Cannabinoid Receptor Type I Modulates Alcohol-Induced Liver Fibrosis
- 6 Downloads
The cannabinoid system (CS) is implicated in the regulation of hepatic fibrosis, steatosis and inflammation, with cannabinoid receptors 1 and 2 (CB1 and CB2) being involved in regulation of pro- and antifibrogenic effects. Daily cannabis smoking is an independent risk factor for the progression of fibrosis in chronic hepatitis C and a mediator of experimental alcoholic steatosis. However, the role and function of CS in alcoholic liver fibrosis (ALF) is unknown so far. Thus, human liver samples from patients with alcoholic liver disease (ALD) were collected for analysis of CB1 expression. In vitro, hepatic stellate cells (HSC) underwent treatment with acetaldehyde, H2O2, endo- and exocannabinoids (2-arachidonoylglycerol (2-AG) and Δ9-tetrahydrocannabinol [THC]), and CB1 antagonist SR141716 (rimonabant). In vivo, CB1 knockout (KO) mice received thioacetamide (TAA)/ethanol (EtOH) to induce fibrosis. As a result, in human ALD, CB1 expression was restricted to areas with advanced fibrosis only. In vitro, acetaldehyde, H2O2, as well as 2-AG and THC, alone or in combination with acetaldehyde, induced CB1 mRNA expression, whereas CB1 blockage with SR141716 dose-dependently inhibited HSC proliferation and downregulated mRNA expression of fibrosis-mediated genes PCα1(I), TIMP-1 and MMP-13. This was paralleled by marked cytotoxicity of SR141716 at high doses (5–10 µmol/L). In vivo, CB1 knockout mice showed marked resistance to alcoholic liver fibrosis. In conclusion, CB1 expression is upregulated in human ALF, which is at least in part triggered by acetaldehyde (AA) and oxidative stress. Inhibition of CB1 by SR141716, or via genetic knock-out protects against alcoholic-induced fibrosis in vitro and in vivo.
Liver fibrosis is a common complication of many chronic liver diseases due to alcohol abuse, chronic hepatitis B and C, nonalcoholic steatohepatitis, prolonged drug exposure and autoimmune causes (1). Upon chronic injury, liver parenchyma is replaced by excess extracellular matrix (ECM) produced by activated hepatic stellate cells (HSC) and myofibroblasts, resulting in cirrhotic remodeling, progressive functional impairment and—potentially—hepatocellular cancer (2, 3, 4).
The endocannabinoid system (ECS) comprises endocannabinoids such as arachidonoyl ethanolamide (anandamide, AEA) and 2-arachidonoylglycerol (2-AG), their corresponding receptors (cannabinoid receptors 1 and 2 [CB1 and CB2]) and certain phytocannabinoids such as Δ9-tetrahydrocannabinol (THC). Recently, an important role of ECS in liver fibrogenesis was described (5). Upregulation of the ECS in several types of chronic liver damage, but not in healthy liver was demonstrated (6,7). Here, CB1 expression was confined to activated myofibroblasts and vascular endothelium, whereas CB2 was expressed by inflammatory cells (8). CB1 and CB2 seem to mediate opposite effects in fibrogenesis, as CB1 knockout mice (CB1−/−) are resistant to fibrogenesis induced by CCl4, thioacetamide (TAA) or bile duct ligation (BDL) (9), whereas CB2−/− mice reveal more progressive fibrosis after CCl4 treatment (10). In addition, pharmacological inhibition of CB1 with rimonabant, a CB1 antagonist, reduced expression of transforming growth factor β1 (TGFβ1) and α-smooth muscle actin (αSMA), and improved histology of BDL-, TAA- or CCl4-treated mice (8,9). Clinically, patients with chronic hepatitis C and daily cannabis consumption displayed more severe fibrosis progression than non- or occasional consumers (11, 12, 13).
