Role of the Heme Oxygenases in Abnormalities of the Mesenteric Circulation in Cirrhotic Rats

The heme-oxygenase (HO) system is the rate-limiting step in heme degradation, releasing carbon monoxide (CO), biliverdin, and free iron. Three HO isoforms (HO-1, HO-2, and HO-3) have been identified that can carry out this reaction (Abraham et al., 1987; McCoubrey et al., 1992, 1997; Shibahara et al., 1993). HO-1 is a 32-kDa heat shock protein (Shibahara et al., 1987) that is inducible by noxious stimuli (Lutton et al., 1992; Neil et al., 1995; Lu et al., 1998). HO-2, a 36-kDa protein, is constitutively expressed and found in abundance in brain, testis, liver (Maines et al., 1986; Trakshel et al., 1988), and endothelium (Zakhary et al., 1996). HO-3, although related to HO-2, is a product of a different gene, with a lesser ability than HO-2 to catalyze heme degradation (McCoubrey et al., 1997).

HO-2, which is localized in endothelial cells and adventitial nerves of blood vessels (Zakhary et al., 1996), functions in circulatory regulation. Furchgott and Jothianandan (1991) reported that CO, arising from the metabolism of heme by HO, dilated blood vessels, whereas inhibition of HO by tinmesoporphyrin (SnMP) produced vasoconstriction (Kozma et al., 1999) and raised blood pressure (Johnson et al., 1995, 1996). Heme, by activating HO, has been shown to dilate the ductus arteriosus (Coceani et al., 1997). Moreover, contiguous, if not overlapping, spheres of activity within the vasculature of nitric oxide (NO) and CO were suggested by the study of Zakhary et al. (1996). The authors concluded that “the similarity of NOS and HO-2 localizations and functions in blood vessels and the autonomic nervous system implies complementary and possibly coordinated physiological roles for these two mediators”. CO was shown to relax blood vessels via an endothelial-independent pathway linked to guanylate cyclase, as its effects were accompanied by the accumulation of cGMP levels in vascular smooth muscle (Christodoulides et al., 1995). Furthermore, inhibition of guanylate cyclase attenuated the vasodilator effect of CO (Wang et al., 1997). In isolated gracilis muscle arterioles of rats, which express HO-2, HO inhibitors decreased the diameter of pressurized vessels (Kozma et al., 1999), likely by inhibiting potassium channels as a result of decreased CO production, as has been reported for the thick ascending limb in response to inhibition of HO (Liu et al., 1999).

Mesenteric arteriolar vasodilatation and increased portal inflow play a major role in the development of portal hypertension in cirrhosis (Vorobioff et al., 1984; Shah et al., 1998). Various vasodilator compounds, including glucagon, adenosine, prostacyclin, NO, α-calcitonin gene-related peptide, and adrenomedullin have been proposed as mediators of the mesenteric vascular abnormalities that occur both in cirrhosis and in prehepatic portal hypertension (Shah et al., 1998; Gatta et al., 1999). In cirrhotic rats, inhibition of NO synthesis greatly reduced mesenteric blood flow and hyporeactivity of the mesenteric circulation to vasoconstrictor agents (Sieber et al., 1993). In rats with prehepatic portal hypertension, HO inhibition modified the hyporeactivity of the mesenteric circulation to vasoconstrictors (Fernandez et al., 2001).

