Potential conflict of interest: Nothing to report.
Angiogenesis in liver cirrhosis leads to splanchnic hyperemia, increased portal inflow, and portosystemic collaterals formation, which may induce lethal complications, such as gastroesophageal variceal hemorrhage and hepatic encephalopathy. Cannabinoids (CBs) inhibit angiogenesis, but the relevant influences in cirrhosis are unknown. In this study, Spraque-Dawley rats received common bile duct ligation (BDL) to induce cirrhosis. BDL rats received vehicle, arachidonyl-2-chloroethylamide (cannabinoid receptor type 1 [CB1] agonist), JWH-015 (cannabinoid receptor type 2 [CB2] agonist), and AM630 (CB2 antagonist) from days 35 to 42 days after BDL. On the 43rd day, hemodynamics, presence of CB receptors, severity of portosystemic shunting, mesenteric vascular density, vascular endothelial growth factor (VEGF), VEGFR-1, VEGFR-2, phospho-VEGFR-2, cyclooxygenase (COX)-1, COX-2, and endothelial nitric oxide synthase (eNOS) expressions as well as plasma VEGF levels were evaluated. Results showed that CB1 and CB2 receptors were present in left adrenal veins of sham rats, splenorenal shunts (the most prominent intra-abdominal shunts) of BDL rats, and mesentery of sham and BDL rats. CB2 receptor was up-regulated in splenorenal shunts of BDL rats. Both acute and chronic JWH-015 treatment reduced portal pressure and superior mesenteric arterial blood flow. Compared with vehicle, JWH-015 significantly alleviated portosystemic shunting and mesenteric vascular density in BDL rats, but not in sham rats. The concomitant use of JWH-015 and AM630 abolished JWH-015 effects. JWH-133, another CB2 agonist, mimicked the JWH-015 effects. JWH-015 decreased mesenteric COX-1, COX-2 messenger RNA expressions, and COX-1, COX-2, eNOS protein expressions. Furthermore, JWH-015 decreased intrahepatic angiogenesis and fibrosis. Conclusions: CB2 agonist alleviates portal hypertension (PH), severity of portosystemic collaterals and mesenteric angiogenesis, intrahepatic angiogenesis, and fibrosis in cirrhotic rats. The mechanism is, at least partly, through COX and NOS down-regulation. CBs may be targeted in the control of PH and portosystemic collaterals. (HEPATOLOGY 2012;56:248–258)
Angiogenesis, the generation of new blood vessels from the preexisting vessels, is involved in the development of increased portal inflow and pressure as well as of portosystemic collaterals in portal hypertension (PH).1 The portosystemic collateral vascular bed is primarily triggered with an attempt to divert the stagnant portal blood flow to systemic circulation. Complications include hepatic encephalopathy caused by the noxious material draining into systemic circulation, and bleeding from the most prominent portosystemic shunts, gastroesophageal varices ensue. Vascular endothelial growth factor (VEGF) is pivotal in angiogenesis, triggering proangiogenic activities through mainly VEGF receptor (VEGFR)-1 and VEGFR-2.2 In PH rats, VEGFR-2 inhibition significantly decreased portal venous inflow, increased splanchnic arteriolar resistance, and decreased splanchnic cluster of differentiation (CD)31 expression, suggesting the potential of alleviating the hyperdynamic splanchnic circulation by an antiangiogenesis strategy.3
The active components of marijuana and their derivatives, the cannabinoids (CBs), participate in a wide spectrum of central and peripheral effects. Two different CB receptors have been characterized: cannabinoid receptor type 1 (CB1)4 and cannabinoid receptor type 2 (CB2)5 receptors. The former is distributed in the central nervous system and various peripheral tissues, whereas the later is mostly confined to immune tissue and cells. It has been found that CBs reduced VEGF production and VEGFR-2 activation in cultured glioma cells and mouse gliomas, which could be abrogated by CB1/CB2 receptor antagonists.6 JWH-133, a CB2 receptor agonist, reduced proangiogenic factor expression that destabilizes vessel integrity and facilitates vessel sprouting, including angiopoietin-2 (Ang-2), VEGF, and matrix metalloproteinase-2.7 Results indicate the roles of CBs in ameliorating angiogenesis.
