Carbon Monoxide-Mediated Activation of Large-Conductance Calcium-Activated Potassium

Transcription

1 JPET Fast This article Forward. has not Published been copyedited on and January formatted. 17, The final 2007 version as DOI: /jpet may differ from this version. JPET # page 1 Title page Carbon Monoxide-Mediated Activation of Large-Conductance Calcium-Activated Potassium Channels (BK Ca ) Contributes to Mesenteric Vasodilatation in Cirrhotic Rats. Massimo Bolognesi, David Sacerdoti, Anna Piva, Marco Di Pascoli, Francesca Zampieri, Santina Quarta, Roberto Motterlini, Paolo Angeli, Carlo Merkel, Angelo Gatta. Department of Clinical and Experimental Medicine, University of Padova, Italy (M.B., D.S., A.P., M.D.P., F.Z., S.Q., P.A., C.M., A.G.), Department of Surgical Research, Vascular Biology Unit, Northwick Park Institute for Medical Research, Harrow, Middlesex, UK (R.M.). Copyright 2007 by the American Society for Pharmacology and Experimental Therapeutics.

2 JPET # page 2 Running title page Running title: Mesenteric Vasodilatation in Cirrhosis Corresponding author: Massimo Bolognesi, MD, PhD, Clinica Medica 5, Dipartimento di Medicina Clinica e Sperimentale, Policlinico Universitario, Via Giustiniani 2, Padova Italy. Telephone: Fax: massimo.bolognesi@unipd.it Number of text pages: 31 Number of tables: 0 Number of figures: 8 Number of references: 39 Number of words in the Abstract: 202 Number of word in the Introduction: 528 Number of words in the Discussion: 1463 List of nonstandard abbreviations: ACh, acetylcholine; ANOVA, analysis of variance; BK Ca, largeconductance calcium-activated K+ channels; CO, carbon monoxide; CORM, carbon monoxide releasing molecule; COX, cyclooxygenase; CrMP, Chromium mesoporphyrin; EC 50, molar concentration of ACh causing 50% of the maximal vasorelaxant effect; EET, epoxyeicosatrienoic acids; 20-HETE, 20-hydroxyeicosatetraenoic acid; EDHF; endothelial derived hyperpolarizing factor; HO, heme oxygenase; IbTx, iberiotoxin; Indo, indomethacin; L-NAME, N G -nitro-larginine-methyl-ester; NO, nitric oxide; NOS, nitric oxide synthase; ODQ, 1H- [1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one; PE, phenylephrine; pec 50, log EC 50 ; PSS, physiological salt solution; R max, maximum relaxation of the artery; sgc, soluble guanylyl cyclase; SNP, sodium nitroprusside. Recommended section: Gastrointestinal, Hepatic, Pulmonary, and Renal

3 JPET # page 3 ABSTRACT Large-conductance calcium-activated potassium channels (BK Ca ) are important regulators of arterial tone and represent a mediator of the endogenous vasodilator carbon monoxide (CO). As an upregulation of heme oxygenase (HO)/CO system has been associated with mesenteric vasodilatation of cirrhosis, we analyzed the interactions of BK Ca and of HO/CO in the endotheliumdependent dilatation of mesenteric arteries in ascitic cirrhotic rats. In pressurized mesenteric arteries (diameter µm) of ascitic cirrhotic rats, we evaluated: a) the effect of inhibition of BK Ca, HO and guanylyl-cyclase on dilatation induced by acetylcholine and by exogenous CO; b) HO-1 and BK Ca subunits protein expression. Results: Inhibition of HO and of BK Ca reduced acetylcholineinduced vasodilatation more in cirrhotic than in control rats, while inhibition of guanylyl-cyclase had similar effect in the two groups. CO was more effective in cirrhotic than in control rats, and the effect was hindered by BK Ca inhibition. The expression of HO-1 and of BK Ca α subunit were higher in mesenteric arteries of cirrhotic rats compared to those of control animals, while the expression of BK Ca β1 subunit was lower. In conclusion, an overexpression of BK Ca α subunits, possibly due to HO upregulation with increased CO production, participates in the endotheliumdependent alterations and mesenteric arterial vasodilatation of ascitic cirrhotic rats.

4 JPET # page 4 INTRODUCTION Mesenteric arterial vasodilation is a key mechanism in the pathophysiology of the hyperdynamic circulatory syndrome of cirrhosis. This syndrome is responsible for serious complications, such as ascites, hepatorenal syndrome and gastrointestinal hemorrhage. The pathophysiological mechanism that supports the vasodilation of mesenteric arteries in cirrhosis is a decrease in the response of the arteries to vasoconstricting agents (Sieber et al., 1993), caused by an increase in vasodilating substances of endothelial origin, such as nitric oxide (NO), prostacyclin and, as recently demonstrated, carbon monoxide (CO) (Wiest and Groszmann, 1999; Fernandez et al., 2001; Gonzales-Abraldes et al., 2002). We have recently shown that an increased action of the heme oxygenase (HO)/CO system plays a role in the hyporesponsiveness of small resistance mesenteric arteries to phenylephrine (PE) only in the advanced stage of experimental cirrhosis (Bolognesi et al., 2005). Therefore, the increased activity of the HO/CO system may participate in the evolution of cirrhosis from compensated to decompensated. HO is a microsomal enzyme with two main distinct isoforms: the inducible isoenzyme HO-1 and the constitutive one HO-2 (Zhang et al., 2001; Johnson et al., 2003a). It is the rate-limiting enzyme in the degradation of heme to biliverdin, CO and free iron (Motterlini et al., 1998). CO, generated by HO in endothelial and smooth muscle layers of arterial vessels, modulates vascular tone by inducing relaxation of vascular smooth muscle cells through stimulation on soluble guanylyl cyclase (sgc) and opening of large-conductance calcium-activated potassium channels (BK Ca ) (Zhang et al., 2001). It also inhibits the formation of 20-hydroxyeicosatetraenoic acid (20-HETE) (Zhang et al., 2001; Kaide et al. 2004), a vasoconstrictor which inhibits potassium channels. BK Ca are by far the most abundant K + channel expressed in vascular smooth muscle cells; their importance as physiological regulators of blood flow has long been recognized (Clapp et al., 2003, Kotlikoff and Hall, 2003). The opening of these channels, in response to localized intracellular

