© 2005 American Thoracic Society DOI: 10.1165/rcmb.F292
Hemoglobin-Mediated, Hypoxia-Induced Vasodilation via Nitric OxideMechanism(s) and Physiologic versus Pathophysiologic RelevanceCenter for Free Radical Biology, Departments of Anesthesiology, Physiology & Biophysics, and Environmental Health Sciences, University of Alabama Birmingham, Birmingham, Alabama Correspondence and requests for reprints should be addressed to Jack R. Lancaster, Jr., Professor, Departments of Anesthesiology, and Physiology & Biophysics, Center for Free Radical Biology, University of Alabama at Birmingham, 208 Biomedical Research Building II, 901 19th Street S., Birmingham, AL 35294-2172. E-mail: doctorno{at}uab.edu
Structurally, dioxygen (O2) and nitric oxide (NO, nitrogen monoxide) are almost identical, differing by only one atomic number. They also share a few chemical properties, including their physical state at standard temperature and pressure (both gases), existence as dissolved nonelectrolytes in aqueous solution, paramagnetic character, and high affinity for binding to transition metals. This last property dictates that both O2 and NO will bind to the heme iron of hemoglobin, properties which have been known for nearly 150 yr (1). Physiologically, a major function of NO is in regulating vascular tone, thereby directly influencing blood perfusion and consequently tissue O2 delivery. It is tempting to speculate that hemoglobin might be capable of delivering NO as well as O2 to hypoxic tissue. So hemoglobin in red blood cells (RBC) may provide not only the O2 required for tissue respiration but also NO (or related species, NOx), a potent vasodilator.
Intra RBC S-nitroso Hb (SNOHb), and more recently, nitrite have been proposed as principal sources of RBC-derived vasoactive NOx. Unfortunately, at this time no credible proposal exists for the efficient transfer of NOx equivalents, whether via SNOHb or nitrite, from the RBC to the vessel wall. Any such proposal must contend with the extremely rapid and irreversible intraerythrocytic consumption of NO by reaction with dioxygen bound to the heme of oxyhemoglobin (2, 3) or with the unoccupied heme of deoxyhemoglobin. Oxyheme is the major heme species in all but the most severely hypoxic vascular beds. The binding of NO to deoxyheme is also extremely rapid, but less is known about the further reactions of this species. Both reactions proceed at a bimolecular reaction rate of
In 1996, Stamler and colleagues suggested an allosteric positive cooperativity in uptake and liberation of both O2 and NO by hemoglobin, involving heme (O2, NO) and thiol (NO) (8–10). They proposed that RBC deoxyhemoglobin takes up both O2 and NO in the lung and then, upon transit to a site of low pO2, liberates both (Table 1). For such a mechanism to be physiologically meaningful, the production and delivery of NO must occur within the constraints imposed by the system physiology. During the past 3 yr, many laboratories have identified problems with the physiologic, biochemical, biophysical, and chemical underpinnings of the SNOHb hypothesis.
The first problem with the SNOHb hypothesis relates to the mechanism of formation and the production rate of SNOHb. SNOHb was thought to be formed from the reaction of NO with a reduced cysteine thiol on intraerythrocytic hemoglobin (8, 9). Formation of nitrosothiol from NO and thiol requires one-electron oxidation (RS– + NO  RSNO + e–), for which there is no obvious electron acceptor. The prospect of ferriheme (Fe3+) functioning as an electron acceptor (11) is highly unlikely because the rate of this reaction is many orders of magnitude too slow to be physiologically relevant (12). Experiments involving the addition of a small volume of concentrated anaerobic NO have been used to claim that the yield of SNOHb from the infused NO is significant compared with competing reactions such as irreversible NO scavenging (13). The observed SNOHb may not have been produced from NO itself but from other NOx species. Injection of concentrated NO solutions into aerobic solution results in a complex (and essentially uncontrollable) mixture of reactive species, at least some of which (e.g., nitrous anhydride, N2O3) are well-known agents for rapid nitrosation of thiols to produce nitrosothiols directly (14–17). Multiple investigators have challenged the SNOHb hypothesis by reporting undetectable (subnanomolar) levels of SNOHb in the circulation (9, 12, 18–21). All methods for detection of nitrosothiols involve the liberation of NO. One method uses photolysis with detection of the liberated gas-phase NO by chemiluminescence (22). Another major method involves chemiluminescence measurement of total nitrogen oxide species without and then with Hg2+, which breaks down nitrosothiols (23, 24). The Hg2+ method has been shown to reproduce the photolysis method for circulating rodent SNOHb levels, but not for humans (25). In