Prostacyclin and nitric oxide (NO), synthesized by cyclooxygenases (COX-1 and COX-2) and endothelial NO synthase (eNOS), respectively, participate in the hyperdynamic circulation and vascular derangement in PH.8, 9 An enhanced COX-1 expression in regions of ovarian epithelial tumors undergoing extensive angiogenesis has been found, and COX-1 may promote ovarian cancer development through the stimulation of angiogenesis.10 COX-2 has also been found to participate in angiogenesis, and COX-2 inhibitor suppresses angiogenesis in secondary bone tumors.11 Furthermore, NO plays a role in the process of neovascularization in PH rats.12, 13 Among the NO synthases, eNOS has been proven critical in angiogenesis in response to tissue ischemia in vivo.14 To date, whether CB agents influence angiogenesis in cirrhosis by modulating COX or NOS remains unexplored.
The hemodynamic influences of CBs in cirrhosis and PH are worth noting, too. SR141716A, a CB1 antagonist, elevated blood pressure and reduced mesenteric blood flow and portal pressure (PP) in CCl4-induced cirrhotic rats. In addition, monocytes from cirrhotic rats elicited SR141716A-sensitive hypotension in normal rats.15 However, the hemodynamic influences exerted by CB-receptor agonists in cirrhosis are still unknown.
Considering the complexities of CB effects in different pathophysiological conditions, we herein surveyed the presence of CB receptors (e.g., CB1 and CB2) in left adrenal veins of sham rats and derived from which the splenorenal shunts in bile duct ligation (BDL)-induced cirrhotic rats. The portosystemic collateral circulation is mainly represented by the splenorenal shunt in animal models with liver cirrhosis.16 Furthermore, influences of CBs on CB receptors, mesenteric vascular density, portosystemic shunting, and the mechanism of antiangiogenic effect were evaluated. Results show that CB2 agonist JWH-015 reduces PP and alleviates portosystemic shunting and mesenteric angiogenesis in BDL rats, which is, at least in part, mediated by COX and NOS down-regulation. Furthermore, JWH-015 ameliorates hepatic angiogenesis and fibrosis.
For a description of the materials and methods, see the Supporting Materials and Methods, in which are described the following: materials17; animal model18-20; measurement of systemic and portal hemodynamics21; in situ perfusion of liver22-24; intrahepatic resistance calculation; portosystemic shunting ratio analysis25, 26; hematoxylin and eosin (H&E) staining; mesenteric vascular density assessment with immunofluorescence (IF) study27; intrahepatic angiogenesis evaluation with CD31 immunohistochemical (IHC) staining28; hepatic fibrosis determination with Sirius Red staining; real-time quantitative polymerase chain reaction (PCR)29-33; western blotting analysis34; and determination of plasma VEGF level.
CB receptor messenger RNA and protein expressions in left adrenal veins of sham rats, splenorenal shunts of BDL rats, and mesentery of both
Real-time PCR, western analysis, and IHC staining were applied to survey CB1 and CB2 expressions in rats on day 35 after BDL and the corresponding sham rats.
Hemodynamics, mesenteric angiogenesis, and portosystemic shunting modulation in BDL rats with different treatments
On day 43 after BDL, three series of experiments were performed on BDL rats that had randomly received vehicle (dimethyl sulfoxide; DMSO), arachidonyl-2-chloroethylamide (ACEA; CB1 agonist, 3 mg/kg/day, intraperitoneally [IP]), or JWH-015 (CB2 agonist, 3 mg/kg/day, IP) from days 35 to 42 days after BDL. In the first series, body weight (BW), mean arterial pressure (MAP), heart rate (HR), PP, and superior mesenteric arterial (SMA) flow were measured. In the second series, portosystemic shunting was evaluated by the color microsphere method. In the third series, H&E staining, mesenteric window vascular length and area and hepatic vascular density with CD31 IF staining, and Sirius Red staining of liver were performed. Additional experiments were arranged to survey portosystemic shunting and mesenteric angiogenesis in BDL rats treated with the following: (1) AM-630 (CB2 antagonist, 3 mg/kg/day, IP); (2) JWH-015 plus AM630; and (3) another CB2 agonist, JWH-133 (1 mg/kg/day, IP). To survey the acute hemodynamic effects exerted by JWH-015, MAP, HR, PP, SMA flow change, and intrahepatic resistance were measured at 30 minutes after JWH-015 injection (2.4 mg/kg, IP35).
