Interplay of liver-heart inflammatory axis and cannabinoid 2 receptor signalling in an experimental model of hepatic cardiomyopathy

Abstract

Background and rationale:

Hepatic cardiomyopathy, a special type of heart failure develops in up to 50% of patients with cirrhosis and is a major determinant of survival. However, there is no reliable model of hepatic cardiomyopathy in mice. Herein we aimed to characterize the detailed hemodynamics of mice with bile-duct ligation (BDL)-induced liver fibrosis, by monitoring echocardiography and intracardiac pressure-volume (PV) relationships and myocardial structural alterations. Treatment of mice with a selective cannabinoid-2 receptor (CB₂-R) agonist, known to attenuate inflammation and fibrosis, was used to explore the impact of liver inflammation, fibrosis on cardiac function.

Main results:

BDL induced massive inflammation (increased leukocyte infiltration, inflammatory cytokines and chemokines), oxidative stress, microvascular dysfunction, and fibrosis in the liver. These pathological changes were accompanied by impaired diastolic, systolic and macrovascular functions, cardiac inflammation (increased MIP1, interleukin-1, P-selectin, CD45+ cells) and oxidative stress (increased malondialdehyde, 3-nitrotyrosine and NADPH-oxidases). CB₂-R up-regulation was observed both in livers and hearts of mice exposed to BDL. CB₂-R activation markedly improved hepatic inflammation, impaired microcirculation, fibrosis. CB₂-R activation also decreased serum TNF-alpha levels, and improved cardiac dysfunction, myocardial inflammation and oxidative stress underlining the importance of inflammatory mediators in the pathology of hepatic cardiomyopathy.

Conclusion:

We propose BDL-induced cardiomyopathy in mice as a model for hepatic/cirrhotic cardiomyopathy. This cardiomyopathy, similarly to cirrhotic cardiomyopathy in humans, is characterized by systemic hypotension, impaired macro-and microvascular function accompanied by both systolic and diastolic dysfunction. Our results indicate that the liver-heart inflammatory axis has a pivotal pathophysiological role in the development of hepatic cardiomyopathy. Thus, controlling liver and/or myocardial inflammation (e.g. with selective CB2-R agonists) may delay/prevent the development of cardiomyopathy in severe liver disease.

It has been known for over sixty years that patients with end-stage liver disease have distinct cardiovascular pathology in the absence of other cardiovascular diseases. Subsequent studies discovered that cardiac output in these patients failed to adequately increase when subjected to exercise or pharmacological stress. Eventually, these findings were found to be related to decreased myocardial contractile function together with diastolic dysfunction(1).

This condition was described as hepatic/cirrhotic cardiomyopathy, which, by definition, is a chronic cardiac dysfunction in patients with liver failure/cirrhosis characterized by blunted contractile responsiveness to stress and/or altered diastolic relaxation with electrophysiological abnormalities, in the absence of known cardiac disease irrespective of the causes of cirrhosis(1). The gold standard therapy for end-stage liver failure is liver transplantation, however, its availability is largely limited to patients with stable cardiovascular status. Nonetheless, cardiovascular function is a major determinant of the clinical outcome of cirrhotic patients undergoing liver transplantation(2). Therefore, development of drugs to improve liver failure-associated cardiovascular dysfunction is desperately needed, which also requires the availability of a reliable animal model(s). Although several rat and mouse models of hepatic failure/cirrhosis have been described(3–7), to date no ideal animal model for hepatic cardiomyopathy has been proposed. The major limitation of model development lies in the application of different toxins (e.g. carbon tetrachloride(5)) widely used to generate liver failure. These compounds have an unspecific cellular toxicity affecting not only hepatocytes but also other cell types including cardiomyocytes(8). Hence, their use might lead to the development of toxic cardiomyopathy regardless of liver injury. Moreover, the detailed hemodynamic analysis and the underlying pathophysiology of hepatic cardiomyopathy has not yet been fully revealed.

Ejection fraction and fractional shortening are two of the most often used parameters of echocardiography to describe cardiac performance, however, their values are largely dependent on the heart rate and loading conditions of the heart. Pressure-volume (PV) analysis is the gold standard method used for the detailed characterization of cardiovascular function in animals(9). PV analysis-derived special, load-and heart rate-independent parameters of contractility, diastolic function and myocardial stiffness provide more accurate determination of cardiac performance in comparison to conventional echocardiography(10).

The role of endocannabinoid-cannabinoid 1 receptor (CB₁-R) signalling has been proposed in the development of liver injury-related cardiovascular pathologies (5, 11–13). Moreover, the beneficial role of cannabinoid type 2 receptor (CB₂-R) activation has been shown in models of liver injury, inflammation and fibrosis(14–17).

Therefore, we aimed to characterize the detailed hemodynamics and myocardial pathology in mice 2 weeks following the ligation of the common bile duct (BDL), a widely used model of liver fibrosis(18–20). We also investigated the effect of attenuation of the liver inflammation and fibrosis by using the selective CB₂-R agonist HU910(14, 21) on cardiac and vascular function in mice. We found that the development of liver fibrosis was associated with distinct cardiovascular changes including impaired cardiac performance and contractility along with decreased micro-and macrovascular function. Cardiac phenotype resembled the characteristics of heart failure and was characterized by myocardial inflammation and oxidative stress.