The present study was designed to determine whether 1) augmentation of HO contributes to mesenteric vascular abnormalities in cirrhotic rats, 2) gene transfer reproduces these alterations in normal rats, and 3) nitric oxide synthase (NOS)-NO and HO-CO systems interact in the mesenteric circulation to affect vascular tone and reactivity. To achieve these goals, the isolated perfused superior mesenteric vasculature (SMV) was studied in terms of 1) reactivity to constrictor agents in normal rats and in those with carbon tetrachloride (CCl₄)-induced cirrhosis, 2) effects of inhibition of HO with SnMP in each group, and 3) effects on the SMV produced by transfection with the human HO-1 (HHO-1) gene. The participation of NO in the development of hyporeactivity of the mesenteric vasculature in cirrhotic rats was also examined. This issue was addressed by determining enhancement of the vascular responses to constrictor agents as modified by successive inhibition of CO and NO production and then reversing the order of inhibition, with NO preceding CO. We were able to identify a significant component of each system, NOS-NO and HO-CO, that contributed to the vascular hyporeactivity of cirrhotic rats, suggesting “complementary and possibly coordinated physiological roles” (Zakhary et al., 1996) for CO and NO. Our results indicate a key role for HO in the control of the mesenteric circulation, as inhibition of HO activity with SnMP normalized mesenteric perfusion pressure and restored toward normal the response of the mesenteric vasculature to vasoconstrictor agents. We also found, quite unexpectedly, an increase in HO-2, the constitutive enzyme, rather than HO-1, the inducible enzyme in the SMV in cirrhotic portal hypertension. Finally, HO activity augmented by HO-1 gene transfer decreased mesenteric perfusion pressure and reduced the reactivity of the SMV to vasoconstrictor agents.

Materials and Methods

Materials. Phenylephrine (PE) was purchased from Sigma-Aldrich (St. Louis, MO) and dissolved in distilled water. Endothelin-1 (ET-1) was purchased from Peninsula Laboratories (Belmont, CA) and dissolved and stored in 0.1% acetic acid at –20°C. SnMP was purchased from Porphyrin Products, Inc. (Logan, UT). N^G-nitro-l-arginine methyl ester (l-NAME) was purchased from Sigma-Aldrich and dissolved in distilled water.

Animals. Adult male Sprague-Dawley rats (150–160 g) were purchased from Charles River Laboratories, Inc. (Wilmington, MA) and housed in an animal facility with controlled normal day-night cycle. The rats were maintained on a standard diet of Purina chow and were allowed free access to water and food until the night before experiments. Experiments were performed in accordance with the guidelines set forth by The American Physiological Society for ethical treatment of animals.

Induction of Cirrhosis. Cirrhosis was induced in animals (∼170 g b.wt.) by combined treatment with CCl₄ and phenobarbital, the latter given in drinking water (Sacerdoti et al., 1991). After 1 week of phenobarbital administration, CCl₄ was given by gavage, starting with a dose of 30 μl, and then increasing doses were administered according to changes in body weight. CCl₄ was administered once a week and then twice a week for 8 to 10 weeks. Treatment was terminated 1 week prior to conducting experiments. Rats on phenobarbital during the time of treatment were used as controls. Free access to standard chow was allowed throughout the study. In all of the CCl₄-treated rats used for study, portal pressure was measured by a Statham transducer after direct cannulation of the portal vein before the experiments; rats with portal pressure <8 mm Hg were excluded from the study.

Perfusion of Isolated Mesenteric Artery (MacGregor Preparation). The superior mesenteric artery was cannulated with a PE-60 catheter and perfused with 25 ml of warm Krebs-Henseleit buffer. After the superior mesenteric artery was isolated with its mesentery, the gut was excised near its mesenteric border. The mesenteric arterial bed was then brought in a 37°C water-jacketed container and perfused at a constant rate of 0.8 ml/min/100 g b.wt. with oxygenated (95% O₂/5% CO₂) Krebs-Henseleit solution (118.5 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl₂, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 25.0 mM NaHCO₃, and 11.1 mM dextrose, pH 7.4 at 37°C). The effluent was constantly collected, and the pressure was measured by a Statham transducer.

