Inorganic nitrite has recently been recognized to possess vascular activity that is enhanced in hypoxia. This has been demonstrated in humans in the forearm vascular bed. In animal models nitrite reduces pulmonary vascular resistance, but its effects upon the pulmonary circulation of humans have not yet been demonstrated. This paradigm is of particular interest mechanistically since the pulmonary vasculature is known to behave differently to the systemic. To investigate, 18 healthy volunteers were studied in a hypoxic chamber (inspired oxygen, 12%) or while breathing room air. Each received an infusion of sodium nitrite (1 μmol/min) or 0.9% saline. Three protocols were performed: nitrite/hypoxia (n = 12), saline/hypoxia (n = 6), and nitrite/normoxia (n = 6). Venous blood was sampled for plasma nitrite, forearm blood flow was measured by strain-gauge plethysmography, and pulmonary arterial pressure was measured by transthoracic echocardiography. Plasma nitrite doubled and clearance kinetics were similar whether nitrite was infused in hypoxia or normoxia. During hypoxia, nitrite increased forearm blood flow (+36%, P < 0.001) and reduced three separate indirect indexes of pulmonary arterial pressure by 16%, 12%, and 17% (P < 0.01). Pulmonary, but not systemic, arterial effects persisted 1 h after stopping the infusion, at a time when plasma nitrite had returned to baseline. No effects were observed during normoxia. Therefore, in hypoxic but not normoxic subjects, sodium nitrite causes arterial and pulmonary vasodilatation. In addition, hypoxia-induced pulmonary vasoconstriction was attenuated for a prolonged period and not dependent on a simultaneous elevation of plasma nitrite. This finding is consistent with the direct extravascular metabolism of nitrite to nitric oxide to effect hypoxia-associated bioactivity.
- pulmonary hypertension
- nitric oxide
endothelial nitric oxide (NO) is an integral molecule of vascular homeostasis (27). The subsequent metabolism of NO within the vasculature, to nitrite and then largely to nitrate, has traditionally been considered to be a one-way process. However, nitrite has in recent years become the focus of renewed interest.
Inorganic sodium nitrite was first used as a treatment for angina in the 19th century (20) as it has a stable chemical structure that allowed for cheap preparation and easy storage (8). However, its effects were slow and unpredictable, and so it fell out of favor as faster-acting agents such as organic nitrates became available (28). Furthermore, it was recognized in 1936 that oral nitrite administration to humans could result in circulatory collapse and syncope (34). In 1953, this was shown to be due to the relaxation of vascular smooth muscle (13).
These older studies established that high concentrations of nitrite can dilate arteries under normoxic conditions. However, recent experiments demonstrate that a similar effect can be achieved at a much lower concentration of nitrite if the local environment is altered, for example, by hypoxia or ischemia (26). It has been suggested that this potentiation is mediated by the conversion of nitrite to NO within the erythrocyte by deoxyhemoglobin (9).
Modern human studies of nitrite have been broadly limited to the examination of its local vasodilator effect in the forearm following a brachial arterial infusion. In healthy subjects, doses producing a 10-fold increase in plasma nitrite increased forearm blood flow (FBF) (9), as would be expected from the early literature. More recently we have demonstrated that high-dose nitrite has an enhanced effect upon FBF when administered under hypoxic conditions (24).
The implications that can be made from these pharmacological studies are limited, because it is difficult to separate the acute effects of a single dose of nitrite from other effects related to study design. Specifically, the use of cumulative dosing regimes, the fact that hypoxia measurements were recorded on a nonequilibrated baseline (12), and the limitation of observations to a single vascular bed all mean that potentially important mechanistic insights could have been missed. Of particular interest would be the effect of nitrite upon the pulmonary vasculature, since this is known frequently to behave differently from the systemic circulation.
Nebulized nitrite administered to newborn lambs reduced hypoxia-induced pulmonary hypertension by 60% (15). The magnitude of this effect correlated with the reduction of pH and the degree of hemoglobin-oxygen desaturation present. The effect was prolonged compared with the administration of inhaled NO gas. The pulmonary vasodilator action of nitrite has been recorded in rats (7) where it was shown to be partially blocked by the inhibition of xanthine oxidase (XO). Low-dose nitrite has also been demonstrated to reduce pulmonary vascular resistance in a canine model of acute pulmonary embolism (11). However, the effects of nitrite upon the pulmonary circulation in humans have not been examined.
