OUP user menu

The potential role of the red blood cell in nitrite-dependent regulation of blood flow

Rakesh P. Patel, Neil Hogg, Daniel B. Kim-Shapiro
DOI: http://dx.doi.org/10.1093/cvr/cvq323 507-515 First published online: 14 October 2010


Nitrite was once thought to have little physiological relevance. However, nitrite is now being increasingly recognized as a therapeutic or possibly even physiological precursor of nitric oxide (NO) that is utilized when needed to increase blood flow. It is likely that different mechanisms for nitrite bioconversion occur in different tissues, but in the vascular system, there is evidence that erythrocyte haemoglobin (Hb) is responsible for the oxygen-dependent reduction of nitrite to modulate blood flow. Here, we review the complex chemical interactions of Hb and nitrite and discuss evidence supporting its role in vasodilation. We also discuss ongoing work focused on defining the precise mechanisms for export of NO activity from red blood cells and of other pathways that may mediate nitrite-dependent vasodilation.

  • Nitrite
  • Haemoglobin
  • Nitric oxide
  • Red blood cell
  • Vasodilation

1. Introduction

In the 1950s, Furchgott and Bhadrakom1 showed that high concentrations of nitrite (on the order of 100 μM) caused relaxation in pre-constricted aortic rings. Largely due to the fact that this concentration is several orders of magnitude greater than that found under physiological conditions, a role for nitrite as a physiological vasodilator was not considered for some time.2 However, over the last decade, the role of nitrite in the vasculature has evolved from being a relatively inert product of a nitric oxide (NO) metabolism/detoxification to being a specific biomarker for activation of endothelial NO synthase (eNOS) in the human circulation3,4 to the current view that this anion may be a therapeutic or even possibly physiological mediator of NO signalling during situations where oxygen tension is diminished.58 In the latter context, one investigated function is that of stimulating vasodilation and specifically mediating increased blood flow in response to decreased oxygen tension, a fundamental physiological process that ensures increased blood flow (and hence oxygen delivery) to metabolically active (and hypoxic) tissues. In addition, much work has focused on protective effects of nitrite-dependent NO formation against tissue damage in the context of ischaemia–reperfusion injury.810 Although NO-independent mechanisms of nitrite-dependent vasodilation have also been suggested,11,12 much evidence points to NO as an essential intermediate, necessitating the presence of oxygen tension-sensitive mechanisms that can reduce nitrite to NO in the vasculature. In conditions of extreme low pH, nitrous acid (HNO2) will disproportionate to spontaneously generate NO, and this mechanism may occur in the stomach mucosa and very ischaemic tissue.13,14 However, nitrite also acts as an NO-dependent vasodilator at physiological pH where NO formation from disproportionation can be excluded. A variety of enzyme/protein-dependent mechanisms have been proposed to explain nitrite reduction under physiological conditions. These include the involvement of haemoglobin (Hb),15,16 myoglobin,17,18 xanthine oxidoreductase (XOR),19,20 NO synthase,21 carbonic anhydrase,22 cytochrome c oxidase,23,24 cyclooxygenase,12 aldehyde oxidase,25,26 and cytochrome c.27,28 Each of these may contribute in different tissues under different conditions. However, Hb may be the most likely candidate to function at normal and near-normal pH and oxygen saturations in the vasculature and to contribute to hypoxic vasodilation. Experiments showing that oxypurinol does not diminish nitrite-mediated increases in forearm blood flow are contrary to the hypothesis that XOR plays a role in vasodilation.29 However, full elucidation of nitrite's mechanism of action in this and other processes is ongoing. In this review, we focus on the role of Hb and the red blood cell (RBC) in nitrite-mediated vasodilation.

2. Nitrite-dependent vasodilation

Multiple studies have shown the ability of nitrite to stimulate vasodilation of isolated vessels from both systemic and pulmonary vascular beds.15,3035 In fact, this observation was first made by Furchgott and Bhadrakom1 in the 1950s with the demonstration that nitrite at relatively high concentrations (100 µM) could stimulate vasodilation of aortic strips. However, only in the last decade has it been realized that nitrite can stimulate vasodilation in isolated vessels at concentrations that span the physiological (in the range of 100–200 nM36) to therapeutic range (up to 200 μM15). The primary difference between the early seminal studies and those performed more recently has been the incorporation of low oxygen tension and/or low pH into experimental systems. At low oxygen tensions, nitrite-dependent vasodilation is significantly potentiated, and this effect largely occurs via the canonical NO-dependent activation of soluble guanylyl cyclase.12,15,31,32,34,37

