Hydrogen sulfide, an enhancer of vascular nitric oxide signaling: mechanisms and implications

Abstract

Nitric oxide (NO) vascular signaling has long been considered an independent, self-sufficient pathway. However, recent data indicate that the novel gaseous mediator, hydrogen sulfide (H₂S), serves as an essential enhancer of vascular NO signaling. The current article overviews the multiple levels at which this enhancement takes place. The first level of interaction relates to the formation of biologically active hybrid S/N species and the H₂S-induced stimulation of NO release from its various stable “pools” (e.g., nitrite). The next interactions occur on the level of endothelial calcium mobilization and PI3K/Akt signaling, increasing the specific activity of endothelial NO synthase (eNOS). The next level of interaction occurs on eNOS itself; H₂S directly interacts with the enzyme: sulfhydration of critical cysteines stabilizes it in its physiological, dimeric state, thereby optimizing eNOS-derived NO production and minimizing superoxide formation. Yet another level of interaction, further downstream, occurs at the level of soluble guanylate cyclase (sGC): H₂S stabilizes sGC in its NO-responsive, physiological, reduced form. Further downstream, H₂S inhibits the vascular cGMP phosphodiesterase (PDE5), thereby prolonging the biological half-life of cGMP. Finally, H₂S-derived polysulfides directly activate cGMP-dependent protein kinase (PKG). Taken together, H₂S emerges an essential endogenous enhancer of vascular NO signaling, contributing to vasorelaxation and angiogenesis. The functional importance of the H₂S/NO cooperative interactions is highlighted by the fact that H₂S loses many of its beneficial cardiovascular effects when eNOS is inactive.

The Vascular eNOS/sGC/cGMP/PKG Pathway May Not Be Completely Self-Sufficient

Vascular NO production (overviewed in 12, 39, 43, 53, 74, 89, 90, 107) is predominantly due to endothelial NO synthase (eNOS), a calcium-dependent enzyme constitutively expressed in vascular endothelial cells.¹ Various vasorelaxant and angiogenic hormones and factors, as well as shear stress, lead to calcium mobilization in the endothelial cells, which activates eNOS in a calmodulin-dependent manner. In the presence of various co-factors (e.g., NADPH and BH₄), eNOS converts its physiological substrate l-arginine to NO and l-citrulline. In addition to calcium, eNOS is also regulated by phosphorylation/dephosphorylation at several critical regulatory amino acid residues. NO, produced by eNOS, either reaches its targets within the endothelial cell itself, or diffuses to the underlying vascular smooth muscle cells. In turn, NO binds to the heme group of its target enzyme, soluble guanylate cyclase (sGC), and activates it. The sGC-mediated production of cGMP, via stimulation of downstream enzymes (cGMP-dependent protein kinases, PKGs) is primarily responsible for the biological effects of eNOS, such as vascular relaxation and angiogenesis. Vascular cGMP levels are physiologically degraded by phosphodiesterase 5 (PDE5) (19, 26, 105). The vascular eNOS/sGC/cGMP/PKG pathway, one of the most intensively studied signaling pathways in biology, is generally considered a stand-alone, self-sufficient pathway that does not rely on external biochemical enhancers.

Several decades after the discovery of the essential role of the NO/sGC/cGMP pathway in the control of the cardiovascular system, the regulatory roles of another gaseous mediator, hydrogen sulfide (H₂S), started to emerge (overviewed in 55, 57, 60, 61, 66, 85, 108, 109, 125–129, 139, 143, 145). In brief, H₂S is produced in the vascular system by three distinct enzymes, cystathionine-gamma-lyase (CSE), cystathionine-beta-synthase (CBS), and 3-mercaptopyruvate sulfurtransferase (3-MST). The substrates of CBS and CSE are l-cysteine and homocysteine; the substrate of 3-MST is 3-mercaptopyruvate which is produced from l-cysteine. H₂S exerts its biological effects via a variety of mechanisms including posttranscriptional modification of critical cysteines in various enzymes via a novel process entitled S-sulfhydration. Similar to NO, H₂S causes vasorelaxation (143, 144), participates in the physiological maintenance of blood pressure (149), and serves an endogenous stimulator of angiogenesis (15, 106, 124). In addition, similar to NO, which converts into various stable or semi-stable “pools” (e.g., nitrite) and can be regenerated from it under certain conditions (59, 76, 77), H₂S converts into thiosulfate, which can regenerate biologically active H₂S (79, 117, 141).

Not only do NO and H₂S exhibit biological and functional similarities in the cardiovascular system, but several lines of data, most of which has emerged over the last 5 years, indicate that the two pathways, in fact, cooperate with each other. In the vascular system, H₂S, in many respects, is now viewed as an enhancer of the NO/cGMP/sGC/PKG pathway, without which eNOS cannot function to its fullest physiological extent. The biosynthesis, biological effects, metabolism, and physiological and pathophysiological roles of H₂S in a variety of diseases are subject to separate review articles (55, 57, 60, 61, 66, 85, 108, 109, 125, 127, 128, 139, 143, 145). The sole focus of the current review is to summarize the mechanisms by which H₂S acts as an enhancer of the vascular eNOS/sGC/cGMP/PKG system.

H₂S Stimulates NO Release from Its Stable or Semi-Stable Pools

Starting with the work of Moore, Whiteman, and coworkers (1, 147, 150) the notion began to emerge that the vascular effects of NO and H₂S may be interdependent, and may be, at least in part, related to the formation of a combined NO/H₂S species, i.e., a nitrosothiol (147). These findings, together with earlier observations of Kimura and coworkers who demonstrated that H₂S enhances the vascular relaxant effect of NO (47), suggested that H₂S may act as an enhancer of vascular eNOS/sGC/cGMP/PKG signaling, an effect, which, perhaps, is most relevant in the microvasculature (143, 144).

