A recent history of nitroxyl chemistry, pharmacology and therapeutic potential

Abstract

Due to the excitement surrounding the discovery of NO as an endogenously generated signalling molecule, a number of other nitrogen oxides were also investigated as possible physiological mediators. Among these was nitroxyl (HNO). Over the past 25 years or so, a significant amount of work by this laboratory and many others has disclosed that HNO possesses unique chemical properties and important pharmacological utility. Indeed, the pharmacological potential for HNO as a treatment for heart failure, among other uses, has garnered this curious molecule a considerable amount of recent attention. This review summarizes the events that led to this recent attention as well as poses important questions that are still to be answered with regards to understanding the chemistry and biology of HNO.

Linked Articles

This article is part of a themed section on Nitric Oxide 20 Years from the 1998 Nobel Prize. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v176.2/issuetoc

Abbreviations

sGC

soluble GC

IR

ischaemia–reperfusion

EDRF

endothelium‐derived relaxing factor

EDHF

endothelium‐derived hyperpolarizing factor

Introduction

The discovery of nitric oxide (NO) as an endogenously generated signalling agent with particular importance to the cardiovascular system ushered in a new era in biological signalling in the late 1980s and early 1990s (Furchgott, 1999; Ignarro, 1999; Murad, 1999). This paradigm‐shifting finding indicated that a small, freely diffusible molecule could be endogenously synthesized to act as a specific signalling species via unique chemical interactions at a receptor protein [in this case, the enzyme soluble GC (sGC) (Ignarro, 1999)]. Embracing this concept was originally difficult since NO was previously known primarily as a biological toxin and air pollutant and it was being proposed that it could elicit important and specific physiological effects (Moncada et al., 1991). Indeed, this finding was so novel and profound that the Nobel Prize was awarded in 1998 to the three researchers most responsible for this discovery, Louis Ignarro, Robert Furchgott and Ferid Murad. As a result of their work, many others became interested in other possible physiological effects associated with NO‐related and/or derived species that could conceivably be present in a physiological milieu. For example, NO reacts readily with superoxide (O₂ ⁻) giving peroxynitrite (ONOO⁻) and numerous studies (see Ferrer‐Sueta and Radi, 2009) have focused on the biological relevance of this reaction (Reaction (1)).

(1).

Chemical Formula: $\mathrm{NO}+{{\mathrm{O}}_2}^{-}➔{\text{ONOO}}^{-}$

NO can also be oxidized to give nitrosating species that could modify biological molecules, altering their function and/or activity (Thomas et al., 2008). As part of the torrent of work focused on determining the extent of nitrogen oxide‐based physiological signalling, several laboratories, including this one, began to explore the chemical biology and potential physiology associated with the one‐electron reduced and protonated NO species, nitroxyl (HNO, also referred to as azanone and nitrosylhydride, among other names). Many of the important advancements in the area of HNO chemistry and biology occurred contemporaneously with work on the chemical biology of NO and other related or derived species, (such as ONOO⁻ and nitrosation chemistry).

This review intends to present a history of the events (albeit not necessarily in chronological order) leading to our current understanding of HNO chemistry and biology. To be clear, the view presented herein will be biased towards my own personal perspective of the events in the field of HNO chemical biology and physiology and thus focuses on the period of time that coincides with the avalanche of work on mammalian nitrogen oxide chemical biology (i.e. immediately after the discovery of endogenous NO generation for the purposes of physiological signalling – starting in the early 1990s). I apologize to the many investigators who contributed greatly to the field of HNO chemistry, biochemistry and pharmacology if their work is not mentioned or appropriately highlighted. This is not a sign of disrespect but rather the result of a desire to be concise and to provide a story of one person's 25‐year journey through the ‘thickets’ of nitrogen oxide chemical biology and pharmacology. For those wanting more comprehensive and detailed treatments of the chemical biology and/or physiology of HNO, readers should consider the excellent reviews by Miranda (2005) and Irvine et al. (2008) as well as several others from this laboratory (Fukuto et al., 2005; Fukuto et al., 2009; Fukuto et al., 2013). It is also worth noting the recent book that is devoted to various aspects of HNO chemistry and biology (Doctorovich et al., 2017).

HNO – chemical studies

Prior to the discovery of endogenous, mammalian nitrogen oxide biosynthesis and NO‐based signalling, the chemistry of HNO was of interest primarily to inorganic chemists. Numerous labs examined the fundamental chemistry of HNO, and many of the results of these studies serve as the basis for subsequent work and speculation regarding its biological chemistry and its targets/actions. Of particular importance to both chemical and biological studies was the early discovery and characterization of Angeli's salt (sodium trioxodinitrate, Na₂N₂O₃), an important donor of HNO that is amenable to both chemical and biological studies (Angeli and Angelico, 1901). Angeli's salt is capable of releasing HNO (along with one equivalent of nitrite, NO₂ ⁻) in a proton‐dependent manner via Reaction (2) with a rate constant of 4.6 × 10⁻⁴ s⁻¹ (at room temperature) (Bonner and Hughes, 1988; Bonner and Ravid, 1975). Importantly, the release of HNO occurs readily at pH 4–8 allowing its use at physiological pH (Bonner and Ravid, 1975).

(2).

Chemical Formula: ${\mathrm{N}}_2{{\mathrm{O}}_3}^{2-}+{\mathrm{H}}^{+}\to \mathrm{HNO}+{{\mathrm{NO}}_2}^{-}$

For the most part, HNO can only be conveniently studied using donors such as Angeli's salt since HNO is inherently unstable with respect to dimerization–dehydration, giving initially hyponitrous acid (H₂N₂O₂), which ultimately decomposes to nitrous oxide (N₂O) and water (Reaction (3)) (Bazylinski and Hollocher, 1985).

