Hydrogen Sulfide Actions in the Vasculature

Abstract

Hydrogen sulfide (H₂S) is a small, gaseous molecule with poor solubility in water that is generated by multiple pathways in many species including humans. It acts as a signaling molecule in many tissues with both beneficial and pathological effects. This article discusses its many actions in the vascular system and the growing evidence of its role to regulate vascular tone, angiogenesis, endothelial barrier function, redox, and inflammation. Alterations in some disease states are also discussed including potential roles in promoting tumor growth and contributions to the development of metabolic disease.

Introduction

Hydrogen sulfide (H₂S) is an endogenous signaling molecule that has many different functions in diverse cell types. In the cardiovascular system, it has been shown to elicit vasodilation (82), angiogenesis (8), anticoagulation (151), and many other functions. The importance of H₂S in regulating normal and pathological functions appears to be similar but complementary to that of nitric oxide. Indeed, these two molecules intersect in both synergistic and antagonistic ways under tissue-specific conditions suggesting the two systems are coupled to regulate vascular growth and function.

Functions that will be discussed in the following sections include regulation of the synthesis of H₂S, the role of H₂S as a vasodilator, specific functions in the regulating gastrointestinal blood flow, antioxidant properties of H₂S, H₂S regulation of barrier function and inflammation, H₂S functions in mitochondria, and H₂S regulation of angiogenesis.

Brief History

Hydrogen sulfide (H₂S) has been known to exert physiological functions for many years but only in the mid-1990s did definitive studies demonstrate endogenous production and activity (1). A recent review of the timeline of H₂S research documents the evolution of the field from early observations of eye irritation by sewer gases to current studies of developing therapeutics (184). Early studies demonstrated that exogenous H₂S relaxes smooth muscle, including thoracic aorta and portal vein, and this relaxation was, in some cases, potentiated by the presence of a NO-donor (68). In other studies, the H₂S donor sodium hydrosulfide (NaHS) also potentiated NO-induced relaxation (139). Further progress was made when it was clearly demonstrated that endogenous enzymes in vascular tissue generate measurable H₂S and that loss of these enzymes impairs vascular function and elevates blood pressure (238). Studies on the angiogenic properties of H₂S include both animal studies (159) and studies in cultured endothelial cells (ECs) (29), and studies on oxygen sensing were conducted in multiple species from fishes (152) to rodents (154, 158). This is an exciting new area with profound implications for reestablishing perfusion in ischemic tissue and of blocking the pathway to inhibit the growth of tumor cells as discussed below.

Soon after the endogenous activities of H₂S were confirmed, reports of the measurements of physiological concentrations began to be published with values ranging from >100 μM (77) to less than 1 μM (173). This is an area of continued controversy that has increased in complexity as it has become apparent that multiple sulfides have physiological activity (101). As reports of multiple sulfides and polysulfides exerting responses accumulate, understanding of the actions ascribed to H₂S are expanding to include multiple players (101, 105, 141, 259). Part of the controversy of how to measure physiological sulfides is due to the limited sensitivity of early assays which relied on a colorimetric assay utilizing methylene blue (95, 136) that lacked sensitivity and required acidification of samples. More recent work suggests that this approach might overestimate sample concentrations, and other methods such as high-performance liquid chromatography (HPLC) using sulfhydration of monobromobimane (170, 190), nitro-benzoxadiazole compounds or other indicators appear to be more accurate ways to measure physiological sample concentrations. Indeed, the controversy over physiological concentrations has been a flashpoint in this field and is an area that will only be resolved when there is agreement on the appropriate methods for measuring H₂S in physiological solutions. A recent review on the topic outlines potential roles that differing endogenous sulfide pools may play normal and pathological functions (104, 171). These methods are discussed below in some detail.

The ability to accurately measure levels of H₂S in physiological fluids and tissues demonstrates that decreased levels in the plasma correlate with multiple cardiovascular disease states including diabetes (183, 222, 224), liver disease (9, 10, 76, 160, 248), heart disease (146, 157, 181) which may indirectly result from H₂S regulation of vascular function as discussed below.

In addition to acute regulation of vascular tone and blood flow, H₂S has also been linked to regulation of other vascular functions, including barrier function (89, 134, 244), angiogenesis (19, 119), and inflammation (163, 245). Each of these is also discussed below in more detail. Finally, the field of hydrogen sulfide regulation of vascular function in homeostasis and disease is rapidly changing with increases in understanding of the synthesis and actions of this molecule, making new clinical applications of H₂S releasing therapeutics (110, 139, 234) a rapidly expanding field with many more entrants in the field expected as we learn more about this powerful gasotransmitter.

H₂S and Vasodilation

Direct actions

The physical properties of H₂S make it a unique signaling molecule that can act in an autocrine, paracrine, and/or endocrine manner. H₂S is highly volatile and, under physiological conditions (37 °C/pH 7.4), dissociates rapidly to HS⁻ and H⁺ to drive redox reactions. H₂S has been shown to display a high affinity for iron-containing molecules such as heme (18). The propensity to bind to heme and to be released from H₂S acid-labile and bound sulfane sulfur pools (discussed below) demonstrates potential endocrine roles.

Regardless of the source, studies within the vasculature have shown that endogenously produced H₂S is a vasodilator, promotes angiogenesis, inhibits apoptosis, and increases production of endothelial-derived relaxing factors (EDRFs). One of the mechanisms by which H₂S mediates its effects is through direct sulfhydration of proteins. During sulfhydration, H₂S converts a thiol group (—SH) on cysteine residues to a persulfide group (—SSH) (56, 135, 140, 166). Numerous approaches to measure sulfhydration take advantage of the reactivity of the —SSH moiety and include a biotin-switch assay, a cysteine labeling assay, a maleimide assay, a tagswitch assay, and liquid chromatography-mass spectrometry as reviewed by Zhang and colleagues (246). These assays differ somewhat in endpoint readout methods, but all rely on labeling or tagging the —SSH group for detection as well the ability to discriminate —SSH from —SH. By utilizing these methods, studies have shown sulfhydration as the primary way that H₂S signals to regulate cellular vascular processes, including angiogenesis, anti-inflammatory, antioxidant, and vasodilatory pathways.

Some studies of the vasodilatory property of H₂S demonstrate that artery vasodilation is mediated by H₂S acting on ECs to initiate hyperpolarizing signals (39, 82, 142, 247). Endothelial-derived hyperpolarizing pathways are reliant on activation of EC ion channels to hyperpolarize the EC membrane. It is speculated that this hyperpolarization is then transmitted to underlying smooth muscle cells via myoendothelial gap junctions leading to subsequent relaxation of smooth muscle cells and overall vessel dilation. One of the earliest studies to detail H₂S changes in ion channel activity was done by Jiang and colleagues (88). Using whole-cell patch-clamp, they demonstrated increased ATP-sensitive potassium channel (K_ATP) activity in response to H₂S application, which was completely lost after treatment with the —SH oxidant chloramine T (CLT). They also observed a loss of H₂S (100 μM) stimulated K_ATP currents in HEK-293 cells co-expressing the Kir6.1 subunit with a mutated SUR1 C26S or SUR1 C6S subunit. In addition, glibenclamide or a higher concentration of H₂S (200 μM) failed to stimulate double-mutated Kir6.1/SUR1 (C6S and C1057S) channels. These results demonstrate that H₂S can stimulate K_ATP channels and identifies critical cysteine residues. However, CTL can also oxidize methionine, and this study did not directly measure sulfhydration of the K_ATP channel, leaving some doubt of the target of the oxidant.

A later study by Mustafa and colleagues (137) also examined the effect of H₂S sulfhydration of K_ATP channel subunits on channel activity. Using a modified biotin switch assay, Mustafa and colleagues demonstrated increased sulfhydration of wild-type (WT) Kir6.1 subunit, which was lost in Kir6.1 C43S mutants overexpressed in HEK293 cells. This increased sulfhydration of Kir6.1 resulted in augmented Kir6.1-ATP binding. Besides, H₂S (300 μM) failed to induce hyperpolarization in cells expressing the Kir6.1 C43S mutants. Although this study reported suppressed H₂S-induced hyperpolarization in the mutants, it did not examine channel activity as precisely as in the previous study by Jiang and colleagues, so that the relative contributions of sulfhydration at these different sites is still unclear.

In addition to these studies of H₂S regulation of the endothelial K_ATP channel, H₂S-induced hyperpolarization has been shown to activate EC intermediate conductance (IK_Ca), small conductance (SK_Ca) potassium channels, large conductance (BK_Ca) potassium channels, and the transient receptor potential cation channel subfamily V member 4 (TRPV4) as well as smooth muscle cell K_ATP channels (48, 63, 137, 142). However, whether this is through direct effect of sulfhydration of these channels still requires further examination.

Newer approaches have been developed and have allowed for identification of sulfhydration of specific cysteines. A recent study by Bibli and colleagues mapped an EC s-sulfhydrome using liquid chromatography/mass spectrometry approach based on a quantitative persulfide site identification. The authors compared three data sets which included: native human ECs isolated from plaque-free and plaquecontaining carotid arteries (dataset 1: sulfhydration patterns of diseased state); cultured human umbilical vein endothelial cells (HUVECs) exposed to static or fluid shear stress (dataset 2: sulfhydration patterns with endogenous H₂S); and HUVECs exposed to shear stress treated with vehicle or the H₂S donor SG1002 (dataset 3: sulfhydration patterns with exogenous H₂S) (16). They identified 1536 differentially s-sulfhydrated cysteines (587 proteins, dataset 1); 2981 differentially sulfhydrated cysteines (414 proteins, dataset 2) and 929 cysteine residues (590 proteins, dataset 3). Enrichment analysis of sulfhydrated proteins from the three datasets identified integrin-mediated signaling and glycolysis as the major common clusters of endothelial protein sulfhydration. With these new approaches, identification of sulfhydromes of specific cell types and sulfhydration patterns in different disease pathogenesis is a promising future.

Vascular bed specific actions

Interestingly, the vascular effects H₂S exerts appear to be somewhat tissue bed specific. Indeed, H₂S has been shown to initiate anti-inflammatory and antioxidant pathways at many sites, but its vasoactive properties have been reported to differ significantly between specific vascular beds. One point of contention is the role of cross talk between the H₂S and NO pathway to mediate H₂S initiated dilation, with some reports showing two nondiverging pathways and others demonstrating significant cross talk between the two pathways. To add another level of complexity, those who report cross talk between the pathways have conflicting conclusions on which gas feeds into the other pathway and if the two pathways are synergistic or inhibitory to each other.