Mesenteric angiogenic factor expression and plasma VEGF concentrations in BDL rats with different treatments.
Mesenteric angiogenic factor messenger RNA (mRNA) and protein expressions, including VEGF, VEGFR-1, VEGFR-2, phosphoralated VEGFR-2 (pVEGFR-2) Tyr 1214, COX-1, COX2, eNOS, and plasma VEGF concentrations in BDL rats treated with DMSO or JWH-015, were evaluated.
Effects of chronic JWH-015 treatment on sham rats
Sham rats were randomly allocated to receive vehicle or JWH-015 (3 mg/kg/day, IP) from days 35 to 42 after the operation. Because noncirrhotic (i.e., sham) rats do not develop significant portosystemic shunts, only mesenteric vascular length and area were evaluated to survey whether JWH-015 affected the relatively normal vasculature.
Results are expressed as mean ± standard error of the mean. Statistical analyses were performed using an unpaired Student t test or one-way analysis of variance, as appropriate. Results were considered statistically significant at a two-tailed P value less than 0.05.
BW and Hemodynamics After Chronic CB Treatments.
Table 1 depicts BW and hemodynamics in BDL rats after vehicle (n = 10), ACEA (n = 14), or JWH-015 (n = 12) treatment. Compared with the vehicle group, JWH-015-treated rats had lower PP and SMA flow. Furthermore, ACEA-treated rats had lower MAP.
Table 1. BW and Hemodynamics in BDL Rats With Different Treatments
JWH-015 significantly decreased PP (vehicle [n = 5] versus JWH-015 [n = 6] [mmHg]: 16.0 ± 0.7 versus 12.1 ± 1.3; P = 0.037) and SMA flow (SMA flow at 30 minus minus baseline flow (mL/min): −1.5 ± 0.5 versus −5.4 ± 0.7; P = 0.003) at 30 minutes after injection without influencing intrahepatic resistance (mmHg/mL/min/g: 0.0053 ± 0.0008 versus 0.0065 ± 0.0009; P = 0.383).
CB Receptor mRNA and Protein Expressions in Left Adrenal Veins (Sham) and Splenorenal Shunts (BDL Group).
The presence of CB receptors was supported by left adrenal vein and splenorenal shunt CB1 and CB2 mRNA and protein expression. Figure 1 reveals that CB1 and CB2 receptor mRNA expression was not significantly different between sham and BDL rats (sham [n = 8] versus BDL [n = 11] [/β-actin]: CB1: 0.001279 ± 0.000190 versus 0.001470 ± 0.000374, P = 0.706; CB2: 0.002182 ± 0.000410 versus 0.002099 ± 0.000520, P = 0.908). On the other hand, CB2 protein expression was significantly higher in BDL rats (sham [n = 5] versus BDL [n = 5] [/β-actin]: CB1: 1.006461 ± 0.038031 versus 1.135064 ± 0.096415, P = 0.430; CB2: 0.168874 ± 0.071678 versus 0.446750 ± 0.093001, P = 0.045). Moreover, ACEA treatment down-regulated splenorenal shunt CB1 mRNA expression (vehicle versus ACEA: 0.002263 ± 0.000152 versus 0.000566 ± 0.000183; P < 0.001), and JWH-015 down-regulated CB2 mRNA expression (0.003658 ± 0.000827 versus 0.000494 ± 0.000084; P = 0.023).
Mesenteric CB Receptor mRNA and Protein Expression.
The presence of CB receptors was proved by mesenteric CB1 and CB2 mRNA and protein expression. Figure 2 shows that mesenteric CB1 and CB2 receptor mRNA and protein expression were not significantly different between sham and BDL rats (mRNA: sham [n = 4] versus BDL [n = 5] [/β-actin]: CB1: 0.001940 ± 0.000427 versus 0.001804 ± 0.000535, P = 0.887; CB2: 0.006670 ± 0.001866 versus 0.002078 ± 0.000170, P = 0.090; protein [/β-actin]: sham [n = 4] versus BDL [n = 5]: CB1: 2.025018 ± 0.190832 versus 2.091687 ± 0.106562, P = 0.756; CB2: 0.94275 ± 0.197974 versus 1.079451 ± 0.136112, P = 0.579). IHC staining revealed that CB1 receptors were mainly distributed in vascular endothelial cells (ECs) and fibroblasts and mast cells scattered in perivascular connective tissue. CB2 receptors were found in vascular ECs and smooth muscle cells and mast cells.
