|Home | About | Journals | Submit | Contact Us | Français|
Chronic liver disease is associated with endotoxemia, oxidative stress, increased endocannabinoids and decreased cardiac responsiveness. Endocannabinoids activate the tumor necrosis factor-alpha (TNFα)–nuclear factor κB (NFκB) pathway. However, how they interact with each other remains obscure. We therefore aimed to clarify the relationship between the TNFα–NFκB pathway and endocannabinoids in the pathogenesis of cardiodepression of cholestatic bile duct ligated (BDL) mice.
BDL mice with TNFα knockout (TNFα−/−) and infusion of anti-TNFα antibody were used. Cardiac mRNA and protein expression of NFκBp65, c-Jun-N-terminal kinases (JNK), p38 mitogen-activated protein kinase (p38MAPK), extracelullar-signal-regulated kinase (ERK), inducible nitric oxide synthase (iNOS), Copper/Zinc and Magnesium-superoxide dismutase (Cu/Zn- and Mn-SOD), cardiac anandamide, 2-arachidonoylglycerol (2-AG), nitric oxide (NOx) and glutathione, and plasma TNFα were measured. The effects of TNFα, cannabinoid receptor (CB1) antagonist AM251 and the endocannabinoid reuptake inhibitor UCM707, on the contractility of isolated cardiomyocytes, were assessed.
In BDL mice, cardiac mRNA and protein expression of NFκBp65, p38MAPK, iNOS, NOx, anandamide, and plasma TNFα were increased, whereas glutathione, Cu/Zn-SOD, and Mn-SOD were decreased. Cardiac contractility was blunted in BDL mice. Anti-TNFα treatment in BDL mice decreased cardiac anandamide and NOx, reduced expression of NFκBp65, p38MAPK, and iNOS, enhanced expression of Cu/Zn-SOD and Mn-SOD, increased reductive glutathione and restored cardiomyocyte contractility. TNFα-depressed contractility was worsened by UCM707, whereas AM251 improved contractility.
Increased TNFα, acting via NFκB–iNOS and p38MAPK signaling pathways, plays an important role in the pathogenesis of cardiodepression in BDL mice. TNFα also suppressed contractility by increasing oxidative stress and endocannabinoid activity.
In chronic liver disease, several factors including bacterial translocation and release of endotoxins stimulate the release of tumor necrosis factor-α (TNFα) and endocannabinoids to depress cardiac contractility [1–5]. Previous studies confirmed that cirrhotic cardiomyopathy is associated with markedly elevated cardiac levels of TNFα and endocannabinoids [5–8]. TNFα signaling is very complex. It activates many intracellular signaling pathways, including nuclear factor κB (NFκB) and three mitogen-activated protein (MAP) pathways, which include the extracellular-signal-regulated kinases (ERK), c-Jun N-terminal kinase (JNK), and p38 MAP kinase (p38MAPK) .
JNK and ERK are cardioprotective factors, while p38MAPK has negative effects on heart contractility [10,11]. Our recent study showed that inhibition of the NFκB activity improves the contractility of cirrhotic hearts . NFκB activates transcription of inducible nitric oxide synthase (iNOS) to produce nitric oxide (NO) and subsequently cGMP [9,13]. We previously showed that the iNOS–NO–cGMP pathway plays an important role in the development of cirrhotic cardiomyopathy .
It is known that TNFα increases endocannabinoid synthesis in macrophages . However, the pathogenic mechanisms of increased endocannabinoids in the cholestatic heart have not been studied yet. We hypothesized that there are additive or synergistic effects on cardiac inhibition between endocannabinoids and TNFα in the heart of mice with cholestatic fibrosis.
Although evidence has suggested the possible roles of increased TNFα and endocannabinoids in the cirrhotic heart [5,8], the exact cellular mechanism of these factors in the development of cholestasis-induced cardiac dysfunction is not yet completely understood. The present study was therefore designed to (1) explore the pathophysiological roles of TNFα and its signaling pathways, including NFκB–iNOS, ERK, JNK, p38MAPK, and endocannabinoids, and (2) clarify the effects of TNFα in cholestasis-induced cardiac dysfunction by using a BDL-induced liver injury model in genetic TNFα-deficient mice, and wild-type mice receiving neutralizing TNFα antibody.