Endocannabinoid expression is faint in patients with healthy livers, whereas in cirrhosis, fatty liver, acute hepatitis or bile duct obstruction, hepatic and serum levels of AEA and 2-AG were found increased (14, 15, 16, 17). Endocannabinoids exert numerous and sometimes opposite effects on liver target cells which could be CB receptor-dependent or independent, and are involved in control of cell survival and apoptosis (18). Thus, AEA induced cell necrosis via several mechanisms, whereas 2-AG induced apoptosis in HSC (10,19,20).
Jeong et al. recently found CB1−/− mice to be resistant to the steatogenic effects of alcohol (21), while the opposite effect was observed in CB2−/− mice (22). The latter study also investigated the effects of 35 wks of alcohol at 16% in drinking water on fibrosis in mice deficient for either CB1−/− or CB2−/−. Authors demonstrated HSC activation and increased collagen production in all groups after prolonged ethanol intake, with the strongest effect observed in CB2−/− mice, and to a lesser extent in CB1−/− mice. In addition to antifibrotic and antisteatotic effects toward alcohol, CB2 also alleviated hepatic inflammation as evidenced by enhanced inflammatory scores on histology and increased mRNA expression of tumor necrosis factor-α, interleukin-1β and monocyte chemotactic protein-1. The protective effect of patent CB2 receptor status and its activation toward alcohol-mediated liver injury was further underscored by showing that mice fed an alcoholic liquid diet containing 6.2% of alcohol (28% of total calories), and treated with a CB2 agonist shifted their Kupffer cell response toward a M2-type of pattern favoring resolution of inflammation and wound healing (23). However, the extent of and effects on fibrosis by CB1 antagonism in the study by Trebicka et al. were small, albeit significant, which could be due to the animal model in which mice received alcohol via drinking water. This approach is considered suboptimal with regard to studying aspects of fibrosis since mice are notoriously resistant to profibrogenic effect from alcohol without additional hepatotoxic triggers (24). None of the studies focusing on the role of ECS in the evolution of alcohol-mediated fibrosis have included human samples, or attempted to decipher mechanisms underlying this interaction. Therefore, we aimed to explore the endocannabinoid-dependent mechanisms of alcohol-induced liver injury in human ALD, in experimental alcohol-induced liver fibrosis using CBT−/− mice, and in HSC to increase our understanding on how alcohol interacts with CB1.
Material and Methods
Paraffin-embedded and frozen human liver tissues were obtained from consecutive patients with ALD who underwent liver biopsy for diagnostic reasons. Patients consumed at least 80 g/day of alcohol, were predominantly male (74%), had a mean age of 49 ± 7.4 years and a mean BMI of 25 ± 3.5 kg/m2. Histologically normal liver tissue specimens from patients undergoing abdominal surgery for reasons unrelated to liver disease served as controls. The extent of alcoholic liver injury was scored according to Kleiner et al. (25). Liver biopsy specimens used for gene expression revealed the following fibrosis stages: F0, n = 6; F1, n = 6; F2, n = 7; F3, n = 5; F4, n = 3. The human study was approved by the Ethics Committee of the University of Heidelberg, Germany where all liver biopsies were collected. All patients gave written informed consent to have their tissues included in the study.
CB1−/− mice and wild-type littermates (C57/BL6 background) were a gift from Andreas Zimmer, University of Bonn, Germany. All animals were housed under standard conditions. Approval for animal experimentation was granted before by the Regional Government of Lower Franconia, Würzburg, Germany (No. 54-2531.31-18/06).
Fibrosis was induced by a modified protocol of 3×/wk intraperitoneal (i.p.) application of 0.15 mg/kg TAA and continuous feeding of 10% (v/v) ethanol (EtOH) in sweetened drinking water for 18 wks (26). Untreated control groups received regular food and water ad libitum. Each of the four experimental groups consisted of eight animals.
Animals were killed by cervical dislocation and liver samples were snap frozen in liquid nitrogen for isolation of mRNA and protein, or fixed in 4% phosphate-buffered formalin for paraffin embedding. Regular hematoxylin and eosin (H&E) staining was performed to evaluate standard histomorphology of liver tissues and to assess the grade of inflammation. Sirius red staining was used to visualize fibrosis.