Studies by Szijártó and colleagues (188) showed that H₂S scavenges NO to produce nitroxyl (HNO) compounds, which reduce NO bioavailability and NO signaling. Using fluorescent probes to visualize basal endogenous levels of H₂S (Washington State Probe-1; WSP-1), NO (4-amino-5-methylamino-2′,7′-difluoresceine diacetate; DAF-FM-DA) and HNO (CuBOT1), Szijártó and colleagues showed that cystathionine gamma-lyase (CSE) knockout (KO) mice have lower H₂S, higher NO, and lower HNO than WT mice. Interestingly, this study reported that the CSE KO mice have enhanced vasodilatory responses in mesenteric arteries and decreased mean arterial pressures but no change in endothelial nitric oxide synthase (eNOS) phosphorylation. Since elevated arterial pressure and loss of vasodilation have been consistently observed in other CSE KO models (7, 8, 35, 126, 238), it is unclear why the different CSE deletions have such divergent effects.

Furthermore, a study by Altaany and colleagues (8) demonstrated a cooperative role of H₂S and NO in HUVECs and primary aortic ECs. Using DAF-FM to measure NO levels in aortic ECs, Altaany and colleagues showed that both NaHS and the CSE substrate, l-cysteine, increased NO production and that the effects of NaHS were mediated by activation of AKT phosphorylation of eNOS. The divergent findings from the Szijártó study above (188) may be due to differences in the vascular bed studied, the proximity of studied sites to bifurcations or branching points, and the lumen diameter of the vessels that were examined, perhaps all due to differences in shear stress on the EC.

One of the earliest studies to investigate flow regulation of CSE expression, by Yuan and colleagues (245), demonstrated a multifaceted role of H₂S in regulating the NO pathway. Using western blots, Yuan et al. showed human aortic endothelial cells (HAECs) had reduced CSE expression in response to laminar flow and increased CSE expression when exposed to oscillatory flow. Yuan and colleagues verified the regulation of CSE expression by flow patterns in vivo, where CSE immunofluorescence was higher in areas of disturbed flow of the aortic arch as well as higher in ligated carotid arteries to model acute disturbance to flow. Elevated CSE expression in the ligated arteries was associated with increased free H₂S and sulfane sulfur. Ligated arteries from WT but not from CSE KO animals had elevated NOx concentration which is in agreement with the report from Altaany and colleagues (6, 8) that H₂S drives NO production. Interestingly, ligated arteries from CSE KO animals had elevated basal levels of NOx (nonligated WT vs nonligated CSE KO) suggesting H₂S scavenges NO as demonstrated by Szijártó and colleagues (188). The authors conclude H₂S governs areas of the vasculature that experience disturbed flow/low shear and drive NO production to mimic what is seen in vascular areas that have high shear to drive normal NO production. Later studies did confirm differential CSE and eNOS expression along the aorta and attributed the expression variations to different blood flow patterns. Using en face immunofluorescence to examine EC expression of CSE in different areas of the aorta, Bibli et al. (17) demonstrated differential expression of the CSE along the aorta. Endothelial CSE expression was lower in the descending aorta (laminar flow, high shear stress) and higher in the thoracic artery branch point (disturbed flow, low shear stress). Bibli et al. further demonstrated that human EC expressed elevated levels of eNOS in response to 24 h of shear stress and decreased expression of CSE and H₂S levels. The group later demonstrated that in atheroprotected areas, laminar blood flow patterns and high shear stress activate the transcription factor Krüppel-like factor (KLF2), which drives eNOS expression. KLF2 also upregulates miR-27b in these atheroprotected areas, which inhibits CSE expression (15). However, in atheroprone areas near bifurcations with disturbed blood flow and low shear stress, they observed the opposite. In these atheroprone areas, there is a lack of mechanical stimuli to promote KLF2 activation thus leading to decreased eNOS expression and miR-27b to disinhibit CSE expression. While these studies do not address cross talk between the two pathways, they do shed light on the fact that NO is a key factor in areas of laminar flow/high shear stress and that H₂S dominates and protects smaller areas of disturbed flow/low shear stress, which may explain the observed divergent results.

Sources of H₂S in the Vasculature

Enzyme expression in vascular cells

In mammals, H₂S is produced endogenously by multiple enzymes, which include CSE, cystathionine beta-synthase (CBS), and 3-mercaptopyruvate sulfotransferase (3MST) (40, 77, 102, 127, 174, 176, 238). CSE and CBS are pyridoxal 5′-phosphate (PLP)-dependent enzymes, whereas 3MST is PLP-independent. The three enzymes drive conversions between their substrates: cysteine, homocysteine, cystathionine, and 3-mercaptopyruvate (3MP) sequentially to generate H₂S along with additional byproducts. Numerous reports have shown that H₂S-producing enzymes are expressed throughout most organs and tissues; however, the distribution of these enzymes depends on the species employed. For example, CBS and 3MST are expressed in rat brain neuronal tissue but CSE is not. In contrast, rat liver and vasculature, including brain blood vessels, have been shown to express all three enzymes (174, 254). A study conducted by Kabil and colleagues (90) found that in mice, expression patterns in the liver paralleled expression in rats. In addition, Kabil and colleagues also showed expression of both enzymes in the mouse kidney, liver, and brain. Interestingly, the mouse liver expressed higher levels of both enzymes compared to the kidney, and H₂S synthesis capacity was similar for the two enzymes. In the brain and kidney, CBS was the primary source accounting for 50% and 95%, respectively. Interestingly, most studies utilize either cysteine or homocysteine as substrates, but in this study, the authors demonstrated that both cysteine and homocysteine were required to generate maximal H₂S levels in tissue, suggesting either or both can serve as endogenous substrates.

The expression of these enzymes in other organs depends on both species studied and the pathophysiological model examined. In the vascular wall, all three H₂S-producing enzymes are expressed throughout the tissue (103, 192), highlighting an essential role of H₂S in vascular function. The vascular wall is composed of three layers: the tunica adventitia, the outermost layer composed of fibrous connective tissue; the tunica media layer, which contains contractile smooth muscle cells; and the innermost layer, the tunica intima, which is comprised of a monolayer of EC that not only act as a barrier between blood and tissues but also as a sensor for changes within the blood. Several reports show CSE expression and activity within the tunica adventitia (83, 103), while others demonstrate expression of both CSE and CBS within the tunica media layer (26, 28, 168, 201). Finally, although there is evidence for all three H₂S-producing enzymes within EC of the intima layer (6, 174), we and others have demonstrated expression and activity of only CSE within ECs of mesenteric arteries (82, 201).

A number of transcription factors have been recently identified to regulate CSE transcription, which is widely accepted as the primary source of H₂S within the vasculature (160). These include the zinc finger transcription factor, specificity protein-1 (Sp1), oxidation sensitive NF-E2-p45-related factor 2 (Nrf2) and Ca²⁺-calcineurin-regulated transcription factor, nuclear factor of activated T-cells (NFAT) (22, 71, 162, 236). While analysis of transcription factor control of CSE has established multiple candidates and continues to uncover new regulators, acute control of CSE activity is less clear. Two candidate regulators of CSE are the second messenger systems of intracellular calcium concentration [Ca²⁺]_i and phosphorylation. One of the earliest studies to examine [Ca²⁺]_i as a regulator of CSE activity was published by Yang and colleagues (238). Using bovine aortic endothelial cells (BAECs), Yang and colleagues demonstrated increased H₂S production following treatment with the Ca²⁺ ionophore A23187 to increase [Ca²⁺]_i, and this increase was prevented by the intracellular Ca²⁺ chelator BAPTA-AM, the calmodulin inhibitor W7 or a CSE-specific short interfering RNA (siRNA). These findings were paralleled by studies using the muscarinic agonist methacholine (MCh), suggesting that stimuli that increase [Ca²⁺]_i lead to the activation of calmodulin and subsequent CSE activation, increasing H₂S production. In contrast, in a later study, Mikami and colleagues observed inhibitory effects on CSE activity by increased [Ca²⁺]_i (127). Supraphysiological concentrations of calcium (1–2 mM) were used in the experiments reported by Yang and colleagues and Mikami and colleagues re-evaluated the regulation of CSE by [Ca²⁺]_i using physiological levels of calcium (≤3 μM) with purified rat liver CSE. They first verified the viability of the purified CSE by demonstrating that it catalyzed cysteine to H₂S. They then demonstrated that in the presence of the co-factor PLP, an increase in [Ca²⁺]_i suppressed CSE activity while neither calmodulin treatment nor its inhibitor, W7, affected CSE activity (127). These divergent results suggest tissue-specific production of H₂S may vary and highlight the complexity of CSE regulation and that there is a need for studies that examine CSE activity in physiological contexts.

The other identified mechanism regulating CSE activity is through phosphorylation of the enzyme. In HEK-293 cells overexpressing CSE with or without co-expression of heme oxygenase 2 (HO-2), Yuan and colleagues demonstrated decreased H₂S production when CSE was co-expressed with HO-2. The activity results paralleled western blot studies showing an increase in phosphorylated serine CSE in the presence of HO-2. Sequence alignment identified serine 377 as the CSE phosphorylation site affected by HO-2. Using HEK-293 cells overexpressing CSE WT or Ser377A, Yuan and colleagues further demonstrated that the carbon monoxide (CO) donor, CORM-2, like HO-2 overexpression, induced serine phosphorylation of CSE via protein kinase G (PKG) activation. The data demonstrated that HO-2 generated CO activates PKG leading to CSE Ser377 phosphorylation and CSE activity inhibition under normoxic conditions. Whereas under hypoxic conditions, HO-2 signaling is suppressed with subsequent decreases in CSE phosphorylation, leading to increased activity and H₂S production (158). This elegant work suggests a potential role of phosphatases that have yet to be examined.

