Role of β3-adrenoceptors for intrahepatic resistance and portal hypertension in liver cirrhosis

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

Increased intrahepatic resistance and splanchnic blood flow cause portal hypertension in liver cirrhosis. Nonselective β-adrenoceptor (β-AR) antagonists have beneficial effects on hyperdynamic circulation and are in clinical use. In this context, the role of the β3-AR is undefined. Here we investigated their expression and role in portal hypertension in patients and rats with liver cirrhosis. We analyzed cirrhotic human and rat tissues (liver, splanchnic vessels) and primary rat cells. Protein expression of β3-AR was determined by western blot and messenger RNA (mRNA) levels by reverse-transcription polymerase chain reaction (RT-PCR). Activities of Rho-kinase and the nitric oxide (NO) effector protein kinase G (PKG) were assessed by way of substrate phosphorylation (moesin, vasodilator-stimulated phosphoprotein [VASP]). Cyclic 3′,5′ adenosine monophosphate (cAMP) accumulation was determined by an enzyme-immunoassay kit. The effects of selective β3-AR agonists (CGP12177A, BRL37344) and antagonist (SR59230A) were investigated by collagen matrix contraction of hepatic stellate cells (HSCs), in situ liver perfusions, and in vivo hemodynamic parameters in bile duct ligation and carbon tetrachloride intoxication in cirrhotic rats. In cirrhosis of humans and rats, β3-AR expression is markedly increased in hepatic and in splanchnic tissues. Stimulation of β3-AR leads to relaxation of HSCs by way of cAMP accumulation, and by inhibition of Rho-kinase activity; any role of NO and its effector PKG was not observed. β3-AR agonists decrease intrahepatic resistance and portal pressure in cirrhotic rats. Conclusion: There is a marked hepatic and mesenteric up-regulation of β3-ARs in human cirrhosis and in two different animal models of cirrhosis. The β3-AR-agonists should be further evaluated for therapy of portal hypertension. (HEPATOLOGY 2009.)

Increased intrahepatic resistance to portal flow1 and decreased splanchnic vascular resistance2 cause portal hypertension in liver cirrhosis. Dysregulation of intracellular vasoconstrictive Rho-kinase-mediated pathways may contribute to both processes. According to our findings, it is overactivated in the liver and dysfunctional in splanchnic vessels.2–7

Reduction of extrahepatic vascular resistance leads to hyperdynamic circulation characterized by increased cardiac output and decreased systemic vascular resistance and mean arterial pressure. In an attempt to maintain effective circulating volume, endogenous vasoconstrictor systems are recruited, including catecholamines.8, 9 The effects of catecholamines are characteristically mediated by the β1- and β2-adrenoceptor subtypes.10 β3-Adrenoceptors (β3-ARs) are predominantly found in adipocytes, where they are involved in the regulation of lipolysis.11–13 They are also expressed in the myocardium14–16 and in endothelial cells as well as in vascular and gastrointestinal smooth muscle cells.17–21 In contrast to β1- and β2-ARs, the β3-AR generally requires higher levels of catecholamines for stimulation.22 Further, β3-ARs are relatively resistant to desensitization, due to the absence of phosphorylation sites for protein kinase A and G-protein-coupled receptor kinases (GRK).23 The expression of β3-AR is regulated by transcription of messenger RNA (mRNA).24, 25 In blood vessels, β3-AR stimulation leads to slow vasorelaxation26–28 by way of Gs-mediated cyclic 3′,5′ adenosine monophosphate (cAMP),29–31 NO-induced cGMP,19, 32 or inhibition of RhoA/Rho-kinase.33

Thus, it is of interest to study whether β3-ARs take part in the well-described intra- and extrahepatic hemodynamic changes of liver cirrhosis. Here we investigated the effect of the β3-AR agonist, CGP12177A (CGP), and β3-AR antagonist, SR59230A (SR), on systemic and portal hemodynamics in two cirrhotic rat models. In addition, we studied the influence of β3-AR agonists on activated hepatic stellate cells. Further, we assessed β3-AR expression in specimens of livers and hepatic arteries from patients with liver cirrhosis.