There are multiple levels of direct chemical interactions between NO (and its various metabolites) and H₂S (and its various metabolites) such that the field of NO/H₂S “cross-talk” has emerged as a separate area of biochemistry, which investigates (among others) the reaction products of H₂S with nitrosothiols, peroxynitrite, and nitrite (8, 17, 22–25, 37, 63, 73, 134, 146, 147). The biological responses induced by HNO (nitroxyl), one of the products of the H₂S/NO interactions, has recently been reviewed by Nagpure and Bian (91). There are several open areas in this rapidly growing field, which are overviewed in several recent articles (24, 41, 63, 67, 73, 78, 91).

Although the goal of this article is not to outline the complex chemistry of H₂S-related biological species, it must be mentioned that in biological systems, at or near physiological pH, in the presence of various biologically relevant and redox active molecules (e.g., thiols, transition metals, proteins), even when starting out from “pure” H₂S gas (e.g., by “bubbling” H₂S into the culture medium of cells), a complex mixture of sulfur species will be created. Because the dissociation constants of H₂S are pK_a1 = 6.8 and pK_a2 >12, at pH 7.4, only about a quarter of H₂S gas will remain as dissolved gas: H₂S will be predominantly present in the form of hydrosulfide anion (HS⁻). At physiological pH, only minute amounts of H₂S will convert into S₂⁻ (73, 92). In addition to these (primary) H₂S-related species, in biological systems, due to a complex sequence of (bio)chemical interactions that are presently only partially understood, H₂S will also lead to the formation of a complex mixtures of organic persulfides and polysulfides (52, 100). While H₂S and HS⁻ are generally believed to have similar (although probably not identical) biological character, the chemical reactions elicited by polysulfides are drastically different (6, 52, 61, 100, 135). Downstream from H₂S, the biological liberation of sulfur (thiyl) radical, the highly unstable intermediate HS molecule, the reactive S⁻ molecule, as well as the formation of other reactive sulfur species including S₂⁻, is also chemically feasible (24, 92, 94, 103). When this “soup” of H₂S-derived reactive species reacts with NO (or with a “soup” of NO-derived species), a wide variety of bioactive intermediates can form, including nitrosopersulfide (SSNO⁻), and SULFI/NO [ON(NO)SO₃⁻] (24). These molecules are biologically active: therefore, via this interaction, H₂S may create complex NO metabolites that modify (perhaps enhance) the biological action of NO (8, 23) (Fig. 1, arrow 1). The exact (patho)physiological role of the various S/N hybrid species remains to be further elucidated.

Fig. 1.

Enhancement of the eNOS-mediated vascular signaling by H₂S. On the level of direct biochemical interactions, H₂S reacts with NO to produce biologically active S/N species (arrow 1). In addition, H₂S has the capacity to release NO from stable and semi-stable “NO storage pools” (e.g., nitrite, nitrosothiols, peroxynitrite) (arrow 2). In addition, H₂S can stimulate intracellular Ca²⁺ mobilization in endothelial cells (arrow 3), which stimulates eNOS via the calcium-calmodulin system. Importantly, H₂S can activate the PI3K pathway by either by sulfhydrative inactivation of PTEN (arrow 4) or by sulfhydrative inactivation of PTP1B (arrow 5). These processes lead to the enhancement of the PI3K pathway, culminating in the elevation of intracellular PIP3 levels. PIP3, in turn, activates Akt, which, then, acts on various downstream kinases and phosphatases (not shown). The end-result is the phosphorylation on an activating serine of eNOS (Ser1177). Calcium mobilization and Akt activation both stimulate eNOS, resulting in increased eNOS-dependent endothelial NO production upon eNOS-stimulating factors, e.g., binding of various calcium-mobilizing hormones acetylcholine or VEGF to their receptors on the endothelial cell membrane (not shown). H₂S also induce a direct posttranslational modification of eNOS via sulfhydration on Cys443 (arrow 6). This action, in turn, further activates and stabilizes eNOS by promoting its dimerization. There is also some evidence that in some experimental systems H₂S can stimulate the expression of eNOS mRNA (arrow 7), putatively increasing eNOS protein levels. Further downstream, NO produced by eNOS binds to soluble guanylate cyclase (sGC), and induces the formation of cGMP. In endothelial cells, in the context of endothelial cell proliferation and migration, the sGC target of NO is in the endothelial cell itself. In the context of the endothelium-dependent relaxation, NO diffuses from the endothelial cell to the underlying vascular smooth muscle cell, where it binds to its sGC target in the vascular smooth muscle cytosol. (The endothelial cell and the smooth muscle cell sGC targets are not separated in the current scheme). H₂S stimulates the sGC/cGMP pathway via several distinct mechanisms. As a reducing agent, H₂S can induce the stabilization of sGC in its reduced, NO-responsive form (arrow 8). H₂S can also react with cGMP to form 8-SH-cGMP (arrow 9), which is biologically active and less sensitive to enzymatic degradation by PDE5. In addition, H₂S, by inhibiting PDE5 (arrow 10), suppresses the degradation of cGMP, thereby ensuring that the cGMP levels remain high for a longer period of time, thereby enhancing the activation of its downstream protein kinase, PKG. Finally, the degradation of H₂S to intracellular polysulfides can induce the oxidative activation of PKG (arrow 11). These processes culminate in the prototypical vascular biological responses induced by both NO and H₂S such as angiogenesis and vasorelaxation. Depending on the biological context, H₂S may also stimulate vascular responses via pathways that do not involve the eNOS pathway (arrow 12).

