Hepatic blood supply is uniquely regulated by the hepatic arterial buffer response (HABR), counteracting alterations of portal venous blood flow by flow changes of the hepatic artery. Hydrogen sulfide (H2S) has been recognized as a novel signaling molecule with vasoactive properties. However, the contribution of H2S in mediating the HABR is not yet studied. In pentobarbital-anesthetized and laparotomized rats, flow probes around the portal vein and hepatic artery allowed for assessment of the portal venous (PVBF) and hepatic arterial blood flow (HABF) under baseline conditions and stepwise reduction of PVBF for induction of HABR. Animals received either the H2S donor Na2S, DL-propargylglycine as inhibitor of the H2S synthesizing enzyme cystathionine-γ-lyase (CSE), or saline alone. Additionally, animals were treated with Na2S and the ATP-sensitive potassium channel (KATP) inhibitor glibenclamide or with glibenclamide alone. Na2S markedly increased the buffer capacity to 27.4 ± 3.0% (P < 0.05 vs. controls: 15.5 ± 1.7%), whereas blockade of H2S formation by DL-propargylglycine significantly reduced the buffer capacity (8.5 ± 1.4%). Glibenclamide completely reversed the H2S-induced increase of buffer capacity to the control level. By means of RT-PCR, Western blot analysis, and immunohistochemistry, we observed the expression of both H2S synthesizing enzymes (CSE and cystathionine-β-synthase) in aorta, vena cava, hepatic artery, and portal vein, as well as in hepatic parenchymal tissue. Terminal branches of the hepatic afferent vessels expressed only CSE. We show for the first time that CSE-derived H2S contributes to HABR and partly mediates vasorelaxation of the hepatic artery via activation of KATP channels.
- buffer capacity
- gaseous molecules
- ultrasonic flowmetry
- ATP-sensitive potassium channels
perfusion of the liver is unique because of its dual blood supply via the portal vein (PV) and the hepatic artery (HA). Although the PV contributes to 70–80%, the HA covers about 20–30% of the total hepatic blood flow (6). Almost three decades ago, Lautt and colleagues (14) described an intimate relationship between PV and HA, the so-called hepatic arterial buffer response (HABR), which comprises the ability of the HA to produce reciprocal compensatory flow changes in response to changes of PV flow. This regulatory mechanism serves not only to fulfill oxygen and metabolic demands of the liver (17, 27) but also to control overall metabolic well being of the organism by maintaining the hepatic clearance and excretory function (11, 12).
HABR is specific to the unique vascular bed of the liver and is apparently regulated by adenosine (5, 10, 13, 16, 27). Adenosine is thought to be secreted at a constant rate into the fluid space of Mall, and the concentration of adenosine is regulated by washout into the portal venules. In case of reduced PV flow, the washout of adenosine is reduced, and accumulation of adenosine causes dilation of the HA, thus buffering the PV flow change (13). Beside adenosine, regulation of the hepatic vascular resistance has been shown to further involve the gaseous inorganic compounds nitric oxide (NO) and carbon monoxide (CO). Whereas NO serves as a potent vasodilator in the HA circulation and exerts only a minor vasodilatory effect in the PV vascular bed, CO acts to maintain PV vascular tone in a relaxed state and exerts no vasodilation in the HA (24). Recently, a third gaseous mediator, hydrogen sulfide (H2S), has been recognized as an important endogenous vasodilator and neuromodulator (32). H2S is synthesized through degradation of cysteine by cystathionine-γ-lyase (CSE) or cystathionine-β-synthase (CBS) (7, 15, 18, 20, 29, 31). Both enzymes were found to be expressed in many mammalian tissues, including the liver (31). In the vascular system, however, CSE is suggested to be the only H2S-generating enzyme (8, 30). Hereby, vasorelaxation by H2S is shown to be mediated by ATP-sensitive potassium channels (KATP) since effects of H2S are mimicked by KATP openers and abolished by their inhibitors such as glibenclamide (31–33).