Another study examining the regulation of CSE activity by phosphorylation identifies a serine site which increases CSE activity when phosphorylated. Using amperometric H₂S sensors, Xu and colleagues reported enhanced H₂S production following 17B-estradiol (E2) treatment of HUVECs and human arterial endothelial cells (HAECs). In both cell types, E2 treatment increased H₂S production that paralleled an increase in CSE phosphorylation at serine 56 and was sensitive to the PKG inhibitor KT5823 (233). This report by Yuan et al. identified different serine residues phosphorylated by PKG and also reported that phosphorylation of CSE increases activity, suggesting differing phosphorylation patterns within the CSE molecule may either promote or inhibit activity similar to kinase regulation of eNOS, as reviewed by Heiss and colleagues (72). In fact, CSE contains numerous serine/threonine/tyrosine residues that may be phosphorylated, and further examination of these sites is expected to yield additional regulatory mechanisms.

The generation of H₂S acting on the blood vessel is not limited to that produced by the enzymes expressed within the vascular wall. The adipose tissue, which surrounds and anchors blood vessels to tissue, called perivascular adipose tissue (PVAT), expresses CSE and releases H₂S as a signaling molecule (188). A study done by Kohn and colleagues showed that contraction of rat aortic rings in response to serotonin was greater in aortic rings without PVAT compared to rings with intact PVAT, suggesting PVAT releases an anticontractile or vasodilatory factor. They further identified the PVAT-released factor as H₂S by demonstrating that the CSE inhibitor, propargylglycine (PAG, 10 mM) and beta cyanoalanine (BCA, 5 mM), inhibited the vasodilatory effect of PVAT in aortic rings (103). It is important to note that millimolar concentrations of PAG and BCA have been shown to inhibit activity of CBS and other PLP-dependent enzymes (11). These studies demonstrate that the source of H₂S acting on blood vessels is not limited to enzymatic production in the vascular wall, but that paracrine actions of adipose tissue generated H₂S can also modulate vascular tone. These findings also raise questions about the conclusions drawn from studies conducted in isolated blood vessels in the absence of in situ influences of PVAT adipocytes. Most of these studies using isolated blood vessels are entirely devoid of PVAT, and in vivo observations may differ significantly, illustrating the need for additional studies of the physiological roles of H₂S.

Circulating, locally synthesized, and store-released H₂S

The sources of H₂S which initiate signaling cascades are not limited to enzymes expressed within the vascular wall and surrounding PVAT to generate blood-borne, paracrine-acting H₂S. Infiltrating immune cells expressing H₂S-producing enzymes also release H₂S under certain conditions to act on cells within the vascular wall. Xu and colleagues demonstrated CSE expression and basal H₂S production in raw264.7 macrophages (234). Interestingly, they also demonstrated that in macrophages, the statins, fluvastatin, and atorvastatin, stimulate increases in CSE expression and H₂S production, which are reversed by Akt and PI3K inhibition. This study also reported that the CSE/H₂S pathway contributes to the anti-inflammatory effects of fluvastatin by inhibiting IL-1β and MCP-1 in LPS-treated macrophages (more H₂S anti-inflammatory properties discussed below).

Biological H₂S in different pools appears to have distinctive behaviors. These pools include free H₂S generated by enzymatic processes; acid-labile H₂S released under acidic conditions; and bound sulfane sulfur pool that releases H₂S under combined acidic and reducing conditions (27, 173). Plasma measurements of H₂S in both human and animal samples suggest that some free H₂S is readily available in circulating blood (4, 84, 216, 222). Indeed, free H₂S appears to be produced basally in many tissues as a highly volatile molecule that dissipates quickly. Therefore, it is probable that the free H₂S measured in plasma underestimates the amount generated locally. Alternatively, acidifying plasma prior to measuring free H₂S is likely to overestimate free H₂S due to the release from the bound pools (172). Even though it is difficult to know the exact normal concentrations of H₂S, most studies estimate it to be in the high nM to low μM range.

Illustrating the ability of H₂S to reach target tissues from the circulation, a study by Sun and colleagues demonstrated that a slow-releasing H₂S donor, diallyl trisulfide mesoporous silica nanoparticles (DATS-MSN) increases plasma free H₂S at 12 and 24 h post the tail vein injection and increases measurable H₂S in the myocardium at 24 h after tail vein injection (181). In contrast, the administration of a rapid-releasing H₂S donor (NaHS) also elevated plasma H₂S concentrations but failed to increase myocardium levels. Free H₂S appears to be quite volatile and short lived due to binding or metabolism in tissues accessible to the plasma. However, even with the different measurable levels in plasma or tissue H₂S levels, both NaHS and DATS-MSN protected the cardiac tissue from ischemia/reperfusion injury (181). Thus, it is apparent that the contributions of different H₂S donors to the diverse biological pools of H₂S (free, acid-labile, or bound sulfane sulfur) are not uniform. Using an intracellular H₂S fluorescence indicator SF7-AM and the published monobromobibane (MBB) method of measuring free H₂S, Yuan and colleagues showed that donors sodium monosulfide (Na₂S) and GYY4137 increased HUVEC free sulfide while DATS increased bound sulfane sulfur (244). Yuan and colleagues also observed a decrease in total sulfide in CSE KO mouse aortic endothelial cells (MAECs) and bound sulfane sulfur but saw no differences in free or acid-labile pools (244), a contrast to what was found in the plasma of CSE KO mice by Shen and colleagues (172). Additionally, Yuan and colleagues determined that the bound sulfane sulfur pool contributed primarily to membrane-permeable H₂S, further highlighting that the diverse biological pools of H₂S may play different roles in regulating cellular function.

The acid-labile pool includes sulfur present in iron-sulfur clusters in nonheme iron-sulfur proteins such as rubredoxins, ferredoxins, aconitase, and succinate dehydrogenase (27). These iron-sulfur proteins are commonly associated with the mitochondria and aid in electron transfer through the electron transport chain; however, they are also found in the nucleus and cytosol and are required cofactors for numerous processes, including DNA replication (145) and iron sensing functions in the case of iron regulatory protein 1 (IRP1) (170, 251). Additionally, iron-sulfur protein assembly machinery is found in both eukaryotic cytosol and mitochondria, reviewed by Lill (114). Using gas chromatography, Levitt and colleagues observed robust release of acid-labile H₂S from liver, brain, heart, skeletal muscle, kidney, and aorta tissues of mice (192–2160 μg/kg protein) with minimal levels released from esophagus tissue or whole blood (42.1 and 12.6 μg/kg protein) respectively (106). The relevance of the sulfur species released from the iron-sulfur proteins was demonstrated by the increased binding of the released HS⁻ to cysteine residues in target proteins; however, it is not yet clear if the sulfur species liberated reach levels that function as signaling molecules in physiological settings and this area of research is currently under investigation.

Sulfane sulfur or elemental sulfur (S⁰) lacks a charge but readily binds to other sulfur atoms due to having six valence electrons (198). The bound sulfane sulfur pool is comprised of persulfides (R-S-SH) and polysulfides (R-S-S_n-S-R). Persulfides and polysulfides are generated in the presence of oxygen from H₂S as well as from HS⁻ reacting with S⁰, and it has been speculated that polysulfides serve as signaling molecules (100). A study by Kimura and colleagues demonstrated that sodium salts of polysulfides, that is, sodium trisulfide (Na₂S₃) and sodium tetrasulfide (Na₂S₄) activate Ca²⁺ influx in rat astrocytes that is suppressed with TRP channel blockers La³⁺, Gd³⁺, and ruthenium red (RR) or by siRNA selective to TRPA1 (100). These findings demonstrated that polysulfides are also bioactive molecules. Yuan and colleagues (244) further validated the bioactivity of polysulfides and demonstrated that polysulfides’ potency correlates to the length of the sulfur tails. Thus diverse sources of H₂S appear to contribute to the overall regulation of physiological functions by this gasotransmitter.

Knockout studies lead to decreased H₂S generation/action

CSE KO animals have been widely used to examine the contributions or protection of H₂S to numerous diseases and examine potential therapeutic roles of H₂S. Using the MBB measurement method coupled with reversed phase-high performance liquid chromatography (RP-HPLC), Shen and colleagues demonstrated that plasma levels of H₂S in CSE KO mice were significantly lower than in WT mice for free sulfide and acid-labile sulfide, but there was no difference between groups in the level of the bound sulfane-sulfur pool (27). Interestingly, the decrement in free H₂S in the CSE KO mice appeared to be ~50%, whereas the acid-labile pool was decreased by only ~20%. This observation strongly suggests that CSE is the primary source of free H₂S in the circulation and that other H₂S-producing enzymes may play more prominent roles in maintaining the bound sulfur pools. The Shen study also highlights the potential importance of bound H₂S pools, especially the acid-labile pool, as physiological reservoirs to release H₂S upon depleting the free H₂S pool. These findings agree with numerous studies demonstrating decreases of free H₂S in the plasma and tissues of CSE KO mice, including the vascular, heart, pancreatic, pulmonary, and renal tissues (6, 23, 156, 166, 239). These observations also illustrate that the multiple H₂S pools may be responsible for the conflicting reports of physiological H₂S concentrations discussed above.

A variety of methods to measure free H₂S have been used, including a spectrophotometric methylene blue assay, sulfide-sensitive electrodes, H₂S-selective fluorescent indicators, and the previously referenced MBB-HPLC method (170). The differing measurement approaches have spawned reports of a wide range of physiological H₂S concentrations that range from nanomolar to millimolar concentration levels. Because acidic and reducing conditions are required to trap or release H₂S in many of these methods to measure H₂S, it is not surprising that a wide range of H₂S concentrations have been reported. Indeed, the divergent H₂S concentrations reported do not negate previous findings but instead serve to highlight the complexity of this volatile gas and the importance of biological sample handling.

Independent of the actual reported H₂S concentrations, studies using CSE KO animals consistently demonstrate both a significant decrease in tissue and plasma H₂S levels as well as a wide range of cellular and physiological processes affected by the loss of CSE (109). Accordingly, H₂S is recognized as a crucial regulator of cardiovascular control and health. For example, Yang and colleagues demonstrated elevated systolic blood pressure in CSE KO compared to WT mice. The difference in systolic pressure increased further with age. They also observed that CSE heterozygote mice’s systolic pressures were significantly elevated above CSE WT, but CSE KO’s systolic pressures were markedly higher at an earlier time point and remained elevated above both WT and heterozygous mice (238). These findings indicate that the expression of one copy of the WT CSE allele confers protection.