(3).

Chemical Formula: $\mathrm{HNO}+\mathrm{HNO}➔\mathrm{HON}[chemistry\; double\; bond\; solid\; lines]\mathrm{NOH}➔{\mathrm{N}}_2\mathrm{O}+{\mathrm{H}}_2\mathrm{O}$

Thus, due to this instability in the presence of another identical molecule, HNO cannot be conveniently stored and must be generated indirectly via the use of donors such as Angeli's salt. Indeed, among most all other common nitrogen oxides of research interest [i.e. NO, nitrogen dioxide (NO₂), NO₂ ⁻, nitrate (NO₃ ⁻), ONOO⁻, hydroxylamine (NH₂OH) and ammonia (NH₃)], it is the only one that cannot be stored and used directly in chemical/biological experiments. The lack of inherent stability of HNO makes difficult many chemical studies (e.g. determination of kinetics and physical properties) that are otherwise routine with stable species. These experimental difficulties undoubtedly hindered early chemical investigations of HNO and were an obstacle to the elucidation of its unique chemical properties. However, the experimental limitations associated with studies of HNO have been overcome by investigators and many of the physical constants and chemical properties of HNO are now more firmly established (vide infra).

An important early question regarding HNO was with respect to its pK_a. A pulse radiolysis study from 1970 reported that the pK_a for HNO was 4.7 (Gratzel et al., 1970). If this were the correct pK_a value, the primary species at physiological pH would be the anion NO⁻ (the NO⁻/HNO ratio would be over 300/1). Indeed, the HNO pK_a value of 4.7 was generally accepted by most researchers until it became evident that it needed to be revised upward. In our laboratory, a re‐evaluation of the published pK_a was deemed necessary since most of the observed chemistry appeared to be due to HNO and not that of the anion, NO⁻ (e.g. reaction with nucleophilic thiols, vide infra). As alluded to above, experimental evaluation of the HNO pK_a is difficult because determination of the concentrations of the equilibrium species is made intractable by the fact that HNO is unstable and self‐reacts (Reaction (3)). The HNO/NO⁻,H⁺ equilibrium is further complicated by the fact that the anionic equilibrium partner, NO⁻, is isoelectronic with O₂ and thus has different available electronic spin states. That is, NO⁻ can potentially exist as a ground state triplet (³NO⁻, akin to ground state molecular oxygen ³O₂) or an excited state singlet (¹NO⁻, akin to singlet oxygen, ¹O₂). As HNO is a ground state singlet molecule (all the electrons paired in bonds or exist as lone pairs) deprotonation could conceivably occur in two distinct ways. HNO can deprotonate to the corresponding singlet ¹NO⁻ in a spin‐conserved process, which could be followed by intersystem crossing to the ground state triplet ³NO⁻. Alternatively, deprotonation of HNO can occur simultaneously with a spin flip to give directly the more stable triplet ³NO⁻ species (Figure 1).

Figure 1.

Possible deprotonation pathways for HNO. (A) Spin‐conserved deprotonation to the singlet ¹NO⁻ species followed by inter‐system crossing (isc) to the ground state triplet ³NO⁻ or (B) deprotonation coupled to spin conversion to the ground state triplet ³NO⁻.

An initial attempt to address this issue by this laboratory using computational chemistry revised the pK_a upward from 4.7 to 7.4 (Bartberger et al., 2001). Although this revision was in the right direction, this pK_a value was subsequently further revised by us (Bartberger et al., 2002) and others (Shafirovich and Lymar, 2002) to the now generally accepted value of approximately 11.4. Also, it was determined that the deprotonation occurs simultaneously with an electron spin flip (Shafirovich and Lymar, 2003) (Figure 1, pathway b), thus avoiding generation of ¹NO⁻ in the deprotonation pathway. The requirement for a spin state change to occur between singlet ¹HNO and its anionic equilibrium partner ³NO⁻ indicates that this deprotonation (and re‐protonation) reaction will be much slower than other deprotonation reactions. Indeed, the deprotonation of HNO (as well as the protonation of NO⁻) is expected to be too slow to be biologically relevant. That is, if HNO is introduced into a biological system, only the chemistry of HNO will be observed and not the chemistry of its anionic equilibrium partner NO⁻ as other reactions are likely to occur first. Also, if NO⁻ (presumably the triplet ³NO⁻) is introduced into a biological system, protonation will be slow and it is equally unlikely that any chemistry associated with HNO will occur (as other possible reactions will likely occur first). This significant revision of the HNO pK_a as well as the determination of the relevant spin states of the equilibrium partners greatly advanced the understanding of how HNO may be acting biologically. At the very least, it is now thought that the biological activity of HNO donors is a result of primarily HNO chemistry, with a limited possibility that the deprotonated ³NO⁻ species plays any major role.