With respect to these vasodilatory properties of H2S, it was the purpose of the present study to examine whether H2S contributes to the mediation of HABR. Therefore, buffer capacity was studied in a rat model upon both exogenous application of H2S and inhibition of the H2S-producing enzyme CSE. Furthermore, blockade of KATP channels by glibenclamide served to unravel the possible mode of action of H2S. Moreover, the expression of the H2S synthesizing enzymes CSE and CBS was analyzed in vascular and liver tissues.
MATERIALS AND METHODS
All drugs used in the present study were purchased from Sigma (Deisenhofen, Germany) if not stated differently.
Anesthesia and monitoring.
Upon approval by the local animal committee (LALLF M-V:LVL-MV/TSD/7221.3-023/05), the experiments were conducted in accordance with the German legislation on protection of animals and the NIH Guide for the Care and Use of Laboratory Animals [DHEW Publication No. (NIH) 86–23, Revised 1985]. Male Wistar rats (body wt 250–400 g; Charles River Laboratories, Sulzfeld, Germany) were used for the experiments. Animals were housed in standard animal laboratories with a 12-h:12-h light/dark cycle and had free access to water and standard laboratory chow ad libitum. Anesthesia was induced by intraperitoneal injection of pentobarbital sodium (50 mg/kg body wt). Supplemental doses (25 mg/kg body wt ip) were given during the experiment if requested to maintain sufficient anesthetic depth, which was determined by the absence of changes in blood pressure and heart rate on intermittent tail clamp. Anesthetized animals were placed in supine position on a heating pad for maintenance of body temperature (36–37°C) and were tracheotomized to facilitate spontaneous respiration (room air). Polyethylene catheters (PE 50, ID 0.58 mm; Portex, Hythe, UK) were inserted into the right carotid artery and jugular vein for assessment of central hemodynamics, blood sampling, and intravenous application of drugs. Throughout the experiments, mean arterial blood pressure (MAP) and heart rate were continuously monitored.
After laparotomy, microsurgical preparation for assessment of liver blood flow was performed. In brief, ultrasonic flow probes were placed around the HA (0.5V; Transonic Systems, Ithaca, NY) and the PV (1.5R; Transonic Systems) and were connected to a flowmeter (T402 Animal Research Flowmeter, Transonic Systems) for continuous monitoring of HA and PV blood flow values. A tourniquet loop was placed around the superior mesenteric artery (SMA) for reduction of blood flow using a micromanipulator-controlled constrictor. For assessment of hepatic tissue oxygenation, a flexible polyethylene microcatheter Clark type Po2 probe (diameter, 470 μm; length, 300 mm) (LICOX System; GMS, Kiel-Melkendorf, Germany) was placed between the surface of two adjacent liver lobes and fixed with histoacryl glue (B. Braun, Melsungen, Germany). This allowed the probe to integrate local tissue Po2 values over the tissue area in contact with the 5-mm long Po2-sensitive area near the catheter tip without interference of ambient air. Online temperature compensation was performed by an additional temperature probe (type K thermocouple probe, LICOX System), which was also positioned and fixed between two adjacent lobes. When all surgical procedures were completed, the intestine was covered with moist cloths to minimize heat loss and drying, and the animals were allowed to recover from surgical stress for 30 min.
Experimental groups and protocol.
The animals were allocated to the following five experimental groups (n = 10 each): one group of animals received a continuous infusion of sodium sulfide (Na2S, 150 μmol·kg−1·h−1 iv) as donor of H2S. For inhibition of the H2S-synthesizing enzyme CSE, animals of the second group received DL-propargylglycine (PAG, 100 mg/kg iv) as single bolus injection, followed by a continuous infusion of isotonic saline (3.1 ml/h iv). For the blockade of KATP channels, animals of the third and fourth group received glibenclamide (GLB) (KATP channel inhibitor) (40 mg/kg ip) followed by a continuous infusion either of sodium sulfide (GLB + Na2S) or of isotonic saline (GLB) respectively. Animals that received equivalent volumes of isotonic saline only served as controls (control).