Observations in isolated vessel systems have been paralleled by studies showing that nitrite is a vasodilator at low concentrations in humans and animal models. The initial focus on nitrite as a mediator of vasodilation in vivo stemmed from the observation that NO inhalation could affect vascular reactivity in tissues distal to the lung if endogenous NO production was inhibited.38,39 It was also observed that a significant decrease in plasma nitrite levels and erythrocyte nitrosyl haemoglobin (HbNO) levels, but not S-nitrosohaemoglobin (SNO-Hb) levels, occurred upon transit of blood from the artery to the vein in the forearm circulation after NO inhalation. Significant differences in nitrite levels were observed even in the absence of NO inhalation, and nitrite consumption upon artery to vein transit was observed to be increased during exercise. Recently, this arterial to venous gradient in nitrite was confirmed in whole blood (arterial, 176 ± 10 nM vs. venous, 143 ± 7 nM36). These early observations led to the hypothesis that nitrite-dependent NO formation in the microcirculation leads to stimulation of NO-dependent vasodilation in response to oxygen demand and have been supported since by data demonstrating significant nitrite A–V gradients across the human cardiac circulation,40 nitrite-dependent increases in human forearm blood flow at low physiological concentrations during hypoxia but not normoxia,15,41 increased blood flow in the skeletal muscle of mice in ischaemic but not normoxic limbs,42 and increased blood flow in the cerebral43 and intestinal vasculature in rats,44 and the pulmonary vasculature in sheep.45 The observation by Dejam et al.46 that nitrite infusions resulting in plasma nitrite concentrations as low as 350 nM produce vasodilation in humans supports (but does not prove) the notion that nitrite acts as a physiological vasodilator. In addition, changes in plasma nitrite predict exercise capacity,47,48 although whether this relates to effects on blood flow or other metabolic effects (e.g. respiration) remain to be elucidated. In addition, several investigators have found that nitrite derived from oral nitrate resulting in marginal increases in plasma nitrite leads to decreases in blood pressure.4951 For example, drinking 250 mL of beetroot juice raised plasma nitrite from ∼380 ± 70 nM at baseline to 580 ± 90 nM after 2.5 h and reduced systolic blood pressure 5.4 ± 1.5 mmHg.49 The effects of the nitrate are abolished when conversion to nitrite by oral bacteria is blocked, which strongly implicates nitrite as the active agent.50 Moreover, recent studies in humans suggest that nitrite is a key intermediate not only in modulating blood pressure, but also effects on aerobic metabolism during exercise in response to dietary nitrate.5053 Collectively, these data highlight the concept that nitrite may be an endogenous element in the mechanisms that regulate vascular NO homeostasis and underscore the potential of nitrite supplementation to modulate NO signalling in vivo. This concept is now being extended to disease with a focus on dysfunctional RBC–nitrite reactions (discussed more below) including the so-called RBC storage lesion that describes toxicity associated with transfusion of older units of RBC, sickle-cell disease, and sepsis. In the latter case, a recent study showed that in sepsis patients compared with controls, A–V nitrite gradients in erythrocytes are lost, whereas plasma nitrite A–V gradients remain statistically different in the whole study group; however, plasma A–V nitrite gradient was also lost in patients who did not survive.54 Although correlative at this stage, these data do suggest a problem with nitrite utilization in human sepsis but many questions remain. For example, is the loss of A–V gradients due to altered red cell–nitrite reactions and/or due to shunting? Overall, there are many possible contributors to the nitrite levels that may be present during sepsis, so more work is required to establish its role in this condition. Interestingly, however, nitrite therapy was shown to increase survival in a rodent model of acute inflammation,55 which taken together highlights the potential for defects in nitrite metabolism in contributing to acute vascular inflammatory disease and for nitrite therapy in treating these disorders. This hypothesis remains to be tested.

The data discussed above supports the hypothesis that nitrite is a physiological vasodilator, but definitive evidence for this function in vivo remains to be obtained. Challenges to this hypothesis from a biochemical perspective are discussed below, but with regard to in vivo data, it should be noted that although nitrite A–V gradients have been reported, in some cases (e.g. coronary sinus), total NO metabolites across the microcirculation remains the same. Indeed, if nitrite is converted to NO to mediate blood flow, then a decrease in total NO metabolite should be expected. Rogers et al.40 also observed an increase in Hb-bound NO, thus suggesting that NO was made (and consistent with the idea that it was made from nitrite), i.e. NO-metabolite redistribution from nitrite to NO occurred across the microcirculation. However, the total NO-metabolite concentration remained the same. In any case, A–V gradients in nitrite do not pinpoint exactly where increased degradation occurs and thus cannot be definitively linked to function as in contributing to control of hyperaemia. Perhaps, the biggest experimental challenge is the fact that to definitively demonstrate a role for nitrite in physiologic blood flow, one would need to eliminate or at least decrease endogenous nitrite concentrations (as one uses NO synthase inhibitors to show physiological roles of the NO that is derived from NO synthases). To date, no viable way to eliminate endogenous nitrite has been discovered. Thus, overall, the hypothesis that nitrite, or more specifically NO derived from nitrite, acts as a physiological vasodilator remains controversial and more work is needed to address this controversy.