5 JPET # page 5 Ca2+ transients (Ca2+ sparks) (Jaggar et al., 2000), leads to hyperpolarization of smooth muscle cell and, thus, to relaxation. Their conductance is increased by membrane depolarization (Barriere et al., 2001), NO (Barriere et al., 2001; Wu et al., 2002), CO (Wu et al., 2002, Jaggar et al., 2005) and by other substances, such as epoxyeicosatrienoic acids (EET)(Archer et al., 2003). In this study we tried to analyze if the alteration of HO/CO system detected in ascitic cirrhosis is linked to alterations in the mechanisms transducing the CO signal in the smooth muscle cell. Therefore, aim of the study was to analyze the role of BK Ca, of sgc and of HO/CO system in the response to Acetylcholine (ACh) of small resistance mesenteric arteries of rats with CCl 4 induced cirrhosis. In cirrhotic rats with ascites, we preliminarily analyzed the effect of Chromium mesoporphyrin (CrMP), a non selective HO inhibitor, on the endothelium-dependent vasodilation induced by ACh. Then, we examined the effect of inhibition of sgc and of BK Ca, alone and combined with CrMP, on ACh-induced vasodilation. Finally, the hemodynamic effect of a COdonor, alone and after the inhibition of HO, sgc and BK Ca, and the expression of HO-1 and of the two subunits of BK Ca (α and β1) in small resistance mesenteric arteries of cirrhotic rats were evaluated.

6 JPET # page 6 MATERIALS AND METHODS Animals. The study was performed on 65 adult male Wistar rats (Charles River Laboratories, Calco, Italy); body weight g. The experiments were carried out in accordance with the legislation of Italian authorities (D.L. 27/01/1992 no. 116), which complies with European Community guidelines (CEE Directive 86/609) for the care and use of experimental animals. The experimental protocol was approved by the Institutional Animal Care and Use Committee. Cirrhosis was induced with the CCl 4 inhalation method in 35 rats drinking phenobarbital (0.30 g/l in the drinking water) as previously described (Bolognesi et al., 2005). Treatment was followed for 16 weeks, and animals were free of treatment for the last week before the experiment. Under anesthesia with ketamine hydrochloride (100 mg/kg body wt intramuscularly), a midventral laparotomy was performed and a section of small intestine was removed. The presence of ascites was confirmed by visual examination at laparotomy. If the presence of ascites was not certain, the rat was not considered ascitic. After laparotomy, cirrhotic rats were classified as cirrhotic with or without ascites. Age-matched animals were used as controls. Isolated Microvessel Preparation. As the low splanchnic vascular resistance observed in portal hypertension depends mostly on mesenteric resistance arteries, which are pre-capillary arteries with diameters narrower than 500 µm (Mulvany and Aalkiaer, 1990), we studied the endothelial function, the response to CO and protein expression in small mesenteric resistance arteries. The no-flow model was chosen to avoid interference from shear stress. The clamped section of the small intestine was placed in a chilled oxygenated modified Krebs bicarbonate buffer (physiological salt solution: PSS) containing mm NaCl, 4.7 mm KCl, 1.2 mm KH 2 PO 4, 1.2 mm MgSO 4, 2.8 mm CaCl 2, 25 mm NaHCO 3 and 11 mm dextrose.

7 JPET # page 7 Third/fourth-order branches of the superior mesenteric artery ( µm in diameter, 1-2 mm in length), were isolated from surrounding perivascular tissue, removed from the mesenteric vascular bed and mounted on glass micropipettes in a water-jacketed perfusion chamber (Living Systems Instrumentation, Burlington, VT, USA) in warmed (37 C), oxygenated (95% O 2 and 5% CO 2 ) PSS. The vessels were mounted on a proximal micropipette connected to a pressure servo controller. Subsequently, the lumen of the vessel was flushed to remove residual blood and the end of the vessel was mounted on a micropipette connected to a three-way stopcock. After the stopcock was closed, the intraluminal pressure was allowed to increase slowly until it reached 80 mmhg. The vessel was superfused with PSS (4 ml/min) at 37 C gassed with 95% O 2 and 5% CO 2 for a 45 min period of equilibration (Bolognesi et al., 2005). Intraluminal pressure was maintained at 80 mmhg throughout the experiment. After the equilibration period, the vessels were challenged with PE, an alpha1-adrenoreceptor agonist (1 µm). An artery was considered unacceptable for experimentation if it demonstrated leaks or failed to constrict by more than 20% to PE. The presence of a functional endothelium was determined on the basis of a prompt relaxation to ACh (1 µm), in the vessel precontracted with PE (1 µm). To remove the endothelium, 2 ml of air were flushed through the lumen (Sun et al., 1994). In these arteries, the absence of a functional endothelium was confirmed after preconstriction with PE by the absence of response to ACh with a normal response to sodium nitroprusside (SNP), an endothelium-independent vasodilator. The effects of ACh and CO administration were evaluated as variations in the internal diameter of the vessels preconstricted with PE 10 µm; all responses were reported as percent inhibition of the contraction induced by PE. Evaluation of the response to ACh of small mesenteric arteries preconstricted with phenylephrine in CCl 4 cirrhotic rats. Responses to increasing doses of ACh (10-9 to 10-4 M) were determined in arteries superfused with PSS containing vehicles for the inhibitors tested. Inhibitors were added to freshly prepared PSS, and 20 to 30-min drug-tissue contact time was allowed before retesting the response to ACh in the same

