New insight sheds light as to how nitrite might reduce blood pressure height: is this alright?

This editorial refers to ‘Long-lasting blood pressure lowering effects of nitrite are NO-independent and mediated by hydrogen peroxide, persulfides, and oxidation of protein kinase G1α redox signalling’, by M. Feelisch et al., pp. 51–62.

Feelisch et al.¹ describe a novel mechanism by which the nitrite anion (NO₂⁻) produces long-lasting blood pressure reductions through a nitric oxide (NO)-independent pathway. Traditionally, NO₂⁻ was considered virtually inert. At a physiological pH, the amount of NO₂⁻ needed to produce vasodilatation of isolated vessels is ∼100 μMol/l, ∼10 000 times lower than the concentration of NO needed to evoke the same response.² Over the last 15 years, however, it has become apparent that much lower concentrations of NO₂⁻, either generated in vivo or administered systemically, can produce vasodilatation and lower blood pressure. Studies of the mechanisms underlying this effect have focused on at least three pathways. The simplest is conversion of NO₂⁻ to NO in acidic environments such as the stomach or in severely ischaemic tissues.^3,4 In an acidic environment, NO₂⁻ is protonated to form nitrous acid, which can yield NO or can generate nitrosating species to yield nitrosothiols, which in turn can undergo homolytic cleavage to release NO. A related pathway has implicated nitrate (NO₃⁻) being converted to NO₂⁻ by bacterial nitrate reductases in the mouth.⁵ The resultant NO₂⁻ is then converted to NO in by the acid environment of the stomach, and then causes local vasodilatation and blood pressure reduction. A third involves conversion of NO₂⁻ to NO by myriad heme-containing proteins, like deoxyhaemoglobin, myoglobin, xanthine oxidase, carbonic anhydrase, neuroglobin, and even nitric oxide synthase.⁶ The NO derived from each of these ultimately leads to activation of soluble guanylate cyclase, increased production of cyclic guanosine monophosphate (cGMP) and the so-called canonical activation of protein kinase G (PKG), of which the prevalent isoform is PKG1α. This activating step is mediated by binding of cGMP to a regulatory domain of PKG1α. PKG, in turn, promotes vasodilatation via several biochemical mechanisms including phosphorylation of the large-conductance calcium-activated potassium channel and activation of the myosin phosphatase complex, which dephosphorylates myosin and reduces its affinity for actin.

In contrast to this canonical activation pathway involving cGMP binding, several prior studies by the authors of the current manuscript have previously shown that PKG1α can be activated by oxidation, which causes disulfide bond formation between two cysteine 42 residues on adjacent chains in the PKG1α homodimer.⁷ The resultant oxidized cross-bridged protein is catalytically active even in the absence of cGMP. Addition of hydrogen peroxide to vessels stimulates PKG1α dimer formation and vasodilatation in a parallel fashion. This same group has produced mice that have the cysteine 42 replaced by a serine in PKG1α (C42S KI mice) and showed that vessels from these animals are resistant to the vasodilating effects of hydrogen peroxide.⁸

An interesting finding of the current study by Feelisch et al. is that NO₂⁻ produced vasodilatation of resistance (mesenteric) vessels in an apparent NO independent pathway, in that the NO scavenger PTIO had no effect on NO₂⁻ evoked relaxations. Likewise, ODQ, an inhibitor of guanylate cyclase had no effect on NO₂⁻ induced vasodilatation, indicating the cGMP activation of PKG1α is unlikely involved. In contrast, resistance vessels of C42S KI mice were completely resistant to NO₂⁻-induced vasorelaxation. Likewise, NO₂⁻ administered in vivo produced a prolonged decrease in blood pressure in wild-type mice, but not in C42S KI mice. The authors showed that NO₂⁻ inhibited catalase in the liver of mice and proposed this as a mechanism that caused hydrogen peroxide levels to be elevated in resistance vessels for up to 24 h after NO₂⁻ exposure.

While hydrogen peroxide could directly oxidize the cysteine thiols on PKG1α, Feelisch et al. provided evidence that polysulfide species were formed in resistance arteries after nitrite exposure, by measuring the surrogates cysteine hydropersulfide (Cys-SSH), oxidized glutathione (GSSG), and sulfane (HSSH). The first two of these were increased for up to 24 h in resistance vessels after nitrite exposure, while the last was actually decreased. The authors reasoned that HSSH is highly reactive and might lead to formation of other polysulfide species. The authors proposed that such polysulfides might also promote oxidation of other thiols including cysteine 42 on PKG1α. A schematic of their proposed pathway is provided in Figure 1.

Figure 1.

Pathway proposed by which ${\text{NO}}_2^{\text{−}}$ inhibits catalase, leading to accumulation of hydrogen peroxide and downstream polysulfide species and oxidize and dimerize PKG1α.

