Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Potential conflict of interest: Nothing to report.
Carbon monoxide (CO) is a stress-inducible gas generated by heme oxygenase (HO) eliciting adaptive responses against toxicants; however, mechanisms for its reception remain unknown. Serendipitous observation in metabolome analysis in CO-overproducing livers suggested roles of cystathionine β-synthase (CBS) that rate-limits transsulfuration pathway and H2S generation, for the gas-responsive receptor. Studies using recombinant CBS indicated that CO binds to the prosthetic heme, stabilizing 6-coordinated CO-Fe(II)-histidine complex to block the activity, whereas nitric oxide (NO) forms 5-coordinated structure without inhibiting it. The CO-overproducing livers down-regulated H2S to stimulate HCO3−-dependent choleresis: these responses were attenuated by blocking HO or by donating H2S. Livers of heterozygous CBS knockout mice neither down-regulated H2S nor exhibited the choleresis while overproducing CO. In the mouse model of estradiol-induced cholestasis, CO overproduction by inducing HO-1 significantly improved the bile output through stimulating HCO3− excretion; such a choleretic response did not occur in the knockout mice. Conclusion: Results collected from metabolome analyses suggested that CBS serves as a CO-sensitive modulator of H2S to support biliary excretion, shedding light on a putative role of the enzyme for stress-elicited adaptive response against bile-dependent detoxification processes. (HEPATOLOGY 2009;49:141-150.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Carbon monoxide (CO) is generated from inducible heme oxygenase 1 (HO-1) and constitutive heme oxygenase 2 (HO-2), respectively, and has the ability to regulate neurovascular functions,1, 2 apoptotic responses,3, 4 and metabolism of xenobiotics and toxicants.5, 6 This gas is overproduced through increased delivery of heme as a substrate and the HO-1 induction on exposure to stressors such as hypoxia and oxidative stress. Mechanisms by which CO regulates cell functions appear to involve an activation of soluble guanylate cyclase (sGC), the enzyme that allows the gas to bind to the prosthetic heme to synthesize cyclic guanosine monophosphate as a second messenger.1 Distinct from nitric oxide (NO) that forms 5-coordinated NO-Fe(II) complex to trigger full activation of the enzyme, CO activates this enzyme only modestly because the gas binding stabilizes 6-coordinated CO-Fe(II)-histidine complex.7 Mitogen-activated protein kinase has also been shown to serve as a CO-responsive signal transducer.8 Gene disruption of HO-1 increases sensitivity to overproduction of reactive oxygen species, inflammatory mediators or xenobiotic metabolism, whereas the gene transfer or CO inhalation under these circumstances suppresses such pathogenic responses.7–9 However, direct mechanisms for the CO reception to trigger these adaptive responses of metabolism remain unknown.
Because this gas has the ability to inhibit ferrous form of the prosthetic heme of enzymes, tryptophan 2,3-dioxygenase or cytochromes P450 have been considered putative CO-sensitive signal transducers regulating cell functions, including cell proliferation,10 immune responses,11 microvascular tone, xenobiotic detoxification, and biliary excretion in the liver.5, 6, 12 However, ferrous heme of these enzymes is not only sensitive to CO but to NO. In this context, whether mechanisms by which CO regulates cell and organ functions is not shared by those for NO has not fully been studied yet.
This study aimed to mine novel CO-responsive regulators for stress-inducible adaptation of metabolism. To this end, we have used metabolome analyses based on capillary electrophoresis equipped with mass spectrometry (CE-MS) for systematic mining CO-responsive gaseous signal transducers. The current results suggest that cystathionine β-synthase (CBS), the enzyme rate-limiting transsulfuration pathway is such a novel CO-sensitive regulator of metabolism that plays an important role for quality control of bile excretion under disease conditions.
CBS, cystathionine β-synthase; CE-MS, capillary electrophoresis equipped with mass spectrometry; CO, carbon monoxide; CORM, CO-releasing metal carbonyl tricarbonyldichlororuthenium (II); ES, 17α-ethinylestradiol; GSH, glutathione; GSNO, S-nitrosyl glutathione; H12, liver exposed to 12-hour hemin treatment; NO, nitric oxide; RuCl3, CO-free ruthenium (III) chloride; SAM, S-adenosyl methionine; SE, standard error; sGC, soluble guanylate cyclase.