8 JPET # page 8 vessel. ACh was added to the bath (extraluminal application), and cumulative dose-response curves were generated, with 2 to 3 min intervals between doses. After each dose-response test, the tissues were washed with fresh PSS for at least 20 min. Vascular diameters were measured 1 to 3 min after the addition of ACh with the use of a video system consisting of a microscope with a CCD television camera (Eclipse TS100-F, Nikon, Tokyo, Japan), a television monitor (Ultrak Inc., Lewisville, Tx, USA) and a video measuring system (Systems Instrumentation, Burlington, VT, USA). In control rats and in ascitic cirrhotic rats, concentration-response curves to ACh were evaluated: a) before and after 20 min superfusion with the HO inhibitor CrMP (15 µm). b) before and after 20 min superfusion with the sgc inhibitor 1H-[1,2,4]oxadiazolo[4,3- a]quinoxalin-1-one (ODQ) (10 µm) in arteries treated with indomethacin (Indo) (2.8 µm) and N G - nitro-l-arginine-methyl-ester (L-NAME) (1 mm); and then after a further 20 min of CrMP (15 µm) plus ODQ (10 µm) superfusion in arteries already evaluated after ODQ superfusion alone. c) before and after 20 min superfusion with the BK Ca inhibitor iberiotoxin (IbTx) (25 nm) in arteries treated with Indo (2.8 µm), L-NAME (1 mm) and ODQ (10 µm); and then after a further 20 min of CrMP (15 µm) plus IbTx (25 nm) superfusion in arteries already evaluated after IbTx superfusion alone. Concentration-response curves to ACh were also evaluated in small mesenteric arteries of control rats before and after 20 min superfusion with three different amount of the BK Ca inhibitor IbTx: 25 nm, 50 nm, 100 nm. Only one experiment was performed in each artery. Evaluation of the response to a CO donor. Small mesenteric arteries of control rats and of rats with cirrhosis and ascites were prepared and mounted on glass micropipettes in a water-jacketed perfusion chamber as already described. After the equilibration period and after verifying their viability, the arteries were preconstricted with the

9 JPET # page 9 addition of PE to the superfusion PSS. A CO donor (200 µm) (CORM-3 - CO releasing molecule (Clark et al., 2003; Foresti et al., 2004) was then added to the superfusion PSS for 15 minutes and the change in lumen diameter was recorded. Chemical composition of CORM-3 is: Tricarbonylchloro(glycinato)ruthenium(II) ([Ru(CO) 3 Cl(glycinato)]). The concentration of CORM-3 was chosen after testing the hemodynamic effect of four different CORM-3 concentrations (25 µm, 50 µm, 100 µm and 200 µm) (Foresti et al., 2004) in a preliminary experiment on small mesenteric arteries of control rats. Only with the higher dose (200 µm) we obtained a clear and measurable effect. Therefore, 200 µm was the minimum efficacious dose in our experimental system. In this preliminary experiment, we also verified that the hemodynamic effect of CORM-3 on small mesenteric arteries is slow, as it could only be detected a few (3 to 4) minutes after administration, and could be fully evaluated only after an exposure of minutes. For this reason, to evaluate CORM-3 effect in our experimental set, the changes in lumen diameter were recorded for 15 min, during a continuous exposure to CORM-3. Taking all this into account, we decided to test a single dose of CORM-3, using the minimum dose that was efficacious in our control animals: 200 µm. The effect of CORM-3 was also verified: a) in arteries superfused with CrMP (15 µm) for 30 minutes; b) in arteries superfused with CrMP (15 µm) plus ODQ (10 µm) for 30 minutes; c) in arteries superfused with CrMP (15 µm) plus IbTx (25 nm) for 30 minutes. In a second series of experiments, the effect of CORM-3 was verified in arteries preconstricted with PE after removal of the endothelium. In these arteries, after verifying CO effect, CORM-3 was removed by the superfusing PSS and, when a stable baseline was obtained, the NO donor SNP (100 nm to produce a detectible dilation) was added. When a stable baseline was observed, CORM-3 was added again to the superfusing PSS, and the dilating effect was measured again. Chemicals CrMP was obtained from Porphyrin Products (Logan, UT, USA). CORM-3 was synthesized by

10 JPET # page 10 Brian E. Mann (Clark et al., 2003) and stock solutions were prepared using deionized water. All other chemicals were obtained from Sigma Chemical (Saint Louis, MO, USA). PE and L-NAME were dissolved in deionized water and diluted with PSS. CrMP was dissolved in a solution of 50 mm NaCO 3. Western blot analysis of HO-1 and of subunits α and β1 of BK Ca protein expression in mesenteric arteries of CCl 4 cirrhotic rats. Standard techniques were used to evaluate protein expression. After removal of veins and adipose tissue, small mesenteric arteries (30-40 arteries with diameter <500 µm) were collected from each rat, snap frozen in liquid N 2 and stored at -80 C until analyzed. The vessels were homogenized in urea lysis buffer. Protein extracts were assayed for protein content using the BCA protein assay kit (Pierce, Rockford, IL, USA). Sodium dodecyl sulphate - polyacrilammide gel elettrophoresis (SDS- PAGE) and immunoblotting were performed on 50 µg of total protein extracts. HO-1 protein expression was detected using a monoclonal murine antibody against HO-1 (StressGen Biotechnologies Corp., Victoria, BC, Canada). Only HO-1, and not HO-2, was tested in the present study, because we had already demonstrated, in similar experimental condition, that among the two main HO isoforms (HO-1 and HO-2), only HO-1 is overexpressed in small mesenteric arteries of CCl 4 -induced cirrhotic rats (Bolognesi et al., 2005). The protein expression of the two subunits of BK Ca, β1- and α-subunits, were detected using polyclonal rabbit anti-slo β1 (KCNMB1) and anti-k Ca 1.1 (alpha subunit 1) (KCNMA1) antibody (Alomone Labs Ltd., Jerusalem, Israel). The secondary antibodies, anti-rabbit conjugated to horseradish peroxidase, were diluted 1:1,000 in Phosphate-buffered saline containing 2% non-fat dry milk. Antigenic detection was visualized by standard ECL-enhanced chemiluminescence (Amersham, Arlington Heights, IL, USA) with exposure to X-ray film. Control antigens (Alomone Labs Ltd., Jerusalem, Israel) were used as positive controls. Protein expression was determined by densitometric analysis using the VersaDoc Imaging System (Bio-Rad Laboratories, Hercules, CA,

11 JPET # page 11 USA). After stripping, the blots were assayed for β-actin content as standardization of sample loading. The quantitative densitometric values of each protein of interest were normalized to β-actin and displayed in histograms. Protein expression of HO-1 was evaluated in 4 control rats, in 6 cirrhotic rats with ascites and also in 4 cirrhotic rats without ascites. Protein expression of BK Ca subunits was evaluated in 5 control rats, in 4 cirrhotic rats with ascites and also in 4 cirrhotic rats without ascites. Data Analysis Data were expressed as mean±se. Vasorelaxant responses were expressed as percent inhibition of the contraction induced by PE. Concentration-response data derived from each vessel were fitted separately to a logistic function by nonlinear regression and EC 50 (molar concentration of ACh causing 50% of the maximal vasorelaxant effect) was calculated and expressed as log [M] (pec 50 ). From the same regression the maximum relaxation (R max ) was also calculated. Two-way ANOVA was used to compare dose-response curves from controls and treated groups. Other data were analyzed by one-way ANOVA or Student s t test for paired or unpaired observations when appropriate. The n values quoted indicate the number of experiments and animals used. The null hypothesis was rejected at p<0.05.