3. Role of RBC/Hb in nitrite-mediated vasodilation

The RBC and Hb have emerged as candidates for mediating nitrite-dependent vasodilation based on several lines of evidence. The first are observations that hypoxic blood flow in the human skeletal muscle is dependent on the oxygen fractional saturation of Hb and not on dissolved oxygen tension.56,57 Notwithstanding that there are multiple vascular bed- and species-dependent mechanisms which regulate hypoxic vasodilation, the fact that the physiological response is dependent on Hb fractional saturation is an important consideration when assessing mechanisms. Importantly, and as discussed in more detail below, nitrite reduction kinetics by erythrocytic- or cell-free Hb are directly regulated by oxygen fractional saturation exhibiting a bell-shaped dependence with respect to oxygen fractional saturation with maximal nitrite reduction rate being observed around the Hb P50 (oxygen tension at which Hb is 50% oxygenated).31,58,59 The consequence of this is that nitrite reduction becomes faster as RBCs are deoxygenated over the physiological range of fractional saturations in the blood flow-controlling arterioles. Most surprisingly, NO is detected outside of the RBC from nitrite reactions also with a bell-shaped dependence on fractional saturation.31 These studies are complemented by data showing that the combination of deoxygenated RBC and nitrite stimulate NO-dependent signalling, which is reversed by the NO scavenger C-PTIO.31 In summary, nitrite is reduced to NO in a manner that is regulated by the fractional saturation of Hb and hence regulated by allosteric mechanisms that control oxygen affinity. This provides a feasible mechanism that directly couples oxygen demand to changes in blood flow. Several other proteins have been shown to reduce nitrite to NO at low oxygen tensions in vascular tissue (listed in Section 1), but there is no clear direct link to these activities and Hb fractional saturation. Within RBCs, these include carbonic anhydrase22 and RBC-associated XOR.60 The possibility remains that at lower oxygen tensions and pH, these other non-Hb-dependent pathways become more prominent, although this remains to be tested. Other RBC and Hb fractional saturation-dependent mechanisms for stimulating NO signalling have been proposed involving SNO-Hb or adenosine triphosphate (ATP) release.57,61 These compounds have also been suggested to play a role in vasodilatory responses attributed to nitrite.6264 The potential interplay between nitrite, ATP, and SNO-Hb-dependent mechanisms will be discussed below.

Considerations of the mechanisms by which RBC may stimulate NO signalling are dominated by the fact that oxyhaemoglobin (oxyHb) or deoxyhaemoglobin (deoxyHb) also scavenge NO and do so with extreme rapidity. Generally, oxygen concentration alone can affect NO metabolism and vascular reactivity. The biological effect which is mediated by the NO concentration ‘sensed’ by the vascular smooth muscle will be dictated by the balance of NO formation and decay (as illustrated in Figure 1). Indeed, vessel studies with cell-free Hb support this concept32,34,35,65 and show that when oxygenated, Hb inhibits nitrite-dependent vessel dilation through NO scavenging. However, when Hb is deoxygenated, no inhibition is observed. In contrast, both oxyHb and deoxyHb inhibit vessel relaxation when NO is generated from a nitrite-independent source. The rate constant for NO scavenging by oxyHb is ∼5 × 107 M−1 s−1, whereas NO binding to deoxyHb occurs with a rate constant of ∼3 × 107 M−1 s−1.6668 This difference is minor; free NO produced in an RBC is scavenged extremely fast under both oxygenated and deoxygenated conditions. Possible mechanisms whereby NO produced through Hb-mediated nitrite reduction scavenging are discussed below. Regarding oxygen sensing, however, the fact that Hb-mediated nitrite reduction is much faster under partially oxygenated conditions than when fully oxygenated (by at least an order of magnitude69) supports a role for Hb in oxygen tension-sensitive nitrite-mediated vasodilation. Assuming a bimolecular rate constant for nitrite reduction of 10 M−1s−1 when the Hb is 50% oxygen saturated, nitrite would be reduced at a rate of 1 nM/s by Hb in whole blood. It should, however, be noted that some laboratories have obtained data from which they conclude that nitrite-mediated vasodilation is Hb-independent, so work in this area is ongoing.26,32

Figure 1

Balance of NO production and consumption by red cells. This illustration qualitatively describes how red cells, due to reactions involving intraerythrocytic Hb, both consume NO and produce it (via the reaction with nitrite). NO consumption is slightly less efficient at lower oxygen pressures due to the fact that NO reacts with oxygenated Hb a little faster than with deoxygenated Hb. NO production is substantially increased at lower oxygen tensions due to the reaction of nitrite and deoxyHb. The net effect is that relatively more NO is produced than consumed as red cells deoxygenate along physiologically relevant oxygen gradients. It is important to note that red cell Hb is always present in the blood vessels (so experiments quantifying NO in the absence of Hb are not directly applicable to physiology). Although some have shown more NO is produced in other tissues (such as the liver) or in the absence of Hb, it is hard to see how this would be relevant to physiological control of blood flow.