Therefore, this study was designed to test the physiological reaction of both systemic and pulmonary circulations to a modest elevation in plasma nitrite concentration in both hypoxia and normoxia. We investigated healthy volunteers during stable hypoxia in an environmental chamber and during normoxic conditions breathing room air. We gave an intravenous infusion of a low dose of sodium nitrite, one traditionally thought to be physiologically inert, or a control infusion of normal saline. We studied both systemic and pulmonary vascular responses.
Eighteen healthy male volunteers with a mean age of 24.0 ± 1.0 yr were recruited from staff and students at the Universities of Cardiff and Glamorgan. The mean weight was 78.6 ± 2.2 kg, and mean body mass index was 24.3 ± 0.6. All subjects had a structurally normal heart, confirmed at enrollment by transthoracic echocardiography (TTE). The mean left ventricular ejection fraction (measured by the biplane method of disks) was 64.8 ± 1.1%.
Exclusion criteria were hypertension, hypercholesterolemia, diabetes mellitus, a family history of premature coronary artery disease, being a current smoker, or taking regular prescription drugs. Women have cyclical variability in NO metabolites and pulmonary arterial vasoreactivity (18, 25); they were not included as these factors could have interfered with interpretation of the results. The protocol was approved by the local Research Ethics Committee, and all subjects gave informed consent.
Subjects were asked to refrain from alcohol, caffeine, and foods with high-nitrite content (such as red meat or green vegetables) for 12 h before each visit. Each had a light breakfast before and then fasted for the duration of the study.
Hypoxic studies were performed in an environmental chamber measuring 6 × 5 × 4 m (Design Environmental, Ebbw Vale, UK) maintained at 21 ± 0.3°C. An inert, nitrogen-rich, hypoxic gas mixture was delivered to the room at a high flow rate to create a normobaric hypoxic environment, while continuous gas exchange ensured that CO2 levels did not alter. The fraction of inspired oxygen (FiO2) was reduced down to 12 ± 0.1% for each hypoxic experiment. Normoxic studies were performed in a temperature-controlled room.
Three physiological studies were conducted: one as the index intervention (hypoxia/nitrite) and two as control studies (hypoxia/saline; normoxia/nitrite). Figure 1 is a schematic of the protocols. Recipients, but not the investigator, were blinded to the content of each infusion given: sodium nitrite (Martindale Pharmaceuticals, Brentwood, UK) or 0.9% saline. However, the subsequent analysis of the data was performed blindly.
In 12 subjects, baseline investigations were completed during normoxia, and the oxygen saturation in the environmental chamber was then reduced from 21% to 12% over a 1-h period. It was maintained at 12% for 2 h more, so that subjects experienced a total of 3 h of stabilization before the preinfusion measurements were obtained. An intravenous infusion of sodium nitrite was then given, single-blind, into the dominant arm at 1 ml/min (1 μmol/min) for 30 min before peak infusion measurements were taken. Subjects remained in the hypoxic chamber for 1 h further after the infusion was completed, before the final measurements were made (1 h after infusion).
Another six subjects were studied in the hypoxic chamber after the same period of equilibration but with the infusion of 0.9% saline (1 ml/min for 30 min) instead of sodium nitrite, again given single-blind. The physiological and biochemical measurements were identical to the previous protocol.
This control study was performed in 6 of the 12 subjects who had been investigated according to the first protocol (hypoxia/nitrite), but this time while they breathed atmospheric air (FiO2, 21%). There was no equilibration period, only a period of rest before the baseline recordings. Thereafter, the protocol for physiological/biochemical measurements was identical to the other experiments. The same infusion regime of sodium nitrite was used as in protocol 1.
Biochemical and Physiological Measurements
At every stage in all protocols, the same measurements were made. FBF was measured in the nondominant arm. Venous blood samples were also taken from this same arm (the opposite arm from that into which nitrite or saline was infused). TTE was performed.
Assessing systemic arterial flow: strain-gauge plethysmography.
Relative changes in FBF were measured using mercury-filled silastic strain-gauge plethysmography (EC-6, Hokanson; Bellevue, Washington) as described previously (35). For specific details, see supplemental data (note: all supplemental data may be found with the online version of this article).
Assessing pulmonary arterial pressure: TTE.