addition, results with human erythrocytes using the same method (Hg2+ method with DAF detection) in two laboratories differ by a factor of 40 (18, 25). There is no doubt these inconsistencies originate methodologically. A valid concern (26) is that the stability of nitrosothiols can vary widely, and that some methods and sample handling could result in Hg2+-independent breakdown. These nitrosothiols would thus go undetected using the Hg2+ method, whereas photolytic release of NO (9, 18) could detect these species. The Hg2+ method has been validated with SNOHb. However, the SNOHb used in the validation assay is prepared by chemical modification with transnitrosating or nitrosating agents that produce a multiplicity of nitrogen oxide–containing species in addition to the cysß93 nitrosothiol (21, 27). These species may react differently to the photolysis and Hg2+ methods (26). In particular, for example, sensitivity of a nitrosothiol to the acid/sulfanilamide treatment used to remove nitrite would mask the existence of such a species (although in general nitrosothiols are quite stable in acid conditions). On the other side, photolysis of samples containing thiols and nitrate (which is by far the most abundant nitrogen oxide species biologically) with high intensity light results in NO formation and thus erroneously high "RSNO" levels (28). In addition, recent work indicates that the origin of the differential presence of circulating SNOHb in rodent compared with human hemoglobin (rodent levels are much higher [70–800 nM] with no [below detection limit, 1 nM] SNOHb in human blood) may be due to differential thiol reactivity of these species (21). Artifactually high SNOHb levels (and also Hb radical signals) (29) can occur from incomplete alkylation or nitrite contamination of reagents used in the assay. In summary, the dramatic differences in measurements of hemoglobin nitrosothiol in different hands are difficult to reconcile; what is needed is for laboratories to work together to identify the origins of the discrepancy, e.g., perform side-by-side measurements on the same sample using the different techniques. The mechanism of NOx transfer from SNOHb to the target (the vascular smooth muscle cell) is unclear. More recent studies (12, 30, 31) have prompted a reevaluation of the original proposed mechanism that the SNOHb nitrosonium moiety is transferred to glutathione via transnitrosation (8, 9). Subsequent studies suggested that the erythrocyte AE1 membrane protein is a major transnitrosation target (10). However, these experiments suffer from methodologic problems, because many involved anaerobic exposure to highly supraphysiologic NO levels (64 µM, NO:heme ratio 1:250) followed by O2 exposure. Such treatment would generate a heterogeneous collection of nitroso species (including nitrite) that could be responsible for nitrosation. Transnitrosation provides a mechanism for the transfer of the nitrosonium equivalent to the intracellular surface of the RBC but does not address how the nitrosonium equivalent is transferred to the vessel wall. Attention should be focused on the precapillary arterioles of 10–60 µm diameter because this is the location of the major drop in intravascular pO2 (32). Blood flow within an unbranched vessel of this size is not uniform. A steep radial velocity distribution exists such that flow is zero at the vessel wall and greatest at the center of the vessel. This bullet-shaped velocity distribution concentrates the RBC in the center of the vessel and creates a cell-free plasma zone at the vessel wall, thus hampering direct transnitrosation from the RBC to the wall. This cell-free layer disappears at the level of the capillaries, where the vessel diameter is similar to the erythrocyte diameter. Here, this might potentially allow direct transfer of the nitroso group from AE1 to the capillary wall (25). It is unclear how such a mechanism would cause hypoxia-induced vasodilation because capillaries lack smooth muscle to respond to NO. Evidence supporting a central tenet of the SNOHb hypothesis, that hypoxia causes Hb- or erythrocyte-induced vasodilation via induction of the R to T allosteric transition with concomitant NO group liberation, has not been supported by EPR studies of NOHb (29). Perhaps the most damaging studies for the SNOHb hypothesis are experiments (see below) suggesting that nitrite, not SNOHb, is the major nitrogen oxide species responsible for hemoglobin-induced hypoxic vasodilation. Ongoing SNOHb work will need to address important issues relating to (1) production—whether SNOHb is a significant source of NOx under hypoxic conditions, and its mechanism of production; and (2) transfer—escape of SNOHb derived NOx from the RBC and transfer to the vessel wall. What is decidedly needed is for different groups to collaborate and perform measurements together comparing different measurements side by side.