The fact that nitrite can mitigate the inhibitory effects of Hb vessel dilation under partially deoxygenated conditions supports the notion that the additional NO generated by Hb-dependent nitrite reduction is sufficient to increase the level of NO present in the smooth muscle. This concept has been demonstrated further with cross-linked cell-free Hb's that have been developed as blood substitutes. One of the major concerns with administration of cell-free Hb is rapid scavenging of NO resulting in hypertension. Using a model of haemorrhage shock and resuscitation with HBOC-201 (a glutaraldehyde cross-linked bovine Hb), the concomitant administration of nitrite during resuscitation attenuated hypertension65 consistent with the model described above. Interestingly, these studies taken together with others using Hb's modified with different cross-linking protocols70 have shown that kinetics and oxygen dependence for nitrite reduction (discussed below) can be altered, raising the potential of engineering a spectrum of Hb's that possess properties optimal for nitrite reduction to NO.

Recent studies have shown that the rate of nitrite reduction is approximately two- to three-fold slower with erythrocytic vs. cell-free Hb.35,71 Nitrite is a charged species at pH 7.4 (pKa ∼3), raising the question of how nitrite is transported in/out of the RBC and if transport modulates its reduction by erythrocytic Hb. HNO2, which is presumably more permeable to plasma membranes than nitrite, will be present in low amounts at pH 7.4, and it is conceivable that nitrite could cross membranes by simple diffusion coupled with protonation/deprotonation driven by concentration gradients.71,72 Overall, the slow uptake of nitrite by RBCs presents a challenge to the notion that RBCs play a role in vasodilation. Our studies suggest, however, that nitrite transport across the RBC is directly regulated by fractional saturation with deoxyHb, preventing nitrite export from the RBC by inhibiting the anion exchanger 1 protein.35 In this way, fractional saturation regulates nitrite transport and deoxyhaem-mediated nitrite reduction in a concerted process.71 The rate of nitrite entry and export into the RBC also must play a role in the observation that nitrite seems to be more concentrated in the red cell than in the plasma (290 vs. 120 nM),36 although contrary findings have been reported by others.73,74 The role of the RBC membrane in controlling nitrite–Hb reactions remains under investigation, but it is likely that the control of nitrite transport into, and out of, the RBC is an additional factor, affecting the ability of nitrite to mediate NO signalling via the RBC.

4. Reaction between nitrite and Hb

Ever since the discovery that relatively low levels of nitrite can modulate blood flow, potentially by an apparent Hb-dependent mechanism,15 there has been renewed interest in the reactions of nitrite and Hb.16,62,69,7580 The reaction of nitrite with oxygenated Hb is shown in Eq. (1) (also Reaction 3 depicted in Figure 2) Embedded Image (1) Both the reaction stoichiometry and the observed autocatalytic kinetics of this reaction indicate significant mechanistic complexity. Autocatalysis occurs when the product of a reaction is also a catalyst of the reaction. Previous studies have implicated many intermediates including hydrogen peroxide, superoxide, nitrogen dioxide (NO2), peroxynitrate, and ozone. Using a combination of experimental and numerical approaches, we have recently reassessed this mechanism and provided an experimental model that kinetically explains experimental data. The most striking difference between this model and previous ones is that the autocatalyst is solely NO2 and hydrogen peroxide only acts as a chain initiator. Under physiological conditions, this autocatalytic component of this reaction is unlikely to occur and the rate will be limited by the initiation step, with a rate constant of ∼0.2 M−1 s−1.81,82 The reaction of deoxygenated Hb and nitrite (Reactions 1 and 2 depicted in Figure 2) yields methaemoglobin (metHb) and iron HbNO via the following reaction scheme83,84 Embedded Image (2) Embedded Image (3) Nitrite binds to, and is reduced by, the ferrous haem to liberate NO, which can then bind to a vacant deoxygenated haem. This reaction should be rate limited by Eq. (2) and should follow simple pH-dependent second-order kinetics. Initial investigations of this reaction concluded that it followed the pH-dependent second-order kinetics as predicted by Eq. (2).8385 However, more recent investigations have revealed that the situation is considerably more complex (reviewed in more detail elsewhere78,86).