The protocols were approved by the Animal Care Committee of the University of Calgary Faculty of Medicine, under the guidelines of the Canadian Council on Animal Care. Male 22–24 g TNFα knockout (TNFα−/−, C57BL/6J-TNG tm1GK1) mice and age-matched C57BL/6J wild-type (WT) controls were obtained from the Jackson Laboratories (Bar Harbor, ME, USA). The animals were maintained on a 12-h light/dark cycle under controlled temperature (18–21 °C) and humidity and they had free access to food and water. Mice were divided randomly into sham-operated control groups (sham) and bile duct ligation (BDL) groups. In total, 15 TNFα−/− mice (9 for BDL and 6 for sham-operation) and 53 TNFα+/+ (wild-type) mice (28 for BDL and 25 for sham-operation) were used.
Bile duct ligation was performed under sterile conditions as described previously . Sham animals underwent the same surgery except bile duct ligation and section. Animals were studied two weeks after BDL or sham surgery. Previous studies showed that 4–6 weeks of BDL fail to induce cirrhosis in mice [16,17]. In our pilot studies, even 8 weeks of BDL failed to induce cirrhosis and markedly increased the mortality rates; thus the 2-week period was chosen for this study.
Anti-TNFα antibody was purchased from BioLegend Inc., (San Diego, CA, USA). UCM707 and AM251 were from Tocris Cookson Ltd. (Elisville, MO, USA). Primary antibodies (NFκBp65, JNK, p38MAPK, iNOS, Cu/Zn-SOD, and G3PDH) and secondary antibodies were purchased from Cell Signaling Technology, Inc. (Boston, MA, USA) and Santa Cruz Biotechnology Inc. (Santa Cruz, CA, USA). Other reagents were purchased from Sigma, Bio-Rad (Hercules, CA, USA), or Fisher Scientific (Pittsburgh, PA, USA).
A total of six groups were studied. Two groups of TNFα knockout mice (TNFα−/−) were used; one group (n = 9) was subjected to bile duct ligation, while the other group (n = 6) was sham-operated. Four groups of TNFα wild-type (TNFα+/+) mice included: sham controls receiving IgG vehicle solution injections (sham-V, n = 13), BDL controls receiving vehicle (BDL-V, n = 16), sham receiving anti-TNFα antibody (sham-anti-TNFα, n = 12), and BDL receiving anti-TNFα antibody (BDL-anti-TNFα, n = 12). The rationale for using the anti-TNFα antibody was to neutralize the excessive amount of plasma TNFα in BDL mice. The anti-TNFα antibody 9 µg was injected i.p. every 4 days after surgery, for two weeks . The same dose of mouse IgG (Sigma, Chemical) was given to BDL-V and sham-V mice serving as controls.
Liver tissue was immediately fixed with 10% formalin in phosphate buffered saline (PBS). Samples were later embedded in paraffin and sectioned (3 µm). For the assessment of hepatic fibrosis, sections were mounted on glass slides and deparaffinised, then immersed for 10 s in saturated aqueous picric acid containing 0.1% Sirius Red F3BA (Polysciences Inc., Warrington, PA, USA), which selectively binds to collagenous proteins.
Ventricular myocytes were isolated from murine hearts using the methods described previously . Cell contraction and relaxation were assessed using a video sarcomere detector (IonOptix Corporation, Milton, MA, USA). Briefly, cardiomyocytes were placed in a Warner chamber mounted on the stage of an inverted microscope (Nikon, Tokyo, Japan) and superfused (~1 ml/min at 25 °C) with a standard Tyrode solution and continuously gassed with 100% O2. The cells within the recording chamber were continuously field stimulated at a rate of 1 Hz, and the contractile response to this stimulation was taken as baseline value. The cardiomyocyte being studied was displayed on a computer monitor using an IonOptix MyoCam camera. Myocytes were allowed to equilibrate to this stimulus for a minimum of 5 min before isoproterenol (10−5 M) stimulation. Calibrated output of the video sarcomere detector was digitized for off-line analysis. Peak cell shortening, maximal contraction, and relaxation velocities, time from basal length to 50% of peak contraction, and time from peak contraction to 50% of basal length in the diastolic phase were recorded and analysed.