For the purpose of the current review, the next relevant reaction is that of H₂S and various semistable pools of NO (e.g., nitrite) leading to the generation of NO (11) (Fig. 1, arrow 2). As overviewed elsewhere (59, 76, 77), nitrite was initially believed an inactive, stable metabolite of NO, but it is now recognized as an important constituent of a biologically active, physiological “NO pool”; in fact, via administration of nitrite, therapeutically relevant, NO-mediated biological responses can be achieved (such as vascular relaxation, cardioprotection, and angiogenesis). The fact that H₂S can stimulate NO release from nitrite, as shown in hypoxic endothelial cells in vitro (11), and furthermore, the fact that the reaction of H₂S with peroxynitrite can produce the NO donor sulfinyl nitrite (37) may suggest that in the presence of H₂S, nitrite may become more biologically active and/or perhaps therapeutically more efficacious. Moreover, peroxynitrite, a reactive NO-derived species (7, 123), may be rendered less noxious by H₂S. In cultured endothelial cells the H₂S-induced release of NO from nitrite is, at least in part, catalyzed by xanthine oxidase (11). In addition to nitrite and peroxynitrite, H₂S can also induce NO release from nitrosothiols and metal nitrosyl complexes (99). It is worth mentioning that not only H₂S can stimulate NO production, but also nitrite can stimulate H₂S production: in a mouse model of chronic heart failure, treatment of the animals with sodium nitrite was found to induce the upregulation of the H₂S-producing enzymes CSE and CBS in the heart; consequently, nitrite treatment was found to elevates circulating H₂S levels (31).

H₂S Stimulates eNOS Activity via Stimulation of Calcium Mobilization

Since eNOS is a calcium-dependent enzyme, intracellular calcium mobilization, in response to endothelium-dependent vasodilatators and angiogenic hormones, is a rapid, potent activator of eNOS. Several lines of studies indicate that this process can be stimulated in the presence of H₂S. First, in rat aorta endothelial cells, Moccia and colleagues demonstrated that H₂S stimulates Ca²⁺ entry in a tetraethylammonium- and glybenclamide inhibitable manner, suggesting the involvement of reverse-mode of K_ATP channels (84). H₂S may also recruit the Na⁺/Ca²⁺ exchanger in a reverse mode (84). Subsequent studies showed that H₂S increased intracellular Ca²⁺ levels in cultured endothelial cells, and this response was inhibitable by dantrolene or xestospongin C (58). Finally, in the HUVEC-derived cell line Ea.hy926, Potenza and colleagues demonstrated H₂S elicited a time- and concentration-dependent intracellular Ca²⁺ response, which was attributed to inositol-1,4,5-trisphosphate-dependent Ca²⁺ release (111). The findings of Potenza and colleagues (showing that the endothelial calcium signal is identical in the presence or absence of extracellular calcium) are in disagreement with the findings of Moccia (suggesting that endothelial cell calcium mobilization in response to H₂S requires the influx of extracellular calcium). Whether the source of calcium is extra- or intracellular, the findings discussed above identify endothelial Ca²⁺ mobilization as the first, most upstream level at which H₂S may enhance the activity of the eNOS/sGC/cGMP/PKG system (84) (Fig. 1, arrow 3). It must be pointed out, nevertheless, that the above-referenced studies investigate calcium mobilization (in response to H₂S), and present functional outcome variables (e.g., endothelial cell migration), but do not directly (e.g., by measuring l-arginine to l-citrulline conversion) show in the endothelial cells that the increased Ca²⁺ signal, is, indeed, directly responsible for the activation of eNOS; further studies, directly testing the link of the H₂S-induced Ca²⁺ signal with eNOS activity, remain to be conducted.

H₂S Stimulates eNOS Activity via AKT-Mediated Phosphorylation

A key level of the regulation of eNOS activity occurs at the level of phosphorylation/dephosphorylation of its critical regulatory amino acids. The best established regulatory amino acid residue is Ser1177 (the phosphorylation of which stimulates eNOS activity). Ser1177 phosphorylation, which increases the specific activity of eNOS, are facilitated by Akt activation in endothelial cells (29, 42, 82). Upstream, the activation of Akt in endothelial cells, is regulated by intracellular PIP3 levels, which are under the control of the PI3 kinase (PI3K) system (92).

The H₂S-induced cytoprotection and cell migration involves the PI3K system which stimulates its downstream effectors such as Akt and Rac-1 (15, 152). In various cultured endothelial cell systems, several lines of studies demonstrate that H₂S stimulates Akt by promoting its phosphorylation at its activating site (Ser493) (15, 21). H₂S also induces Akt activation in other cells types (50, 78, 132). Whether the H₂S-mediated increase in Akt is due to a direct effect of H₂S on PI3K remains to be determined. Part of the stimulatory effect on AKT activity is likely to be related to H₂S-mediated inhibition of Phosphatase and Tensin homolog (PTEN), an essential counterregulatory enzyme of the PI3K pathway, via sulfhydration of Cys124 (78, 98) (Fig. 1, arrow 4). The activity of PTP1B (another counterregulatory enzyme in the PI3K pathway) is also inhibited by H₂S, via sulfhydration at Cys215 (71) (Fig. 1, arrow 5).