Quantitative Real-Time PCR
Total RNA was isolated from snap frozen tissues or from semiconfluent cells using the RNeasy kit (Qiagen, Basel, Switzerland) and converted to cDNA as described (27). Quantitative real-time PCR was performed on a Bio-Rad CFX96 system (Bio-Rad Laboratories, Munich, Germany) and on the ABI 7700 Sequence Detector (Applied Biosystems, Rotkreuz, Switzerland) using QuantiTect Primers (Qiagen, Hilden, Germany) and the LightCycler FastStart DNA Master SYBR Green I kit (Roche Molecular Diagnostics, Basel, Switzerland). For normalization, the housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was amplified in a parallel reaction. Sequences of primers and probes are summarized in Supplementary Table 1.
Hepatic hydroxyproline (HYP) was measured as described previously (27, 28). Briefly, liver tissues (200 mg) were hydrolyzed in 6 mol/L HCl at 110°C for 16 h. Fifty microliters of each sample was incubated with chloramine T (2.5 mmol/L) for 5 min and Ehrlich’s reagent (410 mmol/L) for 30 min at 60°C. Absorption was measured in triplicates at 560 nm. Results are expressed as µg/g of wet liver tissue.
Cell Culture Experiments
Human HSC, isolated from resected human livers (kindly donated by M Pinzani, Florence, Italy) (29,30) and moderately activated rat HSC cell line CFSC-2G, isolated from cirrhotic rat livers (29,31,32) (generously provided by M Rojkind, Washington DC, USA) were cultured in DMEM containing 10% FBS, 2 mmol/L glutamine, 100 U/mL penicillin and 100 U/mL streptomycin. Cells were maintained at 37°C in a 5% CO2 humidified atmosphere.
For all experiments, cells were seeded onto 6- or 12-well plates and, after reaching semiconfluent conditions, starved in serum-free DMEM for 24 h. Treatment with AA (Fluka, Buchs, Switzerland; 200 µmol/L), H2O2 (1–50 µmol/L), 2-AG (Sigma-Aldrich, Munich, Germany; 1–20 µmol/L), THC (200–1000 nmol/L) or SR141716/rimonabant (Sanofi-Aventis SA, Paris, France; 100 nm to 10 µmol/L) was performed for 6 to 48 h under normoxic (24% O2) or hypoxic (1% O2) conditions using a hypoxia workstation (Ruskinn Technology Limited, West Yorkshire, UK).
Cell Viability Analysis
Cells were seeded at 6 × 104 onto 6-well culture plates, and, after reaching semiconfluent state, starved in serumfree DMEM, and treated with different compounds for the following 24 to 48 h. Trypan blue-excluding cells were counted under a light microscope.
For LDH analysis, the supernatants from both the untreated and the treated cells were collected and LDH leakage was assayed with a standard kit (Diatools AG, Villmergen, Switzerland) and an autoanalyzer (Olympus Autoanalyzer AU 2700, Kobe, Japan).
BrdU Incorporation or DNA Synthesis
Cells were seeded at a density of 5 × 104 per well in 96-well plates and after 24 h of starvation, tested compounds were applied for the next 24 h. BrdU was added during the last 4 h and the amount of BrdU incorporation was determined by a colorimetric BrdU cell proliferation enzyme-linked immunoassay (ELISA) according to the manufacturer’s instructions (Roche, Mannheim, Germany). All experiments were done with 6 wells per group and repeated at least three times.
Western Blot Analysis
Protein extracts from liver tissue samples were separated by SDS-PAGE, proteins transferred and membranes probed with mouse monoclonal anti-caspase-3 (cleaved and total), anti-bax and anti-actin antibodies (Cell Signaling Technology Inc., Danvers, MA, USA), followed by incubation with horseradish peroxidase-conjugated goat anti-mouse antibody (Pierce, Rockford, IL, USA). Signals were detected with enhanced chemiluminescence (Western Lightning Chemiluminescence Reagent Plus, Perkin-Elmer Life Science, Boston, MA, USA), imaged and quantified using AIDA software (Raytest, Urdorf, Switzerland).