Materials and methods

For detailed description of Materials and methods applied please refer to the Supplementary material.

Animals

The investigation conformed to the National Institutes of Health guidelines and all animals received humane care according to the criteria outlined in the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication 86–23 revised 1985). Experiments were reviewed and approved by the Institutional Animal Care and Use Committee of the National Institute on Alcohol Abuse and Alcoholism. Young (12–16 weeks old), male C57BL/6J mice (The Jackson Laboratory, Bar Harbor, ME, USA) were used and were kept in a specific pathogen-free animal facility under constant temperature (22±2°C), humidity and with 12-h alternating light cycles and received regular chow diet. Non-fasted mice were used in all experiments. All experiments were performed during the light cycle of the animals.

Procedure of common bile duct ligation and drug treatment regimen

BDL was performed as described previously(19, 22) or in the Supplementary material. Experiments were done at 14 days postoperatively. Sham animals were used as controls (Sham). The CB₂-R agonist HU910 (Institute for Drug Research, The Hebrew University of Jerusalem, Israel) was prepared daily in a mixture of dimethyl sulfoxide, Tween-80, and distilled water in a ratio of 1:1:8(14). The solution was sterile filtered (Millex Syringe Filter, 0.22 μm pore size, Merck-Millipore, Burlington, MA, USA) and administered from day 1 to 14 once a day intraperitoneally (10 mg/kg). Animals were divided into 4 groups including Sham (vehicle-treated), Sham+HU910, BDL (vehicle-treated) and BDL+HU910.

Microvascular flow measurement of the liver

Liver microcirculation was assessed as described previously(22) and in the Supplementary material by laser speckle analysis.

Echocardiography

Echocardiography was executed as described previously(23) and in the Supplementary material.

Hemodynamic measurements

Invasive hemodynamic measurements were performed under general isoflurane (1–2%) anesthesia as described elsewhere(9, 23). Systolic, diastolic, mean arterial pressure, cardiac output, maximal slope of systolic pressure increment (dP/dt_max) and decrement (dP/dt_min), time constant of left ventricular (LV) pressure decay (Tau; Glantz and Weiss method), arterial elastance and total peripheral resistance, ejection fraction and LV end-diastolic pressure (LVEDP) values were obtained. The slope of the LV end-systolic PV relationship (ESPVR; linear model), its maximal value (Emax), the preload recruitable stroke work (PRSW), and the dP/dt_max-end-diastolic volume (EDV) were evaluated as load-and rate-independent contractility indices as previously described(9). After the hemodynamic measurements, animals were euthanized by exsanguination, blood was collected, and serum was prepared, LV and liver samples were flash frozen in liquid nitrogen and stored at −80°C for further experimentation. For histological purposes, LV and liver samples were fixed in 10% neutral buffered formalin and embedded in paraffin.

Serum biomarkers

Levels of alanine aminotransferase (ALT), alkaline phosphatase (ALP) were measured by the Idexx VetTest 8008 (Idexx Laboratories, Westbrook, ME) from the sera. Serum tumor necrosis factor alpha (TNF-alpha) was measured by an ELISA kit according to the manufacturer’s instruction (#MTA00B, R&D Systems, Minneapolis, MN, USA). Serum brain natriuretic peptide (BNP) levels were measured by an EIA kit according to the manufacturer’s recommendation (#RAB0386, Sigma-Aldrich/Millipore, St. Louis, MO, USA).

Cell isolation protocol

Cardiomyocytes, hepatocytes, Kupffer cells, liver sinusoidal endothelial cells (LSECs) and splenocytes were isolated from Sham (n=3) and BDL animals (n=3) as described previously(19, 24) and in the Supplementary materials.

Gene expression studies – quantitative real-time polymerase chain reaction (PCR)

For the RNA isolation protocol see the Supplementary materials. Quantitative real-time PCR experiments were done in heart and liver cDNA samples with SyberSelect PCR Master Mix (Thermo Fisher Scientific) on an ABI 7900HT Realtime PCR Instrument (Applied Biosystems, Foster City, CA). Relative quantification was calculated using the comparative CT method. Reactions were done in duplicates. The average value of housekeeping genes 18s, actin and hypoxanthine-guanine phosphoribosyltransferase (HPRT) was used as reference. The following targets were analyzed: alpha-and beta-myosin heavy chain (α-, β-MHC), atrial natriuretic peptide (ANP), myostatin, angiotensin II receptor 1a (AGTR1a), sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA2), interleukin 1α and β (IL1α, β), IL6, macrophage inflammatory protein 1α (MIP1α), TNF-α, intercellular-and vascular adhesion molecule 1 (ICAM-1, VCAM-1), E-and P-selectin, toll like receptor 4 (TLR4), gp22phox, gp91phox, p47phox, p67phox, cyclooxygenase 1 and 2 (COX1, 2), CB₁-R, CB₂-R, monoacylglycerol, diacylglycerol lipase α and β (MAGL; DAGLα; DAGLβ), fatty acid amide hydrolase (FAAH), transient receptor potential cation channel subfamily V member 1 (TRPV1) and G protein-coupled receptor 55 (GPR55). Primer sequences are described in Suppl. Table 1.