Materials and Methods
Preparation of Mice.
The experimental protocols herein described were approved by our institutional guidelines provided by the Animal Care Committee of Keio University School of Medicine. Mice heterozygous for disruption in the CBS gene were purchased from Jackson Labs (Bar Harbor, MI) and bred at our institution. Male heterozygous CBS-deficient mice (CBS+/−) and their littermates (CBS+/+), and wild-type B6J mice, which were purchased from Clea Japan, Inc (Kawasaki City, Japan), were used at 8 to 12 weeks of age. Mice were allowed free access to laboratory chow and tap water, and were fasted for 18 hours before experiments. Mice were anesthetized with an intraperitoneal injection of ketamine at 120 mg/kg, and xylidine at 6 mg/kg. Their common bile ducts were ligated in proximity to the duodenum, and the gallbladder was nicked and cannulated with a polyethylene P-10 tube to collect bile for 20 minutes after a 10-minute stabilization period.6, 13 Biliary constituents such as total bile salts, bilirubin-IXα, pH values, and bicarbonate (HCO3−) were measured according to previous methods described elsewhere.13 When necessary, biliary samples were collected into tubes containing 10% trichloroacetate to measure glutathione through high-performance liquid chromatography.14 Determination of bilirubin-IXα in bile serves as an indicator of HO-mediated heme degradation in the liver that occurs in parallel with endogenous CO generation. Hepatic CO contents were also measured by gas chromatography as described previously,15 except that the flame ionization detector equipped with a methanizer was used in this study instead of a reduction gas detector. Combination of these methods to determine CO allowed us to distinguish endogenous CO generation from the same gas exogenously administered as an intervention as described in the following session.
Administration of Reagents Studied.
Protoheme IX (hemin) was administered at 40 μmol/kg intraperitoneally at 12 hours before surgical preparation for bile collection. This protocol was denoted as liver exposed to 12-hour hemin treatment (H12) treatment in the text. After collecting bile, livers were excised immediately to be snap-frozen in cold methanol, and the lysates served as samples for contrast-enhanced time of flight/mass spectrometry analyses as described later. In separate sets of experiments, liver samples were minced with 10% trichloroacetic acid at 4°C to measure cysteine and glutathione (GSH) through high-performance liquid chromatography to confirm the data collected from contrast-enhanced time of flight/mass spectrometry, when necessary.
A series of protocols were employed to examine roles of HO-derived CO in regulation of H2S-modulated choleresis in the H12-treated mice. First, zinc protoporphyrin, a potent HO inhibitor, was administered intravenously at 12.5 μmol/hour/kg at 30 minutes before the bile collection; this dose was sufficient to block endogenous CO in the liver. When necessary, tricarbonyldichlororuthenium (II) dimer, the CO-releasing metal carbonyl [tricarbonyldichlororuthenium (II): CORM, Sigma-Aldorich]16 was administered intraportally at 30 minutes before the start of bile collection. When necessary, CO-free ruthenium (III) chloride (RuCl3) was used as a negative control reagent. To examine whether the elevation of H2S in the liver could alter biliary HCO3− excretion, sodium hydrosulfide (NaHS) was administered at 20 μmol/hour/kg through the portal vein at 30 minutes before the bile collection; as seen later in Results, this protocol restored the H12-induced decrease in the hepatic H2S contents without altering a reduction of systemic blood pressure that was induced by a systemic bolus of the NaHS injection. S-nitrosyl glutathione (GSNO) was used as an NO donor. The reagent was injected intraportally with a dose of 7 μmol/kg at 30 minutes before the collection of bile; this protocol did not induce a reduction of systemic blood pressure, whereas greater doses caused hypotension and subsequent decrease in the bile output. In these experiments, administration of the reagent was performed through a 30-gauge miniature needle that was inserted into the portal vein to be fixed at the site of puncture. Finally, to examine therapeutic effects of CO, we examined effects of H12 treatment or administration of CORM in the mice exposed to drug-induced cholestasis. To this end, cholestasis was induced by a subcutaneous injection of 17α-ethinylestradiol (ES) at 5 mg/kg daily for 5 consecutive days before the experiments.17