Portosystemic Shunting Ratio.
Figure 3 depicts the severity of portosystemic shunting in rats with different treatments. Compared with vehicle-injected BDL rats, ACEA did not significantly modify the shunting, whereas JWH-015 significantly alleviated, and AM630 aggravated, the severity of shunting (vehicle [n = 12]: 68.6 ± 3.6, ACEA [n = 8]: 57.4 ± 7.7, JWH-015 [n = 8]: 28.3 ± 5.3, AM630 [n = 5]: 81.2 ± 5.3, vehicle versus JWH-015: P = 0.015; vehicle versus AM630: P = 0.025). Furthermore, CB2 antagonist AM630 abolished the alleviation of shunting exerted by CB2 agonist JWH-015 (JWH-015+AM630 versus vehicle (%): 72.9% ± 19.9% versus 68.6% ± 3.6%, P = 0.140; JWH-015+AM630 versus JWH-015: P = 0.002). Another CB2 receptor agonist, JWH-133, decreased shunting, as compared with vehicle (Fig. 3B; JWH-133 versus vehicle [%]: 46.0% ± 13.3% versus 68.6% ± 3.6%; P = 0.046).
Mesenteric Vascular Density.
Figure 4 shows the vascular length and vascular area per unit area of mesenteric window and representative IF images of BDL rats with vehicle, ACEA, JWH-015, AM630, JWH-015+AM630, or JWH-133 treatments. Compared with vehicle-injected BDL rats, ACEA did not significantly modify vascular density (vascular length per unit area [μm−1 × 100]: 0.79 ± 0.21, P = 0.226 versus vehicle; vascular area per unit area [%]: 0.02 ± 3.41, P = 0.540 versus vehicle), whereas JWH-015 significantly decreased the vascular length and area per unit area of mesenteric window (vehicle versus JWH-015: vascular length per unit area: 0.51 ± 0.06 versus 0.15 ± 0.05, P = 0.001; vascular area per unit area [%]: 11.61 ± 2.13 versus 2.13 ± 0.73, P = 0.001). On the other hand, AM630 significantly increased vascular density (vascular length per unit area: 1.38 ± 0.18, P = 0.001 versus vehicle; vascular area per unit area: 30.01 ± 3.12, P < 0.001 versus vehicle) and abolished the JWH-015 effect, if being used concomitantly (JWH-015+AM630: vascular length per unit area: 1.25 ± 0.20, P < 0.001 versus JWH-015; vascular area per unit area: 12.41 ± 0.78, P = 0.001 versus JWH-015.). Another CB2 agonist, JWH-133, significantly decreased the mesenteric window vascular area (vascular length per unit area: 0.33 ± 0.10, P = 0.184 versus vehicle; vascular area per unit area: 5.86 ± 1.39, P = 0.037 versus vehicle).
Morphologically, compared with the vehicle-treated BDL rats, significantly increased vessels that were irregularly and densely distributed were found in the AM630 and JWH-015+AM630 groups. The vessels also looked tortuous. Remarkably, less vessels in JWH-015- and JWH-133-treated rats were noted.
Vascular Effects of JWH-015 on Sham Rats.
Because JWH-015 significantly alleviated the severity of shunting and vascular density in BDL rats, whether JWH-015 also affected rats without cirrhosis were surveyed. JWH-015 treatment did not influence vascular length and area per mesenteric window area (vehicle versus JWH-015: vascular length per unit area [μm−1 × 100]: 0.43 ± 0.12 versus 0.51 ± 0.08, P = 0.599; vascular area per unit area [%]: 1.15 ± 0.41 versus 1.47 ± 0.67, P = 0.728).
Mesenteric Angiogenic Factor mRNA Expressions in BDL Rats Treated With Vehicle, ACEA, or JWH-015.