Abbreviations

β-AR, β adrenergic receptor; BDL, bile duct ligation; BRL, BRL37344, β3-AR agonist; cAMP, cyclic 3′,5′ adenosine monophosphate; CCl4, carbon tetrachloride intoxication; CGP, CGP12177A, β3-AR agonist; CT-value, number of PCR cycles; HSC, hepatic stellate cells; IBMX, 3-isobutyl-1-methylxanthine; L-NAME, Nω-nitro-L-arginine methyl ester; MAP, mean arterial pressure; MLCP, myosin light chain phosphatase; NO, nitric oxide; NOS, nitric oxide synthase; PKG, protein kinase G; PP, portal pressure; RT-PCR, reverse transcriptase polymerase chain reaction; SEM, standard error of the mean; SR, SR59230A, β3-AR antagonist; VASP, vasodilator-stimulated phosphoprotein.

Materials and Methods

Patients.

Human hepatic arteries and livers from cirrhotic and noncirrhotic patients were obtained (Supporting Material). The study was approved by the local ethics committee (Ethical Committee of University of Bonn 202/01).

Animals.

We used cirrhosis models of bile duct ligation (BDL) and carbon tetrachloride (CCl4) intoxication in Sprague-Dawley rats, as described35–38 (Supporting Material). The local committee for animal studies approved the study (LANUV NRW, 9.93.2.10.35.07.035).

β3-AR Agonists and Antagonists.

Different β3-AR modulating compounds were used35–38 (Supporting Material). All compounds were provided commercially by Sigma-Aldrich (Taufkirchen, Germany).

Western Blotting and Antibodies.

Samples of shock-frozen livers, human hepatic, and rat mesenteric arteries were subjected to western blot analysis as described.3 Ponceau S staining was performed to ensure equal protein loading.3 For detection of β3-AR we used different primary antibodies (for details, please see Supporting Material and Supporting Fig. 1).33

Assessment of Rho-Kinase and Protein Kinase G Activities.

Livers and mesenteric arteries of cirrhotic BDL rats were harvested after injection of CGP12177A or solvent and homogenized. In these homogenates, Rho-kinase activity was assessed as phosphorylation of the endogenous Rho-kinase substrate, moesin, at Thr-558.3, 4 Protein kinase G (PKG) activity was assessed as phosphorylation of the endogenous PKG substrate, vasodilator-stimulated phosphoprotein (VASP), at Ser-239. The phosphorylation state of VASP serves as a marker for PKG activity.3, 4 This was determined by western blot analysis using site- and phospho-specific antibodies (see Western Blotting and Antibodies, above).

Cell Isolation and Culture.

Hepatic stellate cells were isolated by an in situ liver perfusion of healthy rats as described3, 39 (Supporting Material).

Rat hepatocytes were isolated as described40 with small modifications (Supporting Material).

Quantitative Reverse-Transcription Polymerase Chain Reaction (RT-PCR).

RNA isolation, reverse transcription with 0.5 μg total RNA, and detection by RT-PCR were performed as described (Supporting Material).3 Primers and probes for RT-PCR were obtained as a ready-to-use mix (β1-AR: Rn00824536_s1 Adrb1; β2-AR: Rn00560650_s1 Adrb2; β3-AR: ABI4331348, beta3ar-begi, GeneX AbD, Applied Biosystems, Foster City, CA). The ΔCT-method was used for quantification of the results.

Three-dimensional Stress-Relaxed Collagen Lattice Contraction Mode.

The ability of HSC to contract 3D collagen matrices was measured in hydrated collagen gels, as described3, 39 (Supporting Material). The β3-AR agonists CGP12177A and BRL37344 in different concentrations (10−8 to 10−5M) were added in the absence and presence of prazosin (10−8M). Control cell-free gels provided estimates for the precontraction volume and determination of relative changes in volume (percent contraction). All data are from experiments using at least three sets of three collagen lattices using culture-activated HSCs from three different rat HSC isolations.

Assessment of cAMP Accumulation.

In all experiments, 250,000 HSCs were incubated with the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (IBMX) (5 × 10−4M), without (controls) and with CGP12177A (10−6M) or BRL34377 (10−6M) for 1 hour. IBMX was added in order to inhibit cAMP degradation. cAMP was determined using an enzyme-immunoassay kit (CA-201) (Sigma-Aldrich-Chemie, Munich, Germany) according to the manufacturer's guidelines. All data are from experiments using at least three sets of culture-activated HSCs.

In Situ Liver Perfusion.

In situ liver perfusion was performed as described3, 41 (Supporting Material).

Effect of the β3-AR Agonist CGP12177A on Hepatic Vascular Resistance.