Gene expression studies – droplet digital PCR

Equal amounts of cDNA of heart, liver, spleen, brain, cardiomyocytes, hepatocytes, Kupffer cells, LSECs and splenocytes were used. Droplet Digital PCR (ddPCR) was performed using the QX200 Droplet Digital PCR System (Bio-Rad, Hercules, CA, California) to assess the CB₂-R expression levels. For the determination of CB₂-R copy numbers EvaGreen supermix (Bio-Rad) was applied in combination with CB₂-R-specific primers. Following droplet generation and thermal cycling, the absolute CB₂-R copy number was calculated by the QX200 Droplet Reader and QuantaSoft Software (Bio-Rad), with manual threshold adjustment by using the same threshold for all samples.

Histology

Five μm sections were cut for histological purposes. Histology and immunohistochemistries for hematoxylin-eosin, Sirius Red, CD45, 3-nitrotyrosine (3-NT), malondialdehyde (MDA) were performed as described in the Supplementary material or previously(19, 23, 24).

Statistics

Data are presented as mean±SEM. Shapiro-Wilk test was used to test normal distribution of our data. Depending on data distribution, Student’s t-test or Mann-Whitney test was used appropriately to compare two groups. Kruskal-Wallis ANOVA followed by Dunn`s post hoc test or one-way ANOVA followed by Tukey`s post hoc test were used appropriately depending on data distribution in case of multiple study groups. In case of the cell isolate experiments, statistical comparison of selected groups was done by one-way ANOVA coupled with Bonferroni’s multiple comparisons test. Statistics were done by GraphPad Prism 7 (San Diego, CA, USA). A p value of <0.05 was considered significant.

Results

Development of liver injury, fibrosis and microcirculatory failure in BDL

Liver pathology was investigated in a separate set of Sham and BDL animals. Liver injury was associated with perivascular and periductal inflammatory cell invasion and tissue disintegration on hematoxylin-eosin histology (Figure 1A) together with increased serum biomarkers of liver injury including higher levels of ALT (Figure 1A). Development of hepatic fibrosis was confirmed by Sirius Red histology and quantification (Figure 1b), while CD45 immunohistochemistry confirmed leukocyte accumulation in the liver tissue of BDL animals (Figure 1C). Importantly, the above pathology resulted in severe microcirculatory dysfunction in the liver of BDL mice (Figure 1D).

Figure 1. Liver injury, inflammation, fibrosis and microvascular dysfunction in bile duct ligation.

(A) Representative images of hematoxylin-eosin histology (Magnification: 100x) and alanine aminotransferase (ALT) results (n=8–10/group). (B) Sirius red staining (Magnification: 100x) and quantification results (n=8–11/group). (C) Images of CD45 immunohistochemistry (Magnification: 200x) and CD45+ cell quantification (n=5–6/group). (D) Representative images and analysis of liver microcirculation. (n=9–10/group). BDL: bile duct ligation *p<0.05 vs Sham

Systolic, diastolic and macrovascular dysfunction in BDL

The severe hemodynamic alterations observed in BDL mice included lower systolic, diastolic and mean arterial blood pressure values (Figure 2A and B) as well as the significant reduction of cardiac output both with echocardiography (14.0±1.7 mL/min vs. 8.8±0.4 mL/min; p<0.05) and PV analysis (Figure 2B). Interestingly, dP/dt_max did not change (Figure 2A and B) while ejection fraction (Echocardiorgaphy: 57±5% vs. 71±3%; PV analysis: 62±5% vs. 75±1%; p<0.05) and fractional shortening (Echocardiography: 31±4% vs. 45±4%; p<0.05) increased in BDL. In addition, macrovascular function significantly decreased as shown by the reduction of arterial elastance (Figure 2B). Diastolic dysfunction was characterized by significantly decreased early (E)/late (A) mitral inflow Doppler velocity ratio (1.51±0.12 vs. 1.16±0.06; p<0.05), the marked prolongation of myocardial relaxation (Tau_Glantz and TauWeiss), decline of dP/dt_min and increased LVEDP (Figure 2B). Total peripheral resistance did not change (Figure 2B).

Figure 2. Baseline hemodynamics and myocardial contractility.