Whatever the upstream processes are (direct stimulation of PI3K and/or indirect effects via PTEN or PTP1B or via other mechanisms), elevated PIP3 levels by H₂S would be expected to induce Akt activation, which, in turn, would be expected to produce the characteristic changes in the phosphorylation of regulatory amino acids on eNOS. Indeed, this is the case, as demonstrated by multiple studies in response to administration of various H₂S donors in vitro and in vivo (2, 3, 6, 11, 18, 21, 49, 56, 62, 75, 80, 102, 112, 113). For instance, in endothelial cells incubated with H₂S donors, the activation of Akt (evidenced by phosphorylation at Ser473) was accompanied by the phosphorylation of Ser1177 on eNOS (21). In endothelial cells subjected to shear stress, H₂S was found to stimulate eNOS phosphorylation at Ser1177 (49). Similarly, H₂S was shown in multiple studies (in the heart as well as in the coronary arterioles) to stimulate Ser1177 phosphorylation of eNOS in vivo (62, 103, 113). In endothelial cells subjected to shear stress, Ser1177 phosphorylation was diminished after siRNA-mediated silencing of either CSE, CBS, or 3-MST genes (49); likewise, in CSE^−/− mice the degree of basal eNOS phosphorylation at Ser1177 is markedly lower than the corresponding basal Ser 1177 phosphorylation in wild-type mice (62). These findings indicate that basal, physiological Ser1177 phosphorylation of eNOS is maintained by endogenous H₂S through the basal physiological actions of various H₂S-producing enzymes. The functional relevance of the H₂S-mediated eNOS phosphorylation is highlighted by the results Lefer and co-workers who demonstrated that that the protective effect of H₂S was markedly lower in eNOS phosphomutant S1179A mice, than in the corresponding wild-type control mice (62).²

There may be additional levels at which H₂S stimulates the generation of PIP3, even further upstream from PI3K and PTEN. One example relates to VEGF (a proangiogenic and vasorelaxant hormone). The endothelial cell receptor of VEGF (VEGFR2) is a strong activator of PI3K through VEGRF autophosphorylation (150). Tao and coworkers demonstrated that H₂S can reduce the intramolecular Cys1024-Cys1045 disulfide bond in VEGFR2, which, in turn, increases the PI3K stimulating activity of this receptor, culminating in an increased angiogenic response (133).

H₂S Stimulates eNOS Activity via Direct Sulfhydration

eNOS homodimers represent the physiological state of the enzyme, where it produces NO in the most efficient manner, as well as the least amount of superoxide. On the other hand, monomeric eNOS is prone to superoxide generation (33, 114, 131). Wang and colleagues have demonstrated that H₂S sulfhydrates Cys443 of eNOS, which, in turn, stimulates the catalytic activity of eNOS, thereby keeping the enzyme in the dimeric state and suppressing superoxide generation by eNOS and maximizing physiological NO generation (2) (Fig. 1, arrow 6).

H₂S Stimulates eNOS mRNA Synthesis

H₂S was found to increase the expression of eNOS (Fig. 1, arrow 7) in some systems, for instance in EAhy926 cells in vitro (18). In contrast, in other experimental settings only a slight and nonsignificant trend towards increased eNOS protein expression was noted in response to H₂S exposure (3). In some experimental systems, H₂S can also induce an upregulation of eNOS mRNA in vivo. For instance, in a murine model of thrombus formation induced by phototoxic light/dye-injury, H₂S administration caused the upregulation of eNOS in venules of the ear of hairless SKH1-hr mice (69). Moreover, in kidney reperfusion model in the rat, H₂S administration maintained eNOS expression levels during ischemia (while eNOS was downregulated in the ischemic control group that did not receive H₂S) (51). eNOS mRNA expression also increased after H₂S treatment in the corpus cavernosum (81). While H₂S donation can stimulate eNOS mRNA expression, basal physiological production of H₂S does not appear to be an absolute requirement for eNOS expression: eNOS expression levels were comparable in wild-type mice and CSE^−/− mice (62).

H₂S Maintains sGC in an NO-Activatable State

Cells contain two distinct sGC pools: the reduced, ferrous heme group of the enzyme (Fe²⁺) is the physiological form, which is responsive to NO in the classical NO/sGC/cGMP/PKG pathway. However, the heme group of sGC can also be oxidized (e.g., in various pathophysiological states) to yield the ferric (Fe³⁺) that is no longer NO-activatable (9, 28, 34, 38, 105). Recent studies show that H₂S can act as a reducing agent for the heme group of sGC, catalyzing a ferric to ferrous heme transition (154). ODQ (1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one) is a sGC inhibitor that acts via oxidizing the heme group (46). Treatment of recombinant sGC with ODQ reduced the responsiveness of the enzyme to NO, an effect that was attenuated if in the presence of H₂S. On the other hand, H₂S attenuated the response to the heme-independent activator BAY58-2667 that targets oxidized sGC. The beneficial effect of H₂S on maintaining an NO-activatable state of sGC (Fig. 1, arrow 8) has been confirmed in oxidatively stressed endothelial cells as well as in isolated vascular rings (154). The above-described redox effect of H₂S on the heme group of sGC appears to be most relevant under conditions of oxidative stress, because in the absence of oxidative stimuli, H₂S does not influence sGC activity, neither under basal conditions, nor in response to NO donors (21, 154).

H₂S Reacts with cGMP to Yield 8-SH-cGMP

The next level of cooperation between H₂S and the vascular NO/sGC/cGMP/PKG pathway is at the level of cGMP itself. Modified cyclic nucleotides, including 8-nitro-cGMP, play key roles in vascular regulation (95, 96). Recent data show that H₂S directly reacts with 8-nitro-cGMP, yielding 8-SH-cGMP. This molecule appears to be more resistant to PDE5 than “normal” cGMP (95, 96) and, in turn stimulates PKG activity (Fig. 1, arrow 9). 8-Nitro-cGMP/8-SH-cGMP imbalance has been implicated in a variety of pathophysiological conditions including cardiac ischemia and remodeling (95, 96, 100).