Figure 2

Possible major reactions occurring in the RBC that involve nitrite and other nitrogen oxides. The Hb is represented by the haem group, drawn as an oval containing an iron in either the ferrous or the ferric form. Potential reactions where a vasodilatory signalling molecule might be exported from the red cell (large oval) are shown in purple. Double-headed arrows are used to depict reversible reactions (not a resonance structure). In most cases, reactions are only shown once but could occur at multiple places. For example, nitrite would be able to bind any metHb, not just after Reaction 3. Not all of the equations are balanced, e.g. the stoichiometry of Reaction 3 is believed to follow Eq. (1) in the text, but only one nitrite is shown, so complete balance would have been difficult to draw. In other cases, water or protons are not explicit. Reaction 1 is the initial step in the deoxyHb/nitrite reaction, which could result in the HbFe3+–NO intermediate. This intermediate becomes MetHb and NO via Reaction 2. Published dissociation rates of NO from MetHb are on the order of 1 s1 93 which would make this intermediate transient. Some investigators, however, have proposed that this intermediate is more stable,16,62 whereas others have contested this proposal.76 Strictly speaking, Reaction 2 is reversible, but in the RBC, once NO comes off, it is not likely to have chance to rebind to metHb. The combination of Reactions 1 and 2 is under allosteric control and occurs at a rate with a bimolecular rate constant of ∼12 M−1s−1 for R-state Hb and 0.2 M−1s−1 for R-state Hb.69,78 Reaction 3 depicts that of nitrite and oxygenated Hb. This reaction has both a slow, initiation phase and a fast, autocatalytic phase. However, it is not likely that the autocatalytic phase will be realized in a red cell due to competing reactions of intermediate species like NO2. Reaction 4 depicts the reversible binding of nitrite to metHb. In the text, the bound form is shown as the N-bound ‘nitro' form, but it is likely to also bind in the O-bound ‘nitrito' form. The dissociation constant of metHb–nitrite has been reported to be 1 mM,93,94 but substantially lower (as low as 50 μM or less) under certain conditions (such as lower pH).76,95,96 Reaction 5 depicts the reaction of nitrite-bound MetHb with NO to form reduced deoxygenated Hb and N2O3 as proposed by Basu et al.76 Formation of N2O3 could lead to NO activity export from the red cell through two ways: (i) through diffusion out of the red cell with subsequent formation of NO and nitrite outside the cell (shown in the purple arrow) and (ii) by reaction with a thiol (such as glutathione or cysteine) which via Reaction 6 followed by export of the nitrosothiol formed in the reaction. Reaction 7 depicts the formation of SNO-Hb from the HbFe3+–NO intermediate (which has HbFe2+–NO+ character). Here, NO+ from HbFe2+–NO+ is intramolecularly transferred to the β-93 cysteine.102 Reaction 9 depicts the formation of N2O3 by the reaction of nitrite with the HbFe3+–NO intermediate as proposed by Fernandez and Ford.104 The N2O3 can lead to NO activity export as described for Reactions 5 and 6. Reaction 10 describes the action of some, as yet unspecified, nitrite/Hb product to release glycolytic enzymes from the membrane with subsequent production of ATP, which is then exported.63,64 Reaction 11 represents NO binding to deoxygenated Hb which occurs with a rate constant of 3 × 107 M−1s1 66,68 and a dissociation rate of 10−3–10−5 s−1.111,112 Reaction 12 describes oxidative denitrosylation where intermediates in the oxyHb/nitrite reaction oxidize the ferrous nitrosyl haem, producing NO and a ferric haem.77 Last, certainly not least, Reaction 13 describes the NO dioxygenation reaction where NO reacts with oxygenated Hb to form metHb and nitrate. This reaction occurs at a rate of 5 × 107 M−1s−166,67 and is the primary reaction responsible for limiting export of NO from the RBC.