Another important finding that occurred during the period when the HNO pK_a was being revised was the unequivocal determination of the reduction potential for NO. During this period, there were conflicting values for the NO/NO⁻ and NO,H⁺/HNO redox couples in the literature. Experimental work by Ehman and Sawyer (1968) and by Benderskii et al. (1989) reported a reduction potential of approximately −0.8 V (vs. NHE) for the NO/³NO⁻ couple. However, a theoretical value of +0.39 V (vs. NHE) for this couple was also published (Stanbury, 1989). If the NO reduction potential were +0.39 V, it would indicate that a biological one‐electron reduction of NO to give HNO/NO⁻ would be favourable and likely (and therefore, direct NO reduction could be a major source for biological HNO). However, when the theoretical value was initially determined, the correct pK_a and the true nature of the HNO/NO⁻ equilibrium (i.e. HNO deprotonation to ³NO⁻, not ¹NO⁻) had not been established, thus severely compromising the theoretical assessment. When these previously unestablished factors were considered, theoretical re‐evaluation of the NO reduction potential was in line with the experimentally derived values (Bartberger et al., 2002). It is now generally accepted that the NO/³NO⁻ reduction potential is approximately −0.8 V (vs. NHE). At physiological pH, the NO/HNO couple has been determined to be expectedly more positive (between −0.5 and −0.6 V vs. NHE). Regardless, these values indicate that NO is a very poor oxidant and it seems unlikely that one‐electron reduction to generate HNO in a biological system will occur to any significant extent.

One early study regarding the fundamental chemistry of HNO that ended up being extremely important to the current understanding of HNO chemical biology came from the Doyle laboratory who reported that HNO reacts with thiophenol to give the corresponding disulfide and NH₂OH via the intermediacy of a fleeting N‐hydroxysulfenamide (RSNHOH) (Reactions (4) and (5)) (Doyle et al., 1988).

(4).

Chemical Formula: ${\mathrm{C}}_6{\mathrm{H}}_5-\mathrm{SH}+\mathrm{HNO}➔{\mathrm{C}}_6{\mathrm{H}}_5\text{SNHOH}$

(5).

Chemical Formula: ${\mathrm{C}}_6{\mathrm{H}}_5\text{SNHOH}+{\mathrm{C}}_6{\mathrm{H}}_5-\mathrm{SH}➔{\mathrm{C}}_6{\mathrm{H}}_5-\mathrm{S}-\mathrm{S}-{\mathrm{C}}_6{\mathrm{H}}_5+{\mathrm{NH}}_2\mathrm{OH}$

This report indicated that HNO was electrophilic and could react readily with thiol nucleophiles, a property of HNO that appears to be important to its biological activity (vide infra). Further work from our laboratory confirmed the electrophilicity of HNO and indicated that it was particularly ‘thiophilic’ with limited ability to react with ‘harder’, oxygen‐based nucleophiles (Bartberger et al., 2001). Subsequent evaluation of this chemistry by us (Wong et al., 1998) and others (Shoeman et al., 2000) further established the thiophilicity of HNO. Moreover, in collaboration with the Nagasawa laboratory, we also discovered an alternative pathway for this reaction (Wong et al., 1998). As shown in Figure 2, the N‐hydroxysulfenamide intermediate (formed after initial attack of a nucleophilic thiol on the electrophilic HNO) has two possible fates depending on the conditions of the reaction. One fate has already been presented and is the reaction of the N‐hydroxysulfenamide with excess thiol to give the corresponding disulfide and NH₂OH (Reaction (5) and Figure 2, pathway a). The other fate involves a rearrangement of the N‐hydroxysulfenamide intermediate to the sulfinamide, which will occur when thiol reactant is not in significant excess (Figure 2, pathway b) (Wong et al., 1998; Sherman et al., 2010).

Figure 2.

The reaction of a thiol with HNO. Pathway A: reaction of N‐hydroxysulfenamide to give disulfide (RSSR) and NH₂OH. Pathway B: rearrangement of the N‐hydroxysulfenamide to a sulfinamide.

In a biological system, HNO‐mediated conversion of a thiol to a disulfide is a readily reversible process via reductive processes. That is, disulfides are easily reduced back to thiols via a variety of biological reductive processes. However, conversion of a thiol to a sulfinamide or its hydrolysis product, a sulfinic acid (RS(O)OH), is not readily reversible and for most proteins will represent an irreversible modification (i.e. not readily reduced back to the thiol). Thus, the chemical biology of HNO includes reaction with thiol proteins resulting in both reversible and irreversible modification, depending on the nature of the thiol and proximity of other thiols. From a kinetic perspective, it may make sense that thiol proteins could be targets for the actions of HNO. Importantly, the rate constant for the reaction of HNO with a thiol is fairly high, approximately 10⁶–10⁷ M⁻¹·s⁻¹ (Miranda et al., 2003; Jackson et al., 2009), significantly higher than the rate constants for other reported thiol‐modifying signalling species such as H₂O₂ [0.87 M⁻¹·s⁻¹ (Stone, 2004)] or ONOO⁻ [6 × 10³ M⁻¹·s⁻¹ (Radi et al., 1991)]. Indeed, to date, interaction of HNO with specific thiol proteins represents a primary rationale for explaining its observed biological activity (vide infra).

HNO versus NO chemical biology – is HNO merely an NO donor?

Much of the initial interest in HNO chemical biology was due to its perceived similarity to NO as a biological mediator (vide supra). HNO is the one‐electron reduced and protonated congener of NO. Thus, one‐electron oxidation and deprotonation of HNO will generate NO. As alluded to above, NO cannot be easily reduced to HNO in a biological system due to the unfavourable reduction potential. Thus, it is highly unlikely that the effects observed when NO is administered to a biological system are the result of its direct conversion to HNO. However, the opposite possibility, the conversion of HNO to the biologically active NO is not so clear. HNO can be converted to NO by relatively mild hydrogen atom abstractors since the H—NO bond strength is only approximately 50 kcal·mol⁻¹ (Gomes et al., 2004), making it a better hydrogen atom donor than many other biological antioxidants. Indeed, HNO is a potent antioxidant due to its ability to react with oxidizing radicals (Lopez et al., 2007), generating NO, which can also react with and quench reactive radical species (Reactions (6) and (7)).