A commercially available echocardiography system (GE Vingmed System V, General Electric, Horten, Norway) was used to measure three echo surrogate indexes of pulmonary arterial pressure. The systolic pressure gradient between the right ventricle and the right atrium (in mmHg) was calculated from the velocity of tricuspid regurgitation (TR) using the modified Bernoulli equation (37). Right atrial pressure, estimated from the change in diameter of the inferior vena cava during respiration, was added to give the pulmonary arterial systolic pressure (PASP). Pulmonary acceleration time (PAT; in ms) was measured from the onset to the peak velocity of pulmonary arterial forward flow (29). The isovolumic relaxation time (IVRT) of the right ventricle was measured from pulsed tissue Doppler traces of lateral tricuspid annular velocity recorded in an apical four chamber view (5). Additionally, left ventricular velocity time integral and diastolic function were recorded and cardiac output was calculated. For specific details, see supplemental data.
The investigator performing the TTE was present inside the hypoxic chamber for the duration of each scan but not for the entire study protocol. He exhibited no symptoms of exposure to acute hypoxia.
Assessing blood NO metabolites.
Ozone-based chemiluminescence (NO analyzer, NOA280i, Sievers) was used to detect NO liberated from plasma nitrite by tri-iodide reagent cleavage, as we have described previously (32). For specific details, see supplementary data.
Statistical tests were performed using GraphPad Prism version 4 (GraphPad software, La Jolla, CA). All variables measured were normally distributed, and therefore only parametric comparisons were made. Results are expressed as means ± SE. A repeated-measures ANOVA test together with Newman-Keuls multiple comparison posttest was used to compare differences. The relationships between repeated measurements of variables that were continuously distributed were explored and are reported using the Pearson correlation coefficient. Between-group comparison of values taken at the same time point in the hypoxia/nitrite and hypoxia/saline protocols was performed by unpaired t-test. A value of P < 0.05 was chosen for statistical significance.
Intraobserver reproducibility was assessed by repeating measurements in 30 random samples for each variable measured, 1 mo after the first analysis and blinded to the original results. This is reported as coefficients of variation (CVs; in %), calculated using the formula: CV + (SD/arithmetic mean of measurements) × 100, where SD is the standard deviation of residuals (measurement 1 − measurement 2). Systematic bias between measurements was assessed using Bland-Altman analyses.
The Effects of Hypoxia Alone
When the oxygen content of the environmental chamber was reduced, arterial oxygen saturation fell from 98 ± 0.2% at baseline to 83 ± 1.7% preinfusion (P < 0.001). Arterial oxygen saturation similarly fell during the stabilization period of both hypoxia protocols and did not change significantly thereafter.
Heart rate increased from 63 ± 3 beats/min at baseline to 77 ± 2 beats/min preinfusion (P < 0.001), and thereafter it also did not alter significantly during each protocol. Stroke volume did not change, and so cardiac output increased (Table 1). Blood pressure did not alter, indicating that hypoxia produced systemic arterial vasodilation.
During hypoxia, the velocity of TR increased, PAT shortened, and the right ventricle developed a measurable IVRT (Fig. 2, and Table 2). All these changes are indirect markers of an increase in pulmonary arterial pressure, secondary to hypoxic pulmonary arterial vasoconstriction.
The Pharmacokinetics of Nitrite
Preinfusion plasma nitrite levels were similar in the two hypoxia groups, but both concentrations were greater than the preinfusion plasma nitrite measured in the normoxia protocol (hypoxia/nitrite, 191.0 ± 12.3 nmol/l; hypoxia/saline, 216.0 ± 28.6 nmol/l; and normoxia/nitrite, 125.5 ± 14.8 nmol/l; P < 0.05). However, no difference in plasma nitrite was present between the baseline (FiO2, 21%) and the preinfusion samples in all 18 subjects who underwent the 3-h hypoxia-equilibration period [baseline = 189.0 ± 11.3 nmol/l, and preinfusion = 199.4 ± 12.4 nmol/l; not significant (NS)]. A difference was also present between the baseline plasma nitrite samples taken in room air in the hypoxia/nitrite study and the normoxia/nitrite study in the six subjects who underwent both of these protocols (177.0 ± 12.4 and 125.5 ± 14.8 nmol/l; P < 0.05). These combined results demonstrate that the difference in plasma nitrite observed between the two hypoxia groups and the normoxia group cannot be attributed to hypoxia alone. However, it is in keeping with the expected normal range and day-to-day variability in plasma nitrite concentration (Fig. 3, top).