Specimens were deparaffinized and antigen retrieval performed by boiling in sodium citrate buffer (pH 6.0) for 15 min and in 0.1% Trypsin for 20 min at 37°C. After blocking of endogenous peroxidase by 0.6% H2O2 and unspecific binding by 2% goat serum, primary rabbit polyclonal anti-CB1 (Abcam, Cambridge, UK) or anti-CB2 antibody (Novus Biologicals, Littleton, CO, USA) were applied (1:300) and incubated overnight at 4°C, followed by biotinylated goat anti-mouse secondary antibody (1:350) for 1 h at room temperature. Detection with diaminobenzidine (DAKO, Glostrup, Denmark) was performed after incubation with HRP-streptavidin solution (Vector Labs, Burlingame, VT, USA). Nuclei were stained with Mayer’s Hematoxylin. Histopathological evaluations were performed in a blinded fashion by two investigators (E Patsenker and F Stickel).
Statistical analyses were performed using Microsoft EXCEL software. Data are expressed as mean ± SD. The statistical significance of differences was evaluated using the unpaired Student t test or nonparametric analysis of variance (ANOVA) Kruskal-Wallis test.
All supplementary materials are available online at www.molmed.org .
CB1 and CB2 Expression in Human ALD
CB1 Is Upregulated by Acetaldehyde in HSC
In addition, the marker of HSC activation, αSMA, was strongly induced by 2-AG alone, and in combination with AA by 40% and 100%, respectively (Figure 2F), whereas other fibrosis-associated genes such as PCα1(I), TGFβ1 and TIMP-1 were not affected by THC or 2-AG during the observation interval (data not shown). Notably, there were no toxic or apoptotic effects by AA alone or in combination with 2-AG and THC, as shown by LDH leakage (Supplementary Figure 1A), Trypan Blue stain (not shown) or proapoptotic caspase-3 activation, in spite of bax activation in presence of 1 µmol/L THC (Supplementary Figure 1C). Slight, but not significant inhibition of rat HSC proliferation by AA and THC also was observed (Supplementary Figure 1B).
The effect of ethanol (EtOH) itself was tested in the same experimental setting. However, there was no significant effect found on CB1 mRNA, PCα1(I) or αSMA mRNA in rat HSC using 10–200 mmol/L EtOH (Supplementary Figure 2), whereas at 300–500 mmol/L it revealed toxicity (not shown).
SR141716/Rimonabant Exerts Toxicity on HSC
Effects of SR141716 on rat HSC at concentrations below 1 µmol/L showed no effect on any cellular functions tested (data not shown). However, DNA synthesis was inhibited approximately fourfold by both 5 and 10 µmol/L (Figure 3A), and mRNA expression for αSMA, PCα1(I), TIMP-1 and MMP-13 was downregulated by more than 50% by 5 and 10 µmol/L (Figure 3B). Analysis of cell death revealed significant toxicity of SR141716 at 10 µmol/L after 48 h of incubation (Figure 3C), which was confirmed by LDH leakage and Trypan Blue staining (Figures 3D, E). Notably, neither total caspase-3, nor bax were induced by SR141716, but rather decreased due to toxic effects (Figure 3F) ruling out significant proapoptotic effects from SR141716 on HSC. Similarly, Bcl-2 and cleaved caspase-3 were not induced in any of the groups (data not shown).
CB1−/− Mice Are Resistant to Alcohol-Induced Liver Fibrosis
Hepatic Inflammation Is Mildly Affected in CB1−/− Mice upon Fibrosis Induction
Expression of genes coding for surrogate markers of inflammation including COX-2, iNOS and TNFα were significantly upregulated by TAA/EtOH in WT animals. In CB1−/− mice, iNOS and TNFα expression also were moderately increased by TAA/EtOH treatment, whereas COX-2 expression remained unaffected (Figures 7D–F).