Vascular Responses to Agonists and Antagonists. After an initial 30-min basal period, concentration-response curves were constructed with KCl, –5.22 to –3.6 log mol; PE, –9.3 to –6.3 log mol; and ET-1, –11.7 to –10.5 log mol in the presence and absence of HO inhibition with SnMP (10 μM). This concentration of SnMP does not inhibit NOS activity or activity of soluble guanylate cyclase (Meffert et al., 1994; Zakhary et al., 1996). It does, however, inhibit completely the increase in HO activity produced by either heme or SnCl₂ in endothelial cells as well as smooth muscle cells (Li Volti et al., 2002). SnMP was added to the buffer 10 min before vasopressors were examined. Concentration-response curves were constructed by adding increasing bolus amounts (20–50 μl) of agonists. The maximum response was obtained within 60 s; perfusion pressure returned to baseline values in 2 to 3 min after KCl and PE and 10 to 15 min after ET-1. Increasing doses were added after return to the baseline pressure. Experiments performed to determine the differential effects of inhibition of HO and NOS in cirrhotic animals were done using single doses of PE, –7.3 log mol. After the initial equilibration period, the responses to PE were evaluated without inhibitors and then after inhibition of HO with SnMP (10 μM) (n = 6) or l-NAME (100 μM) (n = 6). After restoration of baseline conditions, the second inhibitor was added to the perfusate, and responses to the agonists were again evaluated. l-NAME was dissolved in saline. In rats transfected with the HHO-1 gene, the vasoconstrictor responses of the SMV to PE (–8 log mol) were evaluated after an equilibration period before and after inhibition of HO with SnMP (10 μM).

Western Blot Analysis. For the detection of HO-1 and HO-2 protein, microsomes were prepared from tissues as previously described (Sacerdoti et al., 1988); twenty micrograms of microsomal protein was analyzed by Coomassie staining of the gels and by immunoblot analysis of proteins transferred to nitrocellulose membrane. Briefly, membrane blots were blocked in 5% nonfat milk overnight to block nonspecific binding of antibodies. The blots were then incubated with mouse monoclonal anti-HO-1 or HO-2 antibodies (1:2000) (StressGen Biotechnologies Corp., Victoria, BC, Canada) for 1 h and washed with Tris-buffered saline containing 0.1% Tween 20 three times. The blot was subsequently incubated with rabbit anti-mouse IgG antibody (1:4000) for 1 h. The blot was washed three times with Tris-buffered saline/Tween 20, and immunoreactive proteins were visualized by means of enhanced chemiluminescence reagent and exposure to an autoradiographic film.

Microassay for Measurement of HO Activity. HO activity was measured as previously described (Chernick et al., 1989). Briefly, tissue homogenates (100 μg of protein) were incubated in a NADPH-generating system (20 μM MgCl₂, 30 μM heme, and 1μCi/ml of ¹⁴C-heme). HO activity was measured as the conversion of heme to bilirubin in the presence of biliverdin reductase. ¹⁴C-bilirubin was extracted with chloroform, and radioactivity was measured. Results were expressed as counts per minute per milligram protein.

Construction of the Adenovirus Vector Containing Human HO-1 cDNA. HHO-1 cDNA with adenovirus vector was constructed using a mammalian transfection kit (Stratagene, La Jolla, CA) as previously described by Abraham et al. (1995b). Briefly, the human cDNA plasmid was constructed as follows: a Xho1-Xba1 fragment of a HHO-1 expression vector, pRc/CMVHO-1, was cloned between Xho1 and Xba1 of the pBacPAC8 to generate pBacPAC8 HHO-1. The Bg1II-BamH1 fragment of the pBacPAC8HHO was then introduced into the Bg1II site of pAdBg1II. The human embryonic cell line 293 (ATCC catalog number CRL-1573) was cotransfected with 10 μg of EcoR1-digested adenoviral HHO-1 and 1 μg of Cla1-digested dL7001DNA by calcium phosphate coprecipitation using a mammalian transfection kit. HHO-1 adenovirus construct was replicated and encapsulated into an infectious virus as described previously (Abraham et al., 1995a). The virus was concentrated by centrifugation by using Ultrafree-MC filters (Millipore Corp., Bedford, MA). The viral titers were determined by plaque assay using human embryonic 293 cells. The viral stock was stored at –80°C until use.

In Vivo Gene Transfer. Eight rats were injected intracardiacally with HHO-1 adenovirus suspension (10¹² pfu/300 g b.wt.). Four rats were treated with the adenovirus without the HO-1 gene. The animals were acclimatized for 1 week prior to the start of the experiment, and experiments were performed 1 week after viral administration. Hemodynamic studies were performed as described above in six rats treated with the gene and four rats treated with the empty virus.