Importantly, there was no difference in the infusion-associated increase in plasma nitrite concentration (peak − preinfusion concentration) between the hypoxia/nitrite and normoxia/nitrite groups (205.5 ± 20.9 vs. 206.3 ± 36.7 nmol/l; NS).
The half-lives for decay of plasma nitrite concentration after stopping the infusion were also similar (in hypoxia, 21.7 ± 3.0 min; and in normoxia, 21.3 ± 3.4 min). Plasma nitrite had returned to preinfusion concentrations by 1 h after stopping the infusion in both nitrite groups.
Plasma nitrite did not vary significantly throughout the whole hypoxia/saline protocol.
The Systemic Arterial Response to Nitrite
FBF was similar before the infusions in all groups (hypoxia/nitrite, 2.2 ± 0.1; normoxia/nitrite, 2.1 ± 0.2; and hypoxia/saline, 2.4 ± 0.1 ml·100 ml−1·min−1; NS). It increased when nitrite was infused during hypoxia (peak, 3.0 ± 0.2 ml·100 ml−1·min−1, P < 0.01), but it did not change when nitrite was infused during normoxia or when saline was infused during hypoxia (Fig. 3, middle). By 1 h after the end of the infusion of nitrite during hypoxia, FBF had fallen significantly (to 2.3 ± 0.3 ml·100 ml−1·min−1, P < 0.01 vs. peak) so that it was no longer different from baseline measurements.
There was a significant difference in the increase in FBF at peak, correcting for the corresponding preinfusion values (peak − preinfusion FBF) between the hypoxia/nitrite group and the hypoxia/saline group (P < 0.01). This difference was not present when measured in the same groups at 1 h after the infusion.
The increase in FBF caused by nitrite during hypoxia correlated with the peak plasma nitrite concentration (Pearson r = 0.31, P = 0.002), whereas FBF during normoxia was unrelated to plasma nitrite (Pearson r = 0.03, P = 0.86) (Fig. 4).
The Pulmonary Vascular Response to Nitrite
Changes in the echocardiographic measurements are shown in Table 2. Changes in PASP estimated from the velocity of TR, which could be measured in 15 out of a total of 24 studies, are shown in Fig. 3 (bottom). The diameter of the inferior vena cava and its respiratory variation did not change significantly throughout all the studies, and so the estimated right atrial pressure was constant at 5 mmHg.
PASP almost doubled in response to hypoxia (to 37.9 ± 1.7 mmHg preinfusion). The infusion of nitrite then reduced the PASP by 16% (P < 0.01). Although plasma nitrite had returned to normal by 1 h after the infusion was stopped, PASP remained significantly reduced compared with preinfusion measurements (Table 2, and Fig. 3, bottom). Similar changes were observed in the two other indirect indexes of pulmonary hypertension that were recorded: the PAT (measured in all subjects) was shortened by hypoxia and then prolonged at peak infusion (by 12%) and 1 h later, and the duration of the IVRT in the right ventricle (measured in all subjects except 1) increased during hypoxia and then was attenuated by nitrite both at the end of the infusion (by 17%) and 1 h later (Table 2).
A similar doubling of PASP was observed in response to hypoxia in the hypoxia/saline study (i.e., hypoxic pulmonary vasoconstriction), but no changes were observed during or after the saline infusion. PASP was unchanged throughout the normoxia/nitrite study.
There were no significant differences between the hypoxia/nitrite and hypoxia/saline groups in all three indexes recorded at either baseline or preinfusion. Between-group analysis confirmed an effect of nitrite at peak (P < 0.01 for all 3 indexes). The persistent benefit of nitrite at 1 h was shown by changes in IVRT (P < 0.01); the effect upon PAT tended to significance (P = 0.07), whereas the changes in PASP were not analyzed between groups because of a low prevalence of a TR jet in the hypoxia/saline group (see marked comparisons and legend in Table 2 for details).
