|Home | About | Journals | Submit | Contact Us | Français|
Hepatic ischemia-reperfusion (I/R) injury continues to be a fatal complication that can follow liver surgery or transplantation. We have investigated the involvement of the endocannabinoid system in hepatic I/R injury using an in vivo mouse model. Here we report that I/R triggers several-fold increases in the hepatic levels of the endocannabinoids anandamide and 2-arachidonoylglycerol, which originate from hepatocytes, Kupffer, and endothelial cells. The I/R-induced increased tissue endocannabinoid levels positively correlate with the degree of hepatic damage and serum TNF-α, MIP-1α, and MIP-2 levels. Furthermore, a brief exposure of hepatocytes to various oxidants (H2O2 and peroxynitrite) or inflammatory stimuli (endotoxin and TNF-α) also increases endocannabinoid levels. Activation of CB2 cannabinoid receptors by JWH133 protects against I/R damage by decreasing inflammatory cell infiltration, tissue and serum TNF-α, MIP-1α and MIP-2 levels, tissue lipid peroxidation, and expression of adhesion molecule ICAM-1 in vivo. JWH133 also attenuates the TNF-α-induced ICAM-1 and VCAM-1 expression in human liver sinusoidal endothelial cells (HLSECs) and the adhesion of human neutrophils to HLSECs in vitro. Consistent with the protective role of CB2 receptor activation, CB2−/− mice develop increased I/R-induced tissue damage and proinflammatory phenotype. These findings suggest that oxidative/nitrosative stress and inflammatory stimuli may trigger endocannabinoid production, and indicate that targeting CB2 cannabinoid receptors may represent a novel protective strategy against I/R injury. We also demonstrate that CB2−/− mice have a normal hemodynamic profile.
The endocannabinoid system is emerging as a promising new therapeutic target in inflammation, cancer, and metabolic, cardiovascular, gastrointestinal, and liver disorders [reviewed in (1–8)]. To date, two cannabinoid (CB) receptors have been identified by molecular cloning: the CB1 receptor, which is highly expressed in the brain (9) but is also present in peripheral tissues, including the heart (10, 11), vascular tissues (12, 13), and liver (14–17), and the CB2 receptor, expressed primarily by immune and hematopoietic cells [(18); reviewed in (7)]. The natural ligands of these receptors are lipid-like substances called endocannabinoids, which include arachidonoyl ethanolamide or anandamide (AEA), 2-arachidonoylglycerol (2-AG), and oleylethanolamide (OEA, reviewed in (7).
Cannabinoid CB2 receptor knockout (CB2−/−) mice are fertile and care for their offspring and have similar phenotype as their wild-type littermates (19–22). Fluorescence-activated cell sorting analysis showed no differences in immune cell populations between CB2−/− and CB2+/+ mice, while the immunomodulatory effects of delta (9) tetrahydrocannabinol were absent in knockouts (19–21). CB2−/−mice had elevated serum TNF-α levels compared with their wild-type littermates following a challenge with a low dose of endotoxin (22) and were characterized by markedly accelerated age-related trabecular bone loss and cortical expansion, although cortical thickness remains unaltered (23).
Organ injury, caused by transient ischemia followed by reperfusion (I/R), may develop in common diseases such as myocardial infarction and stroke, and may also accompany coronary bypass surgery and organ transplantation. The destructive effects of I/R arise from the acute generation of reactive oxygen species subsequent to reoxygenation, which inflict direct tissue damage and initiate a chain of deleterious cellular responses leading to inflammation, cell death, and eventually organ failure [reviewed in (24–29)].
Information about the role of cannabinoid receptor activation in cell protective mechanisms against I/R damage in the heart and brain remains limited and conflicting [reviewed in (7, 30, 31)]. Most published studies were performed by using ex vivo models (e.g., isolated perfused hearts), where the role of very important immunomodulatory effects of cannabinoids could not be evaluated, and/or used non-specific ligands for CB2 receptors such as anandamide (2.3–32.3× less selective to CB2 than to CB1), WIN55212–2 (0.6–34× more selective to CB2 than to CB1), HU210 (1.7–8.6× less selective to CB2 than to CB1) with or without the CB2 receptor selective antagonist/inverse agonist SR144528 to delineate the role of CB2 receptors [reviewed in (7, 32)].