(A) Representative curves of left ventricular (LV) pressure and the slope of pressure increment and decrement (dP/dt_max and dP/dt_min) in Sham and bile duct ligated (BDL) mice. (B) Blood pressure values, vascular functional parameters, conventional contractility parameters (cardiac output, maximal slope of pressure increment (dP/dt_max) and markers of diastolic function (time constant of left ventricular pressure decay (Tau_Glantz and Tau_Weiss), LV end-diastolic pressure (LVEDP) and maximal slope of pressure decrement (dP/dt_min). (C) Representative pressure-volume loops. ESPVR: end-systolic pressure-volume relationship. Arrow indicates the decrease of the slope of ESPVR in BDL. (D) Sensitive load-and rate-independent markers (slope of ESPVR, preload recruitable stroke work (PRSW), dP/dt_max-end-diastolic volume (EDV) relationships and Emax) of contractility. Dashed line rectangle indicates the marker (ESPVR) represented on PV loops. (n=7–8/group) *p<0.05 vs Sham

Cardiac contractility is severely impaired in BDL

Specific load-and heart rate-independent markers of cardiac contractility derived from the shift of PV relations following transient vena cava inferior occlusions, such as the slope of ESPVR, PRSW, dP/dt_max-EDV and Emax showed severe impairment of the systolic/contractile function of the myocardium in BDL animals compared to Sham (Figure 2C and D).

Tissue and cellular expression profile of the CB₂-R and cannabinoid signalling

CB₂-R expression was determined in different tissues and isolated cells from Sham and BDL animals. CB₂-R showed 6.3-fold increase in the liver of BDL animals, whereas heart expression levels were increased only by 1.7-fold (Figure 3A). To further characterize the changes in CB₂-R expression, cells isolated from cardiac and liver tissues were investigated. The ddPCR experiments confirmed a 3.0-fold increase of CB₂-R expression in Kupffer cells of BDL mice, while 8.9-fold increase of CB₂-R expression was observed in LSECs isolated from BDL mice in comparison to Sham (Figure 3B). Hepatocyte and cardiomyocyte CB₂-R copy number profiles of Sham and BDL did not differ and were extremely low and comparable to the levels in brain (negative control to CB₂-R expression) (Figure 3B).

Figure 3. CB₂-R expression profile.

(A) CB₂-R expression in heart and liver tissues with quantitative real-time PCR. (n=7–8/group) (B) CB₂-R copy number in different tissues and cells determined by droplet digital PCR. LSEC: Liver sinusoidal endothelial cell, BDL: bile duct ligation. Positive control for CB₂-R expression: spleen and splenocytes; Negative control for CB₂-R expression: brain. (n=3/group) *p<0.05 vs Sham

The mRNA expression of cardiac CB₁-R, MAGL, DAGLα and β, FAAH, TRPV1, GPR55 were not different between Sham and BDL (Suppl. Figure 1A). Liver mRNA expression of CB₁-R, DAGLα were higher, while MAGL was significantly lower in BDL animals (Suppl. Figure 1B). Liver FAAH, DAGLβ, TRPV1, GPR55 expressions were not different among the experimental groups (Suppl. Figure 1B). HU910 treatment did not change the mRNA expression level of the members of cannabinoid signalling in Sham or BDL (Suppl. Figure 1A and B).

CB₂-R agonist treatment improves microvascular function, inflammation, oxidative stress and fibrosis in BDL liver

The application of the CB₂-R agonist HU910 effectively reduced liver fibrosis (Figure 4A), leukocyte infiltration (number of CD45+ cells; Figure 4B) and the associated oxidative stress (indicated by lower levels of MDA and 3-NT; Figure 4C) in BDL. Importantly, HU910 treatment restored the hepatic microcirculation in BDL animals (Figure 4D). The above changes were supported by significantly lower levels of ALT and ALP (Figure 4E), and by markedly reduced expression levels of multiple pro-inflammatory cytokines/chemokines (IL1β, IL6, MIP1α, TNFα), markers of vascular inflammation (E-selectin, VCAM-1, ICAM-1) (Figure 4F), and significantly lower serum TNF-alpha levels (Figure 4F) in HU910-treated BDL mice.

Figure 4. CB₂-R agonist treatment improves liver pathology.

Representative images and analysis of (A) Sirius red staining (Magnification: 100x), (B) CD45, (C) malondialdehyde (MDA) and 3-nitrotyrosine (3-NT) immunohistochemistries. (n=6–11/group) (Magnification: 200x) (D) Representative images and analysis of liver microcirculation. (n=6–9/group). (E) Serum alanine aminotransferase (ALT) and alkaline phosphatase (ALP) (n=8–10/group). (F) Gene expression results of interleukin 1β (IL1β), IL6, macrophage inflammatory protein 1α (MIP1α), tumor necrosis factor-α (TNF-α), E-selectin, vascular and intercellular adhesion molecule 1 (VCAM-1, ICAM-1). Serum level of TNF-alpha. Groups: Sham, Sham+HU910, bile duct ligation (BDL), BDL+HU910. *p<0.05 vs Sham, #p<0.05 vs BDL