Under pseudo-first-order conditions, with nitrite in excess of deoxyHb, it would be expected that the loss of deoxyHb would follow an exponential time course. However, as shown in Figure 3A, the kinetic progress curve is almost linear but with a slight but reproducible sigmoidal character.69,75 A clue to the mechanistic underpinning of these unexpected kinetics was the effect of N-ethylmaleimide (NEM) and inositol hexaphosphate (IP6) on the reaction kinetics. NEM alkylates Hb cysteinyl residues, in particular the β-93 cysteine, and in so doing stabilizes the Hb in the high oxygen affinity or R-state. In contrast, IP6, which binds to a cleft made by the subunit interface of the Hb tetramer, stabilizes Hb in the low oxygen affinity T-state. It was observed that NEM accelerated nitrite-dependent deoxyHb decay and IP6 decelerated it, pointing to an association between reaction rate and protein conformation. It was also observed that the reaction of nitrite with deoxyHb was faster when a percentage of the Hb was previously oxidized to metHb. The full kinetic curve can thus be explained as follows: at the start of the experiment, all the Hb is deoxygenated and in the T-state (T0) and reacts at a slow rate. However, each nitrite reduced generates both a metHb and an NO molecule which then binds to another deoxyHb molecule to form an HbNO. Both the HbNO and metHb, when present in a tetramer (in a similar way to oxygen binding), have a tendency to promote a conformation change to the R-state. Consequently, the more oxidation and nitrosylation of haems, the more R-state deoxygenated haems will be present. As R-state deoxyhaems react more rapidly with nitrite, this will lead to autocatalysis of deoxyHb decay and hence sigmoidal kinetics.69,75,78 The R-state enhancement of nitrite reactivity has the major biological implication that the NO-producing reaction of deoxyHb will be most rapid under partially oxygenated conditions (illustrated in Figure 3B) where there is a maximal presence of deoxygenated haems on R-state tetramers.78 This allows for a graded modulation of NO formation rate along the physiological oxygen continuum with more NO generated as oxygen is decreased. It is important to note that blood flow is largely controlled at the level of the arteriole. Thus, the oxygen tension and Hb oxygen saturations that are relevant to the nitrite-mediated control of blood flow by red cells are those that occur in the arteriole and will vary depending on the tissue, metabolic activity, and allosteric effectors that regulate Hb oxygen affinity. On the basis of Figure 2, the prediction will be that as Hb deoxygenation occurs, graded NO production will occur until the P50 is reached, and thereafter, the rate of nitrite reduction will decrease. However, albeit slower, significant NO production would still occur over saturations that likely span the physiological spectrum (>30%). It should be noted that although several studies have found substantial Hb desaturation as you move down the arteriolar network (up to 50% and oxygen tensions decreasing down to ∼30 mmHg8789), whether or not oxygen saturations become low enough at the arteriole so that Hb-mediated nitrite reduction would modulate blood flow remains controversial. However, among the many mechanistic candidates, Hb appears to act at the highest oxygen tensions and so the same challenge may have more force when aimed at these other candidates. Under conditions of more extreme hypoxia, it is possible that these other non-Hb nitrite reduction pathways (e.g. XOR) begin to contribute to NO production.

Figure 3

Allosterically controlled kinetics of the nitrite/deoxyHb reaction. (A) Kinetics of decay of deoxyHb when reacted with 500 μM nitrite under anaerobic conditions. The reaction was followed by absorption spectroscopy and concentrations of species present (deoxyHb, MetHb, and HbNO) were determined by spectral deconvolution using basis spectra. (B) The reaction rates of nitrite with deoxygenated Hb are plotted as a function of oxygen saturation for cases where the product of the nitrite and Hb concentrations are 10−6 M2. At each oxygen saturation, the rate of the reaction was calculated as [nitrite] × {kt(4[T0] + 3[T1] + 2[T2] + [T3]) + kR(4[R0] + 3[R1] + 2[R2] + [R3]), where the square brackets refer to Hb concentrations. Here, kR/kT was rounded to 100 and kT was set to 0.2 M−1s−1. The contribution by R-state and T-state molecules was obtained by calculating the products of kR and kT separately (so, for example, the R-state contribution is [nitrite] × kR(4[R0] + 3[R1] + 2[R2] + [R3]). This research was originally published in Kim-Shapiro and Gladwin.78 © the American Society of Hematology.

Although Hb/red cell-mediated reduction of nitrite to NO with concomitant increase in blood flow has not been observed in vivo, the dependence of nitrite reduction by Hb on oxygen tension as illustrated in Figure 3B supports the hypothesis that this plays a role in increasing blood flow at the level of the arteriole. Alternative or intermediate species that form in the reaction of nitrite and deoxyHb have been proposed.16,62,76,90 Rifkind and co-workers90 recently proposed that nitrite initially and rapidly binds to the ferrous haem of deoxyHb to form a relatively stable complex. The species was proposed due to a discrepancy between the nitrite remaining in solution after rapid filtration of Hb from an Hb/nitrite mixture, and the amount of HbFe2+–NO formed.90 The unaccounted nitrite was assumed to be semi-stably bound to the haem.90 However, there is no direct evidence to support the formation of this complex, and significant haem-binding is unlikely as there is no associated rapid change in the visible absorption spectrum that supports this notion. It is possible that the species inferred is actually nitrite that is bound to other parts of the Hb molecule.