(6).

Chemical Formula: $\mathrm{R}·+\mathrm{HNO}➔\mathrm{R}-\mathrm{H}+·\mathrm{NO}$

(7).

Chemical Formula: $\mathrm{R}·+·\mathrm{NO}➔\mathrm{R}-\mathrm{NO}$

Thus, the apparent ease of oxidation of HNO to NO may indicate that the chemical biology of HNO is, at least in part, merely the result of its facile oxidation to NO. As will be discussed later, it appears that the most salient pharmacological effects associated with HNO are not due to its conversion to NO. This is supported by the fact that HNO donors exhibit distinct pharmacological profiles compared to kinetically similar NO donors (vide infra). Also, the chemical properties and reactivity of HNO are extremely different from NO [in some cases considered to be ‘orthogonal’ (Miranda et al., 2003)]. Although the biological targets of NO and HNO are similar (e.g. thiols, metalloproteins and radicals), the mechanisms of interaction are distinct (Fukuto et al., 2013). For example, HNO will react directly with thiols (Figure 1) whereas NO requires prior oxidation to an ‘NO⁺’ (nitrosonium) equivalent species. Also, the products of the reactions of HNO versus NO‐derived species with thiols are completely different, with reaction with HNO giving either the disulfide or sulfinamide (depending on conditions, Figure 2) and reaction with NO‐derived species typically giving an S‐nitrosothiol as the initial product.

The reactions of NO and HNO with metals are also very distinct. For example, NO can coordinate with ferric ion (e.g. in myoglobin) to give a ferric nitrosyl complex, which in many cases will react with a nucleophile (e.g. a thiol) to give a nitrosated nucleophile and the ferrous ion (Reactions (8) and (9)) (e.g. Fukuto et al., 2012). The ferrous ion formed in Reaction (9) can react with another equivalent of NO to give the stable ferrous nitrosyl (Reaction (10)).

(8).

Chemical Formula: ${\mathrm{Fe}}^{3+}+\mathrm{NO}➔{\mathrm{Fe}}^{3+}-\mathrm{NO}$

(9).

Chemical Formula: ${\mathrm{Fe}}^{3+}-\mathrm{NO}+\mathrm{Nuc}➔{\mathrm{Fe}}^{2+}+\mathrm{Nuc}-\mathrm{NO}$

(10).

Chemical Formula: ${\mathrm{Fe}}^{2+}+\mathrm{NO}➔{\mathrm{Fe}}^{2+}-\mathrm{NO}$

On the other hand, the reaction of HNO with ferric ion results in direct formation of a ferrous–nitrosyl (Reaction (11)) without the formation of any intermediates.

(11).

Chemical Formula: ${\mathrm{Fe}}^{3+}+\mathrm{HNO}➔{\mathrm{Fe}}^{2+}-\mathrm{NO}+{\mathrm{H}}^{+}$

HNO is also capable of reacting directly with a ferrous haem (e.g. of deoxymyoglobin) to give the ferrous–HNO complex (Sulc et al., 2003) (Reaction (12)), which can be characterized by the ¹H‐NMR signal for the bound HNO ligand at around 15 ppm.

(12).

Chemical Formula: ${\mathrm{Fe}}^{2+}+\mathrm{HNO}➔{\mathrm{Fe}}^{2+}-\mathrm{N}\left(\mathrm{H}\right)\mathrm{O}$

The primary biological target for NO is the ferrous haem of sGC. Coordination of NO to the ferrous ion of sGC (Reaction (10)) results in activation leading to increases in the second messenger cGMP. It was hypothesized that HNO could also activate sGC similarly via coordination to the oxidized ferric‐haem of sGC as the product would also be the ferrous nitrosyl (Reaction (11)). However, this is not the case, probably due to the ferric‐haem of sGC being substitutionally inert (Miller et al., 2009; Zeller et al., 2009).

The reaction of NO with many radical species (e.g. organic radicals, R· or oxygen‐based radicals RO· or ROO·) results in rapid quenching of the radical resulting, at least initially, in the generation of the corresponding nitroso compound (R‐NO, Reaction (13)).

(13).

Chemical Formula: $\mathrm{R}·+\mathrm{NO}➔\mathrm{R}-\mathrm{NO}$

This chemistry has been proposed to be responsible for the antioxidant properties of NO as it results in the quenching of a possible chain‐carrying radical (e.g. Wink et al., 1993; Rubbo et al., 1995). As mentioned above, HNO also rapidly reacts with R· to give R‐H and NO and thus is also reported to be a potent antioxidant (Lopez et al., 2007). The antioxidant properties of HNO are, however, twofold because the radical quenching first reaction (HNO + R· giving R‐H + NO) results in the formation of an equivalent of NO, which as discussed immediately above, is also a potent antioxidant (Reactions (6) and (7)). Thus, HNO has the potential to quench two radical equivalents.

It should be apparent that HNO chemistry is distinct from NO chemistry, and thus, the differences in the biological actions are not unexpectedly also distinct. That is, HNO is not merely a precursor to NO in biological systems but possesses its own important chemical biology.