Activation of CB2 receptor by JWH133, which has been reported to be approx. 200× more selective to CB2 than to CB1 receptors (33), attenuated the inflammatory pain in rats (34, 35) and experimental colitis in mice (36). In the present study we have used pharmacological agonist JWH-133, and antagonist/inverse agonist of cannabinoid CB2 receptors SR144528, as well as CB2 receptor knockout (CB2−/−) mice to study the role of the endocannabinoid system in an in vivo model of liver ischemia reperfusion. In addition, we studied the effects of JWH133 on TNF-α-induced ICAM-1 and VCAM-1 expression in human liver sinusoidal endothelial cells (HLSECs) and adhesion of human neutrophils to HLSECs in vitro. Furthermore, we have characterized the detailed hemodynamic profile of CB2−/−mice. Our results may also have relevance to the reperfusion injury of other organ systems, since I/R injury share similar pathophysiological mechanisms in many organs.
All animal experiments conformed to NIH guidelines and were approved by the Institutional Animal Care and Use Committee of the National Institute on Alcohol Abuse and Alcoholism. CB2−/− mice and their wild-type littermates were developed as described previously and had been backcrossed to a C57Bl/6J background (10, 19, 21–23). C57Bl/6J mice were obtained from The Jackson Laboratory (Bar Harbor, ME, USA).
Mice were anesthetized with pentobarbital (65 mg/kg i.p.). A midline laparotomy incision was performed to expose the liver. The hepatic artery and the portal vein were clamped using microaneurysm clamps. This model results in a segmental (70%) hepatic ischemia. This method of partial ischemia prevents mesenteric venous congestion by allowing portal decompression throughout the right and caudate lobes of the liver. The liver was kept moist at 37°C with gauze soaked in 0.9% saline. Body temperature was maintained at 37°C using a thermoregulatory heating blanket and by monitoring body temperature with a rectal temperature probe. Sham surgeries were identical except that hepatic blood flow was not reduced with a microaneurysm clamp. The duration of hepatic ischemia was 60 min in all experiments, after which the microaneurysm clamps were removed. The duration of the reperfusion was 90 min or 24 h, as indicated. After reperfusion, blood was collected and liver samples were removed, weighed, and snap-frozen in liquid nitrogen for determining biochemical parameters or fixed in 4% buffered formalin for histopathological evaluation.
SR144528 (SR2) and SR141716 (SR1) were from the National Institute on Drug Abuse Drug Supply Program (Research Triangle Park, NC, USA). JWH133 was synthesized as described (33, 37). AM251 and AM630 were from Tocris (Baldwin, MO, USA). All drugs were emulsified in corn oil–water (1:4) as described (10, 11). All drugs were injected intraperitoneally 60 min prior to the occlusion of the hepatic artery and the portal vein. For cell culture experiments all lipid soluble drugs were dissolved in DMSO. All chemicals were from Sigma (St. Louis, MO, USA) except where mentioned otherwise.
The activities of aspartate aminotransferase (AST) and alanine aminotransferase (ALT), indicators of liver damage, were measured in plasma samples using a clinical chemistry analyzer system (PROCHEM-V; Drew Scientific, Oxford, CT, USA).
Myeloperoxidase activity was measured as described previously (38). Briefly, liver samples were homogenized (50 mg/ml) in 0.5% hexadecyltrimethylammonium bromide in 10 mM 3-(N-morpholino) propanesulfonic acid (MOPS) and centrifuged to separate the supernatant. An aliquot of the supernatant was mixed with a solution of 1.6 mM tetramethylbenzidine and 1 mM hydrogen peroxide. Activity was measured spectrophotometrically as the change in absorbance at 650 nm at 37°C, using a Spectramax microplate reader (Molecular Devices, Sunnyvale, CA, USA). MPO activity was expressed as mU/mg protein. Protein content was determined with the DC Protein assay (Bio-Rad, Hercules, CA, USA).
