H2S is endogenously generated in vascular smooth muscle cells. The signal transduction pathways involved in the vascular effects of H2S have been unclear and were investigated in the present study. H2S induced a concentration-dependent relaxation of rat aortic tissues that was not affected by vascular denervation. The vasorelaxant potency of H2S was attenuated by the removal of the endothelium. Similarly, the blockade of nitric oxide synthase or the coapplication of the Ca2+-dependent K+ channel blockers apamin and charybdotoxin reduced the H2S-induced relaxation of the endothelium-intact aortic tissues. Sodium nitroprusside (SNP)-induced relaxation was completely abolished by either 1H-[1,2,4]oxadiazolo-[4,3-a]quinoxalin-1-one (ODQ) or NS- 2028, two soluble guanylate cyclase inhibitors. Instead of inhibition, ODQ and NS-2028 potentiated the H2S-induced vasorelaxation, which was suppressed by superoxide dismutase. The vasorelaxant effect of H2S was also significantly attenuated when Ca2+-free bath solution was used. Finally, pretreatment of aortic tissues with H2S reduced the relaxant response of vascular tissues to SNP. Our results demonstrate that the vascular effect of H2S is partially mediated by a functional endothelium and dependent on the extracellular calcium entry but independent of the activation of the cGMP pathway.
- guanosine 3′,5′-cyclic monophosphate
h2 s has been generally considered as a toxic gas found in the contaminated environmental atmosphere. Its major toxic effects are the toxication of central nervous system and the inhibition of the respiratory system (2, 10, 28). H2S can also be produced endogenously from l-cysteine by cystathionine β-synthase (CBS) and/or cystathionine γ-lyase (CSE) (23, 24). The expression of these enzymes has been detected in various tissues (13). Our recent study (31) demonstrated that CSE was expressed in the rat aorta, tail artery, mesenteric artery, and pulmonary artery, whereas the expression of CBS was not detectable. Under physiological conditions, tissue content of H2S in the brain has been determined to be between 50 and 160 μM (1). The endogenous production of H2S from rat vascular tissues has also been demonstrated in our previous study (31). With the use of a modified sulfide electrode method, the H2S concentration of rat serum was determined to be ∼46 μM (31). These observations speak for the potential physiological functions of H2S in the cardiovascular system.
The relaxant effects of exogenously applied H2S on intestinal and vascular smooth muscles have been reported (13). In both the aorta and portal vein of rats, H2S induced a dose-dependent relaxation, but other thiol-containing endogenous substances such as cysteine and glutathione did not have any vasorelaxant effect (13). More importantly, we found that H2S relaxed rat aortic tissues within physiologically relevant concentrations and that an intravenous bolus injection of H2S transiently decreased blood pressure of rats (31). Thus the physiological effect of H2S on vascular functions should be entertained.
The reported vascular effects of H2S are not always consistent in terms of the relaxant potency of the gas and its interaction with other vasoactive factors (13, 31). The mechanisms underlying the vascular effect of H2S are not completely understood yet, although the activation of ATP-sensitive K+ (KATP) channels in vascular smooth muscle cells (SMCs) might play an important role (31). Whether the vascular effects of H2S are mediated by vasoactive factors released from the endothelium or peripheral nerve endings, e.g., nonadrenergic, noncholinergic (NANC) nerves, is not clear. The interaction of H2S with different cellular signaling pathways in vascular SMCs, including cGMP and oxygen-derived free radicals, has not been systematically examined. The objective of the present study, therefore, was to characterize the relaxant effect of H2S on rat aortic tissues and study the possible cellular and molecular mechanisms underlying the H2S-induced relaxation. To this end, we systematically examined the roles of the endothelium, peripheral innervation, cellular cGMP signaling pathway, and extracellular Ca2+ concentration ([Ca2+]) in the vasorelaxant response to H2S. The interaction of H2S and nitric oxide (NO) on vascular tone was also examined.
MATERIALS AND METHODS
Measurement of isometric contraction of vascular tissues.
