fax: (39) 0755783883
Liver Biology and Pathobiology
The third gas: H2S regulates perfusion pressure in both the isolated and perfused normal rat liver and in cirrhosis†
Article first published online: 17 AUG 2005
Copyright © 2005 American Association for the Study of Liver Diseases
Volume 42, Issue 3, pages 539–548, September 2005
How to Cite
Fiorucci, S., Antonelli, E., Mencarelli, A., Orlandi, S., Renga, B., Rizzo, G., Distrutti, E., Shah, V. and Morelli, A. (2005), The third gas: H2S regulates perfusion pressure in both the isolated and perfused normal rat liver and in cirrhosis. Hepatology, 42: 539–548. doi: 10.1002/hep.20817
Potential conflict of interest: Nothing to report.
- Issue published online: 22 AUG 2005
- Article first published online: 17 AUG 2005
- Manuscript Accepted: 3 JUN 2005
- Manuscript Received: 22 FEB 2005
The regulation of sinusoidal resistance is dependent on the contraction of hepatic stellate cells (HSC) around sinusoidal endothelial cell (SEC) through paracrine cross-talk of vasoconstrictor and vasodilator agents. Hydrogen sulfide (H2S), a recently discovered gas neurotransmitter, is a putative vasodilator whose role in hepatic vascular regulation and portal hypertension is unexplored. Four-week bile duct–ligated (BDL) rats with cirrhosis and control rats were treated daily with NaHS (56 μmol/kg) for 5 days. Isolated livers were perfused first with NaHS for 20 minutes and then with norepinephrine (NE) and the intrahepatic resistance studied. In normal rats and animals with cirrhosis, administration of NE resulted in a dose-dependent increase of portal pressure. This effect was attenuated by H2S treatment (P < .05). The H2S-induced relaxation of hepatic microcirculation was attenuated by glibenclamide, an adenosine triphosphate (ATP)-sensitive K+ channel inhibitor. L-Cysteine, a substrate of cystathionine-gamma-lyase (CSE), decreased vasoconstriction in normal rat livers (P < .05) but failed to do so in livers with cirrhosis. BDL resulted in a downregulation of CSE mRNA/protein levels and activity (P < .05). Our in vitro data demonstrate that CSE is expressed in hepatocytes, HSCs, but not in sinusoidal endothelial cells (SEC). HSC activation downregulates CSE mRNA expression, resulting in a defective production of H2S and abrogation of relaxation induced by L-cysteine. In conclusion, CSE-derived H2S is involved in the maintenance of portal venous pressure. The reduction of CSE expression in the liver with cirrhosis contributes to the development of increased intrahepatic resistance and portal hypertension. (HEPATOLOGY 2005.)
Gaseous transmitters are a growing family of regulatory molecules involved in regulation of physiological and pathological functions in mammalian tissues.1–3 Although nitric oxide (NO) is the best characterized member of this family, it is increasingly recognized that carbon monoxide (CO) and hydrogen sulfide (H2S) also exert a regulatory role.3 In mammalian cells, H2S is mainly generated through a trans-sulfuration4 pathway that is initiated by the conversion of homocysteine into cystathionine5 by cystathionine β-synthase (CBS). Cystathionine is then converted into L-cysteine and 2-ketobutyrate by cystathionine γ-lyase (CSE).6 CBS and CSE are also required for the metabolism of sulfur-containing amino acids, as well as for the production of H2S from L-cysteine.1–4 The expression of CBS and CSE has been identified in many human tissues, including liver, kidney, brain, skin fibroblasts, and blood lymphocytes.1 Current data suggest that CBS is the predominant H2S-generating enzyme in the brain and nervous system,2 whereas CSE is the only H2S-generating enzyme identified in the vascular system.4, 6
In the cardiovascular system, H2S reduces systemic blood pressure and causes vasodilation of rat mesenteric artery, aorta, and portal vein.7–10 Consistent with these findings, a reduced expression/activity of CSE coupled with a decrease in plasma H2S concentration contributes to the pathophysiology of pulmonary hypertension in rodents.9 In addition, CBS deficiency leads to hyperhomocyst(e)inemia, a condition that includes elevated serum levels of homocysteine, homocystine, or homocysteine-mixed disulfides, and associates with increased blood pressure and endothelial dysfunction.11 In rodents, CBS/CSE deficiency induced by genetic deletion12 or chronic treatment with DL-propargylglycine13 results in a severe endothelial dysfunction and hypertension.