Hepatic hemodynamics, including the hepatic arterial blood flow (HABF) and portal venous blood flow (PVBF) as well as the hepatic tissue Po2 were assessed at the following steps: 1) before drug application (pre), 2) 30 min after drug application immediately before SMA tourniquet (post), and 3) upon maximal reduction of the PVBF by complete tourniquet of the SMA, i.e., the time point at which the buffer response was quantified (SMA occlusion). The reduction of PVBF was kept constant over a time period of about 10 min followed by data assessment. In some experiments, reduction of PVBF for induction of HABR was repeated twice, allowing sufficient recovery times (∼15 min) between the individual measurements for regaining baseline conditions. Arterial blood samples for blood gas analysis were taken at all time points described above. At the end of the experiment, animals were euthanized with an overdose of intravenous anesthesia.
Quantification of the HABR and HA conductance.
In addition to the assessment of HABF and PVBF as absolute values (ml/min), we calculated the total hepatic blood flow as the sum of HABF and PVBF. Furthermore, we determined the buffer capacity as change of HABF divided by the change of PVBF × 100 (26), as well as the HA conductance calculated as HABF per kilogram of body weight divided by MAP (ml·min−1·kg−1·mmHg−1) (22).
Total RNA was isolated from liver and vascular tissues using a RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. RNA concentration was determined spectrophotometrically. First-strand cDNA was synthesized by reverse transcription of 2 μg of total RNA using oligo(dT)18 primer (Biolabs, Frankfurt am Main, Germany) and Superscript II RNaseH-Reverse Transcriptase (Invitrogen, Karlsruhe, Germany) in the presence of dNTPs, 5× first-strand buffer, and dithiolthreitol at 72°C for 10 min and 42°C for 60 min. The reverse transcriptase was inactivated by 95°C for 5 min. Rat CSE and CBS were amplified by 24 cycles of PCR consisting of 94°C (30 s) for denaturation, 58°C (CSE) or 56°C (CBS) (30 s) for primer-specific annealing, and 72°C (30 s) for extension using Taq polymerase (Amersham Bioscience, Piscataway, NJ). For the detection of CSE mRNA, the forward primer sequence used was 5′-CAT GGA TGA AGT GTA TGG AGG C-3′, and the reverse primer sequence was 5′-CGA TTG TTA CCT CTG CTG CCG-3′. The PCR product size was 445 bp. For the detection of CBS mRNA, the forward primer sequence used was 5′-GAT GAC ACC GCA GAG GAG AT-3′, and the reverse primer sequence was 5′-CGG GAT CTA CAC CGA TGA TT-3′. The PCR product size was 150 bp. The reaction without RNA template (nontemplate control) was also performed as a negative control. In a comparable assay, the RNA integrity and cDNA synthesis was tested using rat GAPDH as a housekeeping gene and the following primers: 5′-AAC GAC CCC TTC ATT GAC-3′ and 5′-TCC ACG ACA TAC TCA GCA C-3′. PCR products were separated by electrophoresis on 2.0% agarose gels. Ethidium bromide-stained bands were visualized by UV illumination and densitometrically semiquantified (TotalLab; Nonlinear Dynamics, New Castle upon Tyne, UK). The data represent expression of CSE and CBS gene product in relation to that of GAPDH.
Western blot analysis.