After preconstriction with the α1-adrenoceptor agonist methoxamine (10−4M), cumulative concentration-response curves with CGP12177A (2.5-10 mg/L) were obtained as described in the Supporting Material.

Role of NO in the β3-AR Agonist-Mediated Effect.

In order to investigate the role of NO in the β3-AR-mediated effect, the nonselective nitric oxide synthase (NOS) inhibitor Nω-nitro-L-arginine methyl ester (L-NAME) (1 mM) was applied prior to methoxamine in liver perfusion of BDL rats, as described.3

Hemodynamic Studies.

Hemodynamic studies were performed under ketamine anesthesia (60 mg/kg i.m.) as described3, 4, 42 (Supporting Material).

Microsphere Technique.

Cardiac output was measured using the colored microsphere method as described3, 4, 43 (Supporting Material).

Acute Experiments.

After stabilization, portal pressure (PP) and mean arterial pressure (MAP) were monitored continuously for 20 minutes, followed by application of a microsphere technique using yellow microspheres. CGP12177A (0.5 mg/kg) or SR59230A (0.5 mg/kg) were administered intravenously after stabilization, and PP and MAP were monitored continuously for a further 30 minutes, followed by application of the microsphere technique using violet microspheres. Blue microspheres were injected in a small ileocoelic vein for shunt volume measurement.

The same experiments were also performed using animals that were orally treated with propranolol (1 mg/kg/d) for 2 days.

Statistical Analysis.

Data are presented as means ± standard error of the mean (SEM). Analysis of variance (ANOVA) followed by Mann-Whitney U tests were used for comparison between groups. P-values < 0.05 were considered statistically significant.

Results

Protein Expression of β3-AR in Human and Rat Livers and in Primary Rat Liver Cells.

Western blot analysis revealed a marked increase in hepatic protein expression of β3-AR in cirrhotic human livers compared to noncirrhotic livers (Fig. 1A). Hepatic β3-AR protein expression was also significantly increased in both groups of cirrhotic rats (particularly in BDL rats) when compared to their respective controls (Fig. 1B,C). Expression of β3-AR was significantly increased in fully transdifferentiated HSC after 14 days compared to hepatocytes and HSC after 7 days (Fig. 1F).

Figure 1.

Hepatic protein expression of β3-AR in human liver samples (A) and in cirrhotic BDL (B) and CCl4 (C) rats compared to their respective controls as determined by western blot analysis. mRNA expression of β3-AR in liver samples of cirrhotic BDL and CCl4 rats compared to their respective controls (D) as well as in primary rat hepatocytes and primary rat HSC after 7 days and 14 days of culture activation (E) as determined by quantitative RT-RCR and corrected to 18SrRNA as a housekeeping gene. Protein expression of β3-AR in primary rat hepatocytes and primary rat HSC after 7 days and 14 days of culture activation (F) determined by western blot analysis. For liver samples minimum was n = 6/group (*P < 0.01 versus noncirrhotic control) and for cells n = 5 isolations/group (*P < 0.03 versus hepatocytes, #P < 0.001 versus HSC 7 days). In western blot analysis are shown relative densitometric quantifications of all experiments with values of controls set to 100 d.u. (means ± SEM) and representative western blots.

mRNA Levels of β3-AR in Rat Livers and Primary Cells.

In the livers of control rats, β3-AR mRNA was undetectable. In contrast, in livers of cirrhotic rats β3-AR mRNA levels were increased (Fig. 1D). Cellular analysis comparing hepatocytes, 7 days and 14 days cultured HSC, showed that the lowest mRNA concentrations of β3-AR were found in hepatocytes. mRNA concentrations of β3-AR increased progressively with activation and transdifferentiation of HSC (Fig. 1E).

Hepatic Expression of β1-AR and β2-AR.

The hepatic protein expression of β1-AR was significantly decreased in liver cirrhosis compared to noncirrhotic controls (Fig. 2A). This finding was paralleled by a decrease in β1-AR protein expression and mRNA levels in transdifferentiated HSC (Fig. 2B,C). By contrast, β2-AR expression was significantly increased in cirrhotic livers and HSC when compared to controls and hepatocytes (Fig. 2F). In contrast to β3-AR, expression of β2-AR remained unchanged in HSC during ongoing activation (Fig. 2E).

Figure 2.