It is of interest that most of the above findings were only evident in resistance arteries, but not in conduit arteries. At first glance, this seems peculiar, because conduit arteries contain catalase and PKG1α, and therefore, NO₂⁻ should have a similar effect in these vessels as in the resistance arteries. It is likely this difference reflects the ability of conduit arteries to dispose of hydrogen peroxide via alternate mechanisms, such as via glutathione peroxidases, peroxiredoxins or thioredoxins. In keeping with this, in the present study, the authors showed that hydrogen peroxide levels actually declined from 1 to 24 h after NO₂⁻ exposure in conduit vessels. Likewise, the Eaton group has previously shown that aortas have much higher levels of 2-cys peroxiredoxin and thioredoxin than mesenteric arteries.⁹ Thioredoxins could cleave PKG1α cysteine 42 disulfide bonds or reduce other polysulfide species in larger vessels, inactivating this pathway. They also showed that preactivation of PKG1α via the canonical cGMP pathway attenuates oxidant-induced dimer formation.⁹ Thus, a predominance of the canonical pathway in larger arteries together with increased clearance of hydrogen peroxide and disulfides likely accounts for their reduced response to NO₂⁻.

The role of nitrite and the resultant increase in hydrogen peroxide in resistance vessels is clearly complex. Hydrogen peroxide has been implicated as an endothelium-derived hyperpolarizing factor which opens potassium channels in vascular smooth muscle. This effect of hydrogen peroxide is in large part due to oxidation of cysteine 42 of PKG1α. Prysyazhna et al.⁸ have shown that mesenteric vessels from the C42S KI mice have reduced hyperpolarization in response to hydrogen peroxide. Likewise, while catalase blocked acetylcholine responses in wild-type vessels, it was without effect in the C42S KI mesenteric vessels. Hydrogen peroxide via formation of downstream oxidants (like the hydroxyl radical) or carbonate radicals could oxidize tetrahydrobiopterin, which would, in turn, uncouple nitric oxide synthase, leading to a feed-forward production of more hydrogen peroxide. The interplay of endothelium-dependent hyperpolarization, NO synthase, and other vasodilators remains an interesting direction of investigation.

It should be noted that a recent report conflicts with the idea that PKG1α can be activated by hydrogen peroxide. Kalyanaraman et al.¹⁰ found that hydrogen peroxide did not activate either the purified enzyme or PKG1α when transfected into cells. These investigators showed that a cysteine mutant enzyme was simply less active than the wild-type enzyme and postulated that the reason for impaired vasodilatation in the C42S KI mice might be due to reduced activity of the mutant PKG1α. The reason for these discrepant results remains unclear but might be due to different assay conditions or perhaps differences in properties, such as post-translational modifications, of the recombinant proteins examined.

It is also interesting to consider the possible sources of NO₂⁻in vivo that could contribute to the phenomenon described by Feelisch et al. One major source is the diet. Indeed, various vegetables contain substantial amounts of nitrate, which as mentioned above can be converted to nitrite by bacterial reductases. Indeed, Webb et al.¹¹ showed that ingestion of beet juice, which contains high levels of NO₃⁻, produced systemic vasodilatation and blood pressure-lowering that coincided with the accumulation of plasma nitrite. Interruption of the enterosalivary conversion of NO₃⁻ to NO₂⁻ (by having the subject spit out saliva for the 3 h of beet juice ingestion) prevented the blood pressure-lowering effects of beet juice. A second major source is the nitric oxide synthases, which produce NO that is oxidatively degraded to NO₂⁻ and NO₃⁻. Indeed, in mice lacking the endothelial nitric oxide synthase, the plasma levels of NO₂⁻ are reduced by 70%.¹² While NO is considered a local vasodilator, its oxidation to NO₂⁻might promote systemic vasodilatation and blood pressure-lowering via the mechanism identified by Feelisch et al.

Finally, it is worthwhile to consider the consequences if NO₂⁻ really acts by inhibiting catalase over the long term. A perusal of clinicaltrials.gov shows that NO₂⁻ supplementation is being or has been studied in conditions ranging from metabolic syndrome, lung transplant, cerebral vasospasm, sickle cell disease to oesophageal dysmotility. If indeed NO₂⁻ inhibits catalase and causes accumulation of hydrogen peroxide, its chronic use in such conditions might be deleterious. While hydrogen peroxide can have favourable signalling roles, it can also lead to formation of disulfides, as shown by Feelisch et al., which in turn would lead to an array of changes in redox state of the cells. Oxidized glutathione is exported by cells, leading to substantial metabolic demand.¹³ Hydrogen peroxide and its downstream effects can alter gene transcription, promote apoptosis and perturb mitochondrial function. Catalase deficiency promotes a prediabetic phenotype.¹⁴ Therefore, NO₂⁻ inhibition of catalase and hydrogen peroxide accumulation, while having initial favourable effects on haemodynamics, might be deleterious over the long term.

Conflict of interest: none declared.

Funding

This work was supported by the National Institutes of Health Grants R35 HL140016 and Program Project Grant P01 HL129941 to D.G.H. and F32HL142937 to J.P.V.B.

The opinions expressed in this article are not necessarily those of the Editors of Cardiovascular Research or of the European Society of Cardiology.