The Response of Cardiac Output to Nitrite
Cardiac output increased by 25% from baseline to preinfusion in response to hypoxia, because of an increase in heart rate (Table 1). It then decreased by an average of 8% when nitrite was infused during hypoxia. This was the result of a small reduction in both heart rate and stroke volume. Of the two variables, stroke volume tended toward significance (P = 0.07), consistent with the reduced preload secondary to nitrite-induced venodilation.
The increase in cardiac output related to hypoxia was unaltered by the infusion of saline. When nitrite was infused during normoxia, no changes were observed in heart rate, stroke volume, or cardiac output (Table 1).
There were no significant changes in the E-to-e′ ratio, a marker of left ventricular diastolic function, during any of the protocols (Table 1).
The Reproducibility of Noninvasive Physiological Measures
The CVs ranged from 1.9% (left ventricular stroke volume) to T4.8% (IVRT period measured at the lateral tricuspid annulus). The feasibility and reproducibility of the echocardiographic and plethysmographic measurements are shown in the supplementary data section.
This first study in humans confirms that in healthy subjects, intravenous sodium nitrite given in low concentrations has a significant and prolonged vasodilator effect on the hypoxic pulmonary circulation. The most important new finding of this study is that this effect is not dependent on a concurrent elevation in plasma nitrite.
The hypoxia protocols were conducted in an environmental chamber that eliminated the difficulties of achieving stable control of hypoxia through a face mask. One hour was required in order for the chamber to stabilize at 12%, and consequently the subjects were exposed to a gradual decline in FiO2 rather than a sudden drop. As established previously (5), the physiological effects of hypoxia are not stabilized until after at least 2 h of continued exposure, and so all measurements were conducted after equilibration for 3 h inside the chamber. The subjects experienced a decreased and stable arterial oxygen saturation (∼83%) before the infusion of nitrite or saline and throughout the remainder of the protocol.
In our study, plasma nitrite did not increase as a function of exposure to hypoxia. Other investigators have previously reported the same finding (3). However, we did observe a wide variation in plasma nitrite between subjects. This could have been reduced by the use of a body weight adjusted standard diet before the day of experiment.
Limitation of Physiological Measures Employed
Ideally, we would have measured pulmonary vascular resistance directly, for example, using a Swan-Ganz thermodilution catheter. However, facilities for this were not available in the environmental chamber. Instead, we elected to use three validated surrogate echocardiographic measurements of pulmonary arterial pressure. Without invasive measurements, it is not possible to be certain that a change in pulmonary arterial pressure equates to a change in pulmonary vascular resistance. For instance, if vascular resistance is constant but cardiac output falls, pressure will also reduce. However, in our study, the nitrite-associated reduction in pulmonary arterial pressure was between 12–17%, which was approximately double the reduction in cardiac output observed. This suggests that a vasodilator effect of nitrite upon pulmonary vascular resistance was present, although invasive studies are needed to confirm this finding.
The noninvasive assessment of pulmonary arterial pressure using echocardiography has some limitations. An estimation of PASP from the velocity of TR remains the most widely accepted and reproducible marker, but it is limited by a prevalence of regurgitation of only 30 to 70% (29). In our study, TR was detected in just over half of subjects. However, the other two markers used, PAT and the right ventricular IVRT, gave similar results with a detection rate approaching 100% and a robust CV. The sensitivity of these two markers to demonstrate a difference in this protocol was reliant on their use in the same subject, at a consistent location within the right ventricle for each measurement to minimize variance.
Venous strain gauge forearm occlusion plethysmography is a well-established technique for measuring systemic arterial blood flow, but the assumption is made that the changes observed in a single vascular bed can be generalized to the whole body. The test is highly affected by changes in temperature, which alter the flow of blood to the skin, and therefore an assiduous control of the experimental environment is needed. Third, a relative change in FBF rather than an absolute value is obtained. Invasive monitoring of arterial systemic pressure would have been useful, but facilities to perform this were not available in the environmental chamber.
The Enhanced Arterial Vasodilator Effect of Nitrite in Hypoxia
The pharmacokinetic profile of plasma nitrite was similar when it was infused during hypoxia or normoxia. This is in contrast to the physiological changes observed over the same time period that demonstrate vasoactivity of nitrite in hypoxia which was not present in normoxia.