Hepatic protein expression of β1-AR in cirrhotic BDL and CCl4 rats (A) and in primary rat hepatocytes and primary rat HSC after 7 days and 14 days of culture activation (B) determined by western blot analysis. mRNA expression of β1-AR in primary rat hepatocytes and primary rat HSC after 7 days and 14 days of culture activation (C) as determined by quantitative RT-RCR relative to 18SrRNA as a housekeeping gene. Hepatic protein expression of β2-AR in cirrhotic BDL and CCl4 (D) rats and in primary rat hepatocytes and primary rat HSC after 7 days and 14 days of culture activation (E) determined by western blot analysis. mRNA expression of β2-AR in liver samples of cirrhotic BDL and CCl4 rats compared to their respective controls (F). For liver samples the minimum was n = 6/group (*P < 0.01 versus noncirrhotic controls) and for cells n = 5 isolations/group (*P < 0.03 versus hepatocytes, #P < 0.001 versus HSC 7 days). In western blot analysis are shown relative densitometric quantifications of all experiments with values of controls set to 100 d.u. (means ± SEM) and representative western blots.

β-AR Protein Expression in Human Hepatic Arteries and Rat Mesenteric Arteries.

Similar to the hepatic expression pattern, the β3-AR protein expression was increased in human hepatic arteries from cirrhotic patients compared to vessels from noncirrhotic patients (Fig. 3A).

Figure 3.

Protein expression of β3-AR in human hepatic arteries (A) and in mesenteric arteries of cirrhotic BDL (B) and CCl4 (C) rats compared to their respective controls as determined by western blot analysis. Protein expression of β1-AR (D) and of β2-AR (E) in mesenteric arteries of cirrhotic BDL and CCl4 rats compared to noncirrhotic controls as determined by western blot analysis. Shown are relative densitometric quantifications (means ± SEM) of all experiments with values of controls set to 100 d.u. and representative western blots (at least n = 6/group; *P < 0.01 versus noncirrhotic controls).

In rats, β3-AR expression differed between the animal models of cirrhosis. Western blot analysis revealed no significant difference in protein expression of β3-AR between mesenteric arteries from CCl4 and control rats, although there was a trend toward increased expression in CCl4 rats (Fig. 3C). In contrast, β3-AR protein expression was significantly increased in mesenteric arteries of BDL rats when compared to vessels from sham-operated rats (Fig. 3B).

In mesenteric arteries of cirrhotic BDL rats both β1-AR and (particularly) β2-AR protein expression was significantly increased (Fig. 3D,E). In CCl4 cirrhotic rats, mesenteric β2-AR expression was also increased (Fig. 3E), whereas β1-AR protein expression did not significantly differ from control (Fig. 3D).

Activity of Rho-Kinase and PKG in Cirrhotic Livers and Mesenteric Arteries: Modulation by the β3-AR Agonist In Vivo.

Activity of Rho-kinase, which is at least in part responsible for the increased intrahepatic resistance in cirrhosis, was assessed as phosphorylation of the dividing Rho-kinase substrate moesin, at Thr-558.3, 7 Hepatic Rho-kinase activity in BDL rats decreased 4-fold after stimulation of β3-AR in vivo (Fig. 4A). PKG activity, mediating the vasodilatatory effect of NO, was assessed as phosphorylation of the endogenous PKG substrate, VASP, at Ser-239.3, 4 In contrast to Rho-kinase activity, hepatic PKG activity in BDL rats remained unchanged after β3-AR stimulation in vivo by CGP12177A (Fig. 4B).

Figure 4.

In vivo effects of β3-AR agonist CGP12177A (0.5 mg/kg i.v.) or solvent in cirrhotic BDL rats. Protein expression of phospho-moesin and phospho-VASP in the livers (A,B) and mesenteric arteries (C,D) of these rats was determined by western blot analysis using a phospho- and site-specific antibody. Hepatic protein expression of Gαs (E) as well as of Gαi (F) of cirrhotic BDL and CCl4 rats compared to noncirrhotic controls as determined by western blot analysis. Shown are relative densitometric quantifications (means ± SEM) of all experiments with values of controls set to 100 d.u. and representative western blots (minimum was n = 5/group; *P < 0.01 versus noncirrhotic control; #P < 0.01 versus cirrhotic controls).

In previous studies we could show that decreased Rho-kinase activity and increased PKG activity might be responsible for splanchnic hypocontractility and vasodilation in cirrhotic patients and rats.4, 5 Application of the β3-AR agonist CGP12177A did not change the diminished mesenteric Rho-kinase activity of BDL rats further (Fig. 4C). Similarly, PKG activity remained unchanged as well (Fig. 4D).

s and Gαi Protein Expression in Rat Liver.