Another intermediate that has been proposed to exist is a ferric HbNO. It is likely that NO is initially formed as a ferric nitrosyl species, HbFe3+–NO (shown as the product of Reaction 1 in Figure 2),16,62,90 which dissociates to metHb and NO. The transient formation of such a species would be similar to that proposed to form during nitrite reduction by certain nitrite reductases.91 The dissociation rate of NO from the ferric haem of Hb (1 s−1)9,2 and the fact that NO binds ferrous haems ∼1 million times tighter than ferric ones92 suggest that this species would not be expected to accumulate to detectable levels. However, several investigators have claimed to observe substantial accumulation of the species using absorption spectroscopy and other means.16,62,90

Rifkind and co-workers have suggested that the observed intermediate is chemically different from HbFe3+–NO and have described it as ‘Hb(II)NO+ ↔ Hb(III)NO’ (or HbFe2+–NO+ ↔ HbFe3+–NO using our nomenclature). This nomenclature is profoundly confusing, as (i) it is well known that ferric HbNO has ferrous-nitrosonium character, (ii) it suggests that ferric nitrosyl formed from metHb and NO should also have access to this stabilized state, and (iii) a more stable ferric nitrosyl cannot simply result from electron delocalization/resonance, but must require significant changes in the environment of the haem pocket to dramatically slow the NO off-rate. In all of our extensive spectral characterization of this reaction (published and unpublished), we have not observed any behaviour that suggests the presence of a spectrally distinct intermediate. In addition, we have been unable to detect any accumulation of this putative intermediate using a highly sensitive chemiluminescence-based technique.76

Nitrite bound to metHb, HbFe3+–NO2 (shown as the product of Reaction 4 in Figure 2) has been shown to form in the nitrite/deoxyHb reaction.75,76 However, several studies have reported that the affinity of nitrite for metHb is very low with a dissociation constant for HbFe3+–NO2 of ∼1 mM.93,94 Under certain conditions, such as pH 6.5 or lower, the affinity is much higher.76,95,96 The complex binding of nitrite with metHb may be partially due to its ability to bind in both the more common N-bound ‘nitro' form and the less common O-bound ‘nitrito' forms. This less common O-bound form has recently been observed using X-ray diffraction.97

Under partially oxygenated conditions, nitrite reacts with both oxygenated and deoxygenated haems in parallel.77 The autocatalytic phase of the oxyHb/nitrite reaction is inhibited while accumulation of iron HbNO of the deoxyHb/nitrite reaction is also diminished. These observations are explained in terms of a process called ‘oxidative denitrosylation’ (Reaction 12 in Figure 2) where intermediates in the oxyHb/nitrite reaction such as NO2 oxidize the nitrosyl haem and release NO.77 This action could, in principle, contribute to NO release from the red cell.

5. Export of NO activity from the RBC

One of, if not, the biggest challenges to the notion that nitrite-dependent regulation of blood flow involves a reaction with Hb is that NO itself cannot escape the RBC. This is mainly due to the fast dioxygenation reaction of NO with oxyHb to form nitrate (Reaction 13 in Figure 2). In our 2003 paper, we wrote ‘We realize that the high concentrations of hemoglobin in red cells, coupled with the near-diffusion-limited reaction rates (∼107 M−1s−1) of NO with hemoglobin, seem to prohibit NO from being exported from the red blood cell'.15 With 20 mM Hb in the red cell (in haem), the lifetime of NO would only be ∼1 μs and could diffuse only ∼0.1 µm or less. Applying finite-element computational analysis and ignoring other potential factors such as compartmentalization (and intermediates), the amount of NO exported from a red cell when exposed to a maximal, therapeutically relevant, dose of 200 μM only increased the external NO concentration by 0.1 pM, not nearly enough to effect vasodilation.98 Thus, the ability of red cells to affect nitrite-mediated vasodilation has focused very much on intermediate or alternative species.

One proposed mechanism for NO export involved HbFe3+–NO formed in the nitrite/deoxyHb reaction, where it was argued that since the bound NO here is more labile than when bound to the ferrous haem, it would be easier to deliver.16 However, issues concerning the stability of this species discussed above limit the likelihood of this pathway. Moreover, even if such a species were stable under certain conditions, it is not clear how this would facilitate NO export. Once released from the HbFe3+–NO, the NO would be no better off than NO originally formed in the nitrite/deoxyHb reaction [Eq. (2)] and would likely be scavenged by the dioxygenation reaction.