As would be expected, the vasodilator activity of nitrite resulted in an increase in FBF (arteriodilation), a decrease in cardiac output (venodilation causing a decrease in stroke volume), and no change in heart rate. Unexpectedly, however, we also observed an increase in peripheral systolic blood pressure. A possible explanation why blood pressure did not fall would be that there was peripheral pressure amplification. In healthy young subjects the peripheral blood pressure is greater than the central blood pressure because of the backward reflections of systolic arterial pulse waves traveling at a relatively slow velocity through a compliant vascular tree (36). This effect is exaggerated by the administration of a vasodilator (6), which further slows the reflected pulse wave. In our study, brachial sphygmomanometry may have recorded augmented peripheral pressure, when central pressure might have fallen or remained constant.
The mechanisms responsible for the enhanced activity of nitrite in hypoxia are many and have previously been strongly debated in the literature. One well-explored theory suggests that nitrite is reduced to NO by deoxyhemoglobin in the circulation, which in turn diffuses into tissue to exhibit its effect (14). Alternatively, recent in vitro work proposes that nitrite reduction at a tissue level is the major source of bioactive NO resulting from its supplementation (21). Several tissue nitrite reductases have been proposed, including myoglobin, XO, endothelial NO synthase, and aldehyde dehydrogenase (1, 7), and these all have enhanced nitrite-reducing capacity during hypoxia. In particular, it would be interesting to test whether an inhibitor of XO is capable of blocking the observed effect. Previous work in rats has demonstrated an allopurinol-sensitive vasodilator mechanism of action of nitrite upon the hypoxic pulmonary vasculature (7), but the blockade of XO had no effect on nitrite-associated vasodilatation tested in normoxia (10).
The mechanism of acute vasodilation to nitrite can be explained by either of the above theories. The observation of persistent vasodilation in the pulmonary bed is more consistent with direct tissue reduction of nitrite, as this effect was not dependent on a concurrent elevation in plasma nitrite and interaction with deoxyhemoglobin (expanded in Nitrite is a Pulmonary Vasodilator Whose Effect is Prolonged).
Nitrite is a Pulmonary Vasodilator Whose Effect is Prolonged
In animal studies it has been established that nitrite is a pulmonary vasodilator whose action is potentiated by hypoxia (7, 11, 15). Nitrite also acted as a vasodilator in models of raised pulmonary arterial pressure where an inhaled normoxic gas mixture was used (7, 11), but these models simulated alternative conditions capable of enhancing nitrite reduction to NO, such as ischemia. Hunter showed that nebulized nitrite is a slower pulmonary vasodilator but has a prolonged effect compared with inhaled NO, in keeping with the chemical stability of nitrite salts compared with the more volatile NO gas. This present study has mirrored these findings for the first time in a human model of hypoxic pulmonary hypertension.
However, while we acknowledge that Hunter showed the above effect to be associated with increased deoxyhemoglobin levels and a reduced pH, our data demonstrate that the pulmonary vasodilator effect of sodium nitrite persist even after plasma nitrite has returned to normal. This unexpected finding is at odds with some existing theories of nitrite pharmacodynamics (14), and explaining it requires the involvement of additional concepts. In order for pulmonary vasodilation to occur when plasma nitrite was normal, the initial nitrite infusion must have either primed the pulmonary tissue or alternatively loaded the tissue with nitrite or a nitrite-derived NO metabolite.
It has been previously suggested that stable intermediate metabolites of nitrite in blood are responsible for vascular activity, i.e., NO bound to a thiol group on plasma proteins or hemoglobin (S-NO groups) (31). A transient increase in plasma nitrite increases the population of these NO stores (2). Off-loading of NO from such species may occur to a greater degree in the hypoxic pulmonary artery, a phenomenon we have observed in isolated vessel preparations (30).
The simple fact that oxygen levels are lower in the pulmonary arteries than in the systemic arterial circulation could account for the observed difference by potentiating NO release from nitrite or S-NO. In addition, we have recently shown that isolated pulmonary arteries relax more than aortic rings to nitrite under hypoxic conditions (16). It may follow, therefore, that the pulmonary vasculature vasodilates at lower nitrite concentrations than the systemic circulation. However, given that at 1 h after the nitrite infusion, plasma nitrite had returned to preinfusion levels, the increased sensitivity to nitrite by the pulmonary vasculature cannot explain our findings alone.
Lastly, the pulmonary-specific vasodilator activity of nitrite in hypoxia may simply be attributable to the fact that in hypoxia the pulmonary vasculature constricts more than the systemic vasculature. Thus a sensitized environment is created in which nitrite, or more likely a metabolite of nitrite, can exert an effect.