HNO, EDRF and NO biosynthesis

In an ingenious and watershed paper, Furchgott and Zawadzki reported, in 1980, the existence of a species that could diffuse from endothelial cells and elicit relaxation in neighbouring smooth muscle tissues (Furchgott and Zawadzki, 1980). To refer to this, at the time, unidentified species they used the term ‘endothelium‐derived relaxing factor’ or EDRF. The Ignarro and Furchgott laboratories then reported strong evidence that EDRF is NO (Ignarro et al., 1987; Furchgott, 1988), and it was later determined that EDRF is biosynthesized from the amino acid L‐arginine by a family of enzymes referred to as NOS to give NO and L‐citrulline (see Forstermann and Sessa, 2012). Although it is now generally accepted that EDRF is primarily NO, prior to this revelation, numerous other species were also postulated to be EDRFs including HNO (or at the time, NO⁻). Indeed, our laboratory was involved in the proposal that EDRF could be HNO (Fukuto et al., 1992a). This idea, in part, stemmed from chemical studies involving the oxidation of L‐arginine and arginine analogues. It was postulated and confirmed that L‐arginine could be oxidized to an N‐hydroxy‐L‐arginine intermediate by NOS and was then further oxidized to give NO and L‐citrulline (Figure 3, pathway a) (Stuehr et al., 1991; Wallace and Fukuto, 1991).

Figure 3.

The pathways for the conversion of L‐arginine to NO and citrulline. (A) NOS‐mediated oxidation of L‐arginine to N‐hydroxy L‐arginine, (B) further NOS‐mediated oxidation of N‐hydroxy L‐arginine to NO and citrulline, (C) oxidation of N‐hydroxy L‐arginine to the corresponding cyanamide and HNO. Further hydration of the cyanamide can lead to L‐citrulline and one‐electron oxidation of HNO can give NO.

The oxidation of L‐arginine to give NO and L‐citrulline is a very unusual biochemical transformation since it represents a net five‐electron oxidation of a guanidinium nitrogen atom giving the corresponding urea citrulline and the odd‐electron species, NO. The enzyme NOS is considered to be a type of cytochrome P450 (e.g. Griffith and Stuehr, 1995), which is a family of monooxygenases known to oxidize substrates via two‐electron oxidation processes. Thus, the conversion of the guanidinium nitrogen of L‐arginine to the N‐hydroxyguanidine species would be expected for a P450‐like protein as it represents a simple two‐electron oxidation (Figure 3, step a). However, the next NOS‐mediated step (Figure 3, step b), the conversion of N‐hydroxy L‐arginine to NO and L‐citrulline, represents an odd, three‐electron oxidation. The idea that a P450‐like protein, such as NOS, would perform an odd‐electron oxidation was, at the time, unprecedented. In order to study this biochemically unprecedented step, the oxidation of N‐hydroxyguanidine model compounds by a series of chemical oxidants was examined (Fukuto et al., 1992b; Fukuto et al., 1993a). In almost all cases, the observed product in these chemical studies was the corresponding cyanamide with subsequent generation of HNO. As this process represents a two‐electron oxidation (typical of cytochrome P450 reactions), it was proposed that both the cyanamide and HNO could be precursors for citrulline and NO (Figure 3, pathway c) and that EDRF could be HNO.

The idea that HNO could be a product of NOS catalysis was at least partly supported by subsequent studies of the effects of the enzyme SOD on NOS catalysis. It was previously shown that superoxide (O₂ ⁻) reacts with and destroys EDRF (Gryglewski et al., 1986). Thus, SOD can protect EDRF by catalytically removing O₂ ⁻ from solution before it can react. As NO is known to react rapidly with O₂ ⁻, giving ONOO⁻ (Reaction (1)) and ultimately NO₃ ⁻ (both inactive with respect to EDRF activity), it is entirely reasonable that the effect of SOD in protecting EDRF (aka NO) is due to removal of O₂ ⁻, a highly NO‐reactive species. However, SOD also reacts with HNO to give NO (Murphy and Sies, 1991). Thus, it was conceivable that another effect of SOD is to convert HNO to the biologically active NO species. That is, SOD does not merely protect NO from being destroyed by O₂ ⁻, but it may also convert HNO to the biologically active NO. Indeed, in collaboration with Adrian Hobbs (at the time in the laboratory of Lou Ignarro), we showed that SOD greatly enhances the vasorelaxant activity of HNO (from several HNO donors including Angeli's salt) with only a marginal effect on an NO donor (Fukuto et al., 1993b). Further work using purified and semi‐purified NOS preparations, also showed that SOD was capable of greatly increasing the NO yield from NOS in ways unrelated to its ability to remove O₂ ⁻ (Hobbs et al., 1994). Although this is consistent with the idea that NOS synthesizes HNO (at least under the conditions of these experiments), which can then be converted to NO in the presence of SOD, this remains to be established. It is noteworthy, however, that similar findings were published subsequently by another group who also postulated that HNO (or NO⁻), and not NO, was the actual product of NOS (Schmidt et al., 1996). Rand and colleagues also compared the reactivities of authentic EDRF with NO versus HNO and found the properties to be more like HNO, thus suggesting that a component of EDRF could be HNO (Ellis et al., 2000). It has also been postulated that the product of NOS may be a function of the cellular environment and that numerous nitrogen oxides, including HNO, can be generated (Pagliaro, 2003). Finally, NOS has been reported to be capable of generating HNO from L‐arginine under specific conditions (vide infra).