H₂S Inhibits PDE5 Activity

Vascular cGMP phosphodiesterases (PDEs), predominantly PDE5, are essential negative regulators of cGMP levels. Inhibition of PDE5 activity suppresses cGMP degradation and, by maintaining/prolonging intracellular cGMP levels, enhances downstream signaling. In effect, in the presence of PDE inhibitors, the same amount of NO gets a “bigger bang for the buck” in terms of cGMP-dependent downstream vascular responses. PDE5 inhibitors, by boosting cGMP signaling, have successfully been employed not only in the therapy of male erectile dysfunction, but also pulmonary hypertension and other cardiovascular diseases (19, 26, 105). Two independent laboratories have demonstrated that H₂S inhibits PDE activity (13, 21), which, in turn, boosts NO-stimulated cGMP levels in vascular endothelial cells and vascular smooth muscle cells (Fig. 1, arrow 10). Although H₂S inhibits all known isoforms of both cAMP- and cGMP phosphodiesterases, it is most potent in inhibiting PDE5 (104). The mechanism of the inhibition does not appear to involve PDE sulfhydration (9, 10). The functional consequence of this interaction is that H₂S elevates intracellular cGMP in vascular cells and tissues, thereby promoting and enhancing NO-mediated vascular responsiveness. This is also evidenced by H₂S-induced increases in the phosphorylation of the PKG target VASP (13, 14, 21). There is also evidence for inhibition of the mitochondrial cAMP PDE isoform (PDEA2) by H₂S (88); the functional consequence of this response is the stimulation of mitochondrial electron transport; this effect cooperates with the other mechanisms, including direct electron donation, and activation of ATP synthase (44, 88, 87, 122, 125), by which physiological concentrations of H₂S stimulate mitochondrial electron transport and cellular bioenergetics.

Polysulfides Directly Activate PKG

Another layer of interaction between H₂S and the NO/sGC/cGMP/PKG pathway occurs at the level of protein kinase G (PKG) activation. PKG is the essential effector of eNOS-mediated vasorelaxant, angiogenic, and cardioprotective responses (40, 115). Greiner and coworkers demonstrated that polysulfides (molecules that are formed from H₂S intracellularly via a variety of mechanisms, see above) are capable of catalyzing the formation of an activating interprotein disulfide within PKG Iα, thereby stimulating the activity of PKG (45) (Fig. 1, arrow 11). This has important consequences for the cardiovascular regulatory effect of H₂S: while in control mice H₂S lowers blood pressure, in genetically engineered mice in which PKG's Cys42 redox sensor has been rendered redox-insensitive, the blood pressure effect of H₂S was attenuated (120). PKG may also have other levels of interaction with the NO or H₂S systems, although the information on this subject remains somewhat fragmented. In some cases, evidence of a positive interaction has been presented, for instance PKG was reported to induce CSE expression in the heart (27), whereas in other experimental systems PKG-induced phosphorylation and consequent inhibition of CSE have been observed (4).

Consequences of the NO-H₂S Cooperation for the Control of the Cardiovascular System

If the cooperative biochemical interactions between H₂S and the NO/sGC/cGMP system outlined above and summarized in Fig. 1 are physiologically or pathophysiologically relevant, then the following functional consequences would be expected: 1) H₂S administration would be expected to increase vascular NO production, with a consequent boost of vascular cGMP levels; 2) H₂S administration would be expected to potentiate the vasorelaxant and angiogenic effect of NO; 3) if endogenously produced H₂S plays a cooperative role with the NO system, then inhibition of vascular H₂S biosynthesis would be expected to attenuate the vasorelaxant and angiogenic effect of NO; 4) if the cardiovascular effect of H₂S primarily occurs via enhancement of the NO pathway, then in the absence of functional NO/sGC/cGMP/PKG system the vasorelaxant and angiogenic effects of H₂S would be expected to be diminished. There is, indeed, direct experimental evidence for all of the above-mentioned scenarios, as outlined below.

1) H₂S administration elevates vascular NO and cGMP levels.

In various models of cultured endothelial cells, H₂S has been found to both increase NO levels (3, 18, 112) and to elevate cGMP levels (9, 21). These effects were also associated with an increase in VASP phosphorylation (21), indicative of the fact that H₂S exposure results in the stimulation of the full NO/sGC/cGMP pathway, culminating in an increase in PKG activity. Not only authentic H₂S, but also the CBS/CSE substrate l-cysteine, was found to increase endothelial NO levels, in a manner that was reduced by the CSE inhibitor (5) PAG (3), confirming the ability of the endogenous endothelial H₂S biosynthesis to contribute to NO production. In a cultured rat corpus cavernosum preparation, various H₂S donors were also found to increase NO levels (81). Importantly, the vasorelaxant and angiogenic responses to H₂S are attenuated by the PKG-Iα inhibitor DT-2 (14, 21). In addition, l-cysteine- and H₂S-induced relaxations are less pronounced in aortae isolated from PKG-I KO mice than the corresponding control relaxations in aortae isolated from wild-type mice (14). These data, taken together, support the notion that stimulation of NO production and/or enhancement of the cGMP/sGC system, culminating on PKG activation, represent essential components of the vascular signaling induced by H₂S. There are examples of H₂S donation increasing circulating NO levels as well. For instance, it was demonstrated that the H₂S donor diallyl trisulfide increases circulating NO levels in a pressure overload-induced heart failure model in mice (107). Moreover, Bir and colleagues demonstrated that H₂S donor therapy increases circulating NO levels in mice (11). Finally, Wang and colleagues found that plasma NO levels are lower in CSE-deficient mice than in wild-type mice (2), concluding that endogenous H₂S production contributes to the physiological maintenance of circulating NO levels.

2) H₂S administration potentiates the vascular effect of NO.