Other studies have also used CSE deletion to gain mechanistic insights into the contribution of H₂S in hypertension development. CSE KO mice have been shown to have accelerated vascular remodeling in response to elevated pressure mediated by the proliferation of smooth muscle cells dependent on increased activity of the Ras/Raf-ERK1/2 pathways and inhibition of apoptosis (237). Other groups have shown that CSE KO mice fed an atherogenic diet have elevated plasma lipid profiles, blood pressure, lesion size, oxidative stress, and inflammation compared to their WT counterparts (126). Studies evaluating H₂S in the renal system have similarly shown the importance of H₂S by demonstrating increased mortality and renal damage in CSE KO compared to WT animals in renal ischemia/reperfusion models (23). Numerous studies in multiple cell types isolated from CSE KO and WT mice demonstrate a plethora of impaired signaling cascades verifying the critical roles of endogenous H₂S to regulate flow-dependent vascular remodeling, inflammation downstream of the NLRP3 inflammasome, and oxidation through suppression of NOX 4 (31, 166, 243).

H₂S Regulation of Splanchnic Blood Flow

While H₂S has been shown to be a regulator of blood pressure and flow via effects on the systemic circulation as described above, in the splanchnic circulation, it may play unique roles through the absorption of H₂S synthesized by the gut microbiome. The unique aspects of H₂S in the splanchnic circulation are thus described below in more detail.

Endogenous versus microbiome generation

While the production of endogenous H₂S via enzymes CBS, CSE, 3-MST, and CAT is well established (93, 99, 175), microbial production of H₂S remains a relatively unknown topic. This is in spite of the observation that as much as half of fecal H₂S is microbiome produced (175) with concentrations that can reach low parts per million of H₂S in the rat cecum (180) and up to 30 parts per million in human subjects (62). The gut microbiome plays a critical role in the digestive tract, where it is responsible for digestion, production of vitamins, and resistance to invading pathogens (21, 118, 175). A relatively diverse group of bacteria present in the intestine is responsible for the production of H₂S via the dissimilatory sulfate reduction pathway that oxidizes organic compounds to H₂S. Collectively these enzymes are known as sulfate-reducing bacteria (SRB) (51, 175). This collective group is able to utilize a variety of compounds as electron donors in the dissimilatory reductase pathway, including amino acids, sugars, and carbon compounds (such as carbon monoxide, methanethiol, and methanol) with trithionate and thiosulfate functioning as intermediaries (138, 143). The SRB are present throughout the human intestinal tract, and concentrations ranging from 10³ to 10¹¹/g have been measured in human feces, illustrating their abundance (66). Overall, SRB composition consists of a diverse number of species with varying degrees of sulfate reduction activity. SRB come mainly from five genera, including Desulfovibrio spp. (64–81%), Desulfobacter spp. (9–16%), Desulfobulbus spp. (5–8%), Desulfomonas spp. (3–10%), and Desulfotomaculum spp. (2%) (143).

The gastrointestinal microenvironment is a fluctuating system due to the dietary intake of compounds of varying degrees of absorption and bioavailability, exposing the intestinal microbiome to large and unpredictable concentrations of substrates and enzymes. While SRB are the dominant group responsible for gut H₂S production, several other genera can utilize amino acids to produce small amounts of H₂S. These include Fusobacterium spp., Clostridium spp., Escherichia spp., Salmonella spp., Klebsiella spp., Streptococcus spp., Enterobacter spp. (21). These bacteria generally rely on undigested amino acids from dietary intake, notably cysteine, to generate H₂S as a product of metabolism, along with pyruvate and ammonia via cysteine desulfhydrase activity (21, 59, 175). It remains unclear the exact contribution of this pathway to gut-derived H₂S compared to the SRB dissimilatory reductase pathway.

H₂S is hydrophobic being more soluble in lipid compared to water, with a low barrier to transport across cellular membrane (43, 45). The high partition coefficient associated with H₂S indicates that the molecule does not require a transport protein to facilitate crossing the cell membrane but does accumulate near the source (45). This model does raise the possibility of the paracrine function of H₂S (43). While most H₂S present in the gut lumen from bacterial production is bound to fecal material, a small portion of H₂S is available as unbound (free) in micromolar concentrations that is absorbed into the surrounding mucosa (180, 204). In a healthy state, bacteria-produced H₂S is detoxified by several colonic mechanisms, including oxidation via colonocytes to thiosulfate, and never leaves the gut lumen (143). During dysbiosis, a deleterious alteration in the gut microbiome, and certain gastrointestinal diseases, SRB responsible for H₂S production can increase significantly, and the overall impact of such an increase in bacteria is not well understood (33, 143).

Free nonfecal bound H₂S is reactive and quickly shifts into one of three forms: free H₂S, bound sulfane sulfur, or acid-labile pools (169). The ratio of these biochemical forms and their influence on endogenous H₂S metabolism remains poorly understood. In a study by Shen et al. (169), germ-free mice were shown to have significantly lower levels of free H₂S in plasma and the gastrointestinal tract compared to conventional mice; similarly bound sulfane sulfur levels were reduced in adipose tissue and plasma of the germ-free mice. In this study, CSE expression in adipose tissue was also reduced in the germ-free mice indirectly parallel to the levels of free and bound H₂S (169); thus, microbial metabolism of H₂S appears to impact endogenous CSE expression and activity. However, the full impact of exogenous H₂S on endogenous H₂S is yet to be fully understood.

Effects on portal flow

Endogenous H₂S produced by vascular EC has significant vasoactive properties, as discussed above (93). Although conflicting reports show H₂S can cause both vasodilation and vasoconstriction, this may be somewhat driven by differences in the concentrations and donors investigated. Indeed, several studies have reported concentrations of H₂S below 100 μM promote vasoconstriction, while higher concentrations promote vasodilation in large conduit arteries (254). However, studies by Jackson-Weaver et al. (82) have reported that both endogenous H₂S and exogenous H₂S (from NaHS, an H₂S donor) produce potent mesenteric arterial dilation. In contrast, studies in mouse mesentery by Szijártó et al. (188) showed NO scavenging by mesenteric H₂S to augment vasoconstriction. Thus, the mechanism of action and downstream responses to H₂S in the splanchnic circulation are not entirely understood. Both tissue and concentration-dependent changes in H₂S enzymatic production may underlie the dynamic changes in portal vein pressure and flow under conditions of elevated microbiome generation of H₂S.

The impact of microbial H₂S on portal vasculature has been studied to a minimal extent (74), with most studies focusing on the translocation of bacterial products and resident bacteria from the gut lumen rather than gas metabolite impacts on vascular tone (70, 164, 223). Under physiological conditions, SRB are present with the greatest density in the colon and at much lower concentrations throughout other portions of the gastrointestinal system (51, 196). However, changes in the gut microbiome in a disease state, such as dysbiosis, are known to alter the composition and quantity of microbiota in the small intestine (118, 205). However, the detoxification mechanisms for bacteria-derived H₂S are concentrated in the colon. Therefore, any increase in SRB composition in the small intestine can lead to significant intestinal injury due to inability to detoxify surges in H₂S (21, 143). A report from Huc et al. (79), showed that even the colonic detoxification mechanisms can be overwhelmed with a large bolus of H₂S, leading to systemic hypotension and portal venous hypertension. A more recent publication by Birg et al. (20) demonstrated that administering a much lower concentration of a H₂S donor, NaHS, directly into the small intestine of healthy rats increased portal vein blood flow with no systemic changes in blood pressure but was without effect if administered into the colon. This study was the first to show a splanchnic-specific impact of luminal H₂S suggesting small intestine dysbiosis with SRB overgrowth could selectively impact portal hemodynamics. However, interactions between endogenous and microbiome-derived H₂S remain unexplored.

Alterations in disease

The role of H₂S in inflammation remains poorly understood, with reports showing pro- and anti-inflammatory properties. Part of the dichotomy is likely due to the study of multiple sources of H₂S (endogenous vs. exogenous), but the contrasting outcomes may also be due, in part, to poorly understood interactions between different sources of H₂S. Because imbalance and perturbation of the number and composition of gut bacteria are linked to diseases such as inflammatory bowel disease, obesity, diabetes, liver disease, cancer, and cardiovascular disease (21, 175), roles of H₂S in regulating splanchnic homeostasis may be impacted by elevations in gut-microbiome generated H₂S.

Gut inflammation

H₂S has been previously shown to have a protective effect on the mucosal surface of the gastrointestinal tract. Gut microbiota produces a biofilm that is protective of the bacteria and mucosa and promotes homeostasis by minimizing the “leakiness” of the mucosal layer that would otherwise permit bacterial invasion (205). H₂S promotes the structure of the biofilm and mucus production to reinforce the biofilm layer (196). Disruption of the biofilm layer leads to disorganization of the microbiome and can increase translocation of bacteria into the lamina propria and surrounding vasculature.

Animal studies have shown that H₂S donors can promote colitis resolution and that H₂S donors induce healing in non-steroidal anti-inflammatory agent (NSAID)-induced or ischemia-reperfusion injury in the gut (205). In a healthy gut, endogenous H₂S production is increased at the injury sites, while oxidation of H₂S is significantly reduced, leading to locally higher available concentrations (205). Inhibition of H₂S production conversely promotes the susceptibility to injury of the gastrointestinal mucosa, increases pro-inflammatory cytokine expression, and reduces cyclooxygenase-2 expression (21, 205). SRB, which are generally resistant to antibiotics, can also disrupt the small intestine biofilm at high levels. H₂S thus serves as a defense mechanism to surrounding mucosal insults and the effect of disrupting this protection depends on both site and concentration (175).

In contrast, lipopolysaccharide endotoxins derived from SRB, specifically from D. desulfuricans, cause an increase in the pro-inflammatory cytokine IL-6 (175). The detrimental increased production of IL-6 has been shown with the administration of high concentrations of NaHS into the colon (21), demonstrating that large bolus quantities of H₂S can be pro-inflammatory, leading to increased mucosal inflammation and damage. H₂S generated via oxidation of n-butyrate can thus inhibit colonic epithelium production of the mucosal biofilm leading to increased permeability seen in ulcerative colitis (51, 175).

NaHS has also been used to induce intestinal and mesenteric recovery from ischemic injury via its anti-inflammatory properties (54, 87). Intestinal ischemia causes significant cytokine release with elevated levels of IL-6 and this increase was attenuated by H₂S administration (54, 87). H₂S also increases the expression of both systemic and local antioxidant enzymes to further attenuate ischemic injury through increased activity of superoxide dismutase (SOD) and glutathione peroxidase (GPx) (228). Interestingly, H₂S-mediated healing and anti-inflammatory actions in the gastrointestinal tract appear to be, in part, through NO-dependent mechanisms (54, 87). Again, the contributions of endogenous and exogenous H₂S appear to have contrasting contributions, depending on the studies.