Isolated cardiomyocytes from BDL-V, sham-V, BDL-anti-TNFα, sham-anti-TNFα, BDL-TNFα−/−, and sham-TNFα−/− mice were measured for cell contraction and relaxation function (n = 15 in each group).
The acute effects of TNFα and the cannabinoid CB1 receptor antagonist AM251, and the reuptake inhibitor UCM707 on contractility were examined. Cardiomyocytes were incubated in storage solution (Tyrode solution containing 1.1 mM CaCl2 and bovine serum albumin 50 mg/ml, pH 7.4, at 30 °C) with vehicle (mouse IgG), TNFα (200 pg/ml), UCM707 (3 × 10−6 M), AM251 (10−6 M). In sham-V mice, cell contraction, and relaxation were measured after acute incubation with TNFα, TNFα + UCM707, AM251, and TNFα + AM251 for 30 min [18,19].
Plasma and hearts were collected from the mice two weeks after surgery. An ELISA assay kit for TNFα was purchased from Biosource (Camarillo, CA, USA). The results were expressed as pg/ml.
For cardiac nitrite/nitrate (NOx) measurement, hearts (50 mg) were homogenized in phosphate buffer saline (200 µL) and centrifuged at 4000g for 10 min. The supernatant was measured by commercial available ELISA kits (Cayman Chemical, Ann Arbor, MI, USA). The results were expressed as µmol/mg.
For glutathione detection, an ELISA kit was purchased from BioVision (Mountain View, CA, USA). The reduced cardiac glutathione (GSH) level was calculated as the difference between total glutathione and oxidized glutathione (GSSG). The results were expressed as µg/mg.
Protein expression of cardiac NFκBp65, JNK, p38MAPK, iNOS, and Cu/Zn-SOD were quantified as previously described [15,18,20]. G3PDH (glyceraldehydes-3-phosphate dehydrogenase) expression was used as an internal control.
Cardiac TNFα, NFκBp65, p38MAPKβ, p38MAPKα, and ERK1 mRNA transcription was measured by semi-quantitative RT-PCR as in previous studies [15,18,20]. G3PDH served as an internal control. Primers synthesized by Gibco-BRL Life Technologies (Burlington, ON, Canada) are listed in Table 2.
Further quantification of the iNOS, Cu/Zn-SOD, and Mn-SOD mRNA levels were done by SYBR Green two-step real-time PCR. Sequences of the primers for iNOS and G3PDH are listed in Table 2. Real-time PCR reactions were carried out in a final volume of 25 µl of reaction mixture containing 10 ng of RNA, 12.5 µl of 2× SYBR Green Master Mix (Stratagene), 75 nM of each specific gene primer, and H2O. The samples were run in triplicate, and the mean value was used as the final expression value. A negative control without RNA template was run. Data were analyzed according to the relative standard curve method. CT values obtained for iNOS were first normalized with that of G3PDH prior to analysis. Fold-change in iNOS, Cu/Zn-SOD, and Mn-SOD mRNA was then calculated relative to control (i.e. sham-V) using the comparative CT method as described in the Applied Biosystems User Bulletin #2.
Mice were euthanized, their hearts were excised, and the lipids were extracted. Cardiac levels of anandamide (AEA) and 2-archidonoyl glycerol (2-AG) were quantified by liquid chromatography/in line mass spectrometry by using an Agilent 1100 series LC–MSD, equipped with a thermostat autosampler and column compartment . Values are expressed as fmol or pmol per mg of wet tissue.