RNA Extraction and RT-PCR. SMV were homogenized in liquid nitrogen, and total RNA was extracted with guanine thiocynate, first purified with phenol-chloroform followed by an additional purification with chloroform and then precipitated with isopropanol (Chomczynski and Sacchi, 1987). Total RNA was resuspended in autoclaved diethyl pyrocarbonate-water and quantified using Beckman DU7400 spectrophotometer (Beckman Instruments, Fullerton, CA). To verify the quality of RNA, 10 μg of total RNA was denatured and separated by 1% agarose gel (containing 2.2 M formaldehyde) electrophoresis. The RNA was visualized under UV light with the help of ethidium bromide. Primers were designed to amplify specific fragments of the HHO-1 and/or rat HO-1. The primers used did not recognize HO-2 sequences. The RT-PCR for HHO-1 was performed by using the Advantage RT-PCR kit (BD Biosciences Clontech, Palo Alto, CA) and the primers designed to amplify a RT-PCR product with a predicted size of 555 base pairs. Briefly, 1 μg of total RNA was reverse transcribed using oligo(dT¹⁸) primer to synthesize the second strand and to obtain the cDNA. The reverse transcription was performed at 42°C for 1 h, and the reaction tubes were placed on ice. The PCR of HHO-1 was performed by adding PCR mix prepared with sense primer (5′CAGGCAGAGAATGCT GAGTTC-3′), antisense primer (5′-quenched by heating it at 95°C for 5 min and then the GATGTTGAGCAGGAACGCAGT-3′), deoxynucleoside-5′-triphosphate, Taq polymerase, ³²P-dCTP, and reaction buffer. To amplify the endogenous rat HO-1, the primers were designed from the rat cDNA sequence (sense primer, 5′-TGAAGGAGGCCACCAAGGAG-3′ and antisense primer, 5′-CCCCTGAGAGGTCACCCAGG-3′), which amplifies a 356-base pair fragment. RT-PCR was performed as described above.

Statistical Analysis. Results were expressed as means ± S.E.M. Concentration-response data for PE and KCl derived from each vessel were fitted separately to a logistic function by nonlinear regression, and the maximum asymptote of the curve (R_max) and concentration of agonist producing 50% of the maximal response (EC₅₀), which was expressed as negative log mol, were calculated using commercially available software (Prism 2.01; GraphPad Software, San Diego, CA). Concentration-response data were analyzed by two-way analysis of variance. Differences between groups were evaluated by unpaired Student's t test. The effects of inhibition of HO on the action of vasoconstrictors were evaluated by paired t test. Statistical significance was set at p < 0.05.

Results

We examined the effects of inhibition of HO on perfusion pressure in the isolated perfused SMV obtained from normal and cirrhotic rats. We addressed the effects of inhibition of HO with SnMP (10 μM) on responses of mesenteric arteries to three vasopressor agents: KCl, PE, and ET-1. Figure 1 shows the baseline perfusion pressure in normal and cirrhotic rats with and without SnMP. Mesenteric perfusion pressure was lower in cirrhotic than in normal rats (p < 0.05). Furthermore, inhibition of HO did not significantly increase (6 ± 3%) mesenteric perfusion pressure in control animals (n = 14), whereas in cirrhotic rats a significant increase in pressure of 20 ± 3% was observed (n = 14).

Next, we examined whether inhibition of HO by SnMP (10 μM) can restore normal vasopressor responses to KCl, PE, and ET-1 that were depressed in cirrhotic animals. The vasoconstrictor responses to KCl, PE, and ET-1 were shifted to the right, the maximal response was blunted, and the EC₅₀ for KCl was increased (Figs. 2, 3, and 4). SnMP significantly increased the response to KCl in normal (n = 8) and cirrhotic rats and restored to normal the responsiveness to KCl in the SMV (n = 8) (Fig. 2). SnMP did not affect the vasoconstrictor response to the sympathomimetic agent, PE, in normal rats (n = 8) and restored to normal the responses of the SMV to the vasoconstrictor agent in cirrhotic rats (n = 8) (Fig. 3). SnMP increased the vasoconstrictor responses of SMV to ET-1 in cirrhotic but not in normal rats (n = 8); in cirrhotic rats, the responses to ET-1 in SMV were restored to normal (n = 8) (Fig. 4). These results indicate that SnMP-induced inhibition of HO in cirrhotic rats restores toward normal the responsiveness of the SMV to vasopressor agents.