Portal hypertension is the main complication of cirrhotic liver, which is characterized by increased intrahepatic vasoconstrictive response and a defect in endothelial NO synthase (eNOS) function in sinusoidal endothelial cells (SEC).14, 15 The regulation of sinusoidal resistance is dependent on the contraction of hepatic stellate cells (HSCs) around SEC.16 Substantial evidence shows that a dynamic component, caused by an increased production of vasoconstrictors17 and a decreased production of vasodilators (e.g., NO),14, 15 plays a mechanistic role in increasing intrahepatic resistance.15
Human studies have demonstrated that L-cysteine metabolism is altered in liver cirrhosis and that this defect might be ascribed, at least partially, to a reduction in the expression/activity of the main genes involved in its metabolism.18–22 However, whether the CBS/CSE/H2S pathway contributes to the pathophysiology of portal hypertension is still unknown. The purpose of this study was to assess the role of H2S in the regulation of intrahepatic microcirculation in normal livers and in livers with cirrhosis.
Materials and Methods
NaHS, L-cysteine, glibenclamide, DL-propargylglycine, β-cianoalanine, and NG-nitro-L-arginine methyl ester (L-NAME) were from Sigma Chemical Co. (St. Louis, MO).
All studies were approved by the Animal Study Committee of the University of Perugia. Male Wistar rats (225-275 g) were obtained from Charles River Breeding Laboratories (Monza, Italy) and maintained on standard laboratory rat chow on a 12-hour light/dark cycle. Bile duct ligation (BDL) was performed as previously described.23, 24 All perfusion experiments were performed 4 weeks after BDL. The presence of liver cirrhosis was assessed by histology and Sirius red staining of hepatic collagen24 and by determining the body weight/liver weight ratio. Normal rats and rats with cirrhosis were treated daily with intraperitoneal administration of NaHS (56 μmol/kg), L-cysteine (100 μmol/kg) or vehicle for 5 days. At the end of the treatment, analysis of hepatic vascular responses was performed in normal rats and in rats with cirrhosis, using the isolated perused rat liver preparation.25, 26 The liver was perfused in a recirculating mode with Krebs solution equilibrated with carbogen gas, by using a peristaltic pump as described.26 The perfusion pressure was continuously monitored and recorded with a strain-gauge transducer connected with PowerLab PC (A.D. Instruments, Milford, MA). The preparation was allowed to stabilize for 20 minutes. The global viability of livers was assessed by standard criteria: gross appearance, stable pH of the perfusate, stable perfusion pressure for 20 minutes, and bile flow of >1 μL/min per gram liver. The flow rate during each individual perfusion was maintained at a constant rate of 20 mL/min. Two additional groups of normal and cirrhosis rats were sacrificied under pentobarbital sodium anesthesia (50 mg/kg intraperitoneally) and terminally bled via cardiac puncture. The blood was centrifuged at 7,250g for 20 minutes at 4°C, and the resultant serum was stored at −20°C until analysis. Liver specimens were snap frozen in liquid nitrogen and stored at −70°C.
The dose–response curve was performed by using cumulative doses of norepinephrine (NE) from 10 nmol/L to 10 μmol/L, with a 5-minute interval between them.27 To study the vasoactive effect of endogenous and exogenous H2S, NaHS or L-cysteine (1 mmol/L) were added to the perfusion medium, and the preparation was allowed to stabilize for 20 minutes before NE administration. To investigate the role of adenosine triphosphate–dependent potassium (KATP) channel in the vasodilatory effects of H2S, glibenclamide (1 mmol/L), a KATP channel blocker, or vehicle, was added to the perfusion medium in presence of NaHS (1 mmol/L) 20 minutes before NE administration. In a separate set of experiments, the perfusate was increased from 20 to 60 mL/min over 25 minutes with continuous perfusion pressure monitoring in the presence of DL-propargylglycine, a selective CSE inhibitor, or vehicle. Liver H2S generation was measured in perfusate samples (200 μL) obtained immediately before and after addition of DL-propargylglycine.