All statistical analyses were performed using SPSS 10.0 for Windows (SPSS Inc., Chicago, IL, USA). The results are expressed as means ± SEM An unpaired Student’s t-test was used to compare the differences between two groups, and for more than two groups, comparisons were analyzed by one-way ANOVA followed by Newman–Keuls post hoc test. A p-value of <0.05 was considered to be significantly different.
BDL-V livers compared with those of sham-V mice showed (Fig. 1A), significantly increased collagen deposition and bile duct proliferation in periportal areas (Fig. 1B), organization of fibrotic tissue (Fig. 1C), and mild periportal bridging (Fig. 1D). In contrast, the hearts of BDL-V mice appeared normal and were indistinguishable from sham-V mice on routine light microscopy (not shown).
Compared with sham controls, the maximum systolic and diastolic velocities of cardiomyocytes from BDL mice were significantly reduced. Time to half-peak contractions (TP50) and half-complete relaxations (TR50) were significantly prolonged (Fig. 2). Anti-TNFα antibody significantly improved systolic and relaxation velocities (Fig. 2). Compared with basal conditions, the mean changes in maximal contraction velocity were not significantly different from maximal relaxation velocity as a result of anti-TNFα antibody administration (25.1 ± 2.5 vs. 22.0 ± 8.9%). The results of cardiac function in anti-TNFα-treated BDL mice were similar to those in TNFα knockout mice (BDL-TNFα−/−). Neither anti-TNFα antibody nor TNFα knockout affected the contractility/relaxation from the sham groups.
TNFα incubation with cardiomyocytes from sham mice significantly decreased systolic and relaxation velocities and prolonged TP50 and TR50. Co-incubation of AM251 (CB1 antagonist) and TNFα with cardiomyocytes from sham mice did not impact the contractile and relaxation velocities (Fig. 3). However, the endoannabinoid reuptake inhibitor UCM707 further enhanced the TNFα depressive effect on contractile and relaxation velocities, whereas AM251 partially improved TNFα-inhibited cardiac function.
Compared with sham mice, BDL mice showed higher plasma TNFα, cardiac anandamide, NOx, and lower total and reduced glutathione (Table 1). Anti-TNFα treatment significantly decreased plasma TNFα, cardiac anandamide, NOx, and increased total and reduced glutathione in BDL mice. In BDL-TNFα−/− mice, cardiac anandamide and NOx, and total and reduced glutathione were lower compared to BDL wild-type mice. However, there was no difference in cardiac 2-AG, NOx, and total and reduced glutathione between sham and sham-anti-TNFα mice (Table 1).
Cardiac NFκBp65, p38MAPK, and iNOS were significantly increased and Cu/Zn-SOD was significantly decreased in BDL-V mice (Fig. 4). In anti-TNFα treated BDL and BDL-TNFα−/− mice, lower NFκBp65, p38MAPK, and iNOS; and higher Cu/Zn-SOD protein expression than in BDL-V mice were observed (Fig. 4). However, cardiac JNK expression was not different between any groups of mice (data not shown).
In contrast to the increase in cardiac TNFα mRNA in BDL-V mice, the genotype of the TNFα knockout mice was confirmed by non-detectable cardiac TNFα mRNA in BDL-TNFα−/− mice (Fig. 5A). Compared with the sham group, mRNA expression of cardiac NFκBp65, p38MAPKα, and p38MAPKβ was significantly increased in BDL-V mice (Fig. 5B–D). Cardiac mRNA transcription of NFκBp65, p38MAPKα, and p38MAPKβ was significantly lower in anti-TNFα treated BDL and BDL-TNFα−/− mice compared with the BDL-V group (Fig. 5B–D). Cardiac ERK1 mRNA expression was not different between BDL-V and sham-V mice (data not shown).
Parallel with the increased cardiac iNOS and Cu/Zn-SOD proteins in BDL-V mice; cardiac iNOS, Cu/Zn-SOD, and Mn-SOD mRNA, measured by real-time RT-PCR, was increased in these animals. Anti-TNFα antibody-treated and TNFα knockout BDL mice showed lower expression of these mRNAs (Fig. 6).