Because JWH-015, but not ACEA, significantly alleviated portosystemic shunting and mesenteric angiogenesis, we focused on angiogenic factor expressions in cirrhotic rats treated with vehicle, ACEA, or JWH-015. Figure 5 shows that ACEA or JWH-015 did not significantly influence VEGF and VEGFR-1 mRNA expression (vehicle [n = 9] versus ACEA [n = 8] versus JWH-015 [n = 8] [/β-actin]: VEGF: 0.008942 ± 0.001375 versus 0.006668 ± 0.000992 versus 0.005110 ± 0.001570; VEGFR-1: 0.008131 ± 0.002920 versus 0.005186 ± 0.000924 versus 0.003701 ± 0.000942; all P > 0.05 among groups), except that JWH-015 down-regulated VEGFR-2 expression (VEGFR-2: 0.527257 ± 0.255926 versus 0.187988 ± 0.024062 versus 0.050000 ± 0.092604, P = 0.043 vehicle versus JWH-015). On the other hand, mRNA expression of COX-1 and COX-2 were significantly down-regulated by JWH-015 (vehicle versus ACEA versus JWH-015: COX-1: 0.000937 ± 0.000100 versus 0.000826 ± 0.000072 versus 0.000687 ± 0.000065, P = 0.039 vehicle versus JWH-015; COX-2: 0.013260 ± 0.002862 versus 0.009470 ± 0.001125 versus 0.006986 ± 0.000935, P = 0.018 vehicle versus JWH-015). There was a trend of eNOS down-regulation by JWH-015 (0.001990 ± 0.000274 versus 0.001868 ± 0.000522 versus 0.001073 ± 0.000074, P = 0.073 vehicle versus JWH-015).
Mesenteric Angiogenic Factor Protein Expressions in BDL Rats Treated With Vehicle, ACEA, or JWH-015.
Figure 6 depicts that compared with the vehicle-treated group, ACEA and JWH-015 did not significantly influence VEGF, VEGFR-1, and pVEGFR-2 protein expression (vehicle [n = 6] versus ACEA [n = 6] versus JWH-015 [n = 6]: [/β-actin]: VEGF: 0.5263 ± 0.0359 versus 0.6750 ± 0.0985 versus 0.6992 ± 0.1414; VEGFR-1: 0.4655 ± 0.0814 versus 0.4566 ± 0.0848 versus 0.4015 ± 0.0806; pVEGFR-2 Tyr 1214: 0.8905 ± 0.0287 versus 0.8416 ± 0.0301 versus 0.8219 ± 0.0172; all P > 0.05 among groups), although JWH-015 decreased VEGFR-2 protein expression (VEGFR-2: 0.6236 ± 0.1046 versus 0.4193 ± 0.0653 versus 0.3561 ± 0.0726, vehicle versus JWH-015: P = 0.037). Furthermore, JWH-015 down-regulated COX-1 (0.7698 ± 0.0973 versus 0.6187 ± 0.2439 versus 0.4028 ± 0.0820, vehicle versus JWH-015: P = 0.020), COX-2 (2.3019 ± 0.4113 versus 2.1364 ± 0.7652 versus 0.9256 ± 0.2779, vehicle versus JWH-015: P = 0.020), and eNOS (vehicle versus ACEA versus JWH-015: 1.2430 ± 0.1105 versus 1.4293 ± 0.4712 versus 0.6072 ± 0.1734, vehicle versus JWH-015: P = 0.011) expressions.
Plasma VEGF Levels in BDL Rats With Vehicle or JWH-015 Treatment.
Compared with vehicle-treated BDL rats, JWH-015 significantly decreased the plasma concentration of VEGF (vehicle [n = 6] versus JWH-015 [n = 6] [ρg/mL−1]: 54.4 ± 3.1 versus 45.3 ± 1.1; P = 0.033) in BDL rats.
Figure 7A reveals that compared with the vehicle-treated group, BDL rats treated with JWH-015 had significantly lower mean vascular count (vehicle [n = 6] versus ACEA [n = 6] versus JWH-015 [n = 7]: 87.9 ± 5.2 versus 79.8 ± 3.8 versus 64.3 ± 5.0, P = 0.003 vehicle versus JWH-015 and P = 0.034 ACEA versus JWH-015).