To be sure, it is generally accepted that EDRF is NO and NOS does indeed perform an odd‐electron oxidation to generate NO. The mention of the studies above is not meant to dispute this, because the weight of evidence supports NO as EDRF and the product of NOS (although reconciling these studies with those mentioned directly above has not occurred). However, these early studies regarding the possibility that HNO could be a product of NOS served as the impetus to examine the biology/physiology of HNO, which revealed some fascinating properties that are distinct from those of NO and may lead to the development of potentially important therapeutic agents (vide infra).

HNO pharmacology

Prior to any work pertaining to the vascular activity of HNO (as, e.g. a possible EDRF), the laboratory of Herb Nagasawa at the University of Minnesota published a series of momentous studies delineating important pharmacological properties associated with HNO. They showed that the anti‐alcoholic drug cyanamide (NH₂CN) was actually a prodrug for HNO (DeMaster et al., 1982, 1983, 1984, 1985, 1998; Nagasawa et al., 1990, 1992; Lee et al., 1992). That is, NH₂CN can be oxidized by catalase and H₂O₂ to give an N‐hydroxycyanamide (HONHCN), which in turn spontaneously decomposes to give HNO and cyanide (CN⁻) (Figure 4).

Figure 4.

Catalase‐mediated conversion of cyanamide to N‐hydroxycyanamide and subsequent release of HNO and CN⁻.

HNO generated from cyanamide was found to be a potent inhibitor of the enzyme aldehyde dehydrogenase via modification of the catalytic cysteine thiol (vide supra) leading to an accumulation of the toxic ethanol metabolite acetaldehyde (and thus representing an aversion therapy for excessive alcohol consumption). This was, to my knowledge, the first unequivocal demonstration of a pharmacological activity associated with HNO and the first example of HNO activity due to the interaction with a thiol protein. To a significant extent, these studies provided the framework and basis for our current understanding of HNO biology and pharmacology. Moreover, subsequent work from this group (some in collaboration with our laboratory) provided mechanistic insight into the chemical biology of HNO and led to the establishment of the dual reaction pathway chemistry with thiols previously discussed and shown in Figure 2 (Wong et al., 1998).

As HNO was proposed to be an EDRF, it was important to examine how it affects the cardiovascular system, especially in comparison with NO. As already mentioned previously, one of the earliest studies of this kind was performed using primarily Angeli's salt as an HNO donor and rabbit aorta or bovine intrapulmonary artery as test tissues. This study demonstrated that HNO donors were capable of eliciting vasorelaxation in a cGMP‐dependent manner, similar to NO (Fukuto et al., 1992a). Further work using cyanamide showed that HNO released via cyanamide oxidation by catalase and H₂O₂ (vide supra), also resulted in vasorelaxation in rabbit thoracic aorta (Fukuto et al., 1994). Other groups corroborated these findings and also found vasorelaxant activity associated with HNO (e.g. De Witt et al., 2001; Wanstall et al., 2001; Favaloro and Kemp‐Harper, 2007). Although it is debatable whether HNO is an endogenously generated EDRF, it has also been proposed that HNO is an endothelium‐derived hyperpolarizing factor (EDHF) (see Andrews et al., 2009). As such, HNO donors can have significant effects in decreasing vascular resistance.

A significant event in the history of HNO biology/pharmacology occurred when the laboratories of Wink and Paolocci collaborated and found that HNO (via Angeli's salt) had unique and potentially important effects on the heart. It was first reported that HNO had a positive inotropic effect on the heart that was distinct from the effects of NO (Paolocci et al., 2001). Subsequent work indicated that HNO elicited not only a positive inotropic effect but also a positive lusitropic effect (Paolocci et al., 2003). Importantly, these effects were independent of β‐adrenoceptor signalling and thus represented a potentially novel mechanism for modulating heart function. Further work from various laboratories (particularly those of Paolocci, Wink, Miranda, Kemp‐Harper and Toscano along with many others) indicated that HNO was capable of increasing myocyte Ca²⁺ cycling through activation of both the ryanodine receptor (Cheong et al., 2005; Tocchetti et al., 2007) and the sarcoplasmic Ca²⁺ reuptake pump SERCA2A (via interactions with the SERCA2A binding/regulating protein phospholamban) (Froelich et al., 2008; Sivakumaran et al., 2013), did not exhibit tolerance (Irvine et al., 2011), sensitized myofilaments to Ca²⁺ (Kohr et al., 2010; Gao et al., 2012), presented balanced vascular dilation (Paolocci et al., 2001) and did not induce tachycardia (Paolocci et al., 2003). Taken altogether, these effects represent a near perfect pharmacological profile for the treatment of heart failure. Unlike currently used treatments for heart failure, such as β‐adrenoceptor agonists or Ca²⁺ ionophores, which have significant negative side effects and appear to have only short‐term benefit, HNO appears to elicit only activities that will benefit a patient in heart failure. Current therapies do little to change mortality from heart failure (Arcaro et al., 2014), and it is hoped that HNO donors can change this. Thus, by all indications, HNO (or HNO donors) represent a novel and pharmacologically important strategy for treating heart failure. Due to the recognition of the therapeutic potential and clinical demand for a drug to treat heart failure, clinical studies have been performed using HNO donors with encouraging results (Tian et al., 2015; Parissis et al., 2017). It is safe to say that the discovery of the effects HNO elicits on the cardiovascular system (especially pertaining to heart failure) and its clinical potential took the interest in this molecule to new heights and is largely responsible for any current attention.