The mechanisms underlying cardiac dysfunction in chronic liver disease are only partially known but previous work has suggested a role for increased circulating TNFα and endocannabinoids [8,12]. TNFα, in particular, is thought to be a key mediator of cardiodepression in chronic liver disease [1,4,12,19]. In the current study, pharmacologic/genetic manipulation of TNFα was performed in BDL mice due to availability of genetic TNFα knockout mice as well as anti-TNFα antibody-treated BDL mice. The anti-TNFα antibody we used was able to significantly neutralize the excessive TNF-alpha levels in BDL-V rats. Most importantly, we sought to determine whether TNFα and endo-cannabinoids are involved in the cardiac dysfunction in a murine model of fibro-cholestatic liver disease. In the present study, BDL induced: fibrosis, ductular proliferation, and architectural damage, as previously observed [16,17]; however, the damage stopped short of cirrhosis. In terms of functional depression of cardiac contractility, this BDL mouse model showed similar changes to the chronic BDL rat model which has an established biliary cirrhosis. This suggests that cirrhosis per se is not required to induce cardiac dysfunction—the presence of a moderate degree of liver damage may be enough. Moreover, cholestasis by itself can also induce cardiac dysfunction [21–23], and this may contribute to the overall pathogenesis of cirrhotic cardiomyopathy.
The diverse and numerous signaling pathways of TNFα examined in this study make our paper difficult to cleanly interpret but in our opinion, were unavoidable. Focusing on a single pathway would likely have led to a falsely oversimplified conclusion of the signaling pathways by which this key molecule mediates its effects on the cirrhotic heart. The results of the present study support the hypothesis that TNFα, through the NFκB–iNOS and p38MAPK signaling pathways, plays an important pathogenic role in the development of cholestasis-induced cardiac dysfunction. TNFα elimination, either by genetic deletion or by antibody neutralization, reduced the activities of the NFκB–iNOS and p38MAPK signaling pathways in BDL mice hearts and improved cardiomyocyte contractility. Isolated cardiomyocytes were used in our study to exclude other confounding effects such as the autonomic nervous system and the peripheral vascular tone.
In the present study, acute incubation of TNFα slowed contraction and relaxation velocities in sham cardiomyocytes to values similar to BDL mice. Moreover, we found that acute and chronic pharmacological inhibition of endogenous TNFα by anti-TNFα antibody and genetic deletion of TNFα, corrected the contraction and relaxation dysfunction in isolated cardiomyocytes from BDL mice.
Activated NFκB combines with the TNFα promoter to increase TNFα expression and subsequently trigger NFκB activity . Our previous study had demonstrated that inhibition of the NFκB activity was associated with a decrease in cardiac TNFα levels and an improvement of cardiomyocyte contractility in cirrhotic rat hearts . The current study showed that over-expression of TNFα mRNA levels was accompanied by activation of NFκB in the BDL mouse heart. In concordance with our previous study, the anti-TNFα-related down-regulation of NFκB was associated with the improvement of cardiomyocyte contractility in our BDL mice .
It has been reported that endotoxemia-induced myocardial dysfunction is caused by the release of cardiodepressants such as NO and cGMP [6,25,26]. We recently showed that the synthesis of NO is increased in the cirrhotic heart . In the present study, the increased plasma level of TNFα was associated with a significant elevation of cardiac NOx levels in BDL mice. Furthermore, genetic deletion and pharmacological inhibition of endogenous TNFα decreased NOx levels in BDL mice, and this was accompanied by correction of cardiomyocyte contractile dysfunction. ERK1/2 and JNK had been reported to be involved in the pathogenesis of cardiac 16 hypertrophy . However, there was no difference in cardiac JNK and ERK1 mRNA and protein expression between normal and BDL mice. These results indicated that the JNK and ERK pathways play little or no role in the cardiac dysfunction in our BDL mice. Among four isoforms of p38 (p38α, p38β, p38δ, and p38γ), p38α and p38β are thought to be primarily responsible for regulating inflammation [10,11]. In our current study, mRNA and protein expression of cardiac p38α and p38β MAPK were higher in BDL mice than in sham controls. Meanwhile, inhibition of endogenous TNFα suppressed the p38MAPK pathway in BDL mice. Previous studies indicated that activation of p38MAPK, by triggering NFκB–iNOS signaling pathways, blunts β-adrenergic response to cardiovascular stress, which is the main feature of cirrhotic cardiomyopathy [10,27–29]. In our BDL mice, activation of the p38 pathway might similarly up-regulate the NFκB–iNOS pathway and contribute to myocardial dysfunction. Our observation that the anti-TNFα-related correction of BDL cardiodepression was associated with the suppression of the NFκB–iNOS and p38MAPK pathways is consistent with this notion.