Evidence for a potentiating effect of H₂S on vascular relaxations induced by exogenous NO was first provided by Kimura and co-workers in 1997: H₂S donation was found to enhance the relaxant effect of various NO donors (Na-nitroprusside and morpholinosydnonimine) in rat portal vein and thoracic aortic ring preparations, in an endothelium-independent fashion (47). We have demonstrated in 2012 that the acetylcholine-induced, endothelium-dependent relaxations, as well as the relaxant responses to the NO donor DEA/NO, are enhanced (concentration-response curves shifted to the left) in the presence of exogenously administered H₂S, and this functional synergy was accompanied with a synergistic enhancement of vascular cGMP levels (21). With respect to angiogenesis, vascular overexpression of CSE was found to enhance the stimulating effect of both DEA/NO and VEGF in an aortic ring sprouting assay (21). Moreover, a recent study demonstrated that ZYZ-803 (a novel synthetic H₂S/NO hybrid molecule) induces endothelial cell proliferation, migration, and tube-like structure formation and angiogenesis in rat aortic rings and Matrigel plug assay: all of these actions occur at greater potency than the effect of either H₂S or a NO donor on its own (48). ZYZ-803 also exerts vasorelaxant effects via stimulation of the cGMP pathway (148).

3) Inhibition of vascular H₂S biosynthesis attenuates the vascular effect of NO.

Although the eNOS/sGC/cGMP/PKG system is generally believed to be an independent signaling pathway, surprisingly, it relies, at least in part, on endogenous H₂S production. This component is relatively smaller in the context of vascular relaxation, but it is surprisingly marked in the context of angiogenesis. Wang and Snyder have observed that vasorelaxant effect of metacholine (which, similar to acetylcholine, relaxes blood vessels via activation of cholinergic receptors and stimulation of the eNOS/sGC/cGMP/PKG system) is lower in blood vessels of CSE^−/− mice than the corresponding response in wild-type mice (149). In addition, the vasorelaxant effect of the NO donor SNAP is less pronounced in vascular rings of CSE^−/− mice than the vasorelaxation observed in rings of wild-type control mice (149). Confirming and extending these observations, we have demonstrated that acetylcholine, as well as DE/NO-induced vascular relaxation dose-responses, are shifted to the right in thoracic aortic rings after vascular CSE silencing; this was also associated with a reduction in vascular cGMP levels (21). As far as interactions in angiogenesis: DEA/NO-induced in vitro angiogenic activity (as well as the ability of VEGF to induce angiogenesis) was markedly attenuated after silencing of CSE in endothelial cells; once again, after CSE silencing, the NO donors failed to induce elevations in endothelial cGMP levels (21). H₂S-induced angiogenesis (which served as the control group in this experiment) remained unaffected after CSE silencing. In line with these findings, Wang and coworkers have demonstrated that the angiogenic effect of the eNOS substrate l-arginine is markedly less pronounced in aortic rings of CSE^−/− mice compared with the responses in wild-type mice (3).

4) Inhibition of eNOS attenuates the vascular effect of H₂S.

Moore and colleagues observed in 2006 that the hypotensive effect of intravenous H₂S infusion in rats is suppressed by treatment with the NOS inhibitor l-NAME (1), and similar findings were recently reported in human volunteers: the local vasodilatory response to administration of H₂S was reduced by l-NAME (73). The in vitro vascular relaxant effect of H₂S is also reduced by pretreatment with the NOS inhibitor l-NAME, and it is shifted to the right in vascular rings from eNOS-deficient mice (21), indicating that part (but not all) of the vascular relaxations induced by H₂S require the eNOS/sGC/cGMP/PKG pathway. Since H₂S exerts vascular relaxant effects via a number of additional mechanisms, including K_ATP channel opening, as well as, at higher concentrations, by direct metabolic effects (64, 66), it is not surprising that NOS inhibition only abrogates part of its vascular relaxant responses. However, it appears that the NO system is obligatory for H₂S-induced in angiogenesis: H₂S mediated in vitro angiogenic activity (as well as the ability of H₂S to induce an elevation of endothelial cGMP levels) was completely abrogated by pretreatment of the endothelial cell cultures with the NOS inhibitor l-NAME (21); likewise, the stimulatory effect of H₂S donors in the Matrigel plug assay in vivo was absent in eNOS^−/− mice (21). In agreement with these findings, Wang and colleagues demonstrated that l-NAME, as well as vascular eNOS silencing, abrogates H₂S-induced endothelial cell proliferation, tube formation, and aortic ring sprouting responses in vitro, whereas eNOS overexpression facilitates the angiogenic effect of H₂S (3).³

5) H₂S-induced cardiovascular and therapeutic actions are abrogated in eNOS-deficient systems.

Examples for this are shown in Table 1. Given the fact that (as discussed above) H₂S-induced angiogenic responses (Matrigel plug responses) are suppressed in eNOS-deficient mice (21), it should not come as a surprise that the stimulating effect of H₂S on wound healing (an effect, which, at least in part, relies on the stimulation of angiogenesis) is also suppressed after pretreatment of the animals with the NOS inhibitor l-NAME (21). In isolated perfused hearts, the protective effect of H₂S against isoproterenol-induced cardiac injury is attenuated by l-NAME pretreatment (118). In addition, the well-established (32) ability of H₂S pretreatment to reduce infarct size in a murine model of ischemia-reperfusion injury is absent in eNOS deficient mice (62). Similarly, chronic treatment with SG-1001 (a novel H₂S donor) was no longer protective against the development of heart failure, when the experiments were conducted in eNOS deficient mice (68). There may be some species differences in the extent to which the H₂S-mediated cardioprotection relies on the NOS system: the cardioprotective effect of H₂S pretreatment prior to coronary occlusion/reperfusion is suppressed by l-NAME in a mouse model of myocardial ischemia-reperfusion, while in a similar model, performed in rabbits, the protective effect of H₂S is maintained even after pretreatment with l-NAME (10). In a mouse cardiac arrest/resuscitation model, the H₂S treatment, administered shortly prior to the start of the CPR procedure, prolongs survival in wild-type animals, but is without significant therapeutic effect in eNOS deficient mice (83). Finally, the H₂S-induced preconditioning responses on leukocyte rolling, but not on leukocyte adhesion, are reduced in eNOS^−/− mice, compared with wild-type mice (151). There are, however, examples when eNOS deficiency does not abrogate the beneficial effect of H₂S. For instance, in a model of angiogenesis induced by permanent femoral artery ligation, the beneficial effects of H₂S are maintained even on eNOS-deficient background (11).