Malondialdehyde (MDA) formation was utilized to quantify the lipid peroxidation in tissues and measured as thiobarbituric acid-reactive material as described (38). Briefly, tissues were homogenized (100 mg/ml) in 1.15% KCl buffer. Homogenates (200 μl) were then added to a reaction mixture consisting of 1.5 ml of 0.8% thiobarbituric acid, 200 μl of 8.1% sodium dodecyl sulfate, 1.5 ml of 20% acetic acid (pH 3.5), and 600 μl of distilled H2O and heated at 90°C for 45 min. After cooling to room temperature, the samples were cleared by centrifugation (10,000 g, 10 min), and their absorbance at A532 was measured with 1,1,3,3-tetramethoxypropane as an external standard. The level of lipid peroxides was expressed as nmol MDA/mg protein.
The levels of the inflammatory cytokine TNF-α and the chemokines MIP-1α and MIP-2 in plasma or homogenized liver tissue were determined using commercially available enzyme-linked immunosorbent assays (ELISA, R&D Systems, Minneapolis, MN, USA), according to the manufacturer’s protocol as described previously (38). Expression of TNF-α, MIP-1α, MIP-2, and ICAM-1 in liver tissues was also determined from homogenized liver tissues employing RT-PCR.
Human liver sinusoidal endothelial cells (HLSECs) were obtained from Cell-Systems (Kirkland, WA, USA) and grown in CSC-complete growth medium according to the manufacturer’s recommendations. Cells were grown in 0.2% gelatin-coated 100 mm cell culture dishes and used within 3–6 passages. Polymorphonuclear neutrophil leukocytes (PMN) were isolated from whole blood obtained from a healthy volunteer (NIH Clinical Center, Bethesda, MD, USA) using Ficoll hypaque (GE Biosciences, Piscataway, NJ, USA) density gradient solution. Primary mouse hepatocytes were isolated and cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% fetal bovine serum (FBS) and 100 U/ml penicillin and 100 μg/ml streptomycin.
Cell surface expression of ICAM-1 and VCAM-1 were measured by in situ ELISA as described with modifications (39). In brief, HLSECs were grown in 96-well plates coated with 0.2% gelatin. The cells were treated with either TNF-α (50 ng/ml) ± JWH133 (0–4 μM) for 4 h. In some experiments, the cells were pretreated with either JWH133 (3 μM), SR1 or SR2, or AM 630, each used at the concentration of 1 μM for 1 h followed by incubation with TNF-α for 4 h. Then, cells were washed with phosphate buffered saline (PBS) and fixed in 4% formaldehyde (pH 7.4) and blocked with PBS containing 1% bovine serum albumin containing 100 mM glycine for 2 h at 4°C. The fixed monolayer was probed with either anti-human ICAM-1 or VCAM-1 monoclonal antibodies (R&D Systems, Minneapolis, MN, USA) for 1 h at 37°C and incubated with peroxidase-coupled anti-mouse (1:5000) (Pierce, Rockford, IL, USA) for 1 h at 37°C. Following washing, cells were incubated with 100 μl developing substrate solution (3, 3′, 5, 5′-tetramethyl-benzidine, Sigma) for 10 min and the reaction was terminated with 50 μl of 2N H2SO4, and the absorbance was measured at 450 nm. Each treatment was performed in triplicate and the experiments were repeated three times.
Neutrophil adhesion to endothelial cells was performed as described with modifications in the protocol (40). In brief, HLSECs were grown to confluence in 24-well plates and treated with TNF-α ± JWH133 or pretreated with CB1/CB2 antagonists followed by treatment with TNF-α/JWH133 as described above. Then, PMN were labeled with 2.5 μM Calcein-AM (Molecular Probes–Invitrogen, Carlsbad, CA, USA) for 1 h at 37°C in RPMI 1640 containing 1% FBS. HLSECs were washed twice with HLSECs basal medium and covered with 400 μl of HLSECs basal medium. Then 5 × 104/100 μl labeled PMN cells were added to HLSECs and incubated for 1 h at 37°C. After incubation, the monolayer was carefully washed with PBS to remove the unbound PMN. The adherent PMN were documented by Olympus IX 81 fluorescent microscope using 20× objective (Opelco, Dulles, VA, USA). Three fields were captured/experimental condition. Individual treatments were preformed in duplicate, and the entire set of experiments was repeated twice. The number of adherent PMN cells were counted using NIH Image J software and the values were expressed as PMN adhered/field.