Carbon Tetrachloride (CCl4)-Induced Cirrhosis.
Cirrhosis was induced by administering phenobarbital sodium (35 mg/dL) to the rats with drinking water for 3 days, followed by intraperitoneal injection of 100 μL/100 g body weight of CCl4 in an equal volume of paraffin oil twice a week for 12 weeks.25
Measurement of H2S Concentration in Plasma, Perfusion Media, Liver Homogenate, and Cell Lysates.
To measure H2S, plasma or perfusate samples (250 μL) were added to ice-cold 250 μL NaOH, 5N in a sealed 3-neck reactor.28 The reactor was heated at 37°C and H2S extracted by adding 1 mL 10% trichloroacetic acid. To measure H2S generation in tissues, liver (100-150 mg) and cell samples were homogenized in 1 mL ice-cold tissue protein extraction reagent, and 2 mL of the reaction mixture was introduced in the reactor. The mixture contained 10 mmol/L L-cysteine, 2 mmol/L pyridoxal 5′-phosphate, 100 mmol/L potassium phosphate buffer (pH = 7.4), and 10% (wt/vol) liver homogenates or cell lysates. Calmodulin and calcium chloride were added to a final reaction mixture at concentrations 9.6 μmol/L and 0.6 mmol/L. In same experiments DL-propargylglycine (2 mmol/L) was added to liver homogenates for 5 minutes at 37°C before starting the enzyme reaction. The reactions were initiated by transferring the reactor from the ice bath to a 37°C water bath. A constant stream of nitrogen was passed through the mixture via gas-inlet capillary. The stream of nitrogen carried the sulfide acid in another reactor containing 2 mL sulfide anti-oxidant buffer solution, consisting of 2 mol/L KOH, 1 mol/L salicylic acid, and 0.22 mol/L ascorbic acid at pH 12.8. After 30 minutes, the sulfide anti-oxidant buffer solution was removed and the sulfide concentration measured by a sulfide-sensitive electrode (Model 9616 S2−/Ag+ electrode, Orion Research, Beverly, MA28).
Quantitative Reverse Transciption–Polymerase Chain Reaction of CSE and CBS.
Total RNA was isolated from normal livers and livers with cirrhosis, primary rat liver cell types, and from human liver endothelial cells by using the TRIzol reagent according to the manufacturer's specifications (Invitrogen, Milan, Italy). Livers with cirrhosis were obtained 4 weeks after BDL. RNA was processed directly to complementary DNA by reverse transcription with Superscript III (Invitrogen) as described.24 All polymerase chain reaction (PCR) primers for quantitative and qualitative PCR24 were designed by using the PRIMER3-NEW software and published sequence from the National Center for Biotechnology Information (NCBI) database. Primers (MWG BIOTECH) were as follow (sense and anti-sense): rat GAPDH: ATGACTCTACCCACGGCAAG and TACTCAGCACCAGCATCACC; rat CBS: CCAGGAC- TTGGAGGTACAGC and TCGGCACTGTGTGGTAATGT; rat CSE: GTATTGAGGCACCAACAGGT and GTTGGGTTTGTGGGTGTTTC; human GAPDH: GAAGGTGAAGGTCGGAGT and CATGGGTGGA- ATCATATTGGAA; human CBS: TCGTGATGCCAGAGAAGATG and TTGGGGATTTCGTTCTTCAG; human CSE: CACTGTCCACCACGTTCAAG and GTGGCTGCTAAACCTGAAGC.
Western Blot Analysis of CSE Protein Expression in Liver Tissue.