TNFα induces oxidative stress in adult rat cardiomyocytes and antioxidant (N-acetylcysteine) treatment prevents the depressive effect of TNFα on contraction . Both Cu/Zn-SOD (cytosolic) and Mn-SOD (mitochondrial) are critical determinants in the protection of the heart from oxidative injury . In this study, we estimated the cardiac antioxidative status by measuring the total, reduced, and oxidized glutathione levels, Cu/Zn-SOD, and Mn-SOD activity. We found that antioxidants were significantly decreased in BDL mice. Moreover, inhibition of endogenous TNFα significantly increased the antioxidants (SOD and glutathione) in BDL hearts.
The mechanism of contractile dysfunction in response to endotoxins and/or oxidative stress is complex. For example, endotoxins and reactive oxygen species per se may activate NFκB and iNOS, and not only through TNFα. This may explain why the pharmacological inhibition of endogenous TNFα in our study only partially restored TNFα-induced cardiomyocyte dysfunction.
We also observed that higher plasma TNFα was accompanied by elevated cardiac levels of anandamide in BDL mice. Moreover, anandamide levels were reduced in anti-TNFα-treated BDL and TNFα knockout mice. However, the detailed mechanisms of local production of endocannabinoids and its interaction with the TNFα–NFκB pathway remain unclear. In our study, TNFα-related cardiomyocyte dysfunction was aggravated by simultaneous incubation of the endocannabinoid reuptake inhibitor UCM707 and TNFα in sham cardiomyocytes. What are the possible mechanisms for TNFα to release endocannabinoids in the cirrhotic heart? Previous studies found that macrophages, in response to endotoxin, secrete TNFα which further increases oxidative stress and activates the endocannabinoid system [2,3,31,32]. Our results, combined with previous studies, imply that neutralizing TNFα relieves cardiac oxidative stress and subsequently suppresses the local release of endocannabinoids in fibro-cholestatic hearts (Fig. 7).
The activation of cannabinoid CB1 receptors decreases cardiac response to β-adrenergic agonists [5,8]. Parallel to the inhibition of NFκB activity, CB1 receptor antagonists improve cardiac function in cirrhotic rats [5,8,19]. In order to confirm whether TNFα enhances the release of endocannabinoids, sham cardiomyocytes were incubated with TNFα to mimic TNFα-induced BDL cardiac dysfunction in our study. Although direct incubation of CB1 receptor antagonist (AM251) did not affect the response of cardiomyocytes, concomitant incubation of AM251 and TNFα partially reversed the TNFα-induced blunted response to the ®-adrenergic agonist isopreterenol on cardiomyocytes. These experiments suggest that increased local release of endocannabinoids might be partially responsible for the TNFα-induced β-adrenergic hypo-responsiveness in the fibro-cholestatic heart.
In conclusion, increased TNFα, acting via NFκB–iNOS and p38MAPK signaling pathways, plays an important role in the pathogenesis of cholestasis-induced cardiac dysfunction. Additionally, TNFα-accentuated endocannabinoid action and oxidative stress might also be involved in its negative inotropic effects in fibro-cholestatic hearts. Accordingly, inhibition of TNFα may be potentially therapeutically useful to restore good cardiac contractility in patients with liver disease.
Conflicts of interest
The authors who have taken part in this study declared that they do not have anything to disclose regarding funding or conflict of interest with respect to this manuscript.