Cirrhosis

Several important changes occur in chronic liver disease, including low systemic vascular resistance, reduced responsiveness to vasoconstrictors, and low portal vascular pressure (58, 81). While the vasodilators involved and the precise mechanisms causing these effects are poorly understood, activation of K_ATP channels appears to contribute to the impaired systemic control of vascular tone in liver disease models (133). In a study in an animal model of cirrhosis, K_ATP channel activation in vascular smooth muscle cells (VSMC) was responsible for the increased peripheral vasorelaxation (133). However, blood flow regulation in the liver is unique and likely to be regulated by novel mechanisms. Hepatic stellate cells (HSC) around hepatic sinusoids contribute to vasoconstriction and vasodilation within the liver (60). Changes in endogenous H₂S levels have been shown to correspond to variability in portal pressure in an animal model of cirrhosis by activating K_ATP channels in the vascular smooth muscle (60, 148, 217). An analysis of H₂S regulation of HSC and the presinusoidal splanchnic microcirculation determined that both phenylephrine, an α1-adrenergic agonist, and endothelin-1 (ET-1), a vasopressor, constricted the presinusoidal vessels, while sinusoidal circulation only responded to ET-1, suggesting the HSC are more responsive to ET-1 than phenylephrine (12, 148, 250). Endogenous H₂S attenuated phenylephrine-induced constriction of vessels in a rat liver section, while H₂S had no impact on ET-1 perfusion of the same sections (148). This suggests that H₂S may have regional roles to regulate portal pressure and hepatic vascular resistance, depending on the microcirculatory bed exposed; however, the exact mechanism remains to be completely understood, and the current models only show a correlation rather than direct impact of H₂S as the primary mediator.

The animal model of carbon tetrachloride (CCl₄) induced hepatotoxicity has been reported to cause changes in the activity of serum CSE measurements but different studies have reported contrasting results. After acute induction of hepatotoxicity, measurement of serum CSE showed an increase, which was speculated to contribute to increased H₂S formation (189). In a more recent study by Lei Ci et al., CSE knockout mice had significantly higher levels of hepatic inflammatory markers and lower endogenous H₂S levels during acute hepatitis induced by CCl₄, which was attenuated with pretreatment of H₂S donor (41). On the other hand, chronic CCl₄ administration was reported to decrease serum CSE levels (191). These contrasting results may be caused by concomitant microbial changes in liver disease that alter the concentration of bacteria-derived H₂S in the two conditions.

Interestingly, liver disease can predispose to endotoxemia due to lack of clearance of endotoxins and bacteria from the circulation and to leakiness of the intestinal barrier (10, 58, 205). Endotoxemia in the setting of cirrhosis has been shown to increase plasma H₂S levels (58, 147). It remains unclear if the source of increased plasma H₂S levels is related to endotoxins generation, decreased clearance due to liver disease, or increased endogenous production by CBS or CSE. Irrespective of the source, increased H₂S plasma concentrations lead to increased systemic vasodilation (58, 60).

Energy balance/metabolism/insulin

CSE, and CBS to a lesser extent, are also expressed by insulin-secreting pancreatic β cells, and H₂S has been shown to play an essential role in glucose control (13, 94). Under physiologic conditions, glucose stimulates pancreatic cells by depolarizing the cell membrane and causing an influx of Ca²⁺ into the cell to trigger insulin release (193). H₂S activates K_ATP channels in pancreatic β cells to inhibit insulin secretion (13).

While some reports show that endogenous H₂S plays protective roles during hyperglycemic events (13, 149), the exact mechanism and the contribution of vascular changes to the cytoprotective effects are not entirely understood. It has been suggested that different concentrations of H₂S produce opposing actions directly on pancreatic β cells independent of pancreatic blood flow (149). Indeed, the introduction of exogenous H₂S into this tightly regulated system can significantly impact plasma glucose levels and progression of the metabolic syndrome (13). NaHS can inhibit glucose-induced activation of voltage-dependent calcium channels in the pancreas to decrease glucose-stimulated insulin release while the addition of exogenous H₂S (50–200 μM) can induce apoptosis of pancreatic β cells (13). In addition, decreased blood flow to the pancreas under conditions of endothelial damage may also contribute to inhibitory effects on insulin production and apoptosis of pancreatic β cells while intermittent postprandial delivery of exogenous H₂S derived from gut bacteria may exacerbate the development of insulin resistance under conditions of dysbiosis.

Under conditions of chronic elevations in plasma glucose, decreased pancreatic CSE expression has been correlated with higher insulin secretion independent of vascular effects (i.e., disinhibition of beta-cell secretion) (193). A recent study by Jain et al. (84) reported that diabetic patients have significantly lower levels of H₂S and CSE expression. Interestingly, metformin, the most common medication used to treat diabetes via increasing insulin sensitivity, has also been shown to increase H₂S levels in animal studies suggesting elevated H₂S may contribute to the pleiotropic actions of this medication (224). While the mechanism for reduction in CSE expression and H₂S remains unknown, increased concentrations of IL-8 and MCP-1 (75) suggest vascular inflammation may contribute (84).

H₂S as an Antioxidant

Reduction/oxidation is an important regulator of many physiological functions that can be significantly impacted by controlled and uncontrolled reactive nitrogen, sulfide, and oxygen species. Indeed, the most common endogenously produced reactive oxygen species (ROS), include superoxide and hydroxyl radical (167). Although ROS can serve as essential components of cellular signaling (194), at elevated concentrations ROS modify cellular proteins, carbohydrates, or lipids as well as DNA in an uncontrolled and sometimes irreversible manner. Inappropriate oxidative stress has been linked to the development and progression of multiple human diseases, including cancer (132), cardiovascular disease (124), diabetes (203), and degenerative diseases associated with aging (113).

Effect of H₂S on antioxidant systems

Given the low tissue and plasma concentrations of endogenous H₂S (nM—low micromolar), direct reactions between H₂S and cellular oxidants probably cannot account for the antioxidant properties of H₂S (30, 44, 45). However, there is mounting evidence that H₂S plays several roles in maintaining cellular antioxidant homeostasis by regulating the expression and activity of intracellular antioxidant systems. Several studies have highlighted the role of H₂S in controlling cellular redox homeostasis in the vasculature and in other tissues by increasing the cellular levels of antioxidants, such as glutathione (GSH) and thioredoxin (TrX), through upstream stimulation of expression and activity of the transcription factor NRF2. Much of this evidence comes from studies demonstrating exogenous H₂S restores cellular redox balance. For example, NaHS administration to uranium-intoxicated rats ameliorated the renal biochemical indices and histopathological effects, lowered levels of the oxidative stress marker, malondialdehyde, restored GSH levels, and increased activity of the antioxidative enzymes SOD, catalase (CAT), GPx, glutathione reductase (GR), and glutathione-s-transferase (GST) (258). Moreover, administration of the sulfur donor, diallyl sulfide normalized drug-induced oxidative stress as demonstrated by a significant increase in the renal activity of SOD, CAT, GPx, GR, GST, and lower levels of reduced GSH (91). H₂S has also been shown to be protective from the increased levels of oxidative stress associated with chronic inflammation (129). Lipopolysaccharide treatment of mice reduced CAT and SOD1 activity in kidneys and lungs, but intraperitoneal administration of a H₂S donor prevented the reduction. H₂S administration also resulted in increased GSH/GSSG ratio, reduced H₂O₂, and reduced malondialdehyde levels (38, 249). Given the susceptibility of vascular ECs to oxidative damage (129, 212) and the potential protective effect of H₂S on preserving endothelial function (25, 55, 65, 68), regulation of redox balance in the vasculature is emerging as an important vascular function of H₂S.

In hyperhomocysteinemic mice, decreased brain H₂S levels and increased O₂⁻ production has been observed. Administration of NaHS was shown to normalize O₂⁻ levels parallel to elevated tissue SOD, CAT, GPx activities, and reduced GSH (144). Furthermore, treating osteoblast cells with H₂O₂ increased NADPH oxidase-dependent O₂⁻ production and reduced SOD activity. Pretreating cells with NaHS prevented a fall in SOD activity and reduced O₂⁻ levels (235). In H₂O₂ exposed HUVECs, H₂S levels in the media and CSE activity/expression were decreased. H₂O₂ treatment also increased ROS species levels (DCF and DHE) and reduced SOD1, SOD2, CAT, GST, and GPx expression. Whereas H₂S addition reduced ROS production and restored expression of antioxidant enzymes, inhibiting CSE augmented the H₂O₂ effects on cellular redox status (220). Treatment of HUVEC cells with high glucose increased H₂O₂ production measured by changes in DCF fluorescence, which was blocked by pretreatment of cells with NaHS (69, 116).

Mitochondrial-directed H₂S donors are also able to decrease ROS levels locally in vascular tissues. In ECs exposed to hyperglycemia, there is a significant and local increase in mitochondrial ROS levels (65). H₂S administration or CSE over-expression attenuated the high glucose-induced ROS production (measured as Mitosox^™ fluorescence), whereas CSE siRNA augmented the high glucose-induced ROS (182). ECs exposed to high shear stress decreased CSE expression, and the loss of CSE resulted in increased lipid peroxidation and decreased sulhydration of Prx6 peroxidase, which was associated with hyperoxidation of Prx6 and subsequent enzyme inhibition. Loss of CSE also increased H₂O₂ and O₂⁻ production as well as cellular lipid peroxidation in these vascular cells (15). Finally, high glucose treatment reduced EC levels of glutamate-cysteine ligase and GSH levels in monocytes, which was prevented by pretreating cells with H₂S (85).

Additional studies demonstrate the protective effects of exogenous H₂S on the vasculature in vivo. In a study of retinal tissue in CBS^+/− mice, administration of the slow-release H₂S donor GYY-4137 in vivo normalized ROS and antioxidant levels (64). In a mouse model of unilateral ureteral obstruction, renal CSE/CBS expression, as well as renal and plasma H₂S levels, were reduced. Increasing H₂S levels from an exogenous source restored SOD1 and CAT levels in the kidney (36). In aorta from low-density lipoprotein receptor null mice on a high-fat diet as well as in streptozotocin (STZ)-treated animals, the H₂S donor GY4137 decreased O₂⁻ levels in the vasculature (230). In angiotensin II-induced hypertensive animals, the NADPH-dependent increase O₂⁻ production was prevented by NaHS administration and augmented by CSE inhibition (5). In atherosclerotic rats, SO₂ administration increased circulating levels of H₂S and reduced atherosclerotic plaque size in rats on high cholesterol diet. Following SO₂ treatment, GPx activity and SOD1/2 expression were increased in the aorta along with higher plasma SOD activity (111).