We performed metabolome analyses of tissue lysates collected from snap-frozen livers of mice using contrast-enhanced time of flight/mass spectrometry according to our previous methods.18, 19 Measurements of hepatic H2S contents were based on gas chromatography described in our previous method.14 Biliary flux of bilirubin-IXα (BR-IXα) in bile samples were determined by enzyme-linked immunosorbent assay using the anti–BR-IXα monoclonal antibody as described previously.6, 20 Because BR-IXα is an end product of the HO-mediated degradation of protoheme IX, its measurements in bile serves as an index of endogenous CO generation in the liver.20 The conversion of 15N-methionine to its downstream metabolites was determined by CE-MS to examine different rates of the metabolic flux through CBS in the liver. In these experiments, 15N-methionine was intraperitoneally injected at 150 μmol/100 g body weight, and 15N-homocysteine and 15N-cystathionine were measured by CE-MS using the lysates of liver tissues at 30, 60, and 120 minutes after the methionine challenge. Data were expressed as percentages of the mass-labeled metabolites versus total amounts of metabolites in remethylation cycle [ΣRM: methionine + S-adenosyl methionine (SAM) + S-adenosyl homocysteine (SAH) + homocysteine]. In a separate set of experiments, effects of application of CO on contents of methionine and cystathionine in HepG2 cells were determined in culture. In these experiments, the cells were maintained in Roswell Park Memorial Institute 1640 medium (Invitrogen, Carlsbad, CA) containing 10% fetal bovine serum; the mixture was supplemented with 1× penicillin/streptomycin and maintained at 37°C in an atmosphere of 5% CO2/95% air. The cells were treated with either 50 μmol/L CORM or RuCl3 as a negative control for 16 hours. To measure the metabolites, a frozen pellet of the 1 × 106 cells was homogenized in 10% trichloroacetic acid with 10 mM diethylene triamine pentaacetic acid following brief centrifugation, and the supernatant was used as a sample.
Western Blot Analysis.
Western blot analysis was carried out to examine an induction of heme oxygenase (HO)-1 using the polyclonal antibody SPA896 (Stressgen, Ann Arbor, MI). In these experiments, the blotting against α-tubulin was carried out using the polyclonal antibody (Cell Signaling, Danvers, MA) as an internal control.
Recombinant Full-Length Rat CBS.
The complementary DNA of the full-length rat CBS was a gift from Professor Masao Ikeda-Saito in Tohoku University. Stopped-flow equipment was purchased from Unisoku, Inc. (Tokyo) and used to examine binding of CO or NO to the CBS protein according to previous methods.21 Electron paramagnetic resonance spectrometry to determine 5-coordinated structure of the nitrosylheme complex of CBS was carried out according to previous methods.21, 22
The statistical significance of data among different experimental groups was determined by one-way analysis of variance and Fischer's multiple comparison test. P < 0.05 was considered significant.
CO Overproduction Inhibits Transsulfuration and H2S and Stimulates HCO3− Choleresis.
Metabolome analyses based on CE-MS allowed us to pinpoint metabolic pathways responding to disease conditions. In mouse liver, we detected more than 1800 metabolites, and compared differences between the control and acetaminophen-treated livers.18 This method was used to determine differences in metabolic responses between mouse livers and those overloaded with heme, the stressor inducing oxidative stress and subsequent CO overproduction through increasing the substrate and inducing HO-1 (Fig. 1A). The hepatic CO flux peaked at 6 hours, becoming threefold to fourfold greater during the 6 to 12 hours after challenging with hemin, as judged by BR-IXα, an end product of HO-mediated heme degradation (Fig. 1B).13 Under these conditions, bile output was modestly but significantly increased at 12 to 18 hours after the treatment (Fig. 1C) in parallel with significant elevation of HCO3− to make bile more alkaline (Fig. 1D-F), enhancing solubility of organic anions during the detoxification processes. These results suggest that the heme-elicited choleretic response is not correlated with vasodilatory mechanisms by the gas. Based on these results, we used CE-MS analyses to examine metabolomics in the liver exposed to 12-hour hemin treatment (H12), in which phenotypes of bile remodeling became evident.