A series of measurements by Gladwin and coworkers of forearm blood flow in human volunteers inhaling NO revealed that of the various plasma nitrogen oxide–related species, only nitrite exhibited an arterial–venous (AV) gradient (i.e., higher arterial levels than venous), indicating metabolism during circulatory transit (33). Exercise did not affect the nitrite AV gradient. But, after taking into account an exercise-induced increase in blood flow rate to the arm, the calculated rate of nitrite consumption was found to increase during transit through the working arm. The correlation of increased rate of nitrite depletion with exercise-induced hypoxia suggested that nitrite, and not other circulating nitrogen oxide–containing forms, is somehow converted into NO in the relatively hypoxic forearm. The released NO induces local dilation and increases regional blood flow. Recent studies have provided support for this mechanism.
In 1953 Furchgott and Bhadrakom reported that nitrite causes relaxation of aortic strips (in the absence of hemoglobin); however, the effect is relatively weak (34). More recently, Vleeming and colleagues demonstrated decreased blood pressure in rats from nitrite infusion, although this required relatively high plasma levels ( Nagababu and coworkers (42) and Cosby and colleagues (43) presented data suggesting the deoxyhemoglobin-dependent vasodilator action of nitrite. Nagababu and coworkers (42) showed NO formation when nitrite is added to solutions of deoxyhemoglobin and subjected to chemiluminescence measurement of purged argon through the solution. This is an important observation, because it is the first indication that free NO can be formed from this reaction. The origin of the NO, however, is not clear, and it could arise from several possibilities, only one of which is reduction of nitrite. As reported by Doyle and colleagues (36) and Stepuro and coworkers (37), a substantial fraction of the deoxyhemoglobin is converted to methemoglobin, and quantitation of nitrosylferroheme by EPR spectrometry revealed that more NO is detected by purging and chemiluminescence than nitrosylheme. Nagababu and associates suggested weak complexation of NO to ferriheme and displacement of this NO by argon purging. Such a mechanism, however, is unlikely to explain the origin of the purged NO because the same phenomenon would predict that the NO would quickly (within 0.65–1.5 s, the rate of NO dissociation from ferriheme) (44) be removed by binding to deoxyheme. This is a very rapid reaction and is thus equivalent in effect to purging. Cosby and coworkers (43) found that infusion of nitrite increased forearm blood flow before and during exercise, and this effect occurred with or without inhibition of NO synthase. As with all studies involving the administration of nitrogen oxide, a crucial question is whether the observed effect occurs under physiologically relevant conditions, and in this case, the important parameter is plasma nitrite under normal conditions, which is in the range 0.15–1 µM. Although most of the results presented by Cosby and colleagues were performed with unphysiologically high nitrite (200 µM), significant forearm blood flow increase was observed when the nitrite stock concentration injected in the brachial artery was lowered so that the mean venous nitrite concentration rose from 176 nM to 2.6 µM. This may be reasonably close to the physiologic range; however, the concentration at the site of administration, and in particular the arteriolar side of the capillaries (which will presumably be the site of most influence on forearm blood flow), may well be higher than this as the nitrite is diluted as it moves through the arm and into the circulation; the concentration of NO in the infusate was 6.7 µM (M. T. Gladwin, personal communication). It is puzzling that Lauer and coworkers (45) did not observe significant increase in forearm blood flow under similar conditions with administered nitrite concentrations up to 130 µM (venous concentration). This discrepancy needs to be explained, but one possibility is that the study by Cosby and colleagues performed measurements after 5 min infusion, whereas for Lauer and associates the time was 1 min. Nagababu and coworkers (42) and Cosby and colleagues (43) detected NO in purged samples of nitrite-treated intact erythrocytes, and the extent of NO formation is inversely related to the extent of oxygenation. Reaction of nitrite with oxyhemoglobin has been studied extensively, and involves a complex and autocatalytic sequence of reactions involving nitrogen dioxide (•NO2) and higher oxidation states of heme but apparently does not involve formation of NO (46). With 200 µM infused nitrite, substantial amounts of both nitrosylferroheme and also S-nitrosohemoglobin are formed, and venous levels are higher than arterial (consistent with a reaction involving deoxyhemoglobin). The physiologic significance of these results is unclear due to the high nitrite concentration used. It is unlikely that either species is functionally able to act as a source for NO, the heme-NO because of its stability and SNOHb because of the reasons discussed above. Cosby and coworkers used a rat aortic ring bioassay system to determine the pO2 at which nitrite induces relaxation in the presence of erythrocytes (43). For 2 µM nitrite, relaxation started when the pO2 was decreased to 20–30 mm Hg. This is consistent with the p50 for O2 saturation of 9 mm under these conditions. This shows that appreciable deoxyheme must be present for relaxation to occur. By extension, nitrite-induced vasodilation occurs only in relatively hypoxic regions of the peripheral vasculature. Involvement of deoxyheme is illustrated by substantial vasodilation observed with low nitrite (100 nM) plus hemoglobin using a pO2 of 15 mm Hg in the presence of the allosteric effector inositol hexaphosphate (IHP). This allosteric effector will shift the p50 to higher levels (45 mm Hg), and thus most of the hemoglobin will be in the low-affinity T deoxygenated state. This nicely illustrates the basic requirement for deoxyheme presence.
A crucial question in all these studies is: what is the oxidant that produces metheme? The simplest possibility is that it is nitrite, which upon reduction will generate NO (the "nitrite reductase" mechanism). However, generation of NO within the interior of erythrocytes cannot explain the vasodilatory action of nitrite, because the NO will immediately (< 1 µs) react with either oxy- or deoxyhemoglobin and significant amounts of NO will never escape the red blood cell. Using reported biochemical parameters, we have simulated the flux of nitrite reduction–derived NO escape from the RBC. Details of the simulations will be reported elsewhere. The simulated NO escape flux is on the order of 10–3–10–2 NO molecules µm–2 · s–1, which agrees well with the 47 pM/s rate reported by Cosby and coworkers (43). The current order of magnitude estimate of the average flux of NO derived from the endothelium into the capillary lumen is 103–104 NO molecules µm–2 · s–1 (47). The amount of RBC-produced NO that escapes the RBC (
Nitrite reduction may not be the predominant reaction causing vascular nitrite depletion under hypoxic conditions. Our conception of the nitrite dehydration pathway (or "nitrite anhydrase" mechanism), an alternative pathway for nitrite depletion, begins with a puzzling result originally published by Keilin and Hartree in 1937 (48). It was shown that addition of NO to methemoglobin (where the heme is in the oxidized, Fe3+ state) results in formation of the ferroheme-NO complex (Fe2+NO). This process is called reductive nitrosylation and has also been observed in intact erythrocytes (15). Ford and Ferndandez have recently shown that this process is catalyzed by nitrite, and the mechanism involves oxidation of the NO on the ferriheme to the oxidation state of nitrite with reduction of the heme to ferrous (Fe2+), which binds a second NO. The mechanism of catalysis by nitrite appears to involve reaction of the ferriheme/NO complex with nitrite to generate ferroheme and the important NOx nitrous anhydride (N2O3). In the context of the nitrite/deoxyhemoglobin reaction, it is important to point out that the same iron/nitrogen oxide complex can be formed from binding of nitrite to ferroheme, with a dehydration of the nitrite to the formal oxidation state of nitrosonium (NO+). Thus, overall N2O3 is formed from nitrite by the same overall reaction as in acid medium (2 NO2– + 2H+