Male Sprague-Dawley rats (10–12 wk old) were housed in an animal care facility at the College of Medicine, University of Saskatchewan. Animal experiment protocols were approved by the Committee on Animal Care and Supply of the University of Saskatchewan. The method for measurement of isometric tension development of the isolated rat aortic tissues was the same as that described previously (26). In brief, the aortic rings were mounted on two tungsten wires with one immobilized and the other connected to a transducer (FT 03, Grass Instruments) in organ baths. The baths were filled with 10 ml Krebs bicarbonate solution composed of (in mM) 115 NaCl, 5.4 KCl, 2.5 CaCl2, 1.2 MgSO4, 1.2 KH2PO4, 25 NaHCO3, and 11d-glucose bubbled with 5% CO2 in oxygen. The aortic rings were mechanically stretched to achieve a basal tension of ∼2.0 g and allowed to equilibrate for 1 h before the start of the experiment. The endothelium was kept functionally undamaged. In the experiment where the removal of endothelium was specified, vascular tissues were incubated in Krebs solution containing 0.1% saponin for 55 s and then rinsed with normal Krebs solution (3
Chemicals and data analysis.
The H2S-saturated solution (0.09 M at 30°C) was made by bubbling pure H2S gas (Praxair; Mississauga, Ontario, Canada) into Krebs solution at 30°C for 40 min. Both the H2S gas-saturated solution and NaHS (Aldrich; Milwaukee, WI) stock solution (1 M) were freshly prepared on the day of the experiment. The stock solutions of H2S were directly added to the bath solution to achieve the final concentration of H2S. At 37°C, the concentration of H2S of the bath solution was relatively stable. Direct measurement of H2S concentration with our established assay (31) showed that the stability of H2S in the bath solution varied depending on the initial H2S concentrations. At the highest concentration tested (1 mM), a drop of the H2S concentration around 15% within 30 min was observed (Fig. 1).
PE, charybdotoxin, apamin, superoxide dismutase (SOD), catalase, indomethacin, nifedipine, and sodium nitroprusside (SNP) were purchased from Sigma (St. Louis, MO). 4H-8-bromo-[1,2,4]oxadiazolo-[3,4-d]benz(b)(1,4)oxazin-1-one (NS-2028) was from Calbiochem (San Diego, CA). 1H-[1,2,4]oxadiazolo-[4,3-a]quinalin-1-one (ODQ) was purchased from Tocris (Ballwin, MO).
In total, 45 rats were used in this study. Unless otherwise stated, each experiment used eight aortic rings from two to three rats. Data are expressed as means ± SE and presented as either the percent reduction from total contraction force induced by PE or the absolute values of tension development. All concentration-response curves were fitted with a Hill equation, from which IC50 was calculated. Student's t-test for unpaired samples was used to compare the mean values between the control and tested groups. Multiple comparisons were made with one-way ANOVA followed by a post hoc analysis (Tukey test). Statistical significance was set atP < 0.05.
Vasorelaxation induced by NaHS and H2S gas.
Both NaHS solution and the H2S gas-bubbled solution have been used in the literature to relax vascular tissues. The concentration-dependent relaxant effects of NaHS and the H2S gas-bubbled solution on the PE-precontracted aorta were compared in the present study. The H2S gas-bubbled solution induced vasorelaxation with an IC50 of 124.7 ± 14.4 μM, similar to that induced by NaHS solution with an IC50of 135.5 ± 14 μM (P > 0.05). Both the H2S gas-bubbled solution and NaHS solution had the same threshold concentration of 60 μM at which vasorelaxation was initiated. A maximum relaxation of 55.3% was attained at 600 μM NaHS. In addition, vasorelaxation induced by both NaHS and H2S could be reversed by removing the chemicals (Fig.2 A). Because the majority of experiments were carried out using NaHS solution, for the convenience of description, the term of H2S was used in the rest of this communication. The use of the H2S gas-bubbled solution in each individual experiment was specified.
Effect of denervation on H2S-induced vasorelaxation.