Another significant discovery regarding the pharmacology of HNO is its effects on ischaemia–reperfusion (IR) injury. Once again, the Wink and Paolocci labs (along with the Feelisch lab) collaborated to report that pretreatment of isolated rat hearts with Angeli's salt prior to reperfusion provided significant protection against subsequent reperfusion injury, resembling ischaemic preconditioning (Pagliaro et al., 2003). Importantly, pretreatment with a NO donor (with similar release kinetics) was much less effective. Interestingly, a previous study showed that infusion of Angeli's salt during reperfusion in an ischaemia/ reperfusion (IR) injury model system leads to an exacerbation of reperfusion damage (Ma et al., 1999). Although the mechanism by which HNO either protects or exacerbates IR injury is not known, it does appear that the timing of treatment is an important factor in the outcome.

Endogenous HNO biosynthesis

As discussed above, HNO appears to be a near perfect treatment for a failing heart. HNO elicits increased inotropy, lusitropy (via enhanced Ca²⁺ cycling), increased sensitivity of myofilaments to Ca²⁺ and balanced effects on the vascular system. Importantly, HNO is able to elicit all of these effects without potentially deleterious tachycardia, increased metabolic demand and without any other apparent toxicity. This scenario begs the question – is HNO an endogenously generated signalling molecule important, at the very least, to cardiac function? This combination of effects either represents an unbelievable pharmacological coincidence or HNO donors are actually acting by augmenting an otherwise insufficient endogenous signalling pathway. In many ways, this scenario is reminiscent of NO pharmacology prior to its discovery as an endogenously generated vasorelaxant (i.e. EDRF). Pharmacological studies with NO and/or NO donors revealed it to be an extremely potent and specific vasorelaxant that is relatively non‐toxic (at least at levels that normally regulate vascular function). However, unlike NO (which was found to be biosynthesized by a family of NOS enzymes), there are as yet no established and/or regulated physiological mechanisms for endogenous HNO biosynthesis.

Many of the presumed biological targets for HNO (with respect to its effects on the heart) are thiol proteins (e.g. ryanodine receptor, phospholamban, and myocyte proteins, vide infra), and HNO is likely to affect these proteins via modification of cysteine residues. Considering the vast number of thiol proteins in myocytes, it is remarkable that HNO does not have other significant effects associated with interactions with unrelated thiol proteins. Although the reasons for this apparent observed physiological specificity are unknown, it is highly likely that the proteins that are affected (all of which would be important in a response to a failing heart) are capable of specifically reacting with HNO compared to the plethora of other, physiologically unrelated thiol proteins. A similar conundrum existed for NO. As discussed earlier, the primary physiological target for NO is the regulatory ferrous haem of sGC. Since there are numerous haem proteins, why does NO interact primarily with the ferrous haem of sGC? That is, what allows NO to interact mostly with sGC in the presence of numerous other haem proteins. Also, why does sGC bind primarily only NO and not other diatomic ligands like O₂ or CO? Although addressing these questions regarding NO is beyond the scope of this review [readers are referred to the work of others (e.g. Boon and Marletta, 2005)], it is clear that HNO and NO share many similarities and it would not be at all surprising to discover that HNO is an endogenously generated signalling species.

To be clear, there are several possible ways HNO could be generated endogenously but none of them have been established as being sources for HNO as a signalling species. For example, HNO can be generated via the reaction of a thiol with an S‐nitrosothiol (Reaction (14)) (Wong et al., 1998).

(14).

Chemical Formula: $\mathrm{RS}-\mathrm{NO}+\mathrm{RSH}➔\text{RSSR}+\mathrm{HNO}$

Also, the NO generating enzyme NOS is capable of making HNO when deplete of one of its important prosthetic groups, tetrahydrobiopterin (Rusche et al., 1998; Adak et al., 2000), and has even been indirectly detected from the complete enzyme (Ishimura et al., 2005). However, none of these routes of possible HNO biosynthesis are currently considered to be pathways responsible for HNO generation for the purposes of, for example, responding to cardiovascular distress (such as heart failure). Although purely speculative, due to the highly specific effects HNO has on a failing heart, it would not be surprising to discover an HNO biosynthetic pathway that is active under circumstances of cardiovascular distress. As mentioned previously, NO biosynthesis occurs via enzymatic oxidation of a guanidinium nitrogen on the amino acid L‐arginine (Figure 3, pathway b). Although purely speculative, it is intriguing to consider that HNO biosynthesis, if it does occur, may also come from the oxidation of a nitrogen atom of an amino acid (e.g. arginine, glutamine, and asparagine). Considering the known ability of nitrosocarbonyls (RC(O)NO) to hydrolyse to give HNO and the corresponding carboxylic acid (RC(O)OH) (Miranda et al., 2005), it is also intriguing to consider that four‐electron oxidative conversion of either of the amino acids asparagine or glutamine to the nitrosocarbonyl, followed by hydrolysis can generate HNO and the corresponding acids, aspartate or glutamate respectively (Figure 5).

Figure 5.

Purely speculative pathway for oxidative HNO generation.

The above hypothesis, to be sure, is loosely based on the fact that NO production occurs via a five‐electron oxidation of the amino acid L‐arginine. Also, as mentioned earlier, NOS‐mediated oxidation of L‐arginine has already been shown to generate HNO when the enzyme is depleted of tetrahydrobiopterin (Rusche et al., 1998; Adak et al., 2000). Thus, biological oxidation of L‐arginine with subsequent generation of HNO has precedence and certainly remains a possible pathway for endogenous HNO biosynthesis (as depicted in Figure 3, pathway a–c).

The future of HNO research

Although the field of HNO chemistry and pharmacology/physiology has made significant strides in the past 20 years or so, there are many significant questions that remain to be answered. Some of the more pressing and profound questions are listed below.