Among known metabolites (Table 1), most prominent differences between the control and H12 groups occurred in global decreases in amino acids concurrent with increases in Krebs cycle substrates such as acetyl CoA: the fact that these changes coincided with sustained glutamate, significant increases in glutamine, and high-energy adenosine phosphates appeared to suggest utilization of the amino acid pool for energy substrates. By contrast, several essential amino acids such as methionine, tryptophan and histidine, and serine were maintained. Another important alteration was a global decrease in transsulfuration metabolites such as cystathionine, cysteine, and hypotaurine. These results led us to determine tissue contents of H2S, the terminal product derived from CBS or cystathionine γ-lyase that constitute transsulfuration pathway; this gaseous compound turned out to be suppressed in the H12 group. Based on these measurements, we hypothesized that the H12 treatment limits the activity of CBS so far as judged from maintenance of methionine pool (ΣRM) and serine, a substrate of the enzyme, with suppression of the transsulfuration metabolites residing in the downstream (Fig. 2A). This hypothesis was confirmed by in vivo pulse-chase analysis showing accumulation of 15N-homocysteine and suppression of 15N-cystathionine after the 15N-methionine challenge in the H12 group (Fig. 2B).
Table 1. Comparison of Metabolome Analysis by CE-MS in Liver Extracts Between Control and the Hemin-Treated (H12) Mice
Data indicate mean ± SE of six separate experiments.
Data of metabolites in remethylation cycle and transsulfuration pathway were indicated in Fig. 2A.
Such an inhibitory action of the H12 treatment on the transsulfuration pathway was reproducible when HepG2 cells was treated with CO in culture; contents of cystathionine were significantly suppressed by the application of 50 μmol/L CORM (9.3 ± 1.3 versus 15.9 ± 1.4 nmol/g protein for the vehicle treatment with RuCl3. Mean ± standard error (SE) of three separate experiments, P < 0.03), whereas methionine exhibited no difference (66.3 ± 3.7 versus 80.3 ± 12.2 nmol/g protein for CORM and RuCl3, respectively. Mean ± SE of three separate experiments), suggesting inhibitory action of the gas on CBS.
CO But Not NO Inhibits CBS.
H12-induced metabolomic changes indicating dissociation between remethylation cycle and transsulfuration pathway led us to hypothesize that CBS, a heme-containing enzyme that rate-limits the transsulfuration pathway, is a sensor of the H12-elicited CO overproduction. Rat full-length recombinant CBS were purified (Fig. 3A) to examine whether CO or NO could inhibit the enzyme activities. CO, but not NO, specifically inhibited the enzyme (Fig. 3B). Previous crystallographic studies using a truncated form of CBS showed that the axial ligands for the prosthetic heme were cysteine and histidine, indicating a large peak of absorbance at 448 nm.23 On CO application, the heme formed a 6-coordinated CO-Fe(II)–histidine complex, as judged by a decrease in the absorbance at 448 nm and a reciprocal elevation at 422 nm (Fig. 3C). These results were consistent with previous works using the truncated form of human recombinant CBS.24 Such an inhibitory effect of CO on CBS activity occurred even when sufficient amounts of SAM were present as an allosteric activator,25 whereas the CO concentrations necessary to suppress CBS became greater in the presence of SAM (Fig. 3D). Conversely, NO was able to bind to the heme but with a distinct structure of 5-coordinated nitrosylheme as judged by electron paramagnetic resonance spectrometry (Fig. 3E), suggesting that the enzyme responds specifically to the binding of CO but not that of NO.
CO-Induced HCO3− Choleresis Is Sensitive to H2S and Disappears in CBS+/− Mice.