Intracellular signaling of the β3-AR is mediated through Gαs and might be negatively modulated by Gαi proteins. Therefore, we analyzed the expression of these proteins in cirrhosis. Western blot analysis revealed a significant increase in hepatic protein expression of Gαs in livers of both cirrhosis models compared to controls (Fig. 4E). Gαi expression remained unchanged in cirrhotic livers when compared to controls, however (Fig. 4F).

Three-dimensional Stress-Relaxed Collagen Lattice Contraction Model.

Isolated and culture-activated HSC were loaded in collagen lattices as described.3, 39 Incubation with the β3-AR agonist CGP12177A dose-dependently relaxed the contraction of HSC. The highest concentration elicited the strongest relaxation (Fig. 5A).

Figure 5.

Effect of CGP12177A (A), BRL34377 in the absence and presence of prazosin (B), as well as CGP12177A in the absence and presence of SR59230A (C) on fetal bovine serum-induced HSC contraction in a 3D-stress-relaxed collagen lattice model. cAMP accumulation in HSC after incubation with IBMX alone or together with the β3-AR agonists CGP12177A and BRL34377 (D) as determined using an enzyme immunoassay kit according to the manufacturer's guidelines. All data are from experiments using at least three sets of three collagen lattices with culture-activated HSCs from three healthy rat HSC isolations. Comparisons between two groups were carried out by Mann-Whitney U tests (*P < 0.01 versus positive controls with fetal bovine serum or with IBMX alone; #P < 0.05). (E) The proposed mechanisms involved in β3-AR mediated relaxation.

A different agonist, BRL37344, likewise elicited relaxation of isolated HSC. As shown by coincubation with prazosin, this was α1-AR independent (Fig. 5B).

In order to assess the specificity of CGP12177A, we tested the effect of the β3-AR antagonist SR59230A on CGP12177A-induced HSC relaxation. SR59230A indeed abrogated the relaxation by CGP12177A (Fig. 5C).

cAMP Accumulation in Activated HSC after β3-AR Agonist Incubation.

Incubation of activated HSC with the two β3-AR agonists elicited a significant increase in cAMP accumulation (Fig. 5D). These data strongly suggest the direct involvement of Gαs in causing relaxation of HSC by the β3-AR agonists (Fig. 5A,B). Summarizing the effects of β3-AR agonists in HSC and hepatic tissue in cirrhosis, it is likely that β3-AR-mediated relaxation of HSC is induced by cAMP, directly by lowering Ca2+-entry and Ca2+-mobilization,44 and indirectly by inhibiting the (enhanced) Rho-kinase activity (Fig. 5E).

Isolated Perfused Livers.

Basal perfusion pressure in cirrhotic CCl4 and BDL rats was significantly increased compared to control rats (Fig. 6A). To investigate the effect of β3-AR agonist on hepatic resistance of cirrhotic rats, we studied the dose-dependent response of hepatic resistance to β3-AR agonist CGP12177A after preconstriction with the α1-AR agonist methoxamine. Addition of methoxamine to the perfusate elicited a significant increase in perfusion pressure by increasing hepatic resistance in control and cirrhotic rats, which reached significantly higher levels in cirrhotic rats (Fig. 6A). CGP12177A dose-dependently decreased the hepatic resistance in cirrhotic rat livers (Fig. 6C,D). In livers of noncirrhotic control rats, no effect was observed after administration of CGP12177A (Fig. 6B).

Figure 6.

Basal perfusion pressure and its increase after methoxamine precontraction in controls, as well as in cirrhotic CCl4 and in BDL rats (A). Dose-dependent response to CGP12177A of hepatic resistance after methoxamine precontraction in in situ liver perfusion in controls (B), in cirrhotic CCl4 (C), and in BDL rats with and without concomitant L-NAME (D). The Mann-Whitney U test was used for comparisons between two groups (minimum n = 5/group; *P < 0.05 versus solvent).

In order to investigate a putative role of NO in the β3-AR-mediated intrahepatic vasodilation, the NOS inhibitor L-NAME was administered prior to precontraction, as described.3 The effect of CGP12177A on intrahepatic resistance was not significantly modulated by L-NAME (Fig. 6D). Following L-NAME administration a tendency toward increased resistance was observed, which might not be related to CGP12177A-mediated stimulation of β3-AR, but to inhibition of constitutive eNOS activity in the cirrhotic liver (Fig. 6D).