12 JPET # page 12 RESULTS All rats treated with CCl 4 included in the study had macronodular or micronodular cirrhosis. In 27 out of 35 cirrhotic rats the presence of ascites was confirmed by visual examination at laparotomy. The presence of ascites was dubious in two cirrhotic rats. These intermediate rats were classified as non ascitic, but they were not used for any experiment. Control rats had no appreciable alteration in liver appearance. At the time of the study no difference in body weight between cirrhotic (nonascitic rats: 570±24 g, ascitic rats: 548±11 g) and control rats (575±20 g) was observed. Evaluation of the response to ACh. Comparing dose-response curves to ACh in baseline condition, ascitic cirrhotic rats (n.6) showed a higher sensitivity to ACh in respect of control rats (p=0.036) (Fig. 1). Effect of HO inhibition: Inhibition of HO with CrMP caused a slight shift of the concentration-response curve to ACh in control rats (F=8.61, p<0.001, two-way ANOVA), EC 50 (Fig. 1). On the contrary, a marked rightward shift of the concentration-response curve to ACh was detected after CrMP in cirrhotic rats with ascites (F=4.38, p=0.0027, two-way ANOVA), with a significant increase in EC 50 (Fig. 1), The increase in pec 50 after CrMP was significantly higher in cirrhotic rats with ascites (p=0.03) (Fig. 1). Following CrMP, sensitivities of control and cirrhotic rats with ascites to ACh were the same (Fig. 1). Effect of sgc inhibition and of HO inhibition: After treatment with indo and L-NAME, mesenteric arteries of cirrhotic rats with ascites maintained a higher sensitivity to ACh in respect of control rats (p=0.05) (Fig. 2). In arteries treated with Indo

13 JPET # page 13 and L-NAME, the inhibition of sgc with ODQ provoked a shift of the dose-response curve to ACh, both in control rats (F=3.58, p=0.01, two way ANOVA) and in cirrhotic rats with ascites (F=8.09, p=0.0005, two-way ANOVA) (Fig. 2). The addition of CrMP to ODQ provoked a further decrease in the response to ACh, both in control rats (F=9.77, p<0.001, two way ANOVA) and in cirrhotic rats with ascites (F=15.64, p<0.001, two-way ANOVA) (Fig. 2). The effect of ODQ and of CrMP+ODQ on R max was slightly but not significantly greater in control rats than in cirrhotic rats with ascites (Fig. 2). Effect of BK Ca inhibition: Even after treatment with indo, L-NAME and ODQ, mesenteric arteries of cirrhotic rats with ascites maintained a higher sensitivity to ACh in respect of control rats (p=0.024) (Fig. 3). The addition of IbTx provoked a decrease in the response to ACh both in control rats (F=7.74, p=0.007, two-way ANOVA) and in cirrhotic rats with ascites (F=11.59, p=0.001, two way ANOVA) (Fig. 3). The effect of IbTx was more evident in ascitic cirrhotic rats than in control rats (the increase in pec 50 after IbTx was significantly higher in cirrhotic rats with ascites in respect of control rats: p=0.010, while the decrease in maximum relaxation was not different between the two groups: p: NS). The addition of CrMP to IbTx did not provoke a further decrease in the response to ACh both in control rats and in cirrhotic rats with ascites (p:ns, two-way ANOVA) (Fig. 3). Final sensitivity and maximum relaxation to ACh were similar in controls and in ascitic cirrhotic rats (Fig. 3). In control rats, 25 nm of IbTx did not change the response of mesenteric arteries to ACh, while a significant rightward shift of the concentration-response curve to ACh was evident after incubating the arteries with the higher concentrations of IbTx (50 nm and 100 nm) (F=6.77, p<0.001, two way ANOVA) (Fig. 4). Evaluation of the response to CORM-3 of small mesenteric arteries preconstricted with phenylephrine in CCl 4 cirrhotic rats.

14 JPET # page 14 The addition of CORM-3 to the superfusion caused a slight but not significant dilatation (5±3%) in mesenteric arteries of control rats (p: NS), while it caused a significant vasorelaxation (25±6%) in cirrhotic rats with ascites (p=0.05) (Fig. 5). Vasorelaxation was more evident in ascitic cirrhotic rats than in controls (p=0.012) (Fig. 5). In arteries treated with CrMP and preconstricted with PE, CORM-3 provoked a significantly greater vasorelaxation both in controls and in ascitic cirrhosis (p=0.016 and p=0.022, respectively), but in the latter group the vasorelaxation was more evident (p=0.036) (Fig. 5). In arteries pretreated with CrMP and ODQ and preconstricted with PE, CORM- 3 did not provoke any significant dilatation, both in control and in cirrhotic rats with ascites (Fig. 5). In arteries pretreated with CrMP and IbTx and preconstricted with PE, CORM-3 caused a significant dilatation in controls rats (14±3%, p=0.004), which was slightly but not significantly lower than the dilatation obtained in the arteries pretreated only with CrMP (14±3% vs. 20±3%, respectively, p: NS) (Fig. 5); on the contrary, in cirrhotic rats with ascites, pretreatment with CrMP and IbTx markedly reduced the dilator effect of CORM-3 in respect of the effect obtained after pretreatment with only CrMP (4±1% vs. 46±11, respectively, p=0.003) (Fig. 5). After endothelium removal, CORM-3 provoked a significant but slight dilatation both in controls (p=0.002) and in cirrhotic rats (p=0.056), but more evident in controls (p=0.015) (Fig. 6). After the addition of a low concentration of SNP, which caused per se a modest dilation (25%±8 vs 25%±5, p: NS, in controls and cirrhotic rats, respectively), the vasodilating effect of CORM-3 was more evident, both in controls (p=0.0003) and in cirrhotic rats (p=0.037), but remained more evident in control rats (p=0.0008) (Fig. 6). Western Blot analysis. In small mesenteric resistance arteries, HO-1 and BK Ca α-subunit protein expression was increased in rats with cirrhosis, particularly in those with ascites (Fig. 7, 8). On the contrary, BK Ca β1-subunit protein expression was significantly decreased both in cirrhotic rats with and without ascites (Fig. 8).