Figure 7B discloses that the Sirius Red–stained area per unit area of hepatic tissue was significantly lower in the JWH-015-treated group, but higher in the ACEA-treated group (vehicle versus ACEA versus JWH-015 [%]: 21.39 ± 0.81 versus 25.33 ± 1.49 versus 17.65 ± 1.23, P = 0.046 vehicle versus ACEA; P = 0.043 vehicle versus JWH-015; P < 0.001 ACEA versus JWH-015).
Both CB1 and CB2 receptors are detected in left adrenal veins of sham rats and splenorenal shunts of BDL rats. Furthermore, splenorenal shunt CB2 receptor protein expression was significantly higher in BDL than in sham rats, providing direct evidence of CB2 receptor up-regulation in portosystemic collateral vessels. CB receptors have been identified in various vessels, including rat aorta,36 rat mesenteric vascular bed,37 and human pulmonary artery.38 However, the presence of CB receptors in the collateral vasculature and the up-regulated CB2 receptor expression in cirrhosis have not been proven. In mesenteric tissue, CB1 and CB2 receptors could also be identified in vascular ECs. The current finding suggests that pharmacological agents targeted on CB receptors to control vascular complications in cirrhosis, such as mesenteric angiogenesis and portosystemic collaterals, are feasible. To ensure an effective action mediated by CB receptors, the doses and schedules of CB agonists were determined according to the previous literature.39, 40 Furthermore, down-regulation of CB receptors after corresponding chronic agonist treatments noted in this study suggests the existence of effective treatment and interactions between ligands and receptors.
The cell-surface protein, CD31, is constitutively expressed on ECs and has been used as an indicator of the extent of vascularization and angiogenesis.41 In the current study, CB2 agonist JWH-015, but not CB1 agonist ACEA, decreased the CD31-stained vascular length and area of mesenteric windows (i.e., the extent of mesenteric angiogenesis). To survey whether the antiangiogenesis effect is CB2 receptor mediated, the CB2 antagonist, AM630, was applied, which elicited increased mesenteric angiogenesis and portosystemic collaterals in cirrhotic rats. This is also supported by the finding that combination use of JWH-015 and AM630 abolished the antiangiogenesis effect of JWH-015. Furthermore, another CB2 antagonist, JWH-133, decreased mesenteric vascular density. Another interesting finding is that JWH-015 decreased the hepatic CD31-stained vascular count, as compared with rats treated with vehicle. This could be beneficial, because accumulating evidences suggest that pathological angiogenesis is involved in the development of abnormal angioarchitecture in the cirrhotic liver, which is intimately related to fibrogenesis.42, 43
PP and SMA flow were decreased in both acute and chronic JWH-015-treated cirrhotic rats. Because mesenteric blood flow is the main contributor of portal blood inflow, JWH-015 may alleviate PH through the improvement of splanchnic hyperemia. This is proven by our current finding, showing that JWH-015 injection decreased SMA flow in 30 minutes. Interestingly, acute JWH-015 administration did not alter intrahepatic resistance. Because PP is determined by the net effects of portal inflow, intrahepatic resistance, and portosystemic collateral vascular resistance, the acute portal hypotensive effect of JWH-015 seems to be related to reduced flow. In the chronic setting, in addition to decreased flow, alleviated hepatic fibrosis by JWH-015 may also play a role. This is compatible with the finding that another CB2 agonist, JWH-133, used in CCl4-induced cirrhotic rats promoted the regression of fibrosis.44 Actually, JWH-133 also decreased mesenteric angiogenesis and portosystemic collaterals in the present study. MAP was not influenced by JWH-015, suggesting that use of JWH-015 in cirrhosis is devoid of adverse systemic hemodynamic influence.