Recent studies indicated that H2S derived from cystathionine γ-lyase, an enzyme using cysteine to generate the gas, modulates biliary HCO3− excretion via mechanisms involving glibenclamide-sensitive channels, a putative H2S target.14, 26 We hypothesized that the stress-induced CO stimulates the HCO3− excretion to increase pH in bile through its inhibitory action on CBS-derived H2S. To examine this hypothesis, we chose the dose of the CO-releasing molecule (CORM) that was able to increase hepatic contents of CO comparably to those measured in the H12 treatment: As seen (Fig. 4A), the intraportal administration of CORM at 20 μmol/kg significantly increased hepatic CO contents comparable to those induced by H12 treatment in the intact mice. This dose of CORM suppressed hepatic H2S and stimulated biliary HCO3− flux. Stimulatory effects of CO administration on biliary HCO3− excretion in intact mice were not shared by NO, as judged by observation in the mice administered with GSNO, an NO donor (Fig. 4B): These results were consistent with observation that CBS is sensitive to CO but not to NO in vitro (Fig. 3).
As already seen, H12 treatment increased CO generation (biliary BR-IXα flux), decreased hepatic H2S contents, and stimulated biliary HCO3− flux (Fig. 1). HO blockade by zinc protoporphyrin-IX cancelled these changes elicited by H12 treatment. On the other hand, an administration of NaHS, an H2S donor, abolished the H12-induced suppression of hepatic H2S contents, and significantly attenuated the stimulatory response of biliary HCO3− flux (Fig. 5A), suggesting that H12-inducible CO stimulates biliary HCO3− excretion through modulation of CBS-derived H2S. As previously reported, homozygous CBS knockout mice died of severe hepatic steatosis, whereas heterozygous knockout (CBS+/−) mice survive through compensation without apparent phenotypes.27 In these mice, indeed, the baseline H2S content in livers of CBS+/− mice was comparable to that of CBS+/+ mice, presumably because of compensation of the gas generation through cystathionine γ-lyase. On H12 treatment, CBS+/− mice exhibited an increase in the hepatic CO generation comparably to CBS+/+ mice, but neither decreased H2S contents nor up-regulated biliary HCO3− flux (Fig. 5B), indicating phenotypes distinct from those in CBS+/+ littermates.
CO Protects Against Drug-Induced Cholestasis Through Mechanisms Involving CBS.
We further attempted to investigate whether the administration of CO could improve biliary dysfunction occurring in disease models. To examine this, the mice were treated with ES, a cholestatic reagent suppressing three major osmolites such as HCO3−, glutathione, and bile salts in bile.17 H12 treatment or the administration of CORM significantly increased bile output concurrently with a recovery of HCO3− excretion into bile (Fig. 6A). The anti-cholestatic effects of H12 treatment through stimulation of HCO3− excretion disappeared in the CBS+/− mice (Fig. 6B), suggesting again a pivotal role of CBS for triggering the CO-induced choleresis.
CO administration or HO-1 induction has been shown to protect against tissue injury and considered a potentially useful therapeutic stratagem.8, 16 Serendipitous observation in the liver indicating effects of overproduced CO on metabolism of sulfur-containing amino acids led us to reveal unique physiological actions of this gas on CBS in vivo that are not shared with NO. The current study suggested that stress-inducible CO targets CBS and thereby reduces H2S significantly to stimulate biliary HCO3− excretion that could benefit detoxification processes. Conversely, such a property of stress-inducible CO might jeopardize anti-oxidative defense systems through an overflow of homocysteine or through a shortage of GSH. Under current experimental conditions, however, such a risk seemed little, if any, so far as judged from maintenance of GSH and adenosine triphosphate so far. This appears to result from large difference in amino acid pools between methionine (nmol/g) and thiols including cysteine and GSH (μmol/g). Furthermore, cysteine could be supplied through its uptake from extracellular space by mechanisms involving Nrf2, the transcriptional factor activated in response to oxidative stress or electrophiles such as heme.28, 29 By contrast, the amounts of sulfur-containing amino acids consumed to generate H2S seems relatively smaller than that for synthesizing GSH or hypotaurine, as judged from quantitative information collected by metabolome analysis. Because CBS not only limits synthesis of cystathionine from homocysteine but also directly suppresses H2S generation from cysteine, the inhibitory effects of CO on the enzyme could dictate largely on the action of H2S in the liver, causing a stimulatory effect on bile excretion. Considering recent studies suggesting vasodilatory effects of H2S,26, 30 suppression of CBS-derived H2S by stress-inducible CO might trigger vasoconstriction, but such vasoactive responses did not occur so far as judged from choleretic response of the basal bile flow that is highly dictated by microvascular perfusion. This might result from the fact that stress-inducible CO itself has the ability to maintain the basal microvascular perfusion through multiple vasodilatory mechanisms involving activation of cyclic guanosine monophosphate and modulation of cytochrome P450–derived vasoconstrictors.6, 20, 31
