- Top of page
Cardiac contractility in cirrhosis is normal at baseline but hyporesponsive to stimuli, a phenomenon known as ‘cirrhotic cardiomyopathy’. The pathogenesis remains unclear. Endocannabinoids are vasoactive, but have not previously been examined in the cirrhotic heart. We therefore aimed to systematically clarify a possible role of endocannabinoids in the pathogenesis of cirrhotic cardiomyopathy.
Cirrhosis was induced in Sprague–Dawley rats by bile duct ligation; controls underwent a sham operation. At 4 weeks after operation, isolated left ventricular papillary muscle contractility was studied.
Dose–response curve for a β-adrenergic agonist isoproterenol was constructed in the presence and absence of a CB-1 antagonist AM251 (1 μM). Cirrhotic muscles had a blunted response to isoproterenol, which was completely restored by AM251.
Dose-response curves to anandamide, and CB-1 and CB-2 protein and mRNA expression in Western blot and reverse transcriptase–polymerase chain reaction experiments were not significantly different between cirrhotic and sham muscles.
Force–frequency relationship studies were performed in cirrhotic and normal muscles. At higher frequencies, anandamide reuptake blockers (VDM11 and AM404) significantly enhanced muscle relaxation in cirrhotic muscles, but not in controls. This effect was completely blocked by AM251 and pertussis toxin, whereas tetrodotoxin partially reversed it.
Taken together, these results indicate a pathogenic role for increased local (neuronal) production of endocannabinoids, mediated by a Gi-protein-dependent CB-1-responsive pathway in cirrhotic cardiomyopathy. The increased tachycardia-stress-induced release of endocannabinoids may help explain why contractility is normal at baseline but attenuated with stress.
- Top of page
Although baseline cardiac output is increased in cirrhosis, ventricular contractile responsiveness to stimuli such as hemorrhage, exercise, volume expansion and inotropic drugs is blunted (Caramelo et al., 1986; Ramond et al., 1986; Bernardi et al., 1991; Ingles et al., 1991; Grose et al., 1995). This attenuated ventricular responsiveness in the face of an increased resting cardiac output has been termed cirrhotic cardiomyopathy (Ma & Lee, 1996). We have previously explored a number of possible pathogenic mechanisms for cirrhotic cardiomyopathy from cardiac adrenergic and cholinergic signal transduction pathways to cell membrane fluidity (Ma et al., 1996; 1997; Liu et al., 2000; 2001). However, the exact mechanism of this phenomenon is not yet completely clarified, in particular, the reasons for the depressed responses despite baseline hypercontractility.
Cannabinoids are derived from a tropical plant Cannabis sativa. The active substance in these compounds is Δ9-tetrahydrocannabinol. Cloning of peripheral and central receptors for exogenous cannabinoids suggested the existence of endogenous ligands for these receptors. These natural ligands are lipid-like substances called endocannabinoids, which include arachidonoyl ethanolamide or anandamide, and 2-arachidonoylglycerol (Petrocellis et al., 2004).
It has been shown that anandamide dose-dependently decreases the contractile force of electrically driven isolated rat heart and human atrial muscle strips (Ford et al., 2002; Bonz et al., 2003). There is also evidence supporting local production of endocannabinoids in mammals (Moesgaard et al., 2002), but to our knowledge alterations in local cardiac production of endocannabinoids in pathologic states have not been studied yet.
On the other hand, Batkai et al. (2001) and Ros et al. (2002) reported increased anandamide levels and expression of cannabinoid receptor-1 (CB-1) in hepatic vascular endothelial cells of cirrhotic rats and patients. Since mechanisms in peripheral vessels and cardiac tissue are often similar, we were interested in a possible role of the endocannabinoid system in the pathogenesis of cirrhotic cardiomyopathy in a rat model.