CB₂-R agonist treatment improves cardiovascular function in BDL

HU910 treatment attenuated the baseline diastolic dysfunction in BDL mice (significantly increased E/A ratio on echocardiography; improved diastolic functional parameters such as dP/dt_min, LVEDP, Tau_Glantz and Tau_Weiss) (Figure 5A, B), while the blood pressure values, macrovascular indicators (arterial elastance, total peripheral resistance), cardiac output and dP/dt_max were not affected by the drug treatment (Echocardiography: cardiac output: Sham+HU910: 12.9±1.0 mL/min p=0.87 vs Sham, BDL+HU910: 10.5±0.7 mL/min p<0.05 vs Sham, p=0.42 vs BDL; Figure 5C, D). Importantly, the CB₂-R agonist treatment significantly enhanced intrinsic cardiac contractility as shown by the load-and heart rate-independent parameters of myocardial contractile function such as the increased slope of ESPVR, PRSW, dP/dt_max-EDV and Emax (Figure 5E, F). Ejection fraction (Echocardiography: Sham+HU910: 53±2% p=0.95 vs Sham, BDL+HU910: 76±2% p<0.05 vs Sham, p=0.65 vs BDL; PV analysis: Sham+HU910: 67±1% p=0.66 vs Sham, BDL+HU910: 75±3% p<0.05 vs Sham, p=0.99 vs BDL) and fractional shortening (Sham+HU910: 27±2% p=0.93 vs Sham, BDL+HU910: 45±2% p<0.05 vs Sham, p=1.00 vs BDL) were not affected by the drug treatment in the corresponding study groups (Sham vs Sham+HU910; BDL vs BDL+HU910).

Figure 5. Improved cardiovascular function after HU910 treatment.

(A) Representative motion (M)-mode and Mitral pulse wave Doppler flow images. Red arrow: left ventricular (LV) internal diameter in systole, Yellow arrow: LV internal diameter in diastole. E: early mitral inflow wave, A: late mitral inflow wave. Baseline hemodynamic parameters: (B) diastolic functional indices: E/A ratio on echocardiography, maximal slope of pressure decrement (dP/dt_min), LV end-diastolic pressure (LVEDP), time constant of LV pressure decay (Tau_Glantz and Tau_Weiss), (C) blood pressure values, vascular functional markers. (D) Conventional systolic parameters (cardiac output, maximal slope of pressure increment (dP/dt_max)). (E) Contractility indices: slope of the end-systolic pressure-volume relationship (ESPVR), preload recruitable stroke work (PRSW), dP/dt_max-end-diastolic volume (EDV) relationships and Emax. Dashed line rectangle indicates the marker (ESPVR) represented on PV loops. (n=6–7/group) (F) Representative PV loops. Red line indicates the ESPVR. Arrows indicate the decrease of the slope of ESPVR in BDL and the increase in BDL+HU910. Groups: Sham, Sham+HU910, bile duct ligation (BDL), BDL+HU910. *p<0.05 vs Sham, #p<0.05 vs BDL

Myocardial structural alterations, morphometric data and gene expression profile

BDL was associated with smaller cardiomyocyte size (revealed on hematoxylin-eosin stained sections, Figure 6A and B) and lower body weight, heart weight and heart weight/tibia length ratio (Figure 6C). The drug treatment had no effect on these parameters. Furthermore, elevated serum BNP levels in BDL indicated the development of cardiomyopathy/heart failure (Figure 6D). The presence of cardiomyopathy/heart failure was corroborated by the changes of several cellular markers such as increased levels of ANP, AGTR1a, myostatin and β/α-MHC ratio and decreased expression of SERCA2 (Figure 6E). All these parameters were positively affected by the drug treatment (Figure 6E). BDL animals did not develop myocardial fibrosis (Suppl. Figure 1C and D).

Figure 6. Myocardial alterations in BDL.

(A) Representative images of hematoxylin-eosin histology (Magnification: 400x). (B) Cardiomyocyte diameter results. (C) Body weight (BW), heart weight (HW), HW/BW ratio, HW/tibia length (TL) ratio are shown. (n=10–12/group) (D) Serum brain natriuretic peptide (BNP) levels. (E) Gene expression levels of myostatin, α-, β-myosin heavy chain (MHC) and β/α-MHC ratio, angiotensin II receptor 1a (AGTR1a), sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA2), atrial natriuretic peptide (ANP). (n=5–7/group) Groups: Sham, Sham+HU910, bile duct ligation (BDL), BDL+HU910. *p<0.05 vs Sham, #p<0.05 vs BDL

CB₂-R agonist treatment improves cellular and vascular inflammation in myocardium of BDL mice

Immunohistochemical detection of the leukocyte marker CD45 showed increased leukocyte infiltration of the heart in BDL (Figure 7A). The gene expression levels of different inflammatory markers including MIP1α, IL1α, and TLR4 were significantly elevated in BDL myocardium (Figure 7B). Similarly, markers of vascular inflammation such as P-selectin, VCAM-1 and ICAM-1 were significantly higher in BDL heart (Figure 7B). HU910 treatment significantly reduced the number of CD45+ leukocytes and the expression of these markers (Figure 7A and B).

Figure 7. Myocardial inflammation.