For measuring endocannabinoid levels, mice were euthanized and their livers were removed and extracted. Anandamide, 2-AG, and OEA levels were determined by liquid chromatography/mass spectrometry from liver tissues, isolated hepatocytes, and other cell fractions as described previously (10, 41). Values are expressed as fmol or pmol/mg wet tissue or mg cell protein. In a separate set of experiments isolated primary hepatocytes were treated with TNF-α (100 ng/ml), endotoxin [1 μg/ml; lipopolysaccharide (LPS) from Escherichia coli, 0127:B8], H2O2 (100 μM) and peroxynitrite (50 μM) for 90 min, collected and processed in 5 ml PBS containing 200 μM PMSF, and processed for endocannabinoid measurements.
Liver samples were fixed in 4% buffered formalin. After embedding and cutting 5 μm slices, all sections were stained with hematoxylin/eosin (HE). Myeloperoxidase staining of neutrophils was done by using antimyeloperoxidase antibody according to the manufacturer’s protocol (Zymed Lab., San Francisco, CA, USA), and samples were contrastained with nuclear fast red. Histological evaluation was performed in a blinded manner.
Detailed hemodynamic measurements were conducted in mice anesthetized with 2% isoflurane by using Millar’s pressure-volume system (Millar Instruments, Huston, TX, USA) as described previously (11, 42).
Results are presented as means ± sem. One-way ANOVA followed by Newman-Keuls multiple comparisons post-hoc analysis or unpaired t test for pair-wise comparisons and correlation between the variables (Pearson coefficient test) were calculated using the Graph Pad Prism 4 package (San Diego, CA, USA). P < 0.05 was considered significant.
For assessment of hepatocellular damage of the post-ischemic liver, the serum transaminase AST/ALT activities were measured. After 60 min of ischemia and subsequent 90-min reperfusion (60/90 min I/R), a dramatic increase in liver enzyme activities were observed in vehicle-treated C57Bl6/J mice as compared with sham-operated controls (Fig. 1). Pretreatment with 20 mg/kg of CB2 agonist JWH133 significantly reduced the transaminase levels following I/R, and this effect was not prevented by the CB1 selective antagonists AM251 (not shown), but was largely attenuated by the CB2 selective antagonist SR144528. SR144528 alone showed a tendency to aggravate I/R damage, which did not reach statistical significance (not shown). In agreement with a protective role of CB2 receptors suggested by the pharmacological findings, the liver damage to I/R was significantly more severe in CB2 receptor knockout mice (CB2 −/−) compared with wild-type littermates (CB2+/+) (Fig. 1A, B: right panels). The protective effect of JWH133 against I/R-induced liver damage was abolished in CB2−/− mice (AST: 6440.63611.5 vs. 5900.03526.0 and ALT: 8735.4±716.6 vs. 8543.8±779.5; in CB2 −/− +vehicle vs. CB2 −/− +JWH133, respectively; n=4 in each group) subjected to 60/90 min I/R.
An important factor in the tissue damage following I/R is neutrophil infiltration, an indicator of which is tissue MPO activity. In sham-operated wild-type mice MPO activity was barely detectable (Fig. 2A). 60/90 min I/R induced a marked increase in MPO activity, which was attenuated by JWH133. SR144528 pretreatment prevented the effect of JWH133. Accordingly, the I/R-induced increase of MPO activity was significantly greater in CB2−/− than in CB2+/+ mice (Fig. 2A), which is consistent with the more severe tissue damage observed in the former (Fig. 1A, B: right panels).