We also examined the contributions of NO and CO to attenuating the response to PE in the SMV in view of reports that NO contributes to hyporeactivity of the SMV to pressor agents in cirrhotic rats (Sieber et al., 1993). SnMP was used to inhibit HO because of its demonstrated selectivity at the concentration employed (10 μM) and its solubility (Zakhary et al., 1996). It was important that SnMP did not inhibit NOS, as our study was directed at the vascular effects of CO in the SMV of cirrhotic rats. The relative contribution of HO-CO and NOS-NO to producing hyporeactivity of the SMV of cirrhotic rats to PE was assessed by determining the effects of inhibiting each system separately, i.e., HO-CO with SnMP and NOS-NO with l-NAME, followed by inhibition of the other system, then reversing the order of inhibition of CO and NO production (Fig. 5, A and B). After achieving a significant reduction in vascular hyporeactivity by inhibition of either HO or NOS, additional reduction was obtained on inhibiting the other gas-producing system. For example, in Fig. 5A, the constrictor response of the SMV of cirrhotic rats to PE was increased by inhibition of HO activity with SnMP and was further enhanced by inhibition of NOS activity with l-NAME. Then, the order of inhibition of HO and NOS activities was reversed in Fig. 5B. After the combined inhibition of HO and NOS, the constrictor responses of the SMV of cirrhotic rats to PE were fully restored. Moreover, the effects of either SnMP or l-NAME in restoring to normal the vasoconstrictor responses to PE (Fig. 5) in cirrhotic rats were similar, irrespective of the order of administration of SnMP or l-NAME. These findings are in accord with the findings of Thorup et al. (1999) that CO “may synergize with NO in eliciting vasorelaxation and modulating basal vascular tone” by effects on vascular mechanisms such as, for example, those served by guanylyl cyclase and cGMP.

We examined whether responsiveness of mesenteric arteries of cirrhotic rats to vasopressor agents was associated with increased expression of HO-1 and HO-2 in this vascular bed. Western blot analysis of mesenteric blood vessels obtained from normal and cirrhotic rats (Fig. 6A) demonstrated significantly higher levels of HO-2, but not of HO-1, in vascular tissues obtained from cirrhotic rats (Fig. 6, lanes 5–8) compared with those from normal rats (Fig. 6, lanes 1–4). HO up-regulation was quantified by densitometry-based analysis of HO-1 and HO-2 proteins after blotting the ratio of each sample. As seen in the densitometry analysis (Fig. 6B), HO-2 was increased significantly in SMV obtained from cirrhotic rats when compared with those of control rats.

To assess the effect of the increase in HO-2 protein on HO activity, cell homogenates obtained from SMVs of control and cirrhotic rats were used to measure the enzyme activity as indicated by C¹⁴ heme catabolism to C¹⁴ bilirubin (Fig. 7). Bilirubin formation was significantly increased in SMV samples obtained from cirrhotic rats compared with those obtained from control rats (p < 0.05). These findings help explain 1) the hyporeactivity of the mesenteric vasculature to vasoconstrictor agents and 2) why inhibition of HO helps restore the responsiveness to vasoconstrictor agents of the SMV obtained from cirrhotic animals.