Hemodynamic Effect of β3-AR Modulation In Vivo.

As expected, cirrhotic BDL and CCl4 rats showed the hemodynamic characteristics of portal hypertension when compared to the respective control rats. In cirrhotic rats, PP was markedly increased as a consequence of decreased splanchnic vascular resistance, as well as increased hepatic vascular resistance (Fig. 7). Intravenous application of the β3-AR agonist CGP12177A significantly decreased PP in cirrhotic BDL as well as CCl4 rats (Fig. 7A,B). Surprisingly, this was not accompanied by significant changes in splanchnic vascular resistance (Fig. 7C,D). Thus, this PP-lowering effect of CGP12177A was the result of a reduction of the highly elevated hepatic vascular resistance in cirrhotic BDL and CCl4 rats in vivo (Fig. 7E,F).

Figure 7.

PP (A), splanchnic vascular resistance (C), and hepatic vascular resistance (E) in sham-operated and cirrhotic BDL rats before and after administration of the β3-AR agonist CGP12177A and of the β3-AR antagonist SR59230A. PP (B), splanchnic vascular resistance (D), and hepatic vascular resistance (F) in controls and cirrhotic CCl4 rats before and after administration of CGP12177A and SR59230A. Comparisons between the groups were carried out by Mann-Whitney U tests.

In contrast, intravenous administration of the β3-AR-antagonist SR59230A did not affect PP (Fig. 7A,B), splanchnic vascular resistance, or hepatic vascular resistance of cirrhotic rats (Fig. 7C-F).

As expected, cirrhotic rats showed a fully established hyperdynamic dysfunction, i.e., arterial hypotension, decreased systemic vascular resistance, and increased cardiac output (Fig. 8A -F). Administration of the β3-AR agonist CGP12177A elicited no change in systemic hemodynamics in vivo (Fig. 8A-F). The β3-AR antagonist SR59230A significantly increased systemic vascular resistance in cirrhotic rats of both models (Fig. 8E,F), but significantly decreased the cardiac output only of cirrhotic BDL rats (Fig. 8C). In cirrhotic CCl4 rats the cardiac output showed a tendency to decrease (Fig. 8D). Similarly, MAP showed an insignificant increase (Fig. 8A,B). Shunt flow and heart rate were not influenced by β3-AR modulation (data not shown).

Figure 8.

MAP (A), cardiac output (C), and systemic vascular resistance (E) in sham-operated and cirrhotic BDL rats before and after administration of the β3-AR agonist CGP12177A and of the β3-AR antagonist SR59230A. MAP (B), cardiac output (D), and systemic vascular resistance (F) in controls and cirrhotic CCl4 rats before and after administration of CGP12177A and SR59230A. Comparisons between the groups were carried out by Mann-Whitney U tests.

Hemodynamic Effect of β3-AR Agonism after previous unselective β-AR Blockade by Propranolol.

Importantly, oral administration of propranolol for 2 days prior to the experiments resulted in a significant decrease of the heart rate in BDL rats compared to untreated BDL rats, indicating β1-AR blockade by propranolol (Fig. 9A). This was further confirmed by decreased cardiac output after propranolol treatment of cirrhotic BDL rats (Fig. 9C). In BDL rats, propranolol treatment lowered PP by increasing splanchnic vascular resistance (Fig. 9E,F), confirming β2-AR blockade by propranolol. Nevertheless, MAP and systemic vascular resistance remained unchanged (Fig. 9B,D).

Figure 9.

Heart rate (A), MAP (B), cardiac output (C), systemic vascular resistance (D), PP (E), splanchnic vascular resistance (F), and hepatic vascular resistance (G) in untreated and propranolol-treated cirrhotic BDL rats before and after administration of the β3-AR agonist CGP12177A or of the β3-AR antagonist SR59230A. Comparisons between the groups were carried out by Mann-Whitney U tests.

In addition to the propranolol-induced decrease in PP, β3-AR agonism by CGP12177A induced a further decrease of PP in propranolol-treated cirrhotic BDL rats (Fig. 9E). This was due to marked intrahepatic vasodilation (Fig. 9G), accompanied by splanchnic vasodilation (Fig. 9F). Additional intravenous administration of the β3-AR-antagonist SR59230A did not affect PP, splanchnic vascular resistance, or hepatic vascular resistance of cirrhotic rats (Fig. 9E-G).