For Western blot analysis, 15 mg liver proteins (normal and cirrhotic rats, 4 weeks after BDL) were subjected to 12% SDS-PAGE and electrophoretically transferred to Immobilon-P polyvinylidene difluoride membranes (Millipore Corp., Bedford, MA). Monoclonal anti-CSE antibody, provided by Dr. N. Nishi (Kagawa Medical School, Japan), was used at 1:3,000 dilution overnight at 4°C with shaking.29 The anti-immunoglobulin G horseradish peroxidase conjugate (Bio-Rad, Hercules, CA) was used as the secondary antibody, and specific protein bands were visualized by using enhanced chemiluminescence (Amersham Biotechnology Pharmacia, Piscataway, NJ), following the manufacturer's suggested protocol.24
In Vitro Studies: Isolation and Culture of HSC, Hepatocytes, and SEC.
Human liver endothelial cells (LEC) were from ScienceCell (Seattle, WA). Rat hepatocytes and nonparenchymal cells were isolated as previously described.30, 31 HSCs were cultured on plastic plates in Dulbecco's modified Eagle medium containing 20% fetal calf serum (FCS). SEC were isolated as described elsewhere.32
HSC contraction was assessed as previously described.24, 33 Cell contraction was induced by adding FCS 20% to the HSC monolayers.24 To assess whether H2S or L-cysteine inhibited FCS-induced contraction, NaHS (from 10 μmol/L to 1 mmol/L) or L-cysteine (1 mmol/L) alone or in combination with glibenclamide (100 μmol/L) were added to the HSC monolayers. To allow contraction, lattices were detached by gentle circumferential dislodgment with a 200-μL micropipette tip, and contraction measured by monitoring the change in lattice area over 18 hours.
All values were presented as means ± SD. Statistical tests were performed using INSTAT statistical software (Graphpad Software Inc., San Diego, CA). Differences among the experimental groups were determined by ANOVA followed by the Dunnett's test or Bonferroni's test when appropriate.
H2S Reduces Portal Perfusion Pressure Response to NE.
As shown in the representative ethidium-stained agarose gel in Fig. 1A, we found that both enzymes involved in the synthesis of H2S, CSE and CBS, are detected in normal rat liver by reverse transcription–polymerase chain reaction (RT-PCR). Additionally, these enzymes are catalytically active, generating approximately 10 nmol/min/g of tissue of gaseous H2S in response to L-cysteine, an effect that was inhibited by DL-propargylglycine and β-cyanoalanine, two selective CSE inhibitors (Fig. 1B, n = 6; P < .05).
We examined whether H2S modulates intrahepatic resistance in normal rats. Concentration–response curves to NE were evaluated at a constant perfusion flow rate of 20 mL/min. As shown in Fig. 2A, addition of NE to the perfusate produced a dose-dependent increase in portal perfusion pressure with a median effective concentration (EC50) of 1 μmol/L (n = 7). In contrast, pre-perfusing the liver with 1 mmol/L NaHS or L-cysteine significantly attenuated vasoconstriction induced by NE (Fig. 2A, n = 7; P < .05 NE + H2S or NE + L-cysteine vs. NE alone by ANOVA followed by Dunnett's test). To investigate whether the KATP channels were targets of H2S, we first perfused the liver for 20 minutes with NaHS with or without glibenclamide (a KATP channel blocker) and then with NE. Although glibenclamide alone (1 mmol/L) failed to modulate the intrahepatic resistance in liver perfused with NE alone (data not shown), it reversed the vasodilation caused by NaHS, highlighting the role of KATP channels in the vasodilatory activity of H2S (Fig. 2B; n = 6 per group, P < .05 vs. control and P < .05 H2S alone vs. H2S + glibenclamide).
Modulation of Intrahepatic Resistance by H2S Is NO Independent.