Effect of H₂S on transcriptional regulation of antioxidant pathways

H₂S has been shown to directly activate SOD by binding to the Cu²⁺ catalytic subunit of SOD (165), but growing evidence suggests that H₂S regulation of antioxidant enzymes can also involve activation of the oxidant responsive Nrf2 (42). Nrf2 activity is regulated by the repressor protein, Keap1 (Kelch-like ECH-associated protein). Under basal conditions, Nrf2 is primarily localized in a complex with Keap1 in the cytoplasm via direct protein-protein interactions with two Keap1 molecules, which promote Nrf2 polyubiquitination and subsequent proteasomal degradation by the 26S proteasome (252). In the presence of oxidative stress, Nrf2 dissociates from Keap1 and translocates to the nucleus, where it binds to the antioxidant response element to drive the expression of Nrf2 target genes of proteins regulating critical components of the GSH and TrX antioxidant systems, as well as enzymes involved in NADPH regeneration, ROS detoxification, and heme metabolism (197). Multiple studies have recently demonstrated that the antioxidant effects of H₂S are mediated via Nrf2-dependent increases in expression of antioxidant genes. NaHS administration resulted in Nrf2 activation and Nrf2 expression (3, 108, 230, 256). Nrf2 translocation from the cytosol to the nucleus is required for transcriptional regulation, and an H₂S donor increased translocation of Nrf2 to the nucleus in aortic VSMC as well as in EC and peritoneal macrophages treated with high glucose (3, 230). In vivo, Zheng et al. (258) demonstrated that the H₂S donor sodium hydrosulfide (NaHS) treatment also increased nuclear localization of Nrf2 in renal homogenates of uranium exposed rats. Increased ROS production in HUVEC cells treated with angiotensin II was prevented by pretreatment with NaHS, and this protection was prevented by siRNA targeted against Nrf2 (108). Importantly, H₂S-dependent activation of Nrf2 has been shown to increase expression of the antioxidant genes, NADPH dehydrogenase [quinone]-1 (NQO1), heme oxygenase-1, thioredoxin reductase-1, and glutamate-cysteine ligase catalytic subunit (3, 256). Finally, the antioxidant functions of H₂S were suppressed in Nrf2 null animals (230). Moreover, H₂S increased S-sulfhydration of Keap1, induced Nrf2 dissociation from Keap1, enhanced Nrf2 nuclear translocation, and inhibited O₂⁻ generation in EC expressing a Keap1 c151A mutation (230). Consistent with this finding, Yang et al. demonstrated that NaHS administration increased S-sulfhydrated of Keap1 at cysteine-151, which was associated with induced Nrf2 dissociation from Keap1, enhanced Nrf2 nuclear translocation, and stimulated mRNA expression of Nrf2-targeted downstream genes (240). These observations suggest that regulation of this intrinsic antioxidant cascade may mediate many of the protective effects observed following H₂S treatment.

Barrier Function and Vascular Inflammation

ECs are an important barrier to control the flow of substances and fluid into and out of tissues and to regulate the trafficking of immune cells into tissues. In this section of the review, we will discuss the role of H₂S in regulating these barrier mechanisms.

Endothelial permeability varies across the vasculature, with large to medium arteries being impermeable compared to small vessels and capillaries, which are much more permeable to multiple molecules. Permeability is a complex, regulated function mediated by both transcellular and paracellular pathways. The paracellular pathway involves the contraction of tight and adherens junctions to generate controlled junction disruption and intercellular gaps. The molecular components of adherens junctions are catenins and vascular endothelial (VE)-cadherin. VE-cadherin binds to the actin cytoskeleton through its cytoplasmic domain via both α- and β-catenins (161). Claudins and occludins form tight junctions. In most peripheral tissues with the exemption of the blood-brain barrier (BBB), both adherens and tight junction proteins are intermingled, with VE-cadherin being the predominant controller of vascular permeability (244). Endothelial dysfunction disrupts endothelial barrier function, leading to vascular inflammation, which contributes to the development of cardiovascular diseases. Many circulating factors, such as vascular endothelial growth factor (VEGF), thrombin, cytokines, histamine, and bradykinin activate leukocyte adhesion to ECs leading to phosphorylation of the VE-cadherin cytoplasmatic tail, increasing VE-cadherin/catenin complex dissociation and VE-cadherin internalization with resultant increased permeability (161). On the other hand, counter-mechanisms typically maintain low vascular permeability or restore permeability after inflammation is resolved. The cAMP-elevating factors, angiopoietin-1 and sphingosine-1-phosphate, are among factors that improve barrier function. Recent evidence suggests H₂S might also be among these factors. This is supported by Wang et al.’s work showing that the H₂S donor, NaHS, attenuates urban particulate matter (PM)-induced human lung microvascular EC monolayer disruption (213). It appears that NaHS increases Akt activity to reduce endothelial permeability and exert a protective effect. At the same time, NaHS attenuates PM-induced ROS activation of p38 mitogen-activated protein kinase (MAPK). However, this study did not directly address how p38 activation increases EC permeability or how Akt is protective, but it suggests that MAPK induces actin cytoskeleton reorganization to facilitate paracellular gap formation (212). Akt has also been shown to mediate protective cytoskeletal rearrangement (178).

Direct antioxidant effects of NaHS discussed above suggest it can act as a ROS scavenger to diminish PM-induced increases in 2′,7′-dichlorofluorescein diacetate (DCFDA) fluorescent signal (213). Similarly, NaHS prevented BBB disruption and brain edema in a rat model of subarachnoid hemorrhage by preserving EC function through its antioxidant and anti-inflammatory actions (47). In a prior study, Wang et al. (216) showed similar protection in a model of middle cerebral artery occlusion (MCO) in mice. The authors compared the effect of NaHS to that of 5-(4-methoxyphenyl)-3H-1,2-dithiole-3-thione (ADT). ADT produced a sustained increase in plasma H₂S levels compared with a transient effect of NaHS, but both agents protected mice from MCO-induced BBB disruption. Taken together, these studies suggest that H₂S preserves endothelial barrier function by inhibiting local inflammation and reducing ROS.

Under normal conditions, cellular levels of free sulfide are low (i.e., H₂S, HS⁻ and S₂⁻), but in the presence of ROS, metal ions, and myeloperoxidase, the formation of thiyl radicals is favored producing persulfide and polysulfide compounds. In addition, H₂S could be mobilized from acid-labile sulfur present in amino acid residues of proteins. Evidence suggests these sulfane sulfur reservoirs contribute to some H₂S-mediated biological actions (80, 122). An extensive comparison of the effect of different sulfide donors on endothelial barrier function was reported by Yuan et al. (244). This study compared the effect of Na₂S, GYY4137, DATS, sodium disulfide (Na2S2), Na2S3, and Na2S4 on albumin flux across a HUVEC monolayer at different time points and different concentrations. Contrary to findings in animals discussed above that demonstrated protective effects of NaHS on barrier function, monosulfide donors had little effect on endothelial permeability at low concentrations and increased permeability only at supraphysiological concentrations. Both DATS and polysulfide compounds also significantly increased endothelial permeability with their potency depended on the number of sulfur atoms in the polysulfide compounds, that is, the more sulfurs, the higher the potency.

Interestingly, free H₂S levels detected by loading HUVEC with the H₂S-specific fluorescent probe, SF7-AM, did not significantly increase after treatment with DATS but was increased by both Na₂S and GYY4137. This suggests free H₂S was not responsible for the increased endothelial permeability caused by DATS or the polysulfide compounds. While none of the compounds affected cell viability, most of the studies discussed so far only tested the effect of exogenous H₂S donors, leaving the role of endogenously generated gasotransmitter is still relatively unexplored.

One study from Yuan et al. (244) determined the effect of CSE-mediated production of free H₂S and sulfane sulfur on aortic EC both ex vivo and in vivo. Interestingly, they reported that primary cultured aortic EC from CSE KO mice had decreased permeability compared to EC from WT mice. This difference was associated with an increase in the tight junction protein claudin 5. However, basal solute permeability was not different between genotypes as assessed by Evans blue tissue content in the heart, brain, lungs, liver, kidneys, colon, and spleen. On the contrary, CSE KO mice showed an attenuated VEGF-induced hyperpermeability response in ear pinna compared to WT mice suggesting endogenous CSE-produced H₂S enhances endothelial barrier permeability. These findings further support the studies in cultured cells using H₂S donors, increasing solute permeability (244). The limited permeability in ECs from CSE KO mice was not due to lower levels in free and acid-labile sulfide levels but lower levels of bound and total sulfide levels compared to WT mice. These results are consistent with the finding that DATS restores endothelial permeability to levels comparable to cells from WT mice suggesting that CSE is an enzymatic source of polysulfides. Unfortunately, the authors did not explore the mechanism underlying the dysregulated permeability in mice lacking CSE.

Studies exploring the role of CSE-, 3-MST-, and CBS-produced H₂S in ischemic stroke also support a role for H₂S, primarily produced by CSE and 3-MST, as a contributor to early BBB disruption following transient focal cerebral ischemia (89). Again these findings contrast with a report in which the H₂S donor, NaHS, was shown to protect against ischemia/reperfusion injury (242)—or subarachnoid hemorrhage-induced BBB disruption (47). The difference could be in the timing of the administration of NaHS and the enzyme inhibitors. In studies showing protective effects, NaHS was given either at the onset of ischemia or after 3 h of reperfusion, suggesting the NaHS neuroprotective effects might be due to its vasodilatory and antioxidant actions. However, inhibitors of CSE or 3-MST were applied during the reperfusion phase, in which exaggerated levels of H₂S might be detrimental to barrier function (89). Once again, the mechanism by which endogenous H₂S contributed to BBB disruption was not explored and it is an area that requires further investigation.