The rate of lipid peroxidation was negligible in sham-operated mice as indicated by the low MDA content. MDA content nearly doubled following 60/90 min I/R and this increase was attenuated in mice pretreated with JWH133, an effect preventable by SR144528 pre-treatment (Fig. 2B: left panel). Again, I/R caused a greater increase in MDA in CB2−/− than in CB2+/+ mice (Fig. 2B: right panel).
60/90 min I/R greatly increased the expression of TNF-α, MIP-1α, MIP-2, and ICAM-1 in liver tissue, as documented by RT-PCR (Fig. 3A), with concomitant increase of their levels in both liver homogenates and in the serum, as detected by ELISA (Fig. 3B, C). JWH133 significantly attenuated the I/R-induced increase in cytokine levels in both liver and serum, whereas SR144528 pretreatment largely prevented these effects of JWH133 (Fig. 3B, C). Inflammatory markers were significantly more elevated following 60/90 min I/R in CB2−/− as compared to CB2+/+ mice (Fig. 3A–C).
Sham-operated mice showed normal hepatic histology (Fig. 4A). In mice 24 h following ischemic injury (60 min I), there was a marked degree of reperfusion damage indicated by the necrosis of hepatocytes in the pericentral and midzonal regions, and by the massive neutrophil infiltration in the damaged areas (Fig. 4A, B). The tissue injury and neutrophil infiltration were less pronounced in the JWH133-treated group and more extensive in the CB2−/− mice (Fig. 4A, B).
TNF-α (50 ng/ml) treatment of HLSECs for 4 h resulted in robust activation of ICAM-1 (~5 fold, Fig. 5A) and VCAM-1 (~4-fold, Fig. 5C), respectively, when compared with control. JWH133 (0.5–4 μM) dose dependently inhibited the TNF-α induced increased expression levels of both ICAM-1 and VCAM-1 (Fig. 5A, C), which was attenuated by CB2 antagonists (SR 2 or AM 630; 1 μM), but not CB1 antagonists (SR1; 1 μM, not shown) (Fig. 5B, D). Antagonists had no effect in controls or TNF-α treated cells.
TNF-α (50 ng/ml) treatment, resulted in enhanced PMN adhesion to HLSECs (~3.5-fold) compared with control (Fig. 5E). JWH133 (3 μM) pretreatment markedly inhibited TNF-α induced PMN adhesion to HLSECs and this effect was prevented by CB2 antagonists. Antagonists had no effect in controls or TNF-α treated cells on PMN adhesion.
Ischemia/reperfusion (60/90 min) but not ischemia (60 min) alone led to substantial increases in liver tissue content of the endocannabinoids anandamide (AEA) and 2-arachidonoylglycerol (2-AG) as well as the related lipid oleoylethanolamide (OEA, Fig. 6A). Similarly, the levels of these lipids were increased in the purified hepatocyte fraction (Fig. 6B) as well as in the Kupffer cell/endothelial cell fraction of isolated liver cells (except for 2-AG in the latter; Fig. 6C).
Significant correlation was found between hepatic AEA and 2-AG and the plasma transaminase levels: AEA vs. ALT and AEA vs. AST, r: 0.77 and r: 0.73, (P<0.0001; respectively, Fig. 7) following 60/90 min I/R. Similarly, chemokine-cytokine levels and hepatic AEA and 2-AG content correlated significantly. TNF-α levels correlated with the hepatic content of all three lipids (AEA r: 0.52, P=0.008; 2-AG r: 0.73, P=0.0004; OEA r: 0.7, P=0.0095), whereas MIP-1α and MIP-2 levels correlated only with AEA (r: 0.42, P=0.024 and r: 0.5, P=0.01, respectively, Fig. 7).
TNF-α, endotoxin (LPS), H2O2, and peroxynitrite (ONOO−) exposure of isolated primary hepatocytes for 90 min markedly increased endocannabinoid levels in these cells (especially AEA and OEA; Fig. 8).
JWH133 had no effect on hemodynamic variables of control mice (20 mg/kg i.p. injection, Table 1). Furthermore, CB2−/− mice had normal hemodynamic profile, the parameters in Table 2 are not significantly different from their wild-type littermates.