The interplay between NO and H2S was investigated by perfusing the liver after inhibition of eNOS activity with L-NAME (1 mmol/L) a nonselective eNOS inhibitor. L-NAME was added to the perfusate 20 minutes before performing the perfusion pressure response to NE. In the NE preconstricted livers, L-NAME failed to reduce vasodilation induced by H2S (1 mmol/L), indicating that the vasorelaxant effect of H2S is not mediated by NO at least in this experimental setting (Fig. 3A, n = 7; P > .05). Furthermore, to clarify whether H2S production was regulated by shear stress, an isolated perfused rat liver model with combined high flow rate (20, 30, 40 mL/min) and inhibition of endogeneus H2S biosynthesis with DL-propargylglycine was used. In this experimental setting, incremental increases in flow rate heighten the perfusion pressure and simultaneously activate SEC by shear stress. Compared with control livers (n = 5), studied in parallel, addition of DL-propargylglycine, 1 mmol/L, failed to increase intrahepatic resistance, suggesting that the inhibition of endogenous production of H2S does not modify the shear stress–induced response (Fig. 3B-C; n = 5; P > .05). In addition, the H2S released by normal rat livers was not modified by the increased shear stress, suggesting that sites other than SEC are involved in the synthesis of H2S by the liver. By contrast, H2S release was significantly reduced by DL-propargylglycine (Fig. 3C; n = 5; P < .05). Because these data suggest that SEC are not involved in H2S regulation, the differential expression of CSE/CBS between parenchymal and nonparenchymal cells (SEC and HSC) was investigated. Consistent with previous findings,6 we found that hepatocytes express both CBS and CSE mRNA. As shown in Fig. 3D, expression of CSE, mRNA, and protein was found in HSCs but not in primary rat SEC nor in passaged liver endothelial cells obtained from human liver (LEC). CBS expression was not found in HSCs, SEC, or LEC.
Effects of Endogenous and Exogeneus H2S on the Portal Perfusion Pressure Response to NE in Rat Liver With Cirrhosis.
We then investigated whether H2S modulates hepatic resistance in BDL rats. In livers with cirrhosis, NE produced a dose-dependent increase in portal perfusion pressure compared with sham rats (Fig. 4A, n = 6; P < .05). The livers with cirrhosis showed a significantly increased sensitivity to the vasocontricting activity of NE (Fig. 4A) with an EC50 of 30 nmol/L (n = 6; P < .05 vs. sham operated). Furthermore, although perfusion of livers with cirrhosis with 1 μmol/L of NE elicited maximal vasoconstriction, in the sham-operated rats, the maximal vasoconstrictor response was observed at 30 μmol/L (n = 6; P < .01). The addition of H2S reduced the hyper-responsiveness of livers with cirrhosis to NE, and the vasorelaxant effect was evident at concentrations of NE ranging from 0.1 to 30 μmol/L (Fig. 4A; n = 6; P < .05). In contrast to NaHS, L-cysteine was devoid of any vasoactivity in rats with cirrhosis (Fig. 4B; n = 6; P < .05). To investigate whether H2S exerts a direct vasodilating activity, rats with cirrhosis were treated with NaHS (56 μmol/kg) intraperitoneally for 5 days, and then the isolated livers were perfused with NaHS 1 mmol/L for 20 minutes before performing high flow rate pressure–response curve. Data shown in Fig. 4C demonstrate that perfusion of livers with cirrhosis with H2S significantly shifted the pressure–flow curve to the right, thereby markedly correcting the enhanced vasoconstrictive responses we evidenced in the BDL livers (n = 6; P < .05).
Reduced Hepatic H2S Production Is Due to a Reduced Expression/Activity of CSE in the Liver With Cirrhosis.
Liver obtained 4 weeks after BDL showed histological features of bile duct proliferation and extensive fibrosis as shown by liver histology and Sirius red staining of hepatic collagen (Fig. 5A-D) and increased liver/body weight ratio: from 2.5 ± 0.5 to 7.2 ± 0.6 (n = 6; P < .05). These changed associated with reduced H2S levels in serum, liver homogenates, and liver perfusates (Fig. 5E-F; n = 7; P < .05). As shown in Fig. 6A-C, analysis of CBS and CSE transcripts by qRT-PCR demonstrates that CSE mRNA expression, but not CBS, was significantly decreased in BDL rats compared with sham-operated animals (−50%; P < .05 vs. Sham-operated rats). This was further confirmed by western blot analysis of CSE (Fig. 6D) and the densitometric analysis of CSE protein levels demonstrating approximately 60% reduction of CSE protein expression in the liver of BDL rats (Fig. 6E, n = 6; P < .05 vs. sham-operated). Similarly to the BDL model, we found that CSE expression, mRNA and protein, and H2S generation were significantly reduced in rats administered CCL4 (n = 6; P < .05 vs. control rats), whereas expression of CBS mRNA was unchanged (Fig. 7).