The significance of ECS in modulating fibrosis, steatosis and cell regeneration in experimental liver injury has been demonstrated and reviewed (7,8,33), and CB receptors are induced in acute and chronic liver injury including cirrhosis (14). However, its involvement in alcoholic liver fibrosis was unknown until the present study, which describes the engagement of ECS in the evolution of alcohol-induced liver fibrosis by showing the expression of CB1 in human ALD, its induction in HSC upon challenge with acetaldehyde and reactive oxygen species and a marked resistance of CB1−/− mice toward chronic liver injury modulated further by alcohol. Upregulation of CB1 in human cirrhosis was described by Lotersztajn et al. (9), however, in their study, only four patients with alcoholic cirrhosis were included. Also, no patients with precirrhotic alcoholic liver injury were studied. Interestingly, CB2 expression was not detectable in our study on protein level in humans, further, it remained unaffected by alcohol at mRNA level, and did not correlate with the severity of fibrosis. These findings were confirmed in vitro, demonstrating the same expression pattern for CB1 and CB2 from human HSC. An increased CB2 expression by HSC during activation in culture was shown by Julien et al. (10) but specific treatment of fully activated HSC with alcohol did not have the same effect. Therefore, we hypothesize that there exists an induction of “profibrogenic” CB1 paralleled by an unchanged expression of “antifibrogenic” CB2 which then swings the balance toward net matrix accumulation rather than its degradation.
Endocannabinoid levels, such as of 2-AG and AEA are increased in liver disease (14, 15, 16,20). Here, we show that 2-AG and THC induce CB1 mRNA, which is further enhanced by their combination with AA, jointly favoring the progression of fibrosis in ALD. Moreover, combination of AA and 2-AG resulted in an even stronger upregulation of αSMA in HSC than treatment with either 2-AG or AA alone, indicating its high activation potential. Interestingly, there was no indication of apoptosis following treatment with THC or 2AG at the doses applied, which conflicts with some of the published data from Julien et al. (10) who demonstrated that THC at 1–4 µmol/L and 2-AG at 20 µmol/L exert antiproliferative and proapoptotic effects. However, in the same study, no effect on cell viability by THC at 1 µmol/L in the presence of PDGF was shown. So, it can be concluded that cellular responses toward potentially proapoptotic stimuli are variable and depend on the conditions used. In addition, cell-specific differences between primary and immortalized cells should be taken into account.
The CB1 antagonist SR141716 revealed potent antifibrotic properties at 5 µmol/L without causing significant cell death, before exerting cytotoxic properties on HSC when applied at doses from 5 to 10 µmol/L.
Our data clearly show that CB1−/− mice do not develop liver fibrosis after 18 wks of TAA/EtOH administration. All surrogate markers of fibrosis, such as extracellular matrix deposition, hydroxyproline and mRNA levels of fibrosis-mediating genes consistently demonstrate that CB1−/− mice are effectively protected from fibrosis development. One of the mechanisms for resolution of hepatic fibrosis could be the activation of matrix metalloproteinases, such as interstitial collagenases (human MMP-1/murine MMP-13), stromelysin-1 (MMP-3), and gelatinases A and B (MMP-2 and MMP-9, respectively) which degrade matrix molecules and may contribute to fibrolysis and fibrosis resolution (34, 35, 36). We found a strong upregulation of MMP-2, -3 and -13 in WT mice compared with CB1−/− animals. This seemingly contradictory observation with regards to the balance between fibrogenesis and fibrolysis can be explained by the very complex and different regulation of MMP expression and enzymatic activation during the peak of fibrosis formation and ECM degradation (34,37). Certain MMPs, such as profibrogenic MMP-2, facilitate and promote degradation of basement membrane with subsequent activation, migration and proliferation of HSC/myofibroblasts, resulting in progression of liver injury and fibrosis (38), whereas others, such as MMP-3, -9 and -13 participate in the degradation of extracellular matrix and fibrosis resolution (35,39). In our study, the induction of fibrolytic MMP-3 and -13 in parallel with MMP-2 may reflect a fibrolytic response in an attempt to degrade excessive matrix, which, however, is not sufficient to overcome the profibrogenic stimuli from the toxic treatment in WT mice. In CB1−/− mice induction of MMPs was not so prominent with less induced MMP-2 and elevated MMP-9 and -13, indicating a shift toward enhanced fibrolysis.