To test whether the vascular effect of H2S was mediated by autonomous or NANC (12, 18) nerve endings remaining in the vascular wall, the H2S-induced vasorelaxation was examined using aortic tissues pretreated with a phenol denervation procedure (27) in which tissues were immersed in 0.75% phenol-6.75% ethanol for 20 s. The basal tension and contraction force were not altered by this treatment. The vasorelaxant effect of H2S appeared to be potentiated (Fig. 2 B). H2S (600 μM) induced a 74.5 ± 9.1% relaxation of the denervated vascular tissues (n = 8) compared with the 55.3 ± 3.4% relaxation induced by H2S at the same concentration (n = 8). However, the difference was not statistically significant (P > 0.05). These results suggest that the vasorelaxant effect of H2S is not due to the altered neurotransmitter release from the autonomous or NANC nerve endings.
Interaction of H2S and endothelium on H2S-induced vasorelaxation.
After removal of the endothelium with saponin treatment, the total contraction force induced by 0.3 μM PE was changed from 1.18 ± 0.05 to 1.30 ± 0.10 g (P > 0.05). The H2S-induced maximum relaxation of the precontracted tissues was not affected by the endothelium removal, but the H2S concentration-response curve was shifted to the right, with IC50 changed from 135.5 ± 14 to 273 ± 16 μM (P < 0.05). Thus the H2S-induced vasorelaxation might be partially facilitated by an endothelium-mediated mechanism (Fig. 2 B).
The endothelium releases many vasoactive factors that regulate vascular smooth muscle contractility, including NO and endothelium-derived hyperpolarizing factor (EDHF). Application ofN G-nitro-l-arginine methyl ester (l-NAME; 100 μM), which would limit endogenous NO production, to the endothelium-intact tissues increased the IC50 of H2S from 135.5 ± 14 to 220 ± 12 μM (P < 0.05), indicating that the vasorelaxant effect of H2S was reduced in the presence ofl-NAME. The coapplication of the Ca2+-dependent K+ (KCa) channel blockers apamin and charybdotoxin has been reported to effectively inhibit the vasorelaxation mediated by EDHF (7). This protocol was used in the present study to further test the involvement of EDHF in the vascular effect of H2S. In the presence of apamin (50 nM) and charybdotoxin (50 nM for 30 min), the H2S concentration-response curve of the endothelium-intact aortic tissues was significantly shifted to the right, with IC50 changed from 135.5 ± 14 to 350 ± 13 μM (P < 0.05), an effect similar to the removal of the endothelium (Fig.3 A). Like the removal of the endothelium, the maximal relaxation induced by H2S was not affected by the coapplication of apamin and charybdotoxin. Charybdotoxin (100 nM) alone did not cause any significant shift of the H2S concentration-response curve (IC50 from 128.5 ± 62.4 to 183.2 ± 10.3 μM, P > 0.05). Apamin alone (50 nM) also did not affect the H2S-induced vasorelaxation (Fig. 3 B). The basal contraction force was not significantly affected by these treatments.
Involvement of the cGMP pathway in H2S-induced vasorelaxation.
The cGMP pathway plays an important role in mediating the NO- and carbon monoxide (CO)-induced vasorelaxation (9). To determine whether the H2S-induced vasorelaxation was mediated by this pathway, we studied the vascular effect of H2S in the presence of the soluble guanylyl cyclase (sGC) inhibitors ODQ (10 μM for 20 min) (14, 29) and NS-2028 (1 μM for 20 min) (16), respectively. The basal tension of the endothelium-intact aortic tissue was not significantly affected by these reagents per se. Interestingly, instead of an inhibition, the H2S-induced relaxation was potentiated by ODQ or NS-2028. In the presence of ODQ, the relaxation induced by 1.8 mM H2S was increased from 57.6 ± 3.5% to 94.9 ± 5.6% (P < 0.05) (Fig.4 A). NS-2028 enhanced the H2S-induced relaxation from 70.1 ± 3.5% to 89.7 ± 3.5% (P < 0.05) (Fig. 4 B). On the other hand, the blockade of the cGMP pathway by ODQ (10 μM) or NS-2028 (1 μM) significantly inhibited the SNP-induced vasorelaxation by >90% (Fig. 5 A). These results suggest that the H2S-induced vasorelaxation is not due to the activation of the cGMP pathway.