The hypothesis on the role of the endocannabinoid system in I/R is controversial, as it has been derived from a few studies using isolated organs and is based on results obtained mostly through the use of pharmacological agents (7, 30, 31). In the present study, we used agonists and antagonists of the cannabinoid CB2 receptor as well as CB2 receptor knockout mice to delineate the role of the endocannabinoid system in an in vivo model of liver ischemia reperfusion (I/R).
We demonstrate that pretreatment of mice with a CB2 receptor agonist JWH133 decreases inflammatory cell infiltration, tissue and serum TNF-α, MIP-1α and MIP-2 levels, tissue lipid peroxidation, and tissue expression of adhesion molecule ICAM-1. CB2 activation also attenuates the TNF-α-induced ICAM-1 and VCAM-1 expression in human liver sinusoidal endothelial cells (HLSECs) and adhesion of human neutrophils to HLSECs in vitro.
Importantly, these findings suggest that targeting CB2 cannabinoid receptors may represent a novel strategy in protecting against hepatic I/R injury. In contrast, CB2−/− mice develop increased I/R-induced tissue damage and proinflammatory phenotype. I/R, but not ischemia alone, triggers several-fold increases in the hepatic levels of the endocannabinoids anandamide and 2-AG, which originate from hepatocytes, Kupffer, and endothelial cells. These increases positively correlate with the degree of tissue damage and serum TNF-α, MIP-1α, and MIP-2 levels. Consistently, brief exposure of primary hepatocytes to various oxidants, (H2O2, peroxynitrite) or inflammatory stimuli (TNF-α endotoxin), which are important mediators of reperfusion damage (24–29, 43), triggers marked increase in cellular endocannabinoid levels. Thus, not only inflammatory stimuli (e.g., endotoxin and TNF-α), but also oxidative/nitrosative stress, can modulate endocannabinoid levels in hepatocytes, and most likely in most other cell types, too. Therefore, parenchymal cells may also represent a very significant source of endocannabinoids produced in various pathological conditions associated with increased inflammation and oxidative tissue injury, in addition to the previously reported activated macrophages (reviewed in 7). Our findings also suggest that an I/R-induced activation of hepatic endocannabinoids may limit the extent of tissue injury via stimlation of CB2 receptors.
Increased brain endocannabinoid levels during ischemic brain injury were reported by several studies (44–47). However, the role of endocannabinoids and CB1 receptor activation in cerebral I/R remains controversial. Anandamide, 2-AG, as well as the synthetic cannabinoid WIN 55,212–2, were found to protect cultured cortical neurons against hypoxia and glucose deprivation by a mechanism not involving CB1/2 receptors (48, 49). In contrast, in vivo treatment with WIN 55,212–2 or the CB1 agonist BAY38–7271 reduced infarct size following cerebral ischemia in rats via activation of CB1 receptors (48, 50). Consistent with the CB1-mediated cerebro-protection, infarct size and mortality following cerebral ischemic injury were greater in CB1 knockout mice than in their wild-type littermates (51). The protective role of CB1 receptor stimulation against cerebral ischemia was confirmed by a more recent study using the synthetic cannabinoid HU-210, the protective effect of which could be attributed to CB1-mediated hypothermia (52). Likewise, CB1-mediated hypothermia was responsible for the neuroprotective effects of delta (9)-tetrahydrocannabinol in a mouse ischemic model of cerebral injury (53), and perhaps also in a rat model of global cerebral ischemia (54). In contrast, other studies do not support the neuroprotective role of endocannabinoids and CB1 receptor activation. In fact, the CB1 antagonists SR141716 and LY320135 were found to reduce infarct size and to improve neurological function in a rat model of cerebral ischemia (46, 47).