For Western blot analysis of CSE and CBS, liver tissue was homogenized in lysis buffer (1 M Tris pH 7.5, 5 M NaCl, 250 mM EDTA, 10% Triton-X 100, 4% NaN3, and 100 mM PMSF), incubated for 30 min on ice, and centrifuged for 15 min at 10,000 g. Before use, all buffers received a protease inhibitor cocktail (1:100 vol/vol, Sigma). Protein concentrations were determined using the bicinchoninic acid (BCA) protein assay (Pierce, Rockford, IL) with bovine serum albumin as standard. Equal amounts of whole protein extracts (20 μg) were separated discontinuously on 12% SDS-PAGE gels and transferred to a polyvinyl difluoride membrane (Immobilon-P transfer membrane; Millipore, Billerica, MA). After blockade of nonspecific binding sites, membranes were incubated for 2 h at room temperature with mouse monoclonal anti-CSE (1:1,000; ABNOVA, Taipei, Taiwan) or goat polyclonal anti-CBS (1:2,000, Santa Cruz Biotechnology, Europe), followed by peroxidase-conjugated goat anti-mouse IgG antibody for CSE (1:2,000; LSAB+Systems-HRP; Dako, Hamburg Germany) and donkey anti-goat IgG antibody for CBS (1:5,000, Santa Cruz Biotechnology, Europe) as secondary antibodies. Protein expression was visualized by means of luminol enhanced chemiluminescence (ECL plus; Amersham Pharmacia Biotech, Freiburg, Germany) and exposure of the membrane to a blue light-sensitive autoradiography film (Kodak BioMax Light Film; Kodak-Industrie, Chalon-sur-Saone, France). Signals were densitometrically assessed (TotalLab) and normalized to the β-actin signals (monoclonal mouse anti-β-actin antibody, 1:20,000, Sigma Aldrich) followed by peroxidase-conjugated rabbit anti-mouse IgG antibody for β-actin (1:60,000, Sigma Aldrich).
Liver and vascular tissue (PV, HA, aorta, and vena cava) was fixed in 4% phosphate-buffered formalin for 2–3 days and embedded in paraffin. For immunohistochemical demonstration of CSE and CBS, sections collected on poly-l-lysine-coated glass slides were treated by microwave for antigen unmasking. Mouse monoclonal anti-CSE antibody (1:500, ABNOVA) and goat polyclonal anti-CBS antibody (1:100, Santa Cruz Biotechnology, Europe) were used as primary antibodies and incubated for 18 h at 4°C. After being equilibrated to room temperature, sections were incubated with horseradish peroxidase-conjugated secondary antibodies (LSAB+Systems-HRP, Dako). 3-Amino-9-ethylcarbazole was used as chromogen (Dako). The sections were then counterstained with hemalaun and examined by light microscopy (Axioskop 40; Zeiss, Göttingen, Germany).
After being tested for normality and equal variance across groups, differences between groups were assessed using one-way ANOVA test. A P level of <0.05 was considered significant. All data are given as means ± SE. Analysis was performed using the software package SigmaStat (Jandel, San Rafael, CA).
Infusion of Na2S resulted in a transient (∼60 s) decrease of MAP indicative for the vasoactive property of H2S (Fig. 1A). Astonishingly, MAP regained baseline values at latest after 2 min (Fig. 1A) although Na2S was infused continuously. All other drugs did not even transiently affect systemic hemodynamics (Table 1). Animals exhibited no signs of transient awareness attributable to insufficient anesthetic depth as assessed by negative tail clamp testing or lack of withdrawal of paw pinch. Blood gas analysis revealed no significant differences with an arterial Po2 of 95.5 ± 1.0 mmHg, Pco2 of 50.4 ± 0.5 mmHg, and pH values of 7.37 ± 0.01 among the groups studied.
The application of the respective drugs did not significantly influence the PVBF and the HABF compared with the respective values of the control animals (Fig. 2). However, it is important to mention that the initial Na2S-induced drop in blood pressure was paralleled by a transient decrease in HABF, which was followed by a transient elevation of blood flow, representing hepatic arterial vasodilation (Fig. 1C).
As expected, the reduction of PVBF initiated a pronounced HABR, indicated by a significant increase of HABF in all groups studied. The degree of PV flow reduction did not differ between the individual animals or between the groups and approximated 70% of the baseline values (Fig. 2, A–E). The maximal increase of HA inflow could be observed in the Na2S group (Fig. 2G), whereas the least pronounced HA increase occurred upon the inhibition of H2S-producing enzyme by PAG (Fig. 2H).