CSE Expression Is Downregulated in Activated HSCs.
To investigate the involvement of CSE in the pathogenesis of portal hypertension, the modulation CSE gene expression was assessed in HSCs isolated from normal livers and livers with cirrhosis. The qRT-PCR analysis and Western blot analysis (Fig. 8A) shows that, in comparison with HSCs isolated from sham-operated rats (day 1), HSCs isolated from BDL rats have a decreased expression of CSE, mRNA, and protein (n = 5; P < .05). This correlates with a decreased ability of activated HSCs to release H2S in the presence of L-cysteine (Fig. 8B).
We then assessed whether H2S directly modulates HSCs contractility in vitro. HSCs isolated from normal rats that had undergone spontaneous activation on collagen lattices were exposed to NaHS and L-cysteine. As shown in Fig. 8C, contraction of culture-activated HSCs was significantly decreased by H2S (100 μmol/L) compared with controls (n = 5; P < .05 vs. 20% FCS alone). Additionally, glibenclamide (100 μmol/L) almost completely abrogated the relaxation induced by H2S (Fig. 8D; n = 5; P < .05 vs. 20% FCS + NaHS), In contrast to H2S, L-cysteine (100 μmol/L) failed to inhibit HSCs contraction in BDL-derived HSCs, indicating that HSCs generated from rats with cirrhosis are unable to directly metabolize L-cysteine because of the reduced expression/activity of CSE (Fig. 8E; n = 5; P > .05 vs. 20% FCS alone).
We have provided evidence that a dysregulated production of H2S represents a novel pathway responsible for the intrahepatic vasoregulatory defects observed in the liver with cirrhosis. Previous studies have shown that H2S causes a direct relaxation of portal vein,10 suggesting a hypothetical role for this gaseous mediator in regulating resistance of intrahepatic microcirculation. We demonstrated that exposure to H2S and L-cysteine resulted in a significant reduction of vasoconstriction caused by NE. The fact that in the normal liver the same reduction of intrahepatic resistance was obtained with L-cysteine and NaHS (i.e., endogenous and exogenous source of H2S) suggests that endogenously generated H2S is a physiologically relevant mediator of intrahepatic resistance.
Zhao et al.10 have reported that glibenclamide antagonizes the vasorelaxant effect of H2S in the NE-precontracted rat aorta, leading to the suggestion that H2S opens KATP channels in vascular smooth muscle cells to bring about smooth muscle relaxation. Similarly to these in vitro findings, we found that the H2S-induced decrease of intrahepatic resistance was antagonized by glibenclamide, indicating that the hypotensive effect of H2S is likely mediated by the relaxation of resistance blood vessels of the hepatic microcirculation, through the opening of KATP channels. These findings are consistent with the observation that KATP channels regulate intrahepatic resistance in extrahepatic models of portal hypertension.34, 35
NO is a major regulator of intrahepatic resistance in normal rat liver and modulates basal vascular tone as well as vascular responses to vasoconstrictors.14, 15 Because of the physiological relevance of NO in portal circulation, a number of experiments were carried out in an attempt to determine the role of endothelium-released NO in the H2S-induced vasodilation of the intrahepatic circulation. However, results of these experiments seem to exclude a major contribution of NO to vasodilation induced by H2S. Support to this concept comes from the observation that exposure to L-NAME failed to reverse the vasodilatory effect exerted by H2S on vasocontraction induced by NE. Further information on the reciprocal of regulation of NO and H2S have been obtained in the isolated perfused rat liver model in which CSE inhibition with DL-propargylglycin was combined with a high flow rate of perfusion. This model allowed us to investigate the hemodynamic effect of the concomitant reduction of the endogenous of H2S and the increased generation of NO caused by shear stress. Despite the fact that DL-propargylglycine reduced H2S generation, it failed to exert any hemodynamic effect in the liver microcirculation stressed by high flow rates, suggesting that H2S does not influence the endogenous release of NO in the normal rat liver. In parallel, the increased shear stress failed to stimulate release of H2S. Based on these experimental observations, it appears that NO and H2S are released by different cellular sources and that their hemodynamic effects involve different cellular target(s). Consistent with this view, we demonstrated that H2S exerted the same portal pressure–lowering effect in normal rats and rats with cirrhosis, which are characterized by endothelial dysfunction of intrahepatic circulation.14, 15 An important observation of this study was the demonstration that in contrast to NaHS, which exerts a direct vasodilatory effect, L-cysteine failed to reverse vasoconstriction caused by NE in rats with cirrhosis, suggesting that an impaired activity of the hepatic enzyme(s) involved in the H2S release is operating in the liver of rats with cirrhosis. Here we have provided evidence that a defective generation of H2S is present in BDL rats. This defect is mostly related to a dysfunction of the CBS/CSE system. Supporting this view, we demonstrated that BDL causes a 50% to 60% reduction in the liver expression of CSE, mRNA, and protein, which associates with an approximately 60% reduction in the ability of BDL liver homogenates to generate H2S. The reduction of H2S generation is likely to contribute to portal hypertension, because its replacement reverses enhanced vasoconstriction caused by NE.