Endocannabinoids acting via CB1/CB2-dependent, or CB1/CB2-independent mechanisms have also been implicated in the protection conferred by various forms of preconditioning (including ischemic) of the myocardium [reviewed in (7, 30, 55)]. However, a major limitation of these studies is the use of buffer-perfused isolated heart preparations, in which the effects of endocannabinoids and synthetic agonists on immune cells, which are pivotal in reperfusion damage, cannot be studied, as well as the use of nonselective cannabinoid ligands (7, 30). In a more relevant whole animal model of myocardial I/R injury induced by coronary occlusion/reocclusion in anesthetized mice, the published evidence points to the protective role CB2 but not CB1 receptor activation [reviewed in (7, 55)]. In a recent study using nonselective CB agonist WIN55212–2 and CB2 antagonist AM630 in a mouse model of myocardial I/R, the reduction of leukocyte-dependent myocardial damage could be attributed to CB2 receptor activation since the protection afforded by WIN55212–2 could be prevented by AM630, but not by CB1 antagonist AM251 (56). In agreement with those findings, pretreatment of mice in the present study with the selective CB2 agonist JWH133 afforded protection against hepatic I/R damage by decreasing inflammatory cell infiltration and consequent inflammatory damage. This protective effect as well as the decreased inflammatory cell infiltration were attenuated by a selective CB2 receptor antagonist and in CB2 knockout mice, further confirming the role of CB2 receptors in these events. Consistently, CB2 receptor knockout mice developed increased injury and proinflammatory phenotype following I/R.
It is well established that liver I/R injury is dependent on PMN infiltration, Kupffer cell activation, and cytokine response (57–60). Adhesion molecules mediate the initial attachment of neutrophils to the activated endothelium (58, 61). On reperfusion, TNF-α acts as a continuous stimulator for neutrophil infiltration in the liver and it also up-regulates the production of cell-type specific leukocyte chemoattractants, known as chemokines, which have also been shown to cause up-regulation of cell adhesion molecules and neutrophil activation (59). The increased inflammatory response further aggravates oxidative stress and initiates a chain of deleterious events eventually culminating in cellular dysfunction and death. As demonstrated by the present findings and illustrated in the proposed schematic diagram (Fig. 9), CB2 receptor stimulation limits hepatic injury by decreasing the expression of ICAM-1, VCAM-1, neutrophil infiltration, TNF-α, chemokine (MIP-1α and MIP-2) levels, and lipid peroxidation. The decrease in lipid peroxidation can be secondary to the decrease of inflammatory cell infiltration and activation.
Hepatic ischemia/reperfusion or exposure of primary hepatocytes to various inflammatory stimuli or oxidants triggered marked increases in the cellular endocannabinoid levels, which were positively correlated with the degree of tissue damage and inflammatory markers. Our findings suggest that these endocannabinoids may limit hepatic injury by modulating the expression of adhesion molecules and the infiltration and activation of inflammatory cells by CB2-dependent mechanisms, which is also consistent with the emerging role of CB2 receptors in regulating microglial cell function and neuroinflammation (62, 63). Both mononuclear and polymorphonuclear leukocytes are known to express CB2 receptors (29, 64), which could be activated on these infiltrating cells through a paracrine mechanism by endocannabinoids generated in and released from the various cell types in the liver.
It is noteworthy that the endocannabinoid anandamide can promote stellate cell and hepatocyte apoptosis in vitro by a mechanism not related to CB receptors (65, 66). Thus, the role of endocannabinoids in liver injury is still a very controversial issue requiring further clarification [reviewed in (3, 4, 6, 7)].
Collectively, our results suggest that treatment with selective CB2 cannabinoid agonists may be useful in protecting the liver as well as other tissues against I/R injury. This is particularly encouraging since CB2 receptor stimulation is not associated with psychoactive effects and was also found to mediate antifibrogenic effects in the liver (67). Even though the protection afforded by JWH133 in hepatic I/R was lost in CB2 knockout mice and attenuated by CB2 antagonist, which clearly indicates CB2 receptor involvement, as with any drugs available, it cannot be excluded that JWH133 has some additional beneficial effects not related to CB2 receptor activation.
This study was supported by the Intramural Research Program of NIH/NIAAA (to P.P.) and NIDA grant DA03590 (to J.W.H.).