Analyzing the buffer capacity, livers of the Na2S-treated animals exhibited the maximal compensation among the groups (Fig. 3A). The lowest buffer capacity was observed in animals that received the CSE inhibitor PAG. GLB administration reversed the Na2S-induced increase of buffer capacity to the level found in the control group, and GLB alone did not influence buffer capacity at all. Calculating the HA conductance, highest values were found after the application of Na2S, whereas PAG administration presented with lowest HA conductance (Fig. 3B). The pretreatment with GLB followed by Na2S infusion resulted in reduction of the HA conduction to the level of the control group (Fig. 3B). The hepatic tissue oxygenation was found comparable in all five groups of animals during both baseline conditions and after the drug supplementation (22.2 ± 0.7 mmHg) but decreased upon maximal reduction of PVBF without significant differences between groups (16.1 ± 0.6 mmHg).
As illustrated in Fig. 4 and Fig. 5, mRNAs encoding for CSE and CBS were expressed in rat liver tissue. Protein of CSE and CBS could also be detected in liver tissue. The application of the respective drugs did not change the basal expression of these two H2S-synthesizing enzymes at either mRNA or protein level (Figs. 4 and 5). In addition, the mRNAs encoding for both enzymes were also detected in the vascular system, including the HA, PV, aorta, and vena cava (Fig. 6, A and E). In line with this, immunohistochemistry of PV and HA (Fig. 6, B and F), as well as of aorta and vena cava (Fig. 6, C and G), revealed positive staining for CSE and CBS. In addition, strong immunoreactivity for CSE and CBS could be observed in hepatic parenchymal tissue (Fig. 6, D and H). Of interest, terminal branches of the HA and PV within the portal triads were found positive for CSE (Fig. 6D) but not for CBS (Fig. 6H).
In the present study, we provide major evidence that H2S contributes to the HABR and partly mediates the vasodilative response of the HA. This conclusion is based on the fact that the supplementation of H2S increased HA conductance and almost doubled the buffer capacity. Vice versa inhibition of the H2S synthesis markedly decreased buffer capacity. Furthermore, we could show that vasorelaxation of HA by H2S was inhibited by application of a selective inhibitor of KATP channels. The two H2S-synthesizing enzymes CSE and CBS were found expressed in vessels of the systemic and hepatic circulation. Intrahepatic terminal branches of the HA and PV, however, lack CBS but not CSE expression.
The magnitude or efficacy of the buffer response varies depending on the technique employed and on the condition of the animal (11, 14). In contrast to previous reports of other groups, we placed the ultrasonic flow probes directly around the HA and the PV without ligation of the splenic artery, the left gastric artery, the gastroduodenal artery, and inlet arteries to the splanchnic system (10, 21, 26). In our case, the maximal SMA occlusion did not cause zero flow of PVBF and thus did not provoke maximal HABR. However, the methodology used here is less traumatic and does not require the ligation of arteries supplying the upper splanchnic area with the positive consequence that buffer response is not preactivated because this has been described by our group in a previous study upon ligation of the arteries mentioned above (27). Data of the total hepatic blood flow are in good accordance with the results of previously published studies (3, 25–27), underlining the reliability of the present methodological approach.
Mathie et al. (16) postulated that PV provides the major blood supply of oxygen to the liver and that its loss could lead to a certain degree of hepatic hypoxia. In line with the present observation that HABR could not fully compensate for reduced PV inflow, hepatic tissue oxygenation was found reduced during SMA occlusion but still in the physiological range of liver tissue Po2 (21, 27). At first sight, these findings contrast previous data of our group, demonstrating that upon reduction of PVBF hepatic tissue oxygenation was maintained (21, 27). This discrepancy might be due to the different methodologies and surgical approaches used in the present study for induction of HABR. The present technique with HA dissection and direct placement of the ultrasonic flow probe might have increased vascular tone, which would further explain the less pronounced buffer capacity compared with the respective values of one of our previous studies (26).