A theoretically more likely way that HbFe3+–NO would be involved in the export of NO activity would be if it led to the formation of S-nitrosothiols. Nitrosothiols are relatively stable and thought to play a substantial role in nitrogen oxide signalling including vasodilation and could be exported from the red cell.99,100 Importantly, SNO-Hb was detected following nitrite infusions.15 The formation of SNO-Hb from HbFe3+–NO was described by Singel and co-workers where NO+ from HbFe2+–NO+ either reacts with OH to form nitrite (as in classic reductive nitrosylation101) or reacts with the Hb β-93 cysteine to form SNO-Hb (Reaction 7 in Figure 2).102 This intramolecular transfer mechanism of NO+ was also suggested by Herold and Rock,103 where they showed that the addition of NO to metHb in the presence of GSH results in much more SNO-Hb being formed than S-nitrosoglutathione, and a similar mechanism was proposed by Rifkind and co-workers.79

Reductive nitrosylation refers to the reaction where Fe3+–NO reacts with OH or water to make nitrite leaving a reduced iron, Fe2+.101 A second NO then binds to the ferrous haem (nitrosylation). In 2003, Fernandez and Ford104 discovered that nitrite catalyses reductive nitrosylation. They suggested that the nitrite reacts with Fe3+–NO to make N2O3 leaving a reduced haem (Reaction 9 in Figure 2). Since N2O3 is a strong nitrosating agent, they also suggested that this reaction may be responsible for SNO-Hb formation. Thus, nitrite would participate both in making Fe3+–NO and in the formation of N2O3 with subsequent nitrosation. An alternative pathway was suggested by Basu et al.76 that involves the reaction of HbFe3+–NO2 with NO to form N2O3 (Reaction 5 in Figure 2). The production of N2O3 represents an interesting possibility with respect to export of NO activity as, not only can it be accomplished through the intermediacy of S-nitrosothiols (Reaction 6 in Figure 2), but it may also be accomplished through diffusion of N2O3 itself out of the RBC which could then dissociate to form NO and NO2, as had been suggested earlier.98,105

Despite measured progress in examining how NO or NO activity could escape from the RBC, this question continues to challenge the hypothesis that the reduction of nitrite by RBC–Hb is responsible for nitrite's vasodilatory properties. Recent reports using mass spectrometry failed to show substantial production of free NO or export from the RBC, or even export of N2O3.106,107 Thus, studies to establish how NO activity is exported from the RBC are ongoing.

Although the above investigations have sought to understand NO export mechanisms, it is possible that the vasoactive agent is not NO. RBCs have been shown to induce vasodilation upon deoxygenation through the release of ATP which then stimulates purinergic receptors leading to increased NO production from eNOS.108110 Recently, Rifkind and co-workers63 presented data suggesting that nitrite enhances this activity. They suggest that, like deoxyHb, Hb that has reacted with nitrite displaces membrane-bound glycolytic enzymes which then increase intracellular ATP with subsequent release under hypoxic conditions (Pathway 10 in Figure 2).63 This is similar to a mechanism proposed by English and co-workers64 who also recently reported nitrite-dependent ATP export from red cells. Rifkind and co-workers63 showed that nitrite-treated red cells decrease blood pressure in a rat model and that this effect was eliminated when an ATP scavenger, apyrase, was added. However, given that red cells contain ∼1 mM ATP and changes in extracellular ATP are on the order of several nanomolars, even a little haemolysis of RBCs could account for apparent ATP export. A small amount (<0.001%) of haemolysis that results in 0.2 μM extracellular Hb would simultaneously release ∼10 nM ATP, which is greater than the amount detected by Rifkind and co-workers due to nitrite exposure. In addition, the involvement of ATP in nitrite-mediated vasodilation is not supported by single-vessel bioassays where nitrite-mediated vasodilation occurs even in the presence of NO synthase inhibitors.31 Also, nitrite infusions into the human forearm caused increased blood flow even in the presence of an NO synthase inhibitor, which is not consistent with the ATP mechanism.15

6. Concluding remarks

Nitrite continues to gain increased attention regarding its role in physiology, pathology, and therapeutics. It has become clear that the bioactivation of nitrite and its consequent action on vascular tone may occur through multiple mechanisms. Any viable pathway must be active at physiologically relevant pH and oxygen saturations and be sensitive to oxygen tension, being more efficient at lower oxygen tensions. Although challenges exist and many details remain to be established, particularly with regard to the role of nitrite as a physiological mediator of blood flow, red cell Hb is a major candidate as a transducer of nitrite-dependent bioactivity in the vasculature.

Conflict of interest: D.B.K.-S. and R.P.P. are listed as co-inventors on a patent application entitled ‘Use of nitrite salts for the treatment of cardiovascular conditions'.


This work was supported by NIH grants HL58091 (D.B.K.-S.), GM55792 and HL090503 (N.H.), and HL092624 (R.P.P.)


  • This article is part of the Review Focus on: Inorganic Nitrite and Nitrate in Cardiovascular Health and Disease


View Abstract