Little is known about the mechanism responsible for the reduced expression of hepatic CSE in the liver with cirrhosis. Previous studies have shown that CSE expression is modulated during development, being detected at very low levels in embryos and followed by gradual increases after birth; thereafter, it remains constant in adulthood and decreases with aging.6 Even if the reason for these changes is unclear, it can be speculated that genes involved in differentiation or proliferation may control the CSE expression. Previous studies have documented that BDL leads to decreased expression of the mRNAs encoding CSE,36 and this event has been has been linked to oxidative stress. The human and rodent CSE gene has recently been cloned.6 The core promoter of the gene, located in the 5′-flanking region proximal to the transcriptional start site, contains several putative transcriptional factor-binding sites, including MZF-1 and Sp1.6 Sp1, a member of the Sp/krüppel-like factor family,37 is a ubiquitously expressed transcription factor that recognizes GC-rich sequences present in regulatory sequences of numerous housekeeping genes and genes involved in growth regulation, and is a protein target for free radical elements.38–40 Thus, Sp1 DNA binding activity is modulated by oxidative stress38–40 and increased free radicals formation associates with reduced expression of CSE.40
The vasorelaxation induced by H2S comprises a minor endothelium-dependent effect and a major direct effect on smooth muscles.1–3, 10 Moreover, unlike NO, which is produced from endothelial cells, H2S-generating enzymes are not expressed in vascular endothelium.1–3, 9, 10 Here we have accurately localized all potential cellular sources of H2S in liver cell populations, performing an RT-PCR of both CBS and CSE in freshly isolated hepatocytes, HSCs and SECs. Of the cell types studied, SECs express neither CSE nor CBS, but given the similarities between SECs and vascular endothelial cells, this result was expected.10 In contrast, we found that HSCs isolated from control rat livers and rat livers with cirrhosis express CSE but not CBS. Moreover, HSCs obtained from livers with cirrhosis had a significant lower expression of CSE than cells isolated from control livers, suggesting that synthesis of CSE and H2S production are transcriptionally regulated during HSC differentiation and proliferation. Our data also suggest that CSE downregulation in HSCs plays a major role in the development of portal hypertension in the BDL model. This observation was confirmed in vitro, using the collagen lattice assay. In this experimental setting, we have shown that H2S, but not L-cysteine, attenuated contraction of BDL-derived HSCs induced by 20% FCS. These findings support the hypothesis that reduced activity of CSE enhances the contraction of activated HSCs around the sinusoids.41, 42 In aggregate, these data suggest that the decreased expression of CSE in livers with cirrhosis may be another, previously unrecognized, factor contributing to the hyper-response of the vasculature in liver with cirrhosis and that H2S is part of autocrine loop that limits HSC hyperresponsiveness.
In conclusion, this study provides evidence that H2S is an autocrine mediator involved in regulating HSCs contraction and that a decreased expression of CSE in HSCs may be responsible for the increased intrahepatic resistance in rodent models of liver cirrhosis. These findings provide a new mechanism contributing to portal hypertension in cirrhosis and establishes the rationale for development of new therapeutic strategies for this condition.