These effects of H₂S on the endothelial barrier can have indirect implications on inflammatory cell recruitment to injured tissues. However, H₂S also has direct actions on inflammatory cells. For example, H₂S directly sufhydrates κβ kinase subunit β (IKKβ) protein at cysteine residue 179. IKKβ is part of a complex of proteins that negatively regulate the pro-inflammatory transcription factor nuclear factor-κβ (NFκβ). IKKβ phosphorylation leads to proteasomal-dependent degradation of IKK complex members, releasing NFκβ, activating it, and allowing its nuclear translocation. The serine residues subjected to phosphorylation are between cysteine 179; therefore, sulfhydration of this residue prevents IKKβ phosphorylation inhibiting NFκβ activation. This mechanism was demonstrated using exogenous H₂S and endogenously produced H₂S in vivo, in cultured cells, and in vitro. The same mechanism prevents ROS-mediated activation of NFκβ (247).

H₂S also promotes naïve CD4⁺ T cell differentiation into T regulatory (Treg) cells (241) and the production of the anti-inflammatory cytokine IL-10 (37), both of which have implications in blood pressure regulation and hypertension development (221). Hypertension is associated with a reduction in Tregs and an increase in pro-inflammatory CD4⁺ T cells subset 17 (Th17) (221). A recent report by Cui et al. (46) demonstrated that CSE-mediated synthesis of H₂S in CD4⁺ T is required for angiotensin II-induced hypertension, vascular and renal inflammation in mice. In addition, mice lacking CSE only in CD4⁺ T cells present with a significantly increased blood pressure (5–8 mmHg and is sex-dependent) under physiological conditions. Adoptive transfer of Tregs rescued the outcomes mentioned above, present in mice harboring null alleles for CSE, specifically in CD4⁺ T cells.

Interestingly, the authors demonstrated that the mechanism of Treg disfunction in these mice is via direct sulfhydration of a kinase, activated protein kinase (AMPK). As in the case of IKKβ, sulfhydration of AMPK prevents its phosphorylation, promoting Th subset polarization toward Treg and reducing inflammation. This report also includes data from normotensive and hypertensive subjects demonstrating that lymphocytes from untreated hypertensive subjects produce less H₂S and express less CSE than normotensive and controlled hypertensive individuals. Besides, systolic and diastolic pressures negatively correlated with lymphocyte-produced H₂S while IL-10 levels positively correlated with H₂S levels. Taken together, these studies support a role for H₂S as a critical regulator of the immune system and, consequently, blood pressure regulation. This study also suggests that autoimmune processes might develop due to a reduction in H₂S as reviewed elsewhere (52).

It is currently unresolved whether H₂S enhances or reduces EC permeability, which is primarily illustrated by significant differences in the actions and potencies of the differing forms of sulfide compounds. Despite the controversy on barrier function, collective evidence supports a role for H₂S as an anti-inflammatory molecule that can preserve barrier function and the cellular targets for the endogenous sulfide signaling system is an area of intense current research.

H₂S and Angiogenesis

Angiogenesis is the physiological process of forming new blood vessels from preexisting blood vessels (2). Minimally, this process requires EC proliferation, migration, and organization into tube-like networks. During development, the cellular programs that mediate these functions are active; however, in adulthood, they are primarily relegated to conditions of tissue damage and pathology.

Both in vitro studies and in vivo studies have provided abundant evidence of the angiogenesis-promoting role of H₂S (93). Using various types of ECs, H₂S has been shown to promote EC viability, proliferation, migration, adhesion, and tube formation (8, 19, 25). Studies using chicken chorioallantoic membranes (CAMs), Matrigel plug assays, and various rodent models of tissue injury have also demonstrated that H₂S increases blood vessel formation (19, 24, 86, 92, 97, 104, 115, 125, 179, 195, 206, 208, 209).

Commonly, sulfide salts Na₂S and NaHS are administered to cells or subcutaneously at injury sites (19, 86, 117, 119, 206, 207, 209, 226, 227, 253) to promote angiogenic activity. Other H₂S donors that have been used to promote angiogenesis include naturally occurring substances, such as diallyl disulfide (DADS), DATS, and 4-hydroxythiobenzamide (HTB), and chemically synthesized compounds like the slow-releasing H₂S donor GYY413 (19, 34, 49, 53, 78, 98, 107, 112, 115, 120, 125, 157, 177, 218, 232). H₂S-producing enzymes can be directly modulated through overexpression, knockdown, or pharmacological intervention (14, 185, 215). Frequently used inhibitors to diminish angiogenesis include propargylglycine (PAG), which inhibits CSE, and aminooxyacetic acid (AOAA), which inhibits both CSE and CBS (119, 185, 208, 215). Finally, targeted gene KO animals have been utilized to demonstrate that H₂S production promotes angiogenesis (19, 24, 104, 125).

H₂S appears to stimulate angiogenesis through a variety of different mechanisms, including K_ATP channel activation (97, 153, 186, 199, 254), PI3K/Akt, and other kinase signaling pathways (53, 86, 120, 207, 209, 226, 227, 257), and VEGF-dependent pathways (78, 199, 206–209, 215). Because the general role of H₂S in promoting angiogenesis has been extensively reviewed elsewhere (93, 97, 186), here we will primarily discuss angiogenesis as it relates to two disease-related processes: wound healing and tumor progression.

Tissue repair and wound healing

Angiogenesis is an important element in the tissue repair or wound healing process, as the site of injury requires rapid infiltration of immune cells and nutrients to repair the injured tissue. In several models of tissue ischemia, such as hindlimb, myocardial, and cerebral ischemia, the studies discussed below demonstrate that H₂S plays a positive role to stimulate angiogenesis. In addition, H₂S has been hypothesized to improve outcomes for vascular complications in diabetes, as shown in studies on diabetic mouse models and hyperglycemia discussed below.

In mouse models of hindlimb ischemia, multiple mechanisms were shown to contribute to H₂S-mediated stimulation of angiogenesis as discussed below. NaHS treatment in mice is also discussed as a way to increase vessel growth and blood flow in association with greater VEGF expression in hindlimb skeletal VSMC and VEGFR2 phosphorylation and Akt signaling in EC (209). The role of H₂S in tissue repair following hindlimb ischemia was also demonstrated using CSE KO mice. Vessel density, blood flow, and monocyte infiltration were decreased in CSE KO mice compared to WT following femoral artery ligation; the H₂S donor DATS and nitrite-mediated NO therapy rescued these deficits in the KO mice. In eNOS KO mice (104), NaHS increased hindlimb blood flow and angiogenesis by enhancing EC HIF-1α and VEGR2 expression and function in a NO-dependent manner (19). These studies further highlight H₂S-NO cross talk during vascular remodeling and vessel formation, similar to what has been observed in vasodilation studies discussed above. In CBS HET mice with moderate hyperhomocysteinemia, impaired blood flow and angiogenesis were associated with decreased phosphorylation of Akt (24). In a second study also using CBS HET mice, the H₂S donor GYY4137 improved vessel density and blood flow via the PPARγ-VEGF axis (31).

The role of H₂S was also investigated in angiogenesis studies that address cardiac failure and injury. Following transverse aortic coarctation (TAC), DATS treatment improved left ventricular (LV) remodeling and function by creating a pro-angiogenic environment with increased VEGF expression and decreased expression of angiostatin. In addition, there were higher levels of eNOS phosphorylation as well as greater bioavailability of NO (157). In a rat model of acute myocardial infarction, GYY4137 increased blood vessel density and attenuated LV dysfunction compared to untreated and PAG-treated animals (115). S-propargyl-cysteine (SPRC) which promotes CSE activity and endogenous H₂S production, increased angiogenesis in models of both hindlimb and myocardial ischemia (92).

Following MCO, NaHS treatment significantly increased angiogenesis in the peri-infarct area and improved neurological function. This study indicated that functional outcomes were mediated through PI3K/Akt signaling. NaHS-treated astrocytes exhibited higher VEGF and Ang-1 expression, and inhibition of PI3K/Akt signaling significantly reduced the H₂S-induced expression of these molecules in astrocytes. Besides, PI3K/Akt signaling stimulated both EC migration and tube formation (86).

In animal models of diabetes and under conditions of hyperglycemia, H₂S treatment leads to pro-angiogenic responses and supports superior wound healing outcomes. In a well-characterized model of type II diabetes mellitus (T2DM), db/db leptin receptor-deficient mice treated with H₂S donors NaHS and HTB exhibited improved wound healing following punch biopsy wounding on the dorsal skin compared to untreated and PAG-treated controls (119). A more recent study in the same db/db model created round incisions on ischemic and nonischemic limbs. Treatment with NaHS significantly improved wound healing and angiogenesis and increased the expression of VEGF, EGF, PDGF, and eNOS as well as the phosphorylation of EGFR and PDGFR (207). In a different model of T2DM, ob/ob leptin-deficient mice were subjected to dorsal skin wounds and then treated with NaHS. Granulation tissue of ob/ob mouse wounds exhibited decreased CSE expression and H₂S levels. NaHS therapy increased wound healing and the density of vascular-like structures along with decreased inflammatory molecules in the wound site of ob/ob mice compared to controls (253). STZ-induced diabetic mice treated with 3MP, which promotes 3MST-mediated production of H₂S, had improved angiogenesis and wound healing after burn injury (131). In a similar STZ diabetic model, investigators found that NaHS accelerated reepithelization of a round incision wound on the hind dorsum, and increased VEGF and ICAM-1 as well as SOD and HO-1 expression, while decreasing TNF-α at the wound site, indicating pro-angiogenic effects that appear to work in concert with the antioxidant and anti-inflammatory effects discussed above (206).

Recent studies using H₂S-releasing biomaterials have shown promise in accelerating the wound healing process. For example, a hyaluronic acid (HA) hydrogel was used to encapsulate a pH-dependent H₂S donor JK1 and used in wound healing assays. Application of the H₂S-releasing hydrogel into dermal wounds increased wound healing rate compared to controls, exhibiting greater re-epithelialization, collagen deposition, angiogenesis, and cell proliferation. In addition, there was decreased inflammation due to M2 macrophage polarization (229). In a similar study, the pH-dependent H₂S donor JK1 was incorporated into a sodium alginate (SA) sponge. The SA/JK1 sponge was used to treat full-thickness wounds that produce heavy exudates, and compared to controls it significantly improved the wound healing process with enhanced granular tissue formation, re-epithelialization, collagen deposition, and angiogenesis (255).