Systemic hemodynamics (heart rate, MAP, cardiac output, and systemic vascular resistance) remained unchanged after administration of CGP12177A in propranolol-treated cirrhotic BDL rats (Fig. 9A-D). In contrast, the β3-AR-antagonist SR59230A caused an isolated significant increase in systemic vascular resistance, without leading to an additional decrease in cardiac output, which was already lowered by propranolol (Fig. 9C,D). Heart rate and MAP remained unaffected by SR59230A (Fig. 9A,B).

Discussion

Our study showed for the first time a possible role of the β3-AR in modulation of the increased intrahepatic resistance and portal pressure in cirrhosis. β3-ARs are up-regulated (protein and mRNA) in activated hepatic stellate cells as well as in human and rat cirrhotic livers. Interestingly, the stimulation of β3-AR relaxed activated hepatic stellate cells in vitro. In vivo, this was paralleled by a decrease of intrahepatic resistance and subsequent reduction of portal pressure in both rat models.

Hepatic stellate cells play a crucial role in the development of liver cirrhosis.45, 46 Any chronic liver injury leads to activation and subsequent transdifferentiation of quiescent HSC to highly proliferative, myofibroblast-like cells.45, 46 This transition in phenotype is responsible for fibrogenesis and increased contractility of HSC.47 The numeral increase of activated HSC due to their excessive proliferation in cirrhosis makes these changes significant for the whole organ.45, 46 The hepatic protein content of β3-AR was increased in human and experimental cirrhosis when compared to noncirrhotic controls (Fig. 1A-C). Hepatic up-regulation of β3-AR in cirrhosis coincided with increased mRNA levels (Fig. 1D), suggesting transcriptional regulation, a well-known mechanism for these receptors in other tissues.24, 25 In primary cells, we showed that this up-regulation in cirrhotic livers takes place in activated HSC to an important extent (Fig. 1E,F).

Despite mechanical obstruction/narrowing of intrahepatic blood vessels through collagen production by transdifferentiated HSC, these key cells are still involved in the regulation of intrahepatic vascular resistance.46 During activation, HSC acquire contractility, which contributes to increased intrahepatic vascular resistance.46, 47 Besides decreased intrahepatic production of vasodilators, increased intrahepatic resistance in liver cirrhosis is partly a consequence of the hyperresponsiveness of activated HSC to vasoconstrictors, e.g., catecholamines.3, 8, 9 This hyperresponsiveness may at least be partially caused by an overactivated contractile RhoA/Rho-kinase pathway, particularly in activated HSC.3, 7

Based on the high expression in activated HSC, the functional dynamic role of the β3-AR was confirmed in vitro by collagen-matrix contraction experiments, which revealed a clear relaxation of HSC after application of β3-AR agonists (Fig. 5A,B). The earlier and stronger relaxation of HSC by CGP12177A compared to BRL37344 may be explained by higher β3-AR agonistic efficacy of CGP12177A.35, 48 These β3-AR agonists and other β-AR antagonists show weak binding and antagonistic effects at the α1-AR at high concentrations.36, 38, 49 However, in our experiments the β3-AR agonist BRL37344 relaxed HSC irrespective of the presence of the α1-AR antagonist prazosin, demonstrating the β3-AR-specificity of the relaxation by this compound (Fig. 5B). Hence, this experiment endorses the α1-AR-independent effect of β3-AR agonists in HSC. Finally, unequivocal evidence for the β3-AR-specificity of these compounds was obtained with the β3-AR antagonist SR59230A (Fig. 5C). SR59230A abolished the relaxing effect of CGP12177A in HSC (Fig. 5C).

Like in heart failure,15, 50 profound changes in expression of β-ARs occur in liver cirrhosis. Both, in cirrhosis and in activated HSC β1-AR expression (protein and mRNA) is decreased (Fig. 2A-C). These findings confirm previous data, which showed expression of β1-AR in freshly isolated HSC51 and almost no expression in myofibroblastic HSC (after the second serial passage).52 β2-AR expression, by contrast, was increased in cirrhotic livers, which is probably due to an increased number of HSC, because the protein expression did not change during ongoing activation of HSC (Fig. 2D-F), as described in previous studies.51, 52

The β3-AR is stimulated by high catecholamine concentrations, which lead to slow relaxation of vascular smooth muscle cells, a mechanism relatively resistant to desensitization.23 In cirrhosis, the increased endogenous catecholamine concentrations lead to rapid vasocontraction through α1-AR stimulation overriding putative slow β3-AR-mediated vasodilation. Thus, we speculate that up-regulation of the β3-AR in cirrhotic liver and activated HSC may represent a—probably insufficient—mechanism to evade excessive contraction by way of α1-AR by endogenous catecholamines.