The used H2S donor Na2S is an easy and fast soluble substance, which dissociates to Na+ and S2−; the latter then binds H+ to form HS− as well as H2S. To inhibit endogenous production of H2S, PAG was used as an inhibitor of CSE. Though PAG and BCA are well-known CSE inhibitors (19), we preferred to use PAG. It has been reported that the reversible inhibitor BCA is less effective than the irreversible inhibitor PAG in that livers from BCA-pretreated animals exhibited less inhibition of H2S formation than livers from PAG-pretreated animals (19). Notably, in the present experiments, the supplementation of the respective drugs did not influence basal PVBF, HABF, or systemic hemodynamic parameters. Thus it is reasonable to state that the observed changes upon PV flow reduction are exclusively due to the local, i.e., hepatic action of the drugs.
HABR and H2S.
It has been shown that H2S relaxed rat thoracic aorta, PV, and mesenteric arteries in vitro (4). Moreover, an intravenous bolus injection of H2S dose dependently decreased MAP in anesthetized rats (33). In line with our present data, this H2S-induced vasodilative effect was transient in nature (33). Zhao et al. (33) further reported on a direct effect of H2S on vascular smooth muscle cells, thereby mediating vasodilatation. However, the mechanism of H2S action is not fully understood. The ability of H2S to relax vascular smooth muscle cells most likely occurs through activation of ATP-sensitive K+ channels (31). KATP channels are present in almost all tissues, including vascular smooth muscle. The change in potassium conductance of the smooth muscle cell membrane produces a relaxation of blood vessels (9, 23, 28). To elucidate whether KATP channels were the target of H2S in the HA, GLB was administered. H2S-induced relaxation of the HA was found abolished by GLB, which confirms the notion that KATP channels in vascular smooth muscle cells are the target of H2S (33).
Preliminary experiments of our group with application of the adenosine receptor-1 antagonist 8-phenyltheophylline for blockage of HABR indicated that concomitant infusion of Na2S was capable to partly reverse the HABR blockage (data not shown). These observations imply that H2S is working rather independently of adenosine in mediation of HABR. However, further investigations are necessary to unravel the complex interactions between H2S, adenosine, and other gaseous mediators in the context of regulation of liver blood supply.
It has been postulated that the mediators responsible for the HA dilation were acting upstream from the hepatic sinusoids (2). We studied the expression of H2S-synthesizing enzymes in aorta and vena cava, as well as HA and PV, including their intrahepatic terminal branches within the portal triads. Present data suggest that CBS is the predominant H2S-generating enzyme in brain and nervous system, whereas only CSE was identified in the vascular system, including the rat aorta and mesenteric artery (4, 8, 30). In line with this, we could find the expression of CSE in aorta and vena cava, as well as in HA and PV, at both mRNA and protein level. However, we extend the present knowledge of other groups (1, 31, 33) in that we could observe the second enzyme CBS also expressed in the large vessels studied. In addition, we can report that intrahepatic terminal branches of the HA and PV were found positive for CSE but not for CBS. With the assumption that the space of Mall is the potential site of the mediator-driven communication between the HA and the PV (10), it is thus reasonable to speculate that HABR-associated vasodilatation of HA is partly mediated by H2S, which is predominantly synthesized by CSE. There is evidence that endogenous production of H2S depends on the vessel type. So, Zhao et al. (33) have found about fivefold higher sensitivity of mesenteric artery beds to H2S than of aortic tissues, whereas Hosoki et al. (7) have shown that the homogenates of thoracic aortas yielded more H2S than that of portal vein of rats. The mechanisms for vascular type selective sensitivities to H2S are not fully understood and remain to be clarified. However, we provide first evidence for a differential expression of CSE and CBS in the intrahepatic vascular system.
In summary, our data strongly underscore that H2S increases the HA buffer capacity via activation of KATP channels. In addition, H2S seems to be released from CSE in the terminal hepatic afferent vessels.
The study is supported by a grant from the Deutsche Forschungsgemeinschaft (Vo 450/7-3; Bonn, Bad Godesberg).
The authors cordially thank Berit Blendow, Doris Butzlaff, Dorothea Frenz, and Maren Nerowski (Institute for Experimental Surgery, University of Rostock) for excellent technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society