15 JPET # page 15 DISCUSSION This study demonstrates that in small mesenteric arteries of rats with CCl 4 induced cirrhosis, the response to ACh is increased and it is normalized by the inhibition of HO and BK Ca. CO was more effective in cirrhotic than in control rats, and the effect was hindered by BK Ca inhibition. In mesenteric arteries of cirrhotic rats there is an overexpression of the α-subunit of BK Ca, which, together with the increased expression of HO-1, and an increased production of CO, may cause the increased response to ACh. BK Ca channels are composed of the pore-forming α-subunit and of the auxiliary regulatory β- subunits that modulate channel gating (Tanaka et al., 2004). A change in the expression of BK Ca subunits has been reported in experimental arterial hypertension. Indeed, a decrease in β1-subunit expression has been reported in genetic (Amberg and Santana, 2003a) and in acquired Angiotensin II-induced hypertension (Amberg et al., 2003b), suggesting that it may contribute to the development of hypertension. Bratz et al. (2005) reported a decrease in the expression of BK Ca α- subunit in superior mesenteric arteries from rats made hypertensive with N(omega)-nitro-Larginine. Liu et al. (1997; 1998) reported an increase in the expression of the pore-forming α- subunit in aortas (Liu et al., 1997) and in cerebral arteries (Liu et al., 1998) of spontaneously hypertensive rats, and they suggested that such an increase may be a compensatory vasodilatory reaction in systemic hypertension (Liu et al., 1998). Our data demonstrate that an altered expression of BK Ca may participate in the exaggerated mesenteric vasodilatation of cirrhosis. BK Ca are stimulated by the CO locally produced by endothelial HO (Naik et al., 2003a), which is also overexpressed in experimental cirrhosis (Bolognesi et al., 2005). We showed that inhibition of HO was more effective on the endothelium-dependent vasodilatation in cirrhotic rats with ascites in respect of controls rats (Fig. 1). To analyze the mechanisms that could link HO to the ACh-induced mesenteric vasorelaxation, we studied the relationship between

16 JPET # page 16 HO and the effectors of CO on smooth muscle cells, i.e., scg and BK Ca (Zhang et al., 2001; Jaggar et al., 2002; Wu et al., 2002). In arteries treated with Indo and L-NAME, to inhibit any possible interfering action from cyclooxygenase (COX) and nitric oxide synthase (NOS) respectively, both the inhibition of sgc and, afterwards, of HO had a similar effect on ACh-induced vasorelaxation in control and in ascitic cirrhotic rats. Hence, the action of HO on smooth muscle cells is not exclusively due to the activation of sgc, and moreover, the action on sgc is not what differentiates the HO action in ascitic cirrhotic rats. We evaluated the effect of inhibition of BK Ca in mesenteric arteries pretreated with the inhibitors of COX, NOS and sgc. Sensitivity to ACh was increased in ascitic cirrhotic rats also after the inhibition of these three systems. The inhibition of BK Ca caused a decrease in the maximum relaxation to ACh, particularly in control rats. But it provoked a marked decrease in the sensitivity to ACh only in cirrhotic rats with ascites, suggesting a role of BK Ca in the altered vasoactive response that occurs in this pathological condition. Further inhibition of HO did not lead to any modification in the response to ACh, both in controls and in cirrhotic rats with ascites, excluding other mediators of HO. As CO is the vasoactive molecule produced by HO (Zhang et al., 2001), we investigated the response to exogenous CO in small mesenteric arteries of cirrhotic rats with ascites. We used CORM-3, a CO-releasing water-soluble molecule (Clark et al., 2003; Foresti et al., 2004). In mesenteric arteries of control rats the vasodilating response to CORM-3 was insignificant, while it was evident after elimination of endogenous CO by pre-treatment with the HO inhibitor (Sacerdoti et al., 2006) (Fig. 5). A similar trivial effect of exogenous CO in baseline conditions has been reported by Kozma et al. (1999), who reported that in first-order gracilis muscle arterioles, CO does not produce arteriolar dilatation unless the preparation is exposed previously to CrMP. One possible explanation for the ineffectiveness of exogenous CO as a vasodilator in preparations not exposed to an inhibitor to HO is that, in such a setting, the vasodilatory mechanism mediated by endogenous

17 JPET # page 17 CO is maximally active (Kozma et al., 1999; Zhang et al. 2001). Another explanation may be that inhibition of endogenous CO results in increased NOS activity (Johnson and Johnson, 2003b), which, in turn, permits the exogenous CO to cause dilatation (Barkoudah et al., 2004). At any rate, in control rats the vasodilatory effect of exogenous CO after HO inhibition was completely abolished after inhibition of sgc, but only partially and not significantly reduced after inhibition of BK Ca (Fig. 4). These result seem in agreement with Naik and Walker (2003b), who have shown that the effect of exogenous CO is sensitive to ODQ and IbTx, while the effect of endogenous CO is cgmp-independent, via activation of BK Ca. On the other hand, the lack of CO effect after sgc inhibition, reported also by other authors (Villamor et al. 2000; Naik and Walker, 2003b), could be explained by the finding of Barkoudah et al. (2004), who hypothesized a permissive role of cgmp for effect of CO on BK Ca. This interpretation is confirmed by the scarce effect of CO detected in arteries after removal of the endothelium (Barkoudah et al., 2004; Foresti et al., 2004), an effect that was restored after the administration of NO, which probably produces the minimum background level of cgmp necessary for CO to cause vasodilatation by activating BK Ca channels (Barkoudah et al., 2004). The administration of CORM-3 to mesenteric arteries of cirrhotic rats with ascites caused a vasodilation evident even in baseline condition, which was amplified by pretreatment with CrMP. The dilator effect of CO was abolished not only after scg inhibition, but also after the inhibition of BK Ca (the final response was significantly lower in respect of control rats, p=0.015) underlining the pivotal role of these channels for the action of CO in this condition. The increased vasodilating effect of CO in ascitic cirrhotic rats was reversed in mesenteric arteries without endothelium, a condition in which CO effect was lower in respect of control rats, even after SNP pretreatment. These results, taken together, not only support the hypothesis that the HO/CO system plays a role in the vasodilatation of mesenteric arteries in ascitic cirrhosis, but also suggest that in ascitic cirrhotic rats the effect of CO is mainly through BK Ca. They also underline that in cirrhotic rats there are other endothelial factors, apart from NO, permitting the vasodilating effect of CO on mesenteric