As previously described, the β3-AR stimulate adenylyl cyclase in different cells or tissues through activation of Gαs (Fig. 5D).29–31 Our experiments with activated HSC demonstrated enhanced cAMP accumulation following application of both β3-AR agonists, which led to relaxation of HSC, and a decrease of intrahepatic resistance. These effects are amplified by the up-regulation of Gαs, which was observed in cirrhotic livers, both from CCl4 and BDL rats. The β3-AR also mediated inhibition of hepatic Rho-kinase activity (Fig. 4A) using CGP12177A as agonist in BDL rats by dampening Rho-kinase-mediated contraction. Because in cirrhosis intrahepatic RhoA/Rho-kinase is up-regulated and activated,3, 7 this may represent a further protective counterregulation, against increased intrahepatic vascular resistance.

In situ liver perfusion experiments confirm our in vitro findings (Fig. 6). The strong β3-AR agonist CGP12177A significantly decreased intrahepatic resistance in cirrhotic rats, even at low concentrations (Fig. 6C,D). The explanation lies in the up-regulation of these receptors in the cirrhotic liver. By contrast, in noncirrhotic livers the effect of CGP12177A was minimal (Fig. 6B), which mirrors the low expression of β3-AR in noncirrhotic livers (Fig. 1). The decrease in intrahepatic resistance and portal pressure by the β3-AR agonist CGP12177A was seen in the two different rat models of cirrhosis (Fig. 7). These functional in vivo experiments support the suggestion that increased hepatic expression of β3-AR in human and experimental cirrhosis might be a target to decrease intrahepatic resistance.

Besides intrahepatic expression, the increased splanchnic vascular expression of β2-AR and β3-AR alter extrahepatic hemodynamics in both models of cirrhosis (Figs. 1, 3). Thus, our cirrhotic rats showed typical splanchnic and systemic vasodilation (Figs. 7C,D, 8E,F). Decreased splanchnic Rho-kinase activity—opposed to hepatic—as well as increased NO-mediated vasodilatatory effects contribute to this phenomenon.2, 4–6 The observed up-regulation of β3-AR in such vessels (Fig. 3) suggests an involvement of these receptors in the splanchnic vasodilation. Because the β3-AR are resistant to desensitization, their up-regulation, together with increased levels of circulating catecholamines, could be partially responsible for continuous splanchnic vasodilation in liver cirrhosis. Application of the β3-AR agonist CGP12177A alone had no further effect on splanchnic resistance (Fig. 7C,D). Thus, CGP12177A neither decreased the activity of Rho-kinase in mesenteric arteries, nor increased the activity of the NO effector PKG (Fig. 4C,D). These observations are not surprising, because in cirrhosis splanchnic resistance and Rho-kinase activity are already very low. Interestingly, the intrinsic vasodilatory effect of β3-AR stimulation by CGP12177A could be shown after blocking β2-AR with propranolol. This is due to the fact that CGP12177A also has antagonistic effects on β1- and β2-AR.35 This effect is shown in Fig. 9F: Propranolol increased splanchnic vascular resistance and subsequent administration of CGP12177A decreased it. These data also showed that propranolol did not possess significant β3-AR-blocking properties at the dosage given (Fig. 9).

In summary, our data demonstrate that pharmacological stimulation of β3-ARs by selective β3-AR agonists, in combination with nonselective β-AR blockers, results in a stronger beneficial effect on portal pressure in cirrhosis than nonselective β-AR blockers alone. By contrast, the effect of β3-AR is minor in the splanchnic vascular compartment.

In conclusion, β3-ARs are up-regulated in cirrhosis and its stimulation lowers intrahepatic resistance. They may represent a new target for the therapy of portal hypertension in cirrhosis.

Acknowledgements

The authors thank S. Dentler for critical reading, and G. Hack, C. Conrad, S. Bellinghausen, and D. Bammer for excellent technical support.

Ancillary