1. Clearly, one of the most intriguing and important questions in the field of HNO pharmacology/physiology remains – is HNO endogenously generated as a physiological effector/mediator? As discussed above, an important aspect of this question may be that it would be reasonable to search for endogenous biosynthesis during cardiovascular stress (or at least it would be a good starting point). As HNO has effects on a failing heart and is proposed to be an EDHF, it becomes clear where and when it may be generated.

2. Of course, it is entirely possible, in spite of the remarkable pharmacological coincidence mentioned above, that HNO is merely an effective pharmacological agent that can be used to treat heart failure, alcoholism, prevent IR injury, etc. With respect to its effects on a failing heart, current drug development is primarily for treating acute episodes (i.e. emergency room situations). It will, therefore, be important to address the question: can HNO be developed as a long‐term treatment for heart failure? In order to fully appreciate the potential pharmacological utility of HNO, development of HNO donors that can be used for chronic heart failure will be important.

3. Currently, it is proposed that thiol proteins are targets for HNO activity. In the case of heart failure, it needs to be appreciated that HNO appears to interact primarily with specific myocardial thiol proteins involved in Ca²⁺ cycling and sensitivity (vide supra). If it is true that HNO has high selectivity for interaction with these thiol proteins, considering the immense number of thiol proteins available, there must be something unique about these proteins in terms of HNO reactivity. Thus, an essential question to ask is – how can HNO discriminate between thiol proteins? Or what structural motif is important in determining which thiol protein reacts with HNO? Answering this question will also provide insight into other possible functions of HNO.

4. It appears that HNO has the ability to react with thiol proteins, and these interactions may be responsible for much of its observed biological activity. However, it remains unclear exactly what these interactions are. That is, what is the exact nature of the chemistry of HNO interaction with its targets? As discussed previously, the reaction of HNO with thiols can take at least two pathways (Figure 2), one pathway leading to a disulfide and the other leading to a sulfinamide. In a system that requires ready reversibility, the pathway leading to the disulfide would be the most reasonable, whereas the pathway leading to the sulfinamide would be considered to be essentially irreversible (or at least very slowly reversible). Addressing the question above regarding the nature of the HNO‐reactive structural motif may help answer this question.

5. It has been speculated that HNO may be an endogenous signalling species. Regardless, HNO is clearly a potentially important pharmacological agent. In either case, as a possible regulator of physiological activity, it will be important to determine how the system is regulated. That is, what turns the signal off? Of course, answering this question will be highly dependent on the nature of the chemistry of HNO signalling (see above). If HNO‐mediated disulfide formation is involved, reductive reversibility would be expected. However, if sulfinamide (or the hydrolysed sulfinic acid) is formed, ready reversibility [although possible with select proteins (e.g. Woo et al., 2005)] would not be generally expected.

6. An intriguing aspect of HNO pharmacology relates to its effects on IR injury. As mentioned previously, whether HNO protects against or exacerbates IR injury appears to be a function of when it is administered (pretreatment affords protection, treatment coincident with reperfusion exacerbates injury). In either case, the mechanism by which HNO affects IR injury is not known. As protection from IR injury may be another important aspect of HNO pharmacology, it will be important to address the question: how does HNO protect against or exacerbate IR injury? Although HNO is known primarily for its ability to react with thiol proteins, in this regard, it will also be important to consider possible interactions with metalloproteins.

7. It has been proposed that HNO is an EDHF. The physiological importance of EDHFs (there may be numerous species that fill this role) cannot be overlooked and are likely to be just as important to vascular health/function as EDRF (i.e. NO) (e.g. Bryan et al., 2005). Thus, further characterization of HNO in this regard represents an important aim. Of course, determination of the source of HNO (see above), if it is indeed an endogenously generated EDHF, becomes an overwhelmingly important aspect of this characterization.

The questions posed above represent only a few that are currently worth asking and primarily deal with the biology and pharmacological application of HNO. Clearly, other more chemical questions remain to be addressed as well. However, due to the reported pharmacological properties of HNO and its potential therapeutic utility, the focus is understandably on these aspects of the future of HNO research. Finally, with the recent discovery of other thiol‐derived species, particularly the highly nucleophilic hydropersulfides (RSSH), as prevalent and potentially important physiological forms of sulfur, the interaction of HNO with these species may be expected and important. Indeed, it is conceivable that the apparent ability for HNO to interact with only select thiol proteins may be due to the presence of these highly reactive sulfur species (Alvarez et al., 2017), providing a kinetic advantage for reaction with HNO over other thiol sites. Again, this is speculative, but the intimate relationship between HNO and biological sulfur nucleophiles indicate that the nature of this potential interaction is worth considering in the elucidation of HNO biology. It is certain that other biological targets will present themselves in the future and that the pharmacological (and possibly the physiological) utility of HNO will continue to expand.

Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2017/18 (Alexander et al., 2017a,b,c,d).

Conflict of interest

The author served as a consultant for Cardioxyl Pharmaceuticals, and is named in a patent regarding HNO pharmacology; Nitroxyl progenitors in the treatment of heart failure, United States Patent No. 20040039063 filed 2004. [Correction note: The version of record of this article was published online on the 1st of July 2018. The Conflict of interest section was updated in October 2018 to include the disclosure of the authors relation to the HNO patent.]

Fukuto, J. M. (2019) A recent history of nitroxyl chemistry, pharmacology and therapeutic potential. British Journal of Pharmacology, 176: 135–146. 10.1111/bph.14384.