Although the inhibitory action of stress-inducible CO on the transsulfuration pathway has first been shown in the heme-overloading detoxification model of mice in the current study, a similar event occurred in acetaminophen-induced acute liver injury model of mice in which CO was overproduced through degradation of cytochrome P450–derived heme.5, 18 Our previous study in rats suggested that another HO-derived product bilirubin but not CO has the ability to improve bile acid–dependent bile output of the post–cold ischemic liver grafts through its anti-oxidative action.32 However, such an effect of bilirubin appears to be distinct from the stimulatory action of CO on biliary fluid excretion indicated in the current study. CO has been shown to exert diverse actions on biliary function through multiple mechanisms: First, stress-inducible levels of CO have the ability to elongate the intervals of bile canalicular contraction, which helps increase the stroke volume for promoting bile excretion; this process appears to involve mechanisms mediated by modulation of cytochrome P450 epoxygenases and intracellular Ca2+ mobilization.12 Second, suppression of endogenous CO activates bile acid–dependent bile excretion through accelerated vesicular transport of taurocholate, while inducing no significant elevation of the bile acid–independent fraction.33 Conversely, CO overproduction by the HO-1 induction or exogenous administration of CO stimulates bile acid–independent choleresis concurrently with increased mrp2-dependent excretion of bilirubin-IXα and glutathione, while suppressing biliary excretion of bile salts, indicating the effects of the gas for stimulating fluid excretion into bile.34 Of interest is that glibenclamide, an inhibitor of K+ channel that serves as a putative target for H2S,26 acts on Na+-K+-2Cl− cotransporter in bile duct epithelium to stimulate biliary HCO3− excretion in normal and cholestatic livers.35 We showed that inhibition of cystathionine γ-lyase, another H2S-generating enzyme, stimulates basal and glibenclamide-induced fluid output of bile through stimulating HCO3− excretion without altering the baseline vascular resistance of the liver.14 Recent studies provided evidence that such a glibenclamide-responsive channel is present in rodent cholangiocytes36 or in duodenum,37 contributing to stimulation of the HCO3− excretion.36 Based on these observations, it is not unreasonable to speculate that CO stimulates biliary fluid excretion through mechanisms involving H2S-mediated modulation of glibenclamide-sensitive channels on biliary epithelium. Although further investigation is necessary to determine whether these mechanisms are sensitive to H2S, the current results shed light on a possibility that the CO-CBS system serves as a putative mechanism for stimulating bile acid–independent fluid excretion, facilitating excretion of HCO3− and organic anions such as bilirubin to support heme detoxification. Both glibenclamide and CO help biliary fluid excretion in estrogen-induced hepatocellular cholestasis. Exploration of H2S-sensitive molecular targets occurring on biliary epithelium deserves further studies for evidence that HO-1–derived CO serves as a therapeutic stratagem for protecting against cholestasis.
CO has been believed to share varied physiological effects on biological systems with NO. However, through extrapolation of studies in vitro indicating biochemical actions of CO to trigger structural changes in gas-responsive heme proteins (such as sGC, hemoglobin) distinct from those elicited by NO,7, 19, 21, 22, 38 evidence that CO is a unique gaseous regulator distinct from NO has been emerging. In fact, CO itself modestly activates sGC, by which hepatic sinusoids are constitutively dilated.2, 20, 39 By contrast, in vascular smooth muscle cells in which NO is sufficiently supplied from arteriolar endothelium (for example, brain microcirculation), the inducible CO inhibits NO-elicited sGC activation.40, 41 Besides these observations suggesting physiologic actions of CO occurring dependently of local NO levels, the current study provided evidence for a novel mechanism functioning irrespective of the NO effects. Furthermore, our results shed light on a metabolic link between CO and H2S, suggesting that different gaseous mediators constitute an intriguing link for regulation of organ functions.
The authors thank Kayo Maruyama for technical support in measuring tissue H2S contents.