18 JPET # page 18 arteries. It is not clear what these factors are. It is possible that a role is played by the endothelial derived hyperpolarizing factor (EDHF) and/or by the myoendothelial gap junctions. Further studies are needed to clarify this issue. Western blot data confirmed and supported the results of the hemodynamic experiments. There was an increase in protein expression of HO-1 and of BK Ca α-subunit in mesenteric arteries of cirrhotic rats, particularly in those with ascites, associated to a decrease in the expression of β1-subunit. This finding supports the hypothesis that CO could play a key-role in the regulation of mesenteric vasorelaxation in cirrhosis, as the α-subunit is the target of CO on BK Ca. CO action on BK Ca has been identified in the capacity of enhancing the coupling of Ca 2+ sparks to BK Ca (Jaggar et al., 2002). More recently, Jaggar et al. (2005) have demonstrated that CO activates BK Ca by binding to heme and modifying its interaction with an α-subunit heme-binding domain. The effect of CO on BK Ca is independent from the regulatory β1-subunit (Wu et al., 2002; Jaggar et al. 2005). Indeed, the presence of BK Ca β-subunit is not necessary for the effect of CO, while, on the contrary, it is essential for the stimulating effect of NO (Wu et al., 2002). The decreased expression of β1-subunit may be due to activation of the Renin-Angiotensin-Aldosterone system present in ascitic cirrhosis (Wilkinson and Williams, 1980; Schrier et al., 1988), as β1-subunit synthesis is inhibited by high levels of Angiotensin II (Amberg et al., 2003b). However, the decreased presence of β1-subunit seems not sufficient to inhibit the action of BK Ca, because of the increased expression of the α- subunits and of the increased expression of HO, which, through CO, stimulates BK Ca independently from β1-subunits. Why the expression of α-subunit is increased in cirrhotic rats remains to be clarified. A few authors (Dubuis et al., 2002; Dubuis et al., 2005; Wu and Wang, 2005) have hypothesized that the expression of BK Ca channels may be induced by high CO levels, a condition that might be present in the mesenteric circulation of cirrhotic rats. In conclusion, an overexpression of BK Ca α subunits, together with HO-1 upregulation with increased CO production, participates in the endothelium-dependent alterations and mesenteric

19 JPET # page 19 arterial vasodilatation of ascitic cirrhotic rats.

20 JPET # page 20 Acknowledgments We thank Doctor Brian E. Mann for the synthesis of CORM-3 and Antonietta Sticca for her expert technical assistance.

21 JPET # page 21 REFERENCES Amberg GC and Santana LF (2003a) Downregulation of the BK channel β1 subunit in genetic hypertension. Circ Res 93: Amberg GC, Bonev AD, Rossow CF, Nelson MT and Santana LF (2003b) Modulation of the molecular composition of large conductance, Ca 2+ activated K + channels in vascular smooth muscle during hypertension. J Clin Invest 112: Archer SL, Gragasin FS, Wu X, Wang S, McMurtry S, Kim DH, Platonov M, Koshal A, Hashimoto K, Campbell WB, Falk JR and Michelakis ED (2003) Endothelium-derived hyperpolarizing factor in human internal mammary artery is 11,12-epoxyeicosatrienoic acid and causes relaxation by activating smooth muscle BK Ca channels. Circulation 107: Barkoudah E, Jaggar JH and Leffler CW (2004) The permissive role of endothelial NO in COinduced cerebrovascular dilation. Am J Physiol Heart Circ Physiol 287:H1459-H1465. Barriere E, Tazi KA, Pessione F, Heller J, Poirel O, Lebrec D and Moreau R (2001) Role of smallconductance Ca2+-dependent K+ channels in in vitro nitric oxide-mediated aortic hyporeactivity to alpha-adrenergic vasoconstriction in rats with cirrhosis. J Hepatol 35: Bolognesi M, Sacerdoti D, Di Pascoli M, Angeli P, Quarta S, Sticca A, Pontisso P, Merkel C and Gatta A (2005) Haeme oxygenase mediates hyporeactivity to phenylephrine in the mesenteric vessels of cirrhotic rats with ascites. Gut 54: Bratz IN, Dick GM, Partridge LD and Kanagy NL (2005) Reduced molecular expression of K+

22 JPET # page 22 channel proteins in vascular smooth muscle from rats made hypertensive with N-omega-nitro-Larginine. Am J Physiol Heart Circ Physiol 289:H1284-H1290. Clapp LH and Jabr RI (2003) The BK Channel. Protective or detrimental in genetic hypertension? Circ Res 93: Clark JE, Naughton P, Shurey S, Green CJ, Johnson TR, Mann BE, Foresti R and Motterlini R (2003) Cardioprotective actions by a water-soluble carbon monoxide-releasing molecule. Circ Res 93:e2-e8. Dubuis E, Gautier M, Melin A, Rebocho M, Girardin C, Bonnet P and Vandier C (2002) Chronic carbon monoxide enhanced IbTx-sensitive currents in rat resistance pulmonary artery smooth muscle cells. Am J Physiol Lung Cell Mol Physiol 283:L120 L129. Dubuis E, Potier M, Wang R and Vandier C (2005) Continuous inhalation of carbon monoxide attenuates hypoxic pulmonary hypertension development presumably through activation of BKCa channels. Cardiovasc Res 65: Fernandez M, Lambrecht RW and Bonkovsky HL (2001) Increased heme oxygenase activity in splanchnic organs from portal hypertensive rats: role in modulating mesenteric vascular reactivity. J Hepatol 34: Foresti R, Hammad J, Clark JE, Johnson TR, Mann BE, Friebe A, Green CJ and Motterlini R (2004) Vasoactive properties of CORM-3, a novel water-soluble carbon monoxide-releasing molecule. Br J Pharmacol 142:

23 JPET # page 23 Gonzales-Abraldes J, Garcia-Pagan JC and Bosch J (2002) Nitric oxide and portal hypertension. Metab Brain Dis 17: Jaggar JH, Leffler CW, Cheranov SY, Tcheranova D, E S and Cheng X (2002) Carbon monoxide dilates cerebral arterioles by enhancing the coupling of Ca 2+ sparks to Ca 2+ -activated K + channels. Circ Res 91: Jaggar JH, Li A, Parfenova H, Liu J, Umstot ES, Dopico AM and Leffler CW (2005) Heme is a carbon monoxide receptor for large-conductance Ca 2+ -activated k + channels. Circ Res 97: Jaggar JH, Porter VA, Lederer WJ and Nelson MT (2000) Calcium sparks in smooth muscle. Am J Physiol Cell Phsyiol 278:C235-C256. Johnson FK, Durante W, Peyton KJ and Johnson RA (2003a) Heme oxygenase inhibitor restores arteriolar nitric oxide function in Dahl rats. Hypertension 41: Johnson FK and Johnson RA (2003b) Carbon monoxide promotes endothelium-dependent constriction of isolated gracilis muscle arterioles. Am J Physiol Regul Integr Comp Physiol 285:R536-R541. Kaide JI, Zhang F, Wei Y, Wang W, Gopal VR, Falck JR, Laniado-Schwartzman M and Nasjletti A (2004) Vascular CO counterbalances the sensitizing influence of 20-HETE on agonist-induced vasoconstriction. Hypertension 44:1-7. Kotlikoff M and Hall I (2003) Hypertension: β testing. J Clin Invest 112:

24 JPET # page 24 Kozma F, Johnson RA, Zhang F, Yu C, Tong X and Nasjletti A (1999) Contribution of endogenous carbon monoxide to regularion of diameter in resistance vessels. Am J Physiol Regul Integr Comp Physiol 276:R1087-R1094. Liu Y, Hudetz AG, Knaus HG and Rusch NJ (1998) Increased expression of Ca2+-sensitive K+ channels in the cerebral microcirculation of genetically hypertensive rats: evidence for their protection against cerebral vasospasm. Circ Res 82: Liu Y, Pleyte K, Knaus HG and Rusch NJ (1997) Increased expression of Ca 2+ -sensitive K + channels in aorta of hypertensive rats. Hypertension 30: Motterlini R, Gonzales A, Foresti R, Clark JE, Green CJ and Winslow RM (1998) Heme oxygenase-1-derived carbon monoxide contributes to the suppression of acute hypertensive responses in vivo. Circ Res 83: Mulvany MJ and Aalkiaer C (1990) Structure and function of small arteries. Physiol Rev 70: Naik JS, O'Donaughy TL and Walker BR (2003a) Endogenous carbon monoxide is an endothelial- derived vasodilator factor in the mesenteric circulation. Am J Physiol Heart Circ Physiol 284:H838-H845. Naik JS and Walker BR (2003b) Heme oxygenase-mediated vasodilation involves vascular smooth muscle cell hyperpolarization. Am J Physiol Heart Circ Physiol 285:H220-H228. Sacerdoti D, Bolognesi M, Di Pascoli M, Gatta A, McGiff JC, Laniado Schwartzman M and

25 JPET # page 25 Abraham NG (2006) The rat mesenteric arterial dilator response to 11,12-epoxyeicosatrienoic acid (EET) is mediated by activating heme oxygenase (HO). Am J Physiol Heart Circul Physiol 291: Schrier RW, Arroyo V, Bernardi M, Epstein M, Henriksen JH and Rodes J (1988) Peripheral arterial vasodilation hypothesis: a proposal fort he initiation of renal sodium and water retention in cirrhosis. Hepatology 8: Sieber C, Lopez-Talavera JC and Groszmann RJ (1993) Role of nitric oxide in the in vitro splanchnic hyporeactivity in ascitic cirrhotic rats. Gastroenterology 104: Sun D, Kaley G and Koller A (1994) Characteristics and origin of myogenic response in isolated gracilis muscle arterioles. Am J Physiol 266:H1177-H1183. Tanaka Y, Koike K, Alioua A, Stefani E and Toro L (2004) β1-subunit of maxik channel in smooth muscle: a key molecule which tunes muscle mechanical activity. J Pharmacol Sci 94: Villamor E, Perez-Vizcaino F, Cogolludo AL, Conde-Oviedo J, Zaragoza-Arnaez F, Lopez-Lopez JG and Tamargo J (2000) Relaxant effects of carbon monoxide compared with nitric oxide in pulmonary and systemic vessels of newborn piglets. Pediatr Res 48: Wiest R and Groszmann RJ (1999) Nitric oxide and portal hypertension: its role in the regulation of intrahepatic and splanchnic vascular resistance. Semin Liver Dis 19: Wilkinson SP and Williams R (1980) Renin-angiotensin-aldosterone system in cirrhosis. Gut 21:

26 JPET # page 26 Wu L, Cao K, Lu Y and Wang R (2002) Different mechanisms underlying the stimulation of K Ca channels by nitric oxide and carbon monoxide. J Clin Invest 110: Wu L and Wang R (2005) Carbon monoxide: endogenous production, physiological functions, and pharmacological applications. Pharmacol Rev 57: Zhang F, Kaide JI, Rodriguez-Mulero F, Abraham NG and Nasjletti A (2001) Vasoregulatory function of heme-heme oxygenase-carbon monoxide system. Am J Hypertension 14:62S-67S.

27 JPET # page 27 Footnotes Source of financial support: this work was supported by grants from the University of Padova, from the Italian Ministry of Education and from the Center for the Study of the Liver Disease of the Veneto Region. Person to receive reprint request: Massimo Bolognesi, MD, PhD, Clinica Medica 5, Dipartimento di Medicina Clinica e Sperimentale, Policlinico Universitario, Via Giustiniani 2, Padova Italy. Telephone: Fax: massimo.bolognesi@unipd.it