Tumor progression

Tumor angiogenesis is a major hallmark in the progression of solid tumors. The initial tumor lesion cannot be sustained beyond a few millimeters due to the hypoxic nature and lack of nutrients at the tumor core (219). This hostile microenvironment triggers an angiogenic switch, in which gene expression is altered, and various cytokines and growth factors such as VEGF are produced by the tumor cells, leading to changes in nearby ECs that promote angiogenesis. Angiogenesis then promotes tumor progression by providing the tumor with a dedicated network of vessels that supplies oxygen and other nutrients, enabling rapid growth of the tumor cell mass.

Many solid tumors, including colon, lung, ovarian, prostate, and renal cell carcinomas, reportedly express higher levels of one or more of the H₂S-producing enzymes CBS, CSE, and 3MST (14, 32, 179, 208, 215). Colon cancer biopsies were shown to contain higher levels of CBS compared to patient-matched normal margin mucosa, and colon cancer-derived cell lines displayed both higher levels of CBS and higher production of H₂S compared to normal colon mucosa cells. Pharmacological inhibition of CBS with AOAA or shRNA-mediated downregulation of CBS reduced blood vessel formation in tumors in colon cancer patient-derived xenograft (PDX) models (185). In non-small-cell lung cancer (NSCLC), tumor biopsies had higher levels of CBS, CSE, and 3MST compared to adjacent nontumor tissue. In addition, cell lines derived from NSCLC also expressed higher levels of the enzymes and produced higher levels of H₂S compared to a normal lung epithelial cell line. The authors of this study showed that H₂S increased HIF-1α and VEGFA expression while AOAA and PAG decreased their expression. Those agents also decreased tumor microvessel density and tumor growth in a NSCLC PDX model (208). In primary epithelial ovarian cancers, CBS upregulation correlated with disease severity and CBS expression was higher in ovarian cancer cell lines than a nonmalignant ovarian surface epithelial cell line. Silencing of CBS in an orthotopic mouse model of ovarian cancer reduced vessel formation in tumors, especially in combination with the chemotherapeutic agent cisplatin (14). In a metastatic prostate cancer model, CSE/CTH was upregulated compared with primary tumor cells, corresponding with data showing higher CTH expression in late-stage cancer patients. In addition, RNA-seq data from The Cancer Genome Atlas revealed that higher CTH expression was associated with poor patient survival. Knockdown of CTH suppressed vessel formation in the tumors in the mouse model (215). Finally, in clear-cell renal cell carcinoma (ccRCC) cell lines, H₂S production was higher in Von Hippel-Lindau (VHL)-deficient cells compared to VHL WT cells. Inhibition of CBS with hydroxylamine in a CAM angiogenesis model using the VHL-deficient ccRCC cells reduced vascularization of the xenografts (179). Thus multiple in vivo, ex vivo and in vitro models demonstrate profound stimulation of angiogenesis by elevations in H₂S.

In several cancers, mechanisms underlying H₂S-mediated vessel formation were explored. As indicated above, H₂S can promote angiogenesis through multiple pathways. In NSCLC cells, H₂S increased HIF-1α expression, and in a positive feedback mechanism, HIF-1α increased expression of H₂S-producing enzymes. While this is protective in mediating improved vasodilation, it promotes rapid growth of multiple solid tumors. This group also showed that in tumor cells, H₂S activated PI3K/Akt signaling, while treatment with AOAA and PAG suppressed it, providing a potential pathway for the upregulation of HIF-1α and the subsequent promotion of VEGFA expression and tumor angiogenesis (208). In prostate cancer, increased CTH expression resulted in H₂S-mediated sulfhydration of cysteine-38 on the NF-κB p65 subunit, promoting NF-κB nuclear translocation, which altered expression of IL-1β, MMP-13, and VEGF and elicited pro-angiogenic effects (166, 215). A study in hepatoma cells also demonstrated the involvement of NF-κB in H₂S-induced proliferation, migration, and angiogenesis (256). A similar study in glioma cells showed that p38MAPK/ERK1/2-COX-2 signaling pathways participate in H₂S-induced proliferation and antiapoptotic effects (257).

Although H₂S-elicited angiogenesis is recognized as a major pro-cancer process, the role of H₂S in cancer is complex, with both pro-cancer and anticancer effects. Some of these discrepancies could be due to differences in the amount of H₂S endogenously produced or exogenously delivered to the tumor. For example, in hepatocellular carcinoma (HCC) xenografts, 25 to 100 μM NaHS increased blood vessel formation in tumors, while 800 to 1000 μM NaHS decreased their formation (227). Similarly, in a PDX thyroid carcinoma model, 1.4 to 2.8 mg/kg/day NaHS promoted tumor angiogenesis, while 11.2 mg/kg/day NaHS inhibited this process (226). These two studies also suggested that some of these H₂S-mediated processes may operate via ERK and PI3K/Akt signaling pathways (226, 227).

As discussed above, knockdown and pharmacological inhibition of H₂S-producing enzymes are being assessed as methods for treating various cancers. However, studies using H₂S donor NaHS have shown differing effects. Interestingly, other compounds that serve as H₂S donors have shown promise as treatment options for improving cancer outcomes. These include DAS, DADS, and DATS, which are also natural histone deacetylase inhibitors. DADS and DATS as well as HA-ADT and HS-ASA, a H₂S donating aspirin, reduced tumor growth and, in some cases, tumor angiogenesis in various xenograft models of triple-negative breast cancer (TNBC) (107).

For example, HA-ADT suppressed PI3K/Akt/mTOR and Ras/Raf/MEK/ERK signaling pathways, inhibiting tumor angiogenesis in human breast cancer xenografts (53). In glioblastoma multiforme (GBM) orthotopic xenograft and PDX models, DATS increased histone acetylation and activated caspase-3, decreasing proliferation and tumor angiogenesis (49, 177). Finally, treatment of osteosarcoma cells with DATS suppressed angiogenesis through the inactivation of Notch-1 signaling and the activation of tumor-suppressive microRNAs (112).

Studies to date suggest that H₂S has both physiological and pharmacological effects on angiogenesis and tumor progression. In the future, it will be important to continue to delineate the molecular mechanisms involved in the H₂S-mediated effects on tissue repair and vascularization of tumors. Enhancing or inhibiting H₂S production, depending on the context and desired effect on angiogenesis, could be an important therapeutic strategy for many disease states and malignancies.

Summary and Future Directions

Since its discovery as an endogenous signaling molecule in the brain by Hideo Kimura’s group in 1996 (1), the gasotransmitter H₂S has been validated to be an important regulator of vascular function, joining NO and CO as small, paracrine molecules that modulate vascular function and many other physiological processes (128, 200, 210). As discussed above, H₂S has been shown to vasodilate most arteries, reduce inflammation and oxidative stress, modulate insulin secretion and hepatic blood flow, regulate vascular permeability, and stimulate angiogenesis. However, many questions remain on how production of H₂S is regulated (15, 95), if the molecule is stored in tissue as sulfide compounds (130, 146, 171), and which are the endogenous targets (15, 65, 94, 231). An important question that is generating conflicting answers is on the contributions of the different sulfide pools in regulating physiological responses and contributing to pathological consequences. Indeed, it is not clear if circulating levels of sulfide compounds contribute to vascular effects or if local generation is the primary source for all or most actions of endogenous H₂S (188).

Recent years have also seen the introduction of therapeutic compounds that release H₂S to modulate parent compound actions. This includes NSAIDS combined with sulfide-releasing adjuvants to protect the gastrointestinal tract from detrimental effects on bile generation and gut microbiota (139, 205). Some drugs combine NO and H₂S releasing compounds, which show promising results to provide better anti-inflammatory, analgesic, antiplatelet, and anticancer properties (96, 150, 187). However, caution is warranted by studies showing that excess H₂S can promote the growth of certain tumors (73, 227, 257) and may contribute to hyperdynamic conditions in the splanchnic circulation under conditions of cirrhosis or other gastrointestinal disorders (58). Therefore, harnessing the promise of H₂S-based therapies requires additional research to answer the many questions that remain on how and where these exogenous compounds are working and how they intersect with endogenous synthesis.

In addition to effects in the vasculature, an emerging area of potential therapy by sulfide compounds is regulation of mitochondrial energetics and longevity. Studies in mice suggest that targeting H₂S therapy to mitochondrial targets can attenuate endothelial senescence (50, 155) and may prevent or even reverse some of the senescent effects of Alzheimer’s disease protecting nerve function through anti-apoptotic, anti-inflammatory, and antioxidant pathways (67, 202). Studies in transgenic mice demonstrated preservation of learning and memory as well as diminished Amyloid β plaque size through a reduction in p38 and ERK (202). Additional studies examining H₂S regulation of energetics and metabolism have demonstrated that transplantable organs are protected from degradation by H₂S perfusion (57, 121, 123). These studies are based on the ability of H₂S to suppress cellular metabolism, decreasing metabolic activity and energy requirements.

The field of longevity research has also found links to H₂S with evidence that increased synthesis of H₂S is required for the benefits of caloric restriction (61, 76, 214) through an effect mediated in part through sirtuin-1 upregulation (42, 225, 231). Studies have also demonstrated that some of the benefits of caloric restriction can be replicated by administration of H₂S donors or precursors (211, 226). These exciting developments suggest H₂S-based therapies may have widespread utility in medicine when targeted drugs are developed that capitalize on the beneficial effects of the compounds without the dangerous off-target effects.

Future directions in the field of H₂S research thus require a better understanding of the physiological pools and targets, reliable methods to measure production and tissue levels of the compound, and increased understanding of how endogenous production is regulated. Impressive gains have been made in recent years in advancing the field, and it is expected that the avalanche of information on this new signaling compound will continue to grow as new and better tools to study its physiological and pharmacological effects are developed.

Didactic Synopsis.

Major Teaching Points

• Hydrogen sulfide (H₂S) is an endogenous signaling molecule that exerts both beneficial and pathological effects in the vasculature including vasodilation and regulation of barrier function.

• Endogenous levels of H₂S are highest in target tissues and it appears to act predominately as a local signaling molecule.

• H₂S promotes angiogenesis which can contribute to wound healing and growth of certain tumors.

• H₂S regulation of the circulation involves both endogenous and microbiome-generated H₂S.

• H₂S reduces oxidative stress through multiple mechanisms, primarily through upregulation of antioxidant pathways.