(A) Representative images and quantification of cardiac CD45 immunohistochemistry (Magnification: 400x). (n=6–11/group). Arrows indicate CD45+ cells. (B) Gene expression of macrophage inflammatory protein 1α (MIP1α), interleukin 1α (IL1α), P-selectin, vascular and intercellular adhesion molecule 1 (VCAM-1, ICAM-1), toll like receptor 4 (TLR4) are shown. (n=5–7/group) Groups: Sham, Sham+HU910, bile duct ligation (BDL), BDL+HU910. *p<0.05 vs Sham, #p<0.05 vs BDL

CB₂-R agonist treatment improves myocardial oxidative stress in BDL

The myocardium of BDL mice showed significantly higher levels of oxidative stress when compared to Sham animals (indicated by increased 3-NT and MDA levels, Figure 8A and B). Correspondingly, markedly increased gene expression levels of different nicotinamide adenine dinucleotide phosphate (NADPH)-oxidase subunits (gp22phox, gp91phox, p47phox, p67phox) and the upregulation of COX1 and 2 have been observed (Figure 8C). On the other hand, HU910 treatment effectively attenuated oxidative stress in BDL mice both at the histological (3-NT, MDA immunohistochemistry; Figure 8A and B) and gene expression (Figure 8C) levels.

Figure 8. Myocardial oxidative stress.

Representative images and quantification of (A) 3-nitrotyrosine (3-NT, Magnification: 200x) and (B) malondialdehyde (MDA, Magnification: 200x) immunohistochemistries. (C) Gene expression levels of gp22phox, gp91phox, p47phox, p67phox, cyclooxygenase 1, 2 (COX1, 2) are depicted. (n=5–7/group) Groups: Sham, Sham+HU910, bile duct ligation (BDL), BDL+HU910. *p<0.05 vs Sham, # p<0.05 vs BDL. (D) Schematic illustration of the proposed mechanism. Liver fibrosis in bile duct ligation is associated with tissue and vascular inflammation, endothelial dysfunction and microcirculatory collapse. Under those circumstances, the overproduction of inflammatory/paracrine mediators by the liver/body (e.g. cytokines, chemokines) might provoke cardiac inflammation and oxidative stress and leads to the development of hepatic cardiomyopathy. CB₂-R activation promotes anti-inflammatory, anti-oxidative and tissue protective effects in the liver and heart. ROS/RNS: reactive oxygen/nitrogen species. Figure was modified from Servier Medical Art, licensed under a Creative Common Attribution 3.0 Generic License. http://smart.servier.com/. Arrows indicate CD45+ cells.

Discussion

In the present study, we characterized in detail the changes of cardiac function and myocardial tissue injury in an animal model of advanced liver fibrosis induced by BDL. We show that 1) BDL is associated with deteriorated myocardial systolic and diastolic performance and with micro-and macrovascular dysfunction and shift in myocardial gene expression characteristic of cardiomyopathy, 2) tissue/vascular inflammation and oxidative stress play pivotal role in the pathophysiology of hepatic cardiomyopathy and 3) the CB₂-R agonist treatment attenuates cardiac dysfunction in liver fibrosis, at least in part, by lowering the extent of hepatic inflammation and fibrosis.

The progression of chronic liver disease incorporates the destruction of liver parenchyma in response to different pathological stimuli (e.g. chronic alcoholic or non-alcoholic liver disease, infectious agents and autoimmune processes), eventually leading to the fibrotic remodelling of the liver and the development of cirrhosis(25). Liver transplantation is the gold standard and the only life-saving treatment option for patients with advanced liver failure. However, the development of hepatic/cirrhotic cardiomyopathy in patients with end-stage liver failure is a major determinant of post-transplantation survival. It has been reported that cardiovascular disease-related death (including heart failure) is the leading cause of mortality (different statistics report ~7–40%) following liver transplantation(26, 27). According to different studies, cirrhotic cardiomyopathy might develop in up to 50% of patients with cirrhosis(28). Hepatic cardiomyopathy is characterized by latent cardiac contractile and diastolic dysfunction coupled with hyperdynamic circulation and by blunted cardiac responsiveness to different stress stimuli and electrophysiological abnormalities(27). Based on the above, there is a critical need for a reliable animal model of hepatic cardiomyopathy for basic preclinical and translational research.

Although common bile duct ligation is a widely accepted animal model for advanced liver fibrosis in mice and rats(19, 22, 25, 29), very little is known on the exact pathophysiological changes/injury in the myocardium and hemodynamic alterations in vivo. Therefore, our study aimed to characterize cardiovascular function and pathology in detail in the mouse model of BDL.

According to several reports in humans, cirrhotic cardiomyopathy is characterized by diastolic dysfunction and preserved systolic function at rest(30) indicated by normal or increased ejection fraction (~56–80% depending on the patient population studied(31–33)). Furthermore, classical imaging modalities including echocardiography might provide limited information about the intrinsic myocardial contractility in vivo and might over-or underestimate the ejection fraction especially in cases where significant changes of the loading conditions happen (such as low blood pressure)(10). Therefore, we performed invasive hemodynamic investigation and PV analysis to describe the detailed cardiovascular function in BDL.