To examine further our hypothesis that increased HO levels in mesenteric arteries in experimental cirrhosis are responsible for vasodilation and decreased reactivity of SMV to vasoconstrictor agents, we injected eight normal rats with adenovirus containing the HHO-1 gene (10¹² pfu/300 g b.wt.). These preparations of adenovirus-mediated HO-1 gene construct were shown to increase HO-1 gene expression within 7 days of infection (Abraham et al., 2000). The ability of adenovirus-mediated HHO-1 to infect mesenteric blood vessels was assessed after 7 days, and total RNA from SMV was assessed for the presence of the HHO-1 by RT-PCR and enzyme activity (Fig. 8). Positive signals were obtained for the HHO-1 gene when amplified with HHO-1 primers within 7 days of infection. Injection of adenovirus-mediated HHO-1 gene transfer resulted in the expression of HO-1 in mesenteric blood vessels. The HHO-1 mRNA signal was not detectable in mesenteric arteries infected with empty adenoviral constructs (Fig. 8, lane 2). To determine the quality of total RNA of SMVs, RT-PCR analysis was performed using GAPDH-specific primers. Amplification of GAPDH/PCR products occurred in all RNA tested (Fig. 8). HO protein in tissues obtained from rats infected with adenoviral constructs was significantly higher (1240 ± 200 versus 187 ± 13 cpm/mg protein, n = 4, p < 0.01) than in tissues obtained from rats infected with empty adenoviral constructs.

Finally, we examined the effect of HHO-1 transfection on the mesenteric circulation. As is evident in Fig. 9, perfusion pressure and the increase in perfusion pressure of the SMV induced by PE were attenuated in animals transfected with the HHO-1 gene. Inhibition of HO with SnMP partially reversed these abnormalities. No difference was evidenced in rats treated with the empty virus. These results support our hypothesis that increased HO expression and activity in cirrhosis contribute to the aberrant responses of the SMV to vasoconstrictor agents.

Discussion

Portal hypertension, a major complication of cirrhosis, is the basis for the development of esophageal varices, ascites, renal failure, and hepatic encephalopathy (Gatta et al., 1999). Mesenteric arterial resistance is normally high, whereas in cirrhosis, dilatation of the mesenteric vasculature is associated with an increase in portal inflow, a major determinant of portal hypertension (Shah et al., 1998; Gatta et al., 1999). Furthermore, in experimental cirrhosis, the responsiveness of the mesenteric circulation to pressor agents is greatly impaired (Sieber et al., 1993; Gagano et al., 1997). A number of endogenous vasoactive compounds, previously enumerated, have been proposed to mediate this lack of vascular responsiveness (Shah et al., 1998; Gatta et al., 1999).

In the present study, we have shown in rats with CCl₄-induced hepatic cirrhosis that the resulting attenuation of mesenteric vascular constrictor responses to diverse agonists could be reversed by inhibition of HO, highlighting the potential importance of the HO-CO system to the regulation of circulatory tone and vascular reactivity. Quite unexpectedly, the HO isoform responsible for mesenteric vascular hyporeactivity was shown to be the constitutive enzyme HO-2, not the inducible enzyme HO-1. This finding is not unique, as HO-2 levels are increased by estrogens in endothelial cells (Tschugguel et al., 2001) and by corticosterone in the brain (Weber et al., 1994). Furthermore, this finding stands in contrast to the induction of HO-1 activity in response to portal hypertension produced by ligation of the portal vein (Fernandez et al., 2001) and helps in distinguishing critical differences between these two experimental approaches, CCl₄-induced cirrhosis versus portal vein constriction, to the production of altered responses of the mesenteric vasculature. For example, in contrast to CCl₄-induced cirrhosis, HO inhibition in rats with ligated portal veins did not alter mesenteric vascular hyporeactivity to a α₁-adrenergic agonist (Fernandez et al., 2001). Blockade of NO production was required to reverse the hyporeactivity to methoxamine in this model of portal hypertension.

We also examined potential interactions of the NOS-NO and HO-CO systems in the mesenteric vasculature in view of the demonstrated contribution of NO to splanchnic vascular hyporeactivity in cirrhotic rats (Sieber et al., 1993). Our findings suggest that these systems are complementary in the mesenteric circulation of the cirrhotic rat, which agrees with the “coordinated functions of NO and CO” proposed by Zakhary et al. (1996) and is in accord with the presence in endothelial cells of both NOS and HO (Zakhary et al., 1996) as well as the ability of each system to activate soluble guanylate cyclase and increase cGMP (Abraham et al., 2002). Alternatively, rather than overlapping spheres of activity of NO- and CO-mediated vascular effects reflecting colocalization of NOS and HO isoforms in the endothelium, segmental separation within a vasculature (eNOS predominating in large arteries and HO-2 in small arteries and microvessels) may also give rise to coordinated and complementary responses of the NOS-NO and HO-CO systems (Zakhary et al., 1996). A recent study demonstrated that despite eNOS gene deletion, mice with portal hypertension developed a hyperdynamic circulation, indicating the importance of other vasodilator systems in the production of splanchnic circulatory changes in portal hypertensive animals (Iwakiri et al., 2002).