N2O3 is a potent nitrosating agent. Under conditions of nitrite accumulation, this mechanism could contribute substantially to nitrosothiol (SNOHb) formation in the red blood cell. N2O3 is also a NO producing species. Although the equilibrium constant is small, N2O3 rapidly homolyzes to •NO + •NO2 (49). Intracellular homolysis of N2O3 in the RBC would provide a supply of NO that would react with deoxyHb to form iron nitrosyl (Hb-NO), as observed. MetHb will be formed by both reaction of NO with oxyheme and also reaction of •NO2 with deoxyheme. Diffusion of N2O3 from the RBC accompanied by extracellular homolysis causing •NO production is a potential means of NO transfer to the endothelium. In this way, NO is produced outside the erythrocyte and so the mechanisms of decreased consumption described above act to conserve the NO as opposed to NO production inside the cell. From the standpoint of nitrite-induced vasodilation, formation from nitrosative reactions of N2O3 of a collection of different nitroso species in the heterogeneous environment of the blood vessel could accomplish movement of nitrosonium equivalents to the vessel wall. These species could function as nitrosonium or NO donors (50). On a technical note, Butler and Ridd (49) point out that purging a solution will gasify any N2O3 present and will immediately cause N2O3 to homolyze to NO + •NO2, thus shifting the equilibrium to homolysis and overestimating the solution NO detected when deoxygenated erythrocytes are exposed to nitrite (43). This also is a complication with a previous study by Demoncheaux and colleagues (51), who measured unrealistically high NO concentrations in solutions of nitrite at pH 7.4 (
Hypoxia-induced vasodilatory responses may be due to lowered oxygen tension alone, independent of hemoglobin. NO survives longer with decreased pO2 due to decreased cellular consumption. This leads to increased NO concentration in the tissue (52). Crawford and coworkers reported an increased sensitivity of the vessel to nitrovasodilators in general under hypoxic conditions (53). However, there is disagreement regarding the role of pO2 in GSNO-stimulated vasodilation. Although Crawford and colleagues (53) showed potentiated GSNO-induced vasodilation upon hypoxia, Stamler and coworkers and McMahon and colleagues have not seen this effect (9, 18). Although NO is important in hypoxia-induced vasodilation, other mediators may be involved (54). From subnanomolar measurements of erythrocyte SNOHb, some investigators have concluded that SNOHb is unlikely to be the major intracorpuscular source of transvascular [NO+]. If, as originally proposed (10), SNOHb were not produced, but only consumed under hypoxic conditions, then the concentration of the donor species must necessarily decrease during blood transit through hypoxic vessels as [NO+] equivalents are transferred to the vessel wall. The vessel wall flux of SNOHb derived [NO+] equivalents would be limited by the arterial concentration of SNOHb. If, however, SNOHb is both produced and consumed (as an [NO+] donor) under hypoxic conditions, then its concentration could remain subnanomolar, or very well increase (43), during blood transit through the hypoxic bed, while supporting a physiologically significant flux of NOx to the vessel wall. The point is that unless one knows both the production and consumption rates of the NOx donor, its concentration away from the point of NO delivery is not particularly informative. Instead, more meaningful conclusions can be drawn from measurements of (or perturbations of) species reaction rates (including diffusional fluxes). Whether SNOHb is an important species in the transduction of NOx equivalents to the vessel wall, or simply an innocent bystander, remains to be determined. Some have expressed concern over the degree of hypoxia apparently necessary for hemoglobin-mediated hypoxia-induced vasodilation to occur. The in vitro requirement for the effect is 7–15 mm Hg pO2 for SNOHb (18) and slightly higher for nitrite (43). In the circulation, the pO2 falls from a value of 100 mm Hg in arterial blood to 40 mm Hg in mixed venous blood. In highly metabolically active tissues, however, the pO2 may well fall to levels where the hemoglobin-mediated vasodilation is active. On paper, other possibilities also exist for the formation of NOx species able to escape the concentrated hemoglobin "trap" inside the red cell; so at this stage it would seem imprudent to classify nitrite metabolism by the RBC as nitrite reduction. Presently, is seems premature to accept the notion that RBC-mediated hypoxic vasodilation (whether by nitrite, SNOHb, or otherwise) plays an important role in the normal maintenance of basal vascular tone. It could quite conceivably be important, however, under conditions of increased NO formation during inflammatory conditions or in tissues rapidly consuming O2.
Conflict of Interest Statement: J.M.R. has no conflicts of interest; J.R.L. has no conflicts of interest. Received in final form January 14, 2005
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Introduction
The SNOHb Hypothesis
107 M–1 · s–1, or assuming RBC hemoglobin heme concentration of 23 mM, a pseudo–first order reaction rate of 
10 µM) (
