How red blood cells process nitric oxide: evidence for the nitrite hypothesis

Basal and shear-mediated blood flow is regulated in large part by nitric oxide (NO) produced by the endothelial NO synthase. This oxygen and L-arginine dependent process regulates 25% of resting blood flow and is further activated by the shear-stress of exercise induced metabolic vasodilation to enhance NO-dependent vasodilation of conduit vessels.¹ NO that is produced under basal conditions and enhanced by exercise-shear-activated NOS not only vasodilates but is also oxidized in plasma to form nitrite and S-nitrosothiol compounds.² How these NO metabolites contribute to blood flow and endocrine NO signaling has been the subject of active research and debate over the last 15 years.³

Two dominant theories have been proposed that both require red blood cell and deoxyhemoglobin-dependent bioactivation. The S-nitroso-hemoglobin (SNO-Hb) hypothesis posits that NO binds covalently to the conserved cysteine 93 residue of hemoglobin and during deoxygenation releases an intermediate (still not defined) that is exported from the red blood cell as a low molecular weight S-nitrosothiol, leading to hypoxic vasodilation.⁴ The nitrite reductase hypothesis posits that the endocrine reservoir is nitrite, and during arterial to venous (A-V) transit the nitrite reacts with the deoxy-heme of hemoglobin to form NO (and N₂O₃), which produces hypoxic vasodilation.^{5, 6} The latter hypothesis has been supported by experimental data showing that other globins signal