The possibility that SnMP inhibits NOS as well as HO under the conditions of our study is unlikely, as weak inhibition of NOS with SnMP can be obtained only at a 10× higher concentration (100 μM) than that used in our study (10 μM) (Meffert et al., 1994; Zakhary et al., 1996). Furthermore, SnMP, in the concentration used in our study (10 μM), was shown to be without effect on eNOS activity in the vasculature despite colocalization of HO-2 and eNOS in the endothelium of the several vascular beds studied (Zakhary et al., 1996). This specificity of SnMP relative to HO when compared with NOS has also been reported for the central nervous system (Meffert et al., 1994). Furthermore, the data presented in Fig. 5, A and B show that prior administration of SnMP does not notably affect the action of l-NAME in augmenting the constrictor response of the SMV to PE, indicating that the activity of NOS is not, to a readily detectable degree, affected by prior treatment with SnMP under the conditions of our study. Nonetheless, an effect of NO on HO gene expression and activity may occur over the long term, as NO has been shown to induce HO-1 in endothelial cells in vitro (Hartsfield et al., 1997). On the other hand, transfection of rat endothelial cells with HHO-1, which induced a 4-fold increase in HO activity, did not affect either eNOS protein or NO production (Abraham et al., 2002).

Transfection of normal rats with HHO-1 provided additional support for the role of CO as an essential component in mechanisms that modulate reactivity of the mesenteric vasculature in experimental cirrhosis. The transfection was highly effective, and HHO-1 mRNA expression was evident in the PCR analysis of the mesenteric vasculature (Fig. 8). Furthermore, HO activity in mesenteric arteries from transfected rats was significantly increased compared with the activity in rats treated with the empty virus. Transfected rats showed hemodynamic characteristics similar to, or even more pronounced than, those of cirrhotic animals; viz., perfusion pressure and reactivity of the mesenteric vasculature to constrictor agents were decreased (Fig. 9). We have reported that adenovirus-mediated HHO-1 gene transfection was associated with a decrease in microsomal heme and release of CO (Abraham et al., 2000). Moreover, retrovirus-mediated HO gene transfer into endothelial cells, as we have shown, increases cGMP and protects against oxidant-induced cell injury (Yang et al., 1999). The increase in HO activity produced by this strategy may prove useful as a means of modifying abnormalities of regional vascular beds.

Taken together, our findings support a role for HO-CO that may synergize with NOS-NO in regulating the mesenteric circulation; they provide a conceptual basis for examining abnormalities of vascular beds in human hepatic cirrhosis. The systemic circulation, in addition to the splanchnic vasculature, is affected by hepatic cirrhosis associated with a decline in systemic vascular resistance in human subjects (Schrier et al., 1998) associated with increased CO production, the magnitude of which was correlated with the severity of the cirrhosis (De Las Heras et al., 2003). Inhibition of HO has also been reported to reverse vascular hyporeactivity to norepinephrine in cirrhotic patients (Sacerdoti et al., 2002). CO lowers blood pressure in rats by participating in the regulation of basal tone in resistance blood vessels (Johnson et al., 1995; Kozma et al., 1999). In normal rats, inhibition of HO increased blood pressure (Johnson et al., 1995), whereas in the spontaneously hypertensive rat but not in the normotensive Wistar Kyoto rat, stimulation of HO with SnCl₂ decreased blood pressure (Sacerdoti et al., 1989). The present study has clinical implications: inhibition of HO may serve as a therapeutic option in cirrhotic patients, possibly improving not only portal hypertension but also systemic hemodynamics and, as a consequence, preserving renal function and reducing portal hypertension.