Two weeks of BDL was characterized by hepatic fibrosis, massive inflammation and microcirculatory collapse of the liver. The latter one indicates the presence of endothelial dysfunction in BDL. BDL was also associated with significantly lower blood pressure and worsening of macrovascular function. Although ejection fraction was slightly increased in our study, it is considered within the physiological range in mice. A possible explanation of this phenomenon might be the observed significant changes in the loading conditions (dramatically decreased blood pressure), which is known to influence ejection fraction(9, 23). In accordance with previous reports about a hyperdynamic state in liver failure(5, 34), we observed the preservation of the classical functional parameter dP/dt_max. Importantly, echocardiography and invasive hemodynamic data showed marked deterioration of diastolic function, which is in accordance with human data(30). On the other hand, PV analysis revealed severe contractile dysfunction (decrease of pre-and afterload independent sensitive contractility indices such as ESPVR, PRSW, Emax and dP/dt_max-EDV) during vena cava inferior occlusions indicating a latent but severe myocardial dysfunction.

Although several drugs are available for the treatment of chronic heart failure, no specific therapy has been developed for hepatic cardiomyopathy yet. Consequently, there is an increased need for novel drug therapies to treat liver failure-associated cardiac and vascular complications. Experimental studies on hepatic cardiomyopathy implicated the importance of oxidative/nitrative stress(35), mitochondrial dysfunction(36), apoptosis(1), angiotensin signalling(37), inflammation(38) and endocannabinoid-CB₁-Rsignalling(5, 11, 12, 38, 39) in the development of the disease.

It is well documented that endocannabinoids or synthetic ligands exert opposing effects via CB₁ or CB₂ receptors on multiple pathological processes in the liver and heart(13, 40–42). CB₁-R promotes hepatic steatosis, inflammation and fibrogenesis. Endocannabinoid-CB₁-R signalling also contributes to peripheral vasodilation and cardiac dysfunction associated with liver cirrhosis(5, 12, 38). In contrast, CB₂-R activation has been shown to reduce the extent of liver injury, inflammation and/or fibrosis(14–16, 18, 24, 43) and also decreased cardiovascular inflammation in models of atherosclerosis or myocardial infarction by modulating interactions of activated endothelium and infiltrating immune cells(13, 42). CB₂-R is primarily expressed in inflammatory cells with the highest expression in B lymphocytes, macrophages and macrophage-derived cells (e.g. Kupffer cells), and to a lesser extent in other immune and activated endothelial cells(40). In macrophages and Kupffer cells CB₂-R activation decreases activation and modulates polarization(40). In human/mouse liver sinusoidal endothelium(15), coronary artery(44) and brain(40) endothelial cells CB₂-R attenuates TNF-α or endotoxin induced activation (e.g. expression of various adhesion molecules) and attachment of various immune cells (e.g. macrophages or neutrophils) to the activated endothelium, rolling and transmigration(15, 40, 44) of these immune cells through the activated endothelium both in vitro and in vivo(40).

Consistently with the literature we have observed increased CB₂-R expression in the livers of BDL mice (6.3-fold)(18) which was originating primarily from activated Kupffer and LSECs cells but not hepatocytes(24, 43). CB₂-R expression in hearts of BDL mice was only increased by 1.7-fold in cardiac tissue, which was originating most likely from infiltrating macrophages and or activated endothelial cells, but not cardiomyocytes.

Based on the above, we further characterized our model by providing a detailed description of distinct molecular events in the development of hepatic cardiomyopathy and explored the role of CB₂-R in the liver-heart inflammatory axis by using HU910, a highly selective CB₂-R agonist (14, 21). The BDL-induced characteristic changes of the liver injury, including fibrotic remodelling, inflammatory cell invasion, massive oxidative/nitrative stress and microcirculatory dysfunction as a sign of endothelial damage and vascular inflammation (elevated levels of different endothelial/inflammatory markers) were significantly attenuated by the treatment with HU910, most likely by modulating Kupffer cells and LSECs (as they expressed the highest levels of CB₂-R in the liver).

The progression of hepatic cardiomyopathy in our BDL model was accompanied not only by functional consequences but also by histological/molecular signs of cardiomyopathy. These included lower cardiomyocyte diameter alongside with smaller heart size and the changes of specific markers such as increased serum BNP, myocardial ANP expression (as a sign of heart failure), pathological switch of MHC-isoforms, increased AGTR1a, and increased myostatin levels. Serum BNP levels were shown to be related to cardiovascular dysfunction and disease severity in cirrhosis(45), while myostatin was proposed to promote cardiomyocyte proteolysis(46) and to have a predictor value for the severity and outcome of heart failure(47). Moreover, SERCA2, a regulator of intracellular Ca²⁺-homeostasis, was downregulated, which might have contributed to the diastolic dysfunction in BDL(48).