28 JPET # page 28 LEGENDS FOR FIGURES Figure 1. Concentration-response curves to Acetylcholine (ACh) obtained in small resistance mesenteric arteries, before (open circle) and after (closed square) the inhibition of heme oxygenase with CrMP (15 µm). In baseline condition, the endothelium-dependent vasodilatation induced by ACh was increased in cirrhotic rats with ascites in respect of control rats (difference in pec 50 : p=0.036). Inhibition of heme oxygenase decreased the effect of ACh more in cirrhotic rats with ascites in respect of controls rats (difference in the change of pec 50 : p=0.03). After CrMP the response to ACh was similar in the two groups. = significantly different (p<0.01) from baseline (two-way ANOVA). #=p=0.036 in respect of control rats; *= p=0.05 from baseline; **= p<0.05 from baseline. Figure 2. Concentration-response curves to Acetylcholine (ACh) obtained in small resistance mesenteric arteries incubated with indomethacin (indo) (2.8 µm) and L-NAME (1 mm), to inhibit COX and NOS, respectively. In the presence of vehicle (open circle) the endothelium-dependent vasodilatation induced by ACh was increased in cirrhotic rats with ascites in respect of control rats (difference in pec 50 : p=0.05). In respect of vehicle (open circle), incubation with the scg inhibitor ODQ (10 µm) (closed square) decreased the effect of ACh both in control rats and in rats with cirrhosis and ascites. The addition of CrMP (15 µm) to ODQ (10 µm) (closed triangle) further reduced the effect of ACh in both groups. = significantly different (p 0.01) from the other two curves (two-way ANOVA). #=p=0.05 in respect of control rats; *= significantly different (p<0.05) from Indo+L-NAME; **=significantly different (p<0.05) from Indo+L-NAME+ODQ. Figure 3. Concentration-response curves to Acetylcholine (ACh) obtained in small resistance mesenteric arteries incubated with indomethacin (indo) (2.8 µm), L-NAME (1 mm) and ODQ (10 µm), to inhibit COX, NOS and sgc, respectively. In the presence of vehicle (open circle) the effect

29 JPET # page 29 of ACh was increased in cirrhotic rats with ascites in respect of control rats (difference in pec 50 : p=0.024). In respect of vehicle (open circle), incubation with the BK Ca inhibitor IbTx (25 nm) (closed square) decreased the effect of ACh both in control rats and in rats with cirrhosis and ascites. The inhibition of BK Ca was more effective on the endothelium-dependent vasodilatation induced by ACh in cirrhotic rats with ascites in respect of controls rats (difference in the change of pec 50 : p=0.010). The addition of CrMP (15 µm) to IbTx (25 nm) (closed triangle) did not further modify the effect of ACh in both groups. = significantly different (p<0.01) from indo+l- NAME+ODQ (two-way ANOVA); #=p=0.024 in respect of control rats. *=significantly different (p<0.05) from Indo+L-NAME+ODQ. Figure 4. Concentration-response curves to Acetylcholine (ACh) obtained in small resistance mesenteric arteries of control rats in the presence of vehicle (open circle) and after incubation with three different amount of the BK Ca inhibitor iberiotoxin (IbTx): 25 nm (closed square), 50 nm (closed triangle), 100 nm (closed circle). = significantly different (p<0.05) from baseline and from IbTx 25 nm. *= significantly different (p<0.05) from baseline; **=significantly different (p<0.05) from IbTx 25nM. The inhibitory effect of high concentrations of IbTx on ACh induced vasodilatation underlines the physiological role of BK Ca in the regulation of endothelium-dependent vasodilatation of mesenteric arteries. Figure 5. Effect of a CO releasing molecule (CORM-3, 200 µm) on the internal diameter of small mesenteric arteries. The effect of CORM-3 was evaluated in the presence of vehicle (baseline, white bars), and after elimination of endogenous CO by pre-treatment with the heme oxygenase inhibitor CrMP (15 µm) (black bars). In arteries pre-treated with CrMP (15µM), the effect of CORM-3 was also evaluated after incubation with the sgc inhibitor ODQ (10 µm) (gray bars) or with the BK Ca inhibitor IbTx (25 nm) (striped bars). In the presence of vehicle (white bars), the addition of CORM-3 caused a vasorelaxation more evident in ascitic cirrhotic rats (n.6) than in controls (n.8)

30 JPET # page 30 (p=0.012). In arteries incubated with CrMP (black bars), CORM-3 provoked a significantly greater vasorelaxation both in controls (n.5) and in ascitic cirrhosis (n.4) (p=0.016 and p=0.022, respectively), but in the latter group the vasorelaxation was more evident (p=0.036). In arteries incubated with CrMP and ODQ (gray bars), CORM-3 did not provoke any significant dilatation, both in control (n.4) and in cirrhotic rats with ascites (n.5). In arteries pretreated with CrMP and IbTx (striped bars), CORM-3 caused a significant dilatation in controls rats (n.6) (p=0.004), which was slightly but not significantly lower than the dilatation obtained in the arteries pretreated only with CrMP; on the contrary, in cirrhotic rats with ascites (n.5), pretreatment with CrMP and IbTx markedly reduced the dilator effect of CORM-3 (p=0.003) in respect of the effect obtained after incubation with only CrMP. #= p<0.05 in respect of control rats; *= p<0.05 in respect arteries pretreated with CrMP. Figure 6. Effect of a CO releasing molecule (CORM-3, 200 µm) on the internal diameter of small mesenteric arteries after endothelium removal. The effect was evaluated before and after the administration of the NO donor SNP (100 nm). After endothelium removal, CORM-3 provoked a significant but slight dilatation both in controls (n.4) (p=0.002) (closed diamond) and in cirrhotic rats with ascites (n.5) (p=0.056) (open triangle); the dilatation was more evident in controls (p=0.015). After the addition of a low concentration of SNP, which caused per se a modest dilation, the vasodilating effect of CORM-3 was more evident, both in controls (p=0.0003) and in cirrhotic rats (p=0.037), but remained more evident in control rats (p=0.0008). The effect of CORM after SNP was measured when a new stable baseline was obtained after the administration of SNP. #= p<0.05 in respect of control rats; *= p<0.05 in respect of values obtained before SNP administration. Figure 7. Western blot analysis of heme oxygenase-1 (HO-1) in the small mesenteric arteries of controls rats and of cirrhotic rats, with and without ascites. The reported blots are representative of

31 JPET # page 31 4 to 6 experiments. Lower panel: densitometric analysis of HO-1. *= p<0.05 in respect of control rats. In cirrhotic rats the expression of HO-1 was increased. Figure 8. Western blot analysis of the α- and of β1-subunits of BK Ca in the small resistance mesenteric vessels of controls rats and of cirrhotic rats, with and without ascites. The reported blots are representative of 3 to 5 experiments; lane 1-2: cirrhotic rats with ascites; lane 3-4: cirrhotic rats without ascites; lane 5-6: control rats. Lower panel: densitometric analysis of α-subunit (graph A) and of β1-subunit (graph B). *= p<0.05 in respect of control rats. In cirrhotic rats with and without ascites, the expression of BK Ca α-subunit was increased, while the expression of BK Ca β1-subunit was decreased.