specific signalling pathway, the EC modulation, in particular through CB1 receptor antagonism, has emerged as a new therapeutic prospective. Rimonabant, the first CB1 receptor selective antagonist, was commercialized to treat obesity, metabolic syndrome and insulin resistance [84, 85]. However, after the release on the market, the American (FDA) and European (EMA) regulatory agencies forced the withdrawal of the drug because of the increased incidence of psychiatric side effects, such as mood disorders that, in susceptible individuals, could also lead to major depression with high suicidal risk [86]. The use of rimonabant in the clinical setting was also associated with neurological problems, such as headaches and vertigo, and gastrointestinal disorders [84, 87, 88]. Subsequently, suspecting the same central side effects, the production of other promising CB1 receptor antagonists, namely taranabant [89] and CP945-598 [90], was interrupted.

Therefore, since then, the pharmaceutical research was focused on the identification of new molecules which do not cross the blood–brain barrier and modulate the EC system only in the peripheral compartment, thus preventing the central neuropsychiatric side effects [8, 91-93].

Preclinical data suggest that peripherally based CB1 receptor inverse agonists, such as JD5037 [94], and CB1 receptor neutral antagonists, such as AM6545 [95], could be an efficient class of molecules.

Like the first generation, the effectiveness of the new peripherally restricted CB1 receptor antagonist AM6545 has been initially investigated in the settings of metabolic disorders and obesity [95-97]. AM6545 treatment of ob/ob mice was associated with an improved glycaemic control and dyslipidemia, together with a reduced hepatic triglycerides content and plasma transaminases [95]. Similar results were also obtained with the peripheral CB1 receptor inverse agonist JD5037 [94].

Based on the fact that charged molecules are usually not able to cross the blood–brain barriers unless they are carried by specific transporters, permanently charged compounds, like alkyl pyridinium salts and N-oxides, as well as compounds with an high topological polar surface area, like sulphonamide and sulphamide, both acting as peripheral selective CB1 receptor antagonists, are currently under development. However, these compounds still await to be tested in in vivo experiments to further investigate their molecular pharmacology [93].

Besides the CB1 receptor antagonists, the pharmaceutical potentiality of CB2 receptor agonists has been revalued [98]. A therapeutic use of the CB2 agonist JWH-015 in the management of atherosclerosis has been already documented [99] and the CB2 agonist HU-308 was evaluated in the management of osteoporosis [100].

A new relevant pharmaceutical target is represented by FAAH and MAGL, the rate-limiting enzymes for the degradation of AEA and 2-AG. FAAH genetic or pharmacological inhibition has been reported as a relevant risk factor in fibrosis progression [101] and liver ischaemia-reperfusion injury [98], while inhibiting MAGL by JZL184 administration, has been already documented to protect against hepatic IR injury [28]. The opposite effect of FAAH and MAGL inhibition could be explained through the different affinities of AEA and 2-AG for the CB-2 receptor as an increased 2-AG bioavailability may result in an enhanced CB2 receptor activation. Taken together, these data suggest that potentiating the 2-AG signalling, but not that of AEA, could represent an effective therapeutic approach during the pathological condition characterized by elevated oxidative stress and inflammation.

In summary, besides their utility as pharmacological tools for understanding the pathophysiology of the EC system, novel compounds currently under investigation in preclinical settings may represent new drugs able to modulate the EC signalling without psychiatric side effects. In patients with cirrhosis, many potential indications can be anticipated for the use of the second-generation, peripherally restricted CB1 receptor antagonists from the chronic administration as antifibrotic agents to the short-term use against the hemodynamic disarrangement and the inflammatory response occurring during bacterial infections.