Table 1.

Examples for abrogation of H₂S-induced cardiovascular responses in eNOS-deficient systems

Experimental System	Mode of eNOS Inhibition	Biological Response to H2S That Is Impaired in the Absence of Functional eNOS	Ref. No.
Blood pressure measurement in anesthetized rats	l-NAME	Hypotension	1
Cultured bEnd3 endothelial cell angiogenesis	l-NAME	Migration, proliferation, tube formation	21
Cultured bEnd3 endothelial cells	l-NAME	cGMP formation	21
Cultured bEnd3 endothelial cell angiogenesis	eNOS silencing	Migration, proliferation, tube formation	3
Aortic ring sprouting assay	eNOS−/−	Endothelial cell migration	21
Isolated murine aortic rings	l-NAME	Vasorelaxation	21
Isolated murine aortic rings	eNOS−/−	Vasorelaxation	21
Normothermic cardiac arrest, followed by resuscitation by cardiac compression and respiration in mice	eNOS−/−	Prolongation of survival	81
Matrigel plug assay in mice	eNOS−/−	Angiogenesis	21
Burn induced wound healing in rats	l-NAME	Wound closure	21
Isoproterenol-induced cardiac injury in mice	l-NAME	Cardiac lactate dehydrogenase and creatine kinase release, cardiac TBARS, glutathione, SOD and catalase activity, TNF expression, histopathological alterations	118
Myocardial infarction induced by transient LAD occlusion in mice	eNOS−/−	Infarct size	62
Myocardial infarction induced by transient LAD occlusion in mice	l-NAME	Infarct size	10
Chronic heart failure induced by aortic constriction in mice	eNOS−/−	Cardiac hypertrophy, BNP plasma levels, myocardial contractility	68
Intestinal ischemia-reperfusion in mice, intravital microscopy	l-NAME	Leukocyte rolling	151

BNP, B-type natriuretic peptide; LAD, left anterior descending coronary artery; LDH, lactate dehydrogenase; l-NAME: N^G-nitro-l-arginine methyl ester; SOD, superoxide dismutase; TBARS, thiobarbituric acid reactive substrates; TNF, tumor-necrosis factor alpha.

Implications

There are several practical/translational implications of the above outlined multiple layers of intricate interplay between the vascular NO and H₂S systems. The first one relates to the fact that the eNOS/sGC/cGMP/PKG system becomes impaired in a variety of pathophysiological conditions (from hypertension to atherosclerosis and diabetic complications). As reviewed elsewhere in more detail (121, 129, 138, 142, 143, 145), in many forms of vascular disease (diabetic vascular complications, preeclampsia, vascular aging, atherosclerosis) H₂S levels are impaired; the absence of endogenous H₂S production (e.g., CSE deficient systems) exacerbates the onset and severity of the vascular dysfunction, while H₂S donation improves vascular function. Under conditions of vascular disease, given the partial or absolute requirement of the NO system for H₂S to be fully effective, one would expect that the therapeutic benefit provided by H₂S donation would be diminished. Under such conditions, one would also expect that simultaneous substitution of both NO and H₂S may be more beneficial than the effect of H₂S alone. H₂S levels are dynamically and delicately regulated by production and consumption (30, 121, 140). Since 1) H₂S levels are diminished in various pathophysiological states such as in diabetic complications (121, 129, 138), and 2) the NO system, in part, relies on the presence of H₂S, one would expect that NO-based therapies will not be fully efficacious, unless not only NO, but also H₂S is also substituted. Multiple mechanisms contribute to the loss of vascular H₂S levels in diabetic animals: part of it relates to mitochondrial dysfunction, excessive ROS production, and consequent excessive H₂S consumption (121); another part relates to the oxidative inactivation of the H₂S-producing enzyme 3-MST (20). Under such conditions, appropriately selected antioxidant therapies may be useful to restore biologically active H₂S (as well as NO) levels.

Recognizing the simultaneous need for the restoration of vascular NO and H₂S homeostasis, several groups started to design and test various “hybrid” molecules. H₂S-donating groups have been placed on a large number of existing drugs (reviewed in 119, 142), and some of these combinations may simultaneously enhance NO and H₂S signaling. For instance, ACS6, a H₂S-donating derivative of sildenafil (116), boosts both NO-induced cGMP signaling and H₂S signaling (via PDE5 inhibition and direct chemical H₂S donation, respectively). The angiogenic and vascular effects of the combined NO/H₂S donor ZYZ-803 (48, 148) have already been discussed above.