One of the main contributors to the BDL-induced liver fibrosis is tissue inflammation, which contributes, as liver failure develops, to the production and excretion of several inflammatory cytokines into the blood culminating in a general inflammatory response by the body. As part of this inflammatory response, we observed tissue and vascular/endothelial inflammation not only in the liver but also in the myocardium of BDL mice represented by increased inflammatory gene expression, increased serum TNF-alpha levels and the accumulation of CD45 positive leukocytes (both in the liver and heart). The role of TNF-alpha in BDL has been nicely described by Yang et al., where the authors showed the direct cardiodepressant effect of TNF-alpha on the cardiomyocyte contractility in vitro(38). Cardiodepressant and pro-oxidant effects of inflammatory cytokines including TNF-alpha have been described in ex vivo isolated working heart experiments(49). In accordance with these results, we observed increased TNF-alpha serum levels in BDL mice which could have contributed to the observed myocardial and vascular dysfunction in vivo. Moreover, inflammatory cell invasion has been implicated in the pathology of hepatic cardiomyopathy by contributing to the dysregulation of endocannabinoid-CB₁-R signalling(12, 38). In agreement with increased inflammation in the heart and liver, elevated levels of different NADPH-oxidase subunits, COX enzymes have been observed in the myocardium of BDL mice. Because of the above processes, the increased level of oxidative and nitrative stress (shown by 3-NT formation and lipid peroxidation) might have contributed significantly to the severe cardiac systolic and diastolic dysfunction through non-specific protein oxidation and/or nitration(50).

We found that the selective CB₂-R agonist HU910 treatment markedly attenuated liver inflammation, oxidative stress, microcirculatory collapse and fibrosis. This beneficial effect of CB₂-R activation in the liver was also associated with mitigation of cardiac dysfunction in BDL animals by improving intrinsic contractile function (increased slope of ESPVR, PRSW, dP/dt_max-EDV) and by attenuating diastolic dysfunction (increased E/A ratio, increased dP/dt_min, lower Tau_Glantz, Tau_Weiss and LVEDP). Correspondingly, the myocardial pathological changes including tissue/vascular inflammation and oxidative stress, the serum BNP and cardiac ANP, myostatin levels and MHC-switch were attenuated by the application of the CB₂-R agonist HU910. The current treatment paradigm might have had beneficial effects, at least in part, by reducing liver and consequently systemic inflammation leading to activation of endothelial cells and inflammatory cell infiltration in the myocardium. On the other hand, it could have directly modulated the cardiac inflammatory response and indirectly its deleterious consequences.

In conclusion, we demonstrate that bile duct ligation-induced advanced liver fibrosis is a suitable mouse model to study the pathophysiology of hepatic/cirrhotic cardiomyopathy at a preclinical level as it resembles the characteristics of the clinical syndrome seen in patients. Our results indicate that hepatic cardiomyopathy is characterized by myocardial systolic and diastolic dysfunction coupled with the significant changes of macro-and microvascular function. Our results also show that the liver-heart inflammatory axis has a pivotal pathophysiological role in the development of hepatic cardiomyopathy

List of abbreviations

3-NT

3-nitrotyrosine

ALT

alanine aminotransferase

ALP

alkaline phosphatase

α-MHC

alpha-myosin heavy chain

ANOVA

analysis of variance

AGTR1a

angiotensin II receptor 1a

AST

aspartate aminotransferase

ANP

atrial natriuretic peptide

BNP

brain natriuretic peptide

β-MHC

beta-myosin heavy chain

CB₁-R

cannabinoid-1 receptor

CB₂-R

cannabinoid-2 receptor

CD

cluster of differentiation

BDL

common bile duct ligation

COX1

cyclooxygenase 1

COX2

cyclooxygenase 2

DAGLα

diacylglycerol lipase α

DAGLβ

diacylglycerol lipase β

ddPCR

droplet digital polymerase chain reaction

EDV

end-diastolic volume

ELISA

enzyme linked immunosorbent assay

FAAH

fatty acid amide hydrolase

GPR55

G protein-coupled receptor 55

HPRT

hypoxanthine-guanine phosphoribosyltransferase

ICAM-1

intercellular adhesion molecule 1

IL1α

interleukin 1α

IL1β

interleukin 1β

IL6

interleukin 6

LV

left ventricle

LSEC

liver sinusoidal endothelial cell

MIP1α

macrophage inflammatory protein 1α

MDA

malondialdehyde

MAGL

monoacylglycerol lipase

PRSW

preload recruitable stroke work

PV

pressure-volume

SERCA2

sarco/endoplasmic reticulum Ca²⁺-ATPase

dP/dt_min

slope of systolic pressure decrement

dP/dt_max

slope of systolic pressure increment

ESPVR

slope of the left ventricular end-systolic pressure-volume relationship

SEM

standard error of mean

Tau_Glantz

time constant of left ventricular pressure decay

TLR4

toll like receptor 4

TRPV1

transient receptor potential cation channel subfamily V member 1

TNF-alpha/TNFα

tumor necrosis factor alpha

VCAM-1

vascular adhesion molecule 1