Dietary Sources of Nitrite
Thirty percent of plasma nitrite is obtained from the diet, with the balance arising from intravascular metabolism of NO produced by the endothelium (17). The major dietary source of nitrite is foods rich in nitrate, especially vegetables. Commensal bacteria from the upper gastrointestinal tract first reduce nitrate to nitrite, which is then further reduced to NO by the acidic environment of the stomach. NO gas is then absorbed into the circulation where it is oxidized to nitrite in the plasma (22).
Dietary supplementation of nitrate for 3 days reduced diastolic blood pressure by 3.7 mmHg (19). This study demonstrated a similar increase in plasma nitrite to that achieved in our study. Another study with a similar reduction in both systolic and diastolic blood pressure 3 h after a 30-min period of oral ingestion of foods rich in nitrate (33) also had comparable plasma nitrite levels to our own. In both of these studies, a large (in mmol) dose of nitrate was given that was gradually absorbed from the stomach. This led to a sustained elevation of plasma nitrite over a longer period than in our current study. Consequently, a large proportion of nitrite would have been absorbed into the surrounding tissue in these studies, from where it would have exhibited a vasodilator effect. Therefore, the levels of tissue nitrite will have been much higher in studies using oral nitrate loading of subjects compared with our study where rapidly bioavailable but lower-dose intravenous nitrite loading was used. This is supported by other studies using intravenous infusions of nitrite which report no vasodilator effects in normoxia at a plasma nitrite concentration greater than those achieved in our study (24).
The Clinical Implications of Nitrite: A Novel Targeted NO Donor
Organic nitrates (e.g., nitroglycerine) will remain the primary pharmacological means of nonselective and systemic delivery of a NO-type effect in clinical practice. Our study has demonstrated that inorganic nitrite can also be given, at a low level that is physiologically inert in normoxia, to provide vasodilation targeted to hypoxic vasculature. Previous literature has demonstrated the safety of sodium nitrite infusion in humans, with doses in excess of those used in this study, causing only a minor increase in methemoglobin levels, which was not clinically significant (23).
Little work has been performed on the in vivo effects of organic nitrates in hypoxia. One study has shown that their vasodilator effects are prolonged (4). This would suggest that nitrite could merely be the vehicle by which NO is provided to the tissue and that any NO donor drug would be equally effective. However, it must be recognized that in many of the pathologies where an increase in regional NO concentration would be desirable (such as myocardial infarction, stroke, or solid organ transplantation), the highly effective nature of even a low dose of organic nitrate could cause variability in blood pressure that would be undesirable. Consequently, a regime for administering nitrite such as that suggested by our study could offer the therapeutic benefits of local vasoactivity delivered to hypoxic tissue without systemic vascular collapse. In addition, nitrite, unlike organic nitrates, is unaffected by tachyphylaxis when given at high doses to primates (10). Further studies to ascertain whether a tolerance to low-dose nitrite develops in humans would be helpful.
In light of the prolonged effect on the pulmonary circulation that we have shown, inorganic nitrites may be particularly effective in the treatment of pulmonary hypertension. A major limitation may be that the etiology of pulmonary hypertension is often structural rather than functional, and therefore some patients may not respond to the selective pulmonary vasodilator effect of nitrite. However, current therapies for pulmonary hypertension include vasodilators (such as nifedipine), which are known to help only a minority of patients. Nonetheless, it is standard practice to test whether patients respond to vasodilators, and future studies would allow nitrite to find its position alongside established therapies. Interestingly, the enhanced vasoconstriction specific to high-altitude pulmonary edema might also respond very well to nitrite.
In conclusion, we have demonstrated that a low dose of sodium nitrite, not more than doubling the plasma nitrite concentration, is capable of causing selective vasodilatation in hypoxic vasculature. Additionally, the vasodilatory effect on the pulmonary circulation occurred even when plasma nitrite had returned to baseline levels.
This work was supported by British Heart Foundation Grants FS/06/088 (to T. E. Ingram) and FS/05/110 (to A. G. Pinder)
No conflicts of interest.
We thank Prof. Frank Dunstan and Keith Morris for statistical advice given during the preparation of the manuscript.
- Copyright © 2010 the American Physiological Society