Some free radicals like superoxide and hydrogen peroxide may reduce the contractility of vascular smooth muscle (8). To examine whether the potentiation of the H2S effect by ODQ involved the production of free radicals, the H2S-induced vasorelaxation was evaluated in tissues preincubated with the superoxide scavenger SOD (160 U/ml) (21) and the hydrogen peroxide scavenger catalase (1,000 U/ml) for 10 min. Basal tension was not affected by these chemicals. None of these treatments altered the H2S-induced vasorelaxation. However, the potentiation of the H2S effect induced by ODQ was suppressed in the presence of SOD (Fig. 5 B). The concentration-dependent response curves from the control tissues and ODQ-treated and SOD + ODQ-treated tissues were compared using one-way ANOVA followed by a post hoc analysis (Tukey's test). There was no significant difference between the response curves of the control and ODQ + SOD groups (P > 0.05). However, a significant difference was detected between the control group and ODQ-alone group (P < 0.05) or between the ODQ-alone and SOD + ODQ groups (P < 0.05). At concentrations of 0.6 and 1.8 mM, the H2S-induced vasorelaxation was significantly stronger in the ODQ-alone group compared with either control or ODQ + SOD groups. These results indicate that the augmentation of the H2S-induced relaxation by ODQ is abolished by SOD. The generation of free radicals may underlie the increase in the H2S-induced vasorelaxation by ODQ.
Influence of extracellular [Ca2+] entry on H2S-induced vasorelaxation.
The transmembrane movement of Ca2+ directly affects the contractility of smooth muscles. To examine whether the H2S-induced relaxation was mediated by an extracellular [Ca2+]-dependent mechanism, the concentration-dependent vasorelaxant effect of H2S was studied either using a calcium-free bath solution or with the normal bath solution but in the presence of nifedipine, a voltage-gated Ca2+ channel inhibitor (19). When calcium-free bath solution was used, PE (0.3 μM) generated a contraction force of 0.53 ± 0.04 g, which was 34% of the contraction force in the presence of calcium. Interestingly, PE induced an initial transient contraction spike in three of eight tissues, but a sustained plateau was always developed after the spike in these tissues. The relative changes of the H2S-induced relaxation were always calculated based on the force generated at the sustained plateau.
A transient increase in basal tension was also observed after the addition of nifedipine to the bath solution. This increase amounted to up to ∼15% of the PE-induced contraction and lasted for ∼5 min. After the nifedipine-induced transient contraction subsided, PE was added to the bath solution, which generated a contraction force that was 88 ± 14% of that in the absence of nifedipine (P < 0.05). These contractions were fairly sustained and maintained at plateaus during the course of experiments.
The H2S (600 μM)-induced relaxation was reduced by 59.3 ± 8.3% by nifedipine (P < 0.05) or by 50 ± 6.3% by removing calcium from the bath solution (P < 0.05) (Fig. 6).
Interaction between H2S and SNP on vascular contractility.
To identify whether there was a synergistic effect of NO and H2S in relaxing vascular tissues, the concentration-dependent effect of SNP, a NO donor, on the H2S-pretreated (30 or 60 μM for 15 min) and PE-precontracted endothelium-intact aortic tissue was examined. No significant relaxation was induced by H2S per se at 30 or 60 μM. As shown in Fig. 7, the presence of 30 μM H2S did not affect the vasorelaxant effect of SNP. However, the presence of 60 μM H2S shifted the SNP concentration-response curve to the right, with IC50changed from 6.8 ± 0.5 to 34.2 ± 3.9 nM (P< 0.05), indicating that H2S inhibited the SNP-induced vasorelaxation. On the contrary, the H2S-induced vasorelaxation was not significantly affected by pretreating tissues with 3 nM SNP (data not shown).