JWH-015 also ameliorated portosystemic shunting, implying that the CB2 agonist may be promising in controlling gastroesophageal varices. JWH-015 did not influence mesenteric vascular density in sham rats, suggesting that the CB2 agonist modulates pathological angiogenesis, but not relatively normal vasculature. This is consistent with the previous finding that JWH-133 made the vessels of treated skin tumors smaller in size, more differentiated, and impermeable, compared to the controls.45
CBs had aroused much attention regarding their effects on angiogenesis in the past few years. WIN-55,212-2, a mixed CB1/CB2 agonist, inhibited VEGF production and the activation of VEGF receptor (e.g., VEGFR-2) in cultured glioma cells and in mouse gliomas, which could be abrogated by CB1/CB2 receptor antagonists.6 In another survey, Casanova et al. found that WIN-55,212-2 or JWH-133, a specific CB2 agonist, decreased the expression of proangiogenetic factors, VEGF, Ang-2, and placental growth factor in skin tumor.45 Further study with complementary DNA array analysis revealed that JWH-133 down-regulated genes related to angiogenesis in mouse gliomas, including VEGF and hypoxia-inducible factor-1 alpha.46 Controversial results also indicated that a new CB anticancer quinine, HU-331 (cannabidiol hydroxyquinone), inhibited angiogenesis by directly inducing apoptosis of vascular ECs without changing the expression of pro- and antiangiogenic cytokines and their receptors.7 The various results may be attributed to different experimental settings and CB ligands on different models. Furthermore, they reflect the complicated interactions exerted by CBs.
Compared with cirrhotic rats injected with vehicle, JWH-015 decreased mesenteric VEGFR-2 mRNA expression and protein expressions in mesentery. However, VEGF, VEGFR-1, and VEGFR-2 phosphorylation were not modified. This implies that the antiangiogenesis effect of JWH-015 may be mediated by inhibition of other proangiogenic factors. In fact, the decreased VEGFR-2 expression could be the result, rather than the cause, of decreased vasculature. Interestingly, JWH-015 reduced the circulating VEGF concentration, which is in contrast with the unaltered mesenteric VEGF protein expression. Because CB2 receptors are mainly distributed in immune tissue and cells, such as spleen and monocytes,47 this finding reflects the systemic VEGF-lowering effect of JWH-015 other than mesentery.
COX-1, COX-2, and eNOS were down-regulated by JWH-015. Prostaglandins and NO are known proangiogenic factors.10-13 COX-2 enhances basic fibroblast growth factor–stimulated angiogenesis in rat sponge granuloma.48 WIN 55,212-2, a CB1/CB2 agonist, decreased neoangiogenesis in granuloma tissue and reduced COX-2 expression.49 On the other hand, CB-receptor activation interacts with NO production by either facilitating50 or inhibiting51 NO production. The exact nature of the interactions, which may be receptor and tissue related, remain to be defined. Nevertheless, the current finding is consistent with the previous study demonstrating that NO-synthesis inhibition ameliorates portosystemic shunting in PH rats.21 Furthermore, because NO and prostacyclin are main vasodilators mediating splanchnic hyperemia and hyperdynamic circulation,8, 52 the decrease of SMA flow and PP by JWH-015 may as well be related to NOS and COX down-regulation.
The CB1 agonist, ACEA, did not influence portosystemic shunting, mesenteric vascular density, and mesenteric angiogenic factors expression, except that ACEA-treated rats had lower MAP. The systemic hypotension elicited by ACEA may be related to CB1-receptor activation and is supported by the previous report that systemic hypotension in rats with biliary cirrhosis could be improved by the CB1-receptor antagonist, SR141716A.15
BDL is a well-established model characterized by cholestatic and hepatocellular injury with inflammation, fibrosis, and cirrhosis.53 We treated BDL rats because at the end of week 5 after surgery, a marked portosystemic collateral vascular bed had been established.53 Our finding suggests that patients may benefit from the treatment even at a late stage of liver injury with full-blown portosystemic shunt formation. Furthermore, the use of CBs has been limited because of the unwanted psychotropic effects: The CB1 receptor is mostly expressed in the brain and is responsible for CB psychoactivity,4 whereas the CB2 receptor is distributed “peripherally” and is unrelated to CB psychoactivity.5 CB2 agonist administration therefore seems appropriate because of the lack of undesirable CB1-mediated psychotropic side effects.
In conclusion, a CB2 agonist alleviates PH, severity of mesenteric and hepatic angiogenesis, portosystemic shunting, and hepatic fibrosis in cirrhotic rats. The vascular effects may be mediated, at least in part, by COX and NOS down-regulation. CB receptors may be targeted in the control of mesenteric hyperemia and portosystemic collaterals in cirrhosis.
The authors gratefully acknowledge Yi-Chou Chen and Shao-Jung Hsu for their excellent technical assistance.