Several publications described the involvement of the ECS in hepatic inflammation. In spite of the primary role of the CB2 receptor, CB1 also has been shown to be involved in inflammatory processes (33). Thus, CB1 was found expressed in almost all human peripheral blood immune cells, including B lymphocytes, natural killer cells, polymorphonuclear neutrophils, CD8 and CD4 lymphocytes and monocytes (40). In experimental models of liver fibrosis, both CB2 receptor agonists and CB1 receptor antagonists exerted variable effects via different mechanisms which may involve modulation of the secondary inflammatory response (5,8). A recent study suggested that an antiinflammatory effect of CB1 antagonist SR141716 also may contribute to the favorable effects of CB1 receptor antagonists in the treatment of obesity, nonalcoholic fatty liver disease and closely related features of the metabolic syndrome (41). However, published data in this regard remain contradictory, and only one study addressed the mechanisms of how the ECS intercepts with inflammatory processes (23). In our study, we observed moderately reduced portal inflammation and a lack of COX-2 upregulation in CB1−/− mice following alcohol exposure, while TNFα and iNOS were not different between CB1−/− and WT mice, suggesting that the absence of CB1 has little impact on well-established triggers of hepatic inflammation.
Two aspects of our experimental set up make it difficult to draw direct conclusions for human ALD: (a) the experimental model we used is certainly not perfect in mimicking ALD in human, but a reliable experimental model for inducing liver fibrosis in mice. Kornek et al. have clearly shown, that the addition of alcohol TAA further increases liver injury compared with TAA alone, although features of ALD such as steatosis, Mallory bodies, and “chicken-wire fibrosis” were missing (26). However, rodents are known to be notoriously resistant to fibrosis induction, and most models require addition of a second fibrosis trigger to alcohol administration (26,42); (b) the acetaldehyde concentration of 200 µmol/L causing significant CB1 upregulation in our study is approximately two-fold higher than hepatic concentrations typically encountered in human ALD, but still in the order of what has been described in experimental ALD after dosing rats with different amounts of alcohol (43, 44, 45, 46).
In conclusion, induction of CB1 receptor by AA due to chronic alcohol exposure, as well as by either endo- or exocannabinoids, and, more strikingly, by their combination, are suggestive of a relevant involvement of the ECS in the evolution of ALD, and suggest additive effects of concomitant cannabis consumption in individuals who consume excess alcohol, and may point to CB1 antagonism for treatment of fibrotic ALD, particularly, as CB1 antagonists have been used in treating alcohol dependence without causing significant depression (47) as was observed in patients with nonalcoholic fatty liver disease (48).
The authors declare that they have no competing interests as defined by Molecular Medicine, or other interests that might be perceived to influence the results and discussion reported in this paper.
We are most grateful to Andreas Zimmer from the University of Bonn, Germany, for providing access to the CB1−/− mice. This work was supported by grant 3100 A0 -122114/1 from the Swiss National Science Foundation and by the European Research Advisory Board (ERAB; grant EA 09 20). MO was supported by a grant of the von-Behring-Röntgen Foundation, Marburg, Germany. HKS received funding from the Manfred Lautenschläger Foundation, Heidelberg, Germany.
- 30.Failli P, et al. (1995) The mitogenic effect of platelet-derived growth factor in human hepatic stellate cells requires calcium influx. Am. J. Physiol. 269: C1133–9.Google Scholar
- 32.Rojkind M, et al. (1995) Characterization and functional studies on rat liver fat-storing cell line and freshly isolated hepatocyte coculture system. Am. J. Pathol. 146:1508–20.Google Scholar
- 48.European Medicines Agency. (2008) The European Medicines Agency recommends suspension of the marketing authorisation of Acomplia [press release]. [cited 2011 Oct 28]. Available from: https://doi.org/www.ema.europa.eu/ema/index.jsp?curl=pages/news_and_events/news/2009/11/news_detail_000244.jsp&murl=menus/news_and_events/news_and_events.jsp&mid=WC0b01ac058004d5c1&jsenabled=true