H2S has been tagged as a toxic gas for hundreds of years (20). As a broad-spectrum toxicant, H2S affects many organ systems, including the lung, brain, kidney, etc. (2). However, the possible physiological effect of this gas on the cardiovascular system has only been known recently. Hosoki et al. (13) reported that NaHS relaxed vascular smooth muscles, including the portal vein and aorta. The maximum relaxation of rat aortic tissues (∼30%) was achieved with 1 mM NaHS. The same concentration of NaHS induced a nearly 100% relaxation in the portal vein (13). The mechanisms for the NaHS-induced vasorelaxation were not further investigated by these authors (13). Our present study demonstrates that both the H2S gas-saturated solution and NaHS relax PE-precontracted rat aortic tissues. In our study, the vasorelaxant potency of H2S is greater than that reported by Hosoki et al. (13). At 180 μM, H2S already relaxes the aortic tissues by 44.5% in our study. This discrepancy could be explained by subtle differences in the handling of vascular tissues in different laboratories. Our study demonstrates that a small increase in H2S concentrations induces a significant relaxation. The steepness of the concentration-relaxation curve indicates that the vascular tissues are extremely sensitive to H2S within a narrow concentration range. Possibly, there are several binding sites for H2S on vascular smooth muscles, as suggested by a Hill coefficient of 4.9 (22). This positive cooperativity indicates that the binding sites on vascular tissues for H2S must be fully occupied and activated before a relaxant response can be elicited (30). Could the intrinsic H2S in the blood vessel wall reach the concentration range between 60 and 600 μM under physiological conditions? It has been shown that the producing rate of H2S in brain tissue was ∼20 nM · min−1 · g protein−1 and its local concentration was ∼100 μM (range from 50 to 160 μM) (1, 28). We have reported that the H2S production rates were 3.6 ± 1.3, 8.7 ± 2.7, and 3.4 ± 0.7 nmol · min−1 · g tissue−1 for the aorta, tail artery, and mesenteric artery, respectively (31). According to these production rates, the concentration of H2S in smooth muscle can be ∼100 μM in certain blood vessel walls. As such, SMCs in vivo may well be exposed to endogenous H2S at concentrations close to the effective concentration range of exogenous H2S used in the present study.
We show in the present study that the vasorelaxant effect of H2S is mainly due to a direct interaction of H2S and SMCs, based on the failure of denervation to alter H2S effect and on the observation that H2S still significantly relaxed vascular tissues after endothelium removal. It is noteworthy, however, that a portion of the H2S-induced vasorelaxation was attenuated either by removal of the endothelium, by the application of l-NAME, or by the combined application of apamin-charybdotoxin to the endothelium-intact vascular tissues. These inhibitory effects were represented by the decrease in the vasorelaxant efficacy of H2S but not the amplitude of the relaxation (Figs. 2 and 3). Putatively, the endothelium might release certain factors in response to H2S stimulation to facilitate the relaxation of smooth muscles. Alternatively, the presence of an intact endothelium might retain H2S in the blood vessel wall so that its vasorelaxant effect can be potentiated and prolonged. The coapplication of apamin and charybdotoxin, a protocol to block the effect of EDHF (7), exerted a similar effect as endothelial removal, i.e., inducing a right shift of the H2S concentration-response curve. It thus appears that H2S may release EDHF from the vascular endothelium. The endothelium-derived NO is also likely released by H2S because the application ofl-NAME to the endothelium-intact vascular tissues partially inhibits the H2S-induced vasorelaxation.
The NO- and CO-induced vasorelaxations are mainly mediated by the cGMP pathway (5, 9, 17). Depending on tissue types, another mechanism may be responsible for the vascular effects of NO and CO, i.e., the activation of large-conductance KCa channels in vascular SMCs (25, 26). Quite different from these two endogenous gases, H2S relaxed vascular tissues independent of the activation of the cGMP pathway or KCa channels. First, the H2S-induced vasorelaxation was not inhibited by ODQ and NS-2028. ODQ (10 μM) or NS-2028 (1 μM) has been shown to specifically inhibit sGC and reduce the production of cGMP (14,16, 29). As shown in Fig. 5 A, the vasorelaxant effect of SNP was inhibited by >90% in the presence of ODQ or NS-2028. In contrast, ODQ and NS-2028 even potentiated the vasorelaxant effect of H2S. The effects of these two agents argue strongly against the notion that the vascular effects of H2S are mediated by the activation of the cGMP pathway. The mechanisms underlying the synergistic effects of these sGC inhibitors and H2S on the vascular tone cannot be fully understood at the moment. Nevertheless, considering the inhibition of this synergistic effect by SOD, a superoxide scavenger, it is postulated that the interaction between ODQ or NS-2028 with H2S may have generated vasorelaxant free radicals (8) that further relaxed vascular tissues. Second, our previous and present studies have shown that neither iberiotoxin (31) nor charibdotoxin nor apamin alone (Fig.3) had an effect on the H2S-induced vasorelaxation. Therefore, KCa channels in vascular SMCs are unlikely the target of H2S.