H₂S has been demonstrated to protect against the development of endothelial dysfunction in various pathophysiological conditions including ischemia-reperfusion, diabetes, and preeclampsia (32, 102, 124, 143). Under conditions of vascular oxidative stress, the biological profile of both NO and H₂S becomes drastically different. For instance, as discussed earlier, NO can form the cytotoxic oxidant peroxynitrite as well as a variety of other potentially cytotoxic or detrimental species, and H₂S can also convert into a variety of reactive molecules. Under such conditions, the antioxidant character of H₂S may become more important. To some extent, such antioxidant effects may be related to direct interactions with ROS species, even though the corresponding reaction rate constants, as determined by stopped flow studies, are widely variable, some of them being rather low (17, 25, 93). Probably more important is the fact that H₂S can also exert antioxidant effects through various indirect mechanisms, including the stimulation of the antioxidant vascular “master switch” Nrf2 and stimulation of intracellular glutathione generation; these effects may also contribute to the phenomenon of H₂S-induced preconditioning (4, 16, 54, 112). Importantly, the absence of vascular H₂S biosynthesis, on its own, induces an elevation of oxidative stress (61). When considering the somewhat controversial and fragmented body of evidence, the role of H₂S/NO interactions, on the background of vascular oxidative stress, may be summarized in the following working hypothesis (Fig. 2). 1) Vascular oxidative stress increases the consumption of NO, and converts it into more deleterious species (e.g., peroxynitrite). 2) Oxidative stress may also consume cofactors of eNOS (e.g., tetrahydrobiopterin, NADPH), which impairs eNOS activity and leads to further oxidant generation from eNOS. 3) Vascular oxidant generation may also oxidize sGC, increasing the pool of sGC that is no longer activatable by NO. 4) These processes may form self-amplifying cycles of injury, increasing oxidant generation from various sources including mitochondrial electron transport chain. 5) The vascular oxidative stress response diminishes cellular H₂S levels, either by directly interacting with them, or by inactivating some of the physiological enzymatic sources of H₂S (e.g., 3-MST). 6) When ambient H₂S levels are diminished, vascular oxidative stress further increases (62, 121). Moreover, the stimulatory effect of H₂S on the eNOS/sGC/cGMP/PKG pathway are diminished, because aging, in turn, causes further impairment in the biological effectiveness of the vascular NO system. 7) All of the above processes may be even further exacerbated in aging blood vessels, which appears to suppress H₂S levels, and H₂S-dependent bioenergetic responses in various tissues (125, 154) (although this has not yet been sufficiently investigated in vascular tissues). Aged blood vessels also have lower levels of endogenous antioxidants and have a diminished ability to mobilize antioxidant responses (e.g., aging blood vessels have a markedly diminished ability to mount an Nrf2-mediated antioxidant response) (136, 137). 8) Antioxidant therapy and/or replacement therapy with H₂S donors, including mitochondrially targeted H₂S donors, which have been shown to protect endothelial cells from oxidative injury (130), may be useful to restore the functionality of the eNOS/sGC/cGMP/PKG pathway. It must be reiterated that the above chain of processes represents a working hypothesis, which remains to be experimentally evaluated in the future.

Fig. 2.

Potential mechanisms of impairment of vascular NO and H₂S signaling processes during oxidative stress: a working hypothesis. Vascular oxidative stress increases the consumption of NO and converts it into more deleterious species (e.g., peroxynitrite). Oxidative stress also consumes various eNOS cofactors (e.g., tetrahydrobiopterin, NADPH), which impairs eNOS activity and leads to further oxidant generation from NOS. Vascular oxidant generation may also oxidize sGC, increasing the pool of sGC that is no longer activatable by NO. These processes may form self-amplifying cycles of injury, leading to more oxidant generation from various sources including mitochondrial electron transport chain. The vascular oxidative stress response diminishes cellular H₂S levels, either by directly interacting with H₂S, or by inactivating its enzymatic sources (e.g., 3-MST). When ambient H₂S levels are diminished, vascular oxidative stress is exacerbated and the stimulatory/balancing roles of H₂S on the vascular eNOS/sGC/cGMP/PKG pathway are reduced, which, in turn, further impairs eNOS-mediated vascular responses.

Some Open Questions

There are numerous areas, some more theoretical, some more practical from either the standpoint of physiology or pathophysiology, that remain severely underdeveloped and open for further investigation. We briefly summarize some of these areas, and hope that the current review will stimulate thought and future experimentation in these areas. The first theoretical area focuses around the following question: what (evolutionary) purpose does it serve to “double down” on NO signaling with H₂S-associated enhancers and facilitators? A related question relates to the relative importance of H₂S in regulating various proteins that are traditionally thought to be regulated by ROS-sensitive mechanisms. Several authors speculate on the evolutionary roots of NO, H₂S, and ROS, and attempt to explain the necessity of these interactions (e.g., 35, 99), but the topic, in the specific context of vascular regulation, remains to be further explored. Another topic relates to the currently unknown levels (and possible regional differences) of H₂S in various parts of the vasculature, as well as the potential regional differences in the expression and activity of the various H₂S-producing enzymes. The studies investigating the role of endogenous H₂S production in various vascular function have, so far, typically only focused on one particular enzyme, such as, in some cases, e.g., the coronary arterial bed, where 3-MST has been proposed as the main H₂S-dependent regulatory system (71). However, systematic studies comparing the relative importance of the three H₂S-producing enzymes in regulating various vascular responses in various vascular regions remain to be conducted. This subject remains to be studied, both as a stand-alone matter as well as in relation to eNOS signaling. Another area that remains to be studied is the role of H₂S in the regulation of vascular permeability (and the potential interplay between NO and H₂S in regulating it). Yet another area that remains to be studied much more extensively is the interactions between the H₂S pathways and the various constituents of the ROS “world,” including the regulation by H₂S of the expression and activity enzymes involved in ROS production and ROS neutralization; the responses of the antioxidant systems to various levels of H₂S exposure, and the pathophysiological interplay between the H₂S and ROS systems. Once again, this subject remains to be studied, both as a stand-alone matter as well as in relation to eNOS signaling. A working hypothesis related to the interactions of NO and H₂S in the context of vascular overproduction is presented in Fig. 2 in the context of vascular dysfunction; however, the various steps of this working hypothesis remain to be further studied.

GRANTS

The work of C. Szabo on various aspects of H₂S biology is supported by grants from the National Institutes of Health (R01-GM-107846, R01-CA-178803, and R21-TR-001734), the Shriners of North America (Grant 85800), and the Cancer Prevention and Research Institute of Texas (CPRIT, Grant DP150074).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

C.S. interpreted results of experiments; C.S. prepared figures; C.S. drafted manuscript; C.S. edited and revised manuscript; C.S. approved final version of manuscript.