- Top of page
Interest in cirrhotic cardiomyopathy has been burgeoning recently. In particular, the increasing use of procedures that stress the heart, such as transjugular intrahepatic portosystemic stent-shunts and liver transplantation, has highlighted the clinical significance of this syndrome. Cardiac failure accounts for 7–15% of postoperative mortality following liver transplantation (Rayes et al., 1995; Myers & Lee, 2000). Moreover, cirrhotic cardiomyopathy may contribute to the pathogenesis of hepatorenal syndrome (Lee, 2003; Ruiz-del-Arbol et al., 2003).
A number of possible mechanisms have been investigated to explain the inability of the cirrhotic heart to respond properly to inotropic stimuli, but it is likely that several mechanisms are involved (Liu & Lee, 1999). Previous studies on endocannabinoids and cardiac function in cirrhosis are practically nonexistent. Ross et al. reported an increase in the circulating level of anandamide and increased expression of CB-1 receptors in hepatic vascular endothelial cells of cirrhotic rats. They showed that i.v. administration of a CB-1 antagonist, SR141716, increases the blood pressure without affecting the cardiac output. From these findings they concluded that CB-1 antagonism abrogates the hyperdynamic circulation in cirrhosis by eliminating the vasodilatory effect of circulating endocannabinoids (Ros et al., 2002). However, no firm conclusions about ventricular function can be drawn from that study because cardiac output is affected by many other factors such as body temperature, cardiac contractile force, blood volume, heart rate, peripheral vascular resistance (afterload) and central venous pressure (preload) (Hayes et al., 1984; Kissling et al., 1993).
We therefore decided to investigate systematically the effects of the endocannabinoid system in the pathogenesis of blunted cardiac function of cirrhosis. We studied isolated left ventricular papillary muscles in vitro in order to eliminate the confounding effects of the autonomic nervous system and peripheral vascular tone that are inevitable in an in vivo study.
The inotropic response of BDL papillary muscles to isoproterenol was significantly blunted compared to that of sham-control animals. This finding agrees with previous studies on BDL-cirrhotic rats (Ma et al., 1997) as well as cirrhotic patients (Moller et al., 2001).
Preincubation of sham papillary muscles with anandamide blunted their contractile response to isoproterenol. This effect was completely blocked by preincubation with AM251, confirming that this effect of anandamide was mediated by CB-1 receptors. These results indicate that increased endocannabinoid signaling through the CB-1 receptor can blunt the ventricular responsiveness to β-adrenergic stimuli. In other words, it suggests that ventricular exposure to endocannabinoids might be partly responsible for the blunted function of the cirrhotic heart.
Incubation of BDL papillary muscles with AM251 restored their blunted responsiveness to the sham-control level. This finding further suggests a role for overactivity of the endocannabinoid system, mediated via CB-1 receptors in the cirrhotic heart. CB-1 receptors may show some degree of constitutional activity even in the absence of any ligand in the environment (Gifford et al., 2000; Christopoulos et al., 2001; Wilson & Nicoll, 2001). Considering that our experiments were performed in an organ bath where circulating endocannabinoids could not affect the cardiac tissue, this suggested an increased density of CB-1 receptors in the cirrhotic heart.
However, expression of CB-1 and CB-2 mRNA and protein were not significantly different between the sham and BDL groups. Moreover, dose–response curves to anandamide were similar in sham and BDL muscles, suggesting no alteration in the density or sensitivity of CB-1 and CB-2 receptors in the cirrhotic heart. Blockade of the relaxant effect of anandamide in sham and BDL muscles after AM251 incubation confirmed that anandamide had a receptor-mediated effect.
The above-mentioned experiments convinced us that although blocking CB-1 receptors could restore the blunted function of cirrhotic papillary muscles, upregulation of ventricular cannabinoid receptors was not the responsible mechanism. An alternative explanation could be increased local production of endocannabinoids in the heart. Anandamide is synthesized in cardiac tissue of mammals including cats, dogs and rats (Moesgaard et al., 2002). We hypothesized that local anandamide production in response to stimuli such as tachycardia or β-adrenergic agonists might be increased in cirrhotic hearts.
Anandamide is normally rapidly transported or diffuses across the cell membrane and then degraded by the enzyme FAAH. In order to discern the effect of any possible difference in local anandamide production, two structurally different anandamide reuptake blockers (VDM11 and AM404) were used. These compounds prolong the effect of locally released endocannabinoids by inhibiting their cellular reuptake and subsequent hydrolysis by FAAH (Bifulco et al., 2004). In our preliminary studies, a standard dose–response curve to VDM11 at a frequency of 1 Hz showed a nonsignificant tendency toward increased relaxation in BDL papillary muscles (data not shown). If our hypothesis of increased local endocannabinoid release in response to stress was to be adequately tested, a stronger stress stimulus would be required. We therefore examined higher contraction frequencies with the idea that relative tachycardia might be sufficiently stressful. Therefore, the effect of the anandamide reuptake blockers was examined using force–frequency response plots. There was no significant difference between the response of sham and BDL muscles to higher frequencies of contraction. However, VDM11 and AM404 significantly increased the relaxation of BDL papillary muscles at higher frequencies, whereas the reuptake blockers had no significant effect in the sham controls. These findings strongly support the hypothesis of enhanced stress-induced endocannabinoid release in the cirrhotic heart.
It has been reported that AM404 also blocks brain FAAH activity (Jarrahian et al., 2000), which may complicate the interpretation of our study. On the other hand, it is shown that in most cases, but not all, VDM11 is a weak inhibitor of FAAH (Fowler et al., 2004). The similarity of effect of these two compounds and their ineffectiveness in sham-control muscles suggest that anandamide reuptake blockade was the main mechanism responsible for these effects, rather than inhibition of FAAH enzymatic activity. This idea is further supported by equal expression of FAAH in sham and BDL ventricles in our experiments.
In order to investigate whether the observed effects were mediated through CB-1 receptors, papillary muscles were pretreated with AM251 before adding VDM11 or AM404. AM251 completely blocked the effect of these compounds, indicating that this was indeed a selective receptor-mediated effect.
It is known that anandamide can act on vanilloid TRPV1 receptors under pathological conditions (Di Marzo et al., 2001; Ross, 2003). Domenicali et al. (2005) have recently reported that the effect of anandamide in mesenteric arteries of CCl4-induced cirrhotic rats is partly mediated by vanilloid TRPV1 receptors. Moreover, Pan & Chen (2004) have recently demonstrated a pathophysiological role for these receptors in the ischemic heart. Considering the depressant effect of TRPV1 receptors on cardiac contractility, it is possible that the relaxant effect of anandamide reuptake blockers in our experiments was mediated through vanilloid receptors. However, complete restoration of the observed effects with AM251, which does not have an antagonistic effect on TRPV1 receptors, makes such an involvement unlikely. Furthermore, AM404, which is also a TRPV1 agonist, and VDM11, which is not (De Petrocellis et al., 2000), appeared to be equally effective in BDL rats.
AM404 and VDM11 in high concentrations may exert a direct agonist effect on CB-1 receptors (De Petrocellis et al., 2000). However, it is unlikely responsible for their effect in our experimental setting as they did not have a similar relaxant effect in sham control rats. Specifically, the dose–response curve to anandamide constructed at different frequencies of contraction (0.5, 1.5 and 2 Hz) were identical in sham and BDL rats, while there was a significant difference in their response to VDM11 and AM404. These findings further ruled out a direct CB-1 action as the mechanism responsible for the effect of reuptake blockers in our experiments.
Further experiments were performed to confirm that the effect of VDM11 was mediated through CB-1 receptors. The cardiac effects of CB-1 receptors are thought to be mediated by inhibitory G-proteins (Randall et al., 2002). PTX treatment blocks Gi-protein activity (Otani et al., 2001) and therefore should attenuate the signaling through cannabinoid receptors. PTX pretreatment blocked the effect of VDM11 on BDL papillary muscles, while it had no significant effect on sham muscles. This suggests that the effect of anandamide reuptake blockers was mediated through Gi-protein. This finding together with the result of the AM251 experiments strongly suggests that the effect of VDM11 and AM404 was selectively mediated through CB-1 receptors and also suggests increased local endocannabinoid production in the cirrhotic papillary muscles.
In order to examine the role of CB-2 receptors, muscles were incubated with AM630 (1 μM) before VDM11 or AM404 administration. This manipulation partially restored the effect of VDM11, but failed to significantly affect the response of BDL muscles to AM404. AM630 (1 μM) not only blocks CB-2 receptors but also inhibits the activity of CB-1 receptors by 23% (Ross et al., 1999). Considering the above-mentioned observations and complete blockade of the effect of VDM11 and AM404 by AM251, which is a selective CB-1 antagonist, we believe that CB-1 and not CB-2 is the main receptor responsible for these effects; however, further experiments are required to elucidate the role of CB-2 receptors in this situation.
The next step was to determine the source of increased endocannabinoid production in the cirrhotic papillary muscle. This local endocannabinoid could originate from cardiomyocytes, neurons or trapped immune cells in the tissue. To rule out a neuronal source, papillary muscles were preincubated with TTX to paralyze the neurons without affecting the function of cardiomyocytes (Campbell et al., 2004). TTX blocks neuronal voltage-dependent Na+ channels, but has no effect on voltage-sensitive Ca2+ channels, which are critical for cardiomyocyte excitation. TTX blocked the effect of VDM11 by approximately 70%, suggesting a predominant neuronal source for increased endocannabinoid production, and also a contribution from other sources.
It has been shown that TTX at 100 nM concentration can completely block cardiac neuronal elements but only weakly inhibits myocardial Na+ channels (Honerjager, 1982; Lewis & Endean, 1986). Based on these findings, we used this dose in our protocol. It was possible that the concentration of TTX in the inner regions of the papillary muscle might be less than that of the organ bath, suggesting an even greater neuronal origin of endocannabinoids. In order to investigate this possibility, higher doses of TTX were tested as well (1 and 10 μM). These higher concentrations did not have any additional effect confirming that non-neuronal sources also contribute, in a minor way, to local endocannabinoid production.
The above-mentioned experiments strongly suggest an increase in local endocannabinoid release in cirrhotic ventricles in response to stress. The enhanced levels could be explained by either an increase in local anandamide synthesis or a decrease in anandamide degradation rate. However, FAAH protein expression in the ventricles did not differ between control and BDL rats, suggesting that the anandamide degradation process was intact in the cirrhotic heart.
Shah (2001) has previously suggested that endocannabinoid effects in the peripheral vasculature of cirrhosis are mediated by an NO-dependent mechanism. Moreover, we have previously demonstrated a significant negative-inotropic role of NO in the cirrhotic rat heart (Liu et al., 2000). We therefore used the NOS inhibitor L-NAME to investigate a possible interaction between the cannabinoid and NO systems. Intact force–frequency relationship in sham papillary muscles after L-NAME pretreatment agrees with previous reports in normal rats (Prabhu et al., 1999). Furthermore, L-NAME did not have a significant effect on the responsiveness of BDL muscles to VDM11. This agrees with previous studies in noncirrhotic rat hearts showing that NOS blockers do not affect anandamide responses (Bonz et al., 2003). These results rule out an involvement of NO-mediated mechanisms in the cardiac endocannabinoid effect and further support a selective and independent CB-1 effect.
In summary, the present study showed that local ventricular endocannabinoid production is increased in cirrhosis and this plays a significant role in blunted contractile responsiveness. These effects are mediated through CB-1 and not CB-2 receptors. These results suggest that the endocannabinoid system may contribute to the pathogenesis of cirrhotic cardiomyopathy.