Because both H2S and NO are vasorelaxant factors, and they have different mechanisms of action, one may predict an additive effect when the two factors are administered together. Hosoki et al. (13) observed that the vasorelaxant effect of NO was enhanced by incubating rat aortic tissues with a low concentration of NaHS. In contrast, pretreating aortic tissues in our study with 60 μM H2S shifted the SNP concentration-response curve to the right (Fig. 7). In other words, H2S inhibits the vasorelaxant effect of SNP. The discrepancy between our results and those from the previous study (13) is not readily explainable. The experimental conditions of these studies are different. Hosoki et al. (13) used helical tissue strips of the aorta from Wistar rats, and we used aortic rings from Sprague-Dawley rats. The tissue damage of helical strips is certainly greater than that of ring preparations (4). This may also explain why, in the study by Hosoki et al. (13), the vasorelaxant potency of NaHS was much greater in portal vein rings than in aorta strips. Furthermore, 0.3 μM PE was used in our study, whereas 1 μM norepinephrine was used in the study of Hosoki et al. (13). It appears that the maximal contraction can be induced by 1 μM norepinephrine, whereas 0.3 μM PE (a submaximal concentration) only induces ∼90% of the maximal contraction of rat aortic tissues. The advantage of using a submaximal concentration of PE is that the tissue can react with the relaxant agent in a more sensitive way (6, 11). Nevertheless, our results can be explained by the following hypothesis. Sulfide interacts with a number of enzymes and other macromolecules (10) and leads to the inhibition of cGMP accumulation (15). The inhibition of cGMP pathway by H2S per se may not suffice to alter the vascular tone, i.e., to evoke contraction. It might exert an inhibitory influence, however, on the SNP-induced vasodilation because the vasorelaxant effect of SNP is mainly mediated by the cGMP pathway.
Our novel observations can be summarized as the following.1) The vasorelaxant effect of H2S is mainly mediated by an interaction of the gas with smooth muscles and partially by a functional endothelium. 2) The H2S-induced vasorelaxation is not mediated by the cGMP pathway, although the specific sGC inhibitors (ODQ and NS-2028) further strengthen the H2S-induced vasorelaxation through a novel mechanism.3) The H2S-induced vasorelaxation is extracellular calcium entry dependent. These observations provide clues pertinent to the mechanisms for the H2S-induced vasorelaxation. Both endothelium and vascular smooth muscles may serve as the targets of H2S. By acting on the endothelium, H2S may facilitate the release of vasorelaxant factors, including NO and EDHF. By directly acting on vascular SMCs, H2S may reduce the extracellular calcium entry and relax vascular tissues. Our previous study has shown that H2S stimulated KATP channels in vascular SMCs (31). The opening of KATP channels leads to membrane hyperpolarization, which in turn may close voltage-gated Ca2+ channels. Alternatively, H2S may directly inhibit voltage-gated Ca2+ channels in vascular SMC, a possibility that merits further investigations.
Our study demonstrates that H2S relaxes vascular tissues in different mechanisms from those of NO and CO, although all of these three gas molecules are endogenously generated potent vasorelaxant factors. Furthermore, we show that a prior exposure to H2S reduces the relaxant response of vascular tissues to NO, but the presence of NO does not alter the vascular effect of H2S. The coordinated actions of these endogenous gases may provide an integrated modulation of vascular contractility.
The authors thank K. Safiniuk and G. Beal for technical assistance.
This study was supported by the Natural Sciences and Engineering Research Council of Canada. R. Wang has been supported by an Investigator Award of Canadian Institutes of Health Research/Regional Partnership Program.
Address for reprint requests and other correspondence: R. Wang, Dept. of Physiology, College of Medicine, Univ. of Saskatchewan, Saskatoon, SK, Canada, S7N 5E5 (E-mail:).
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.
April 18, 2002;10.1152/ajpheart.00013.2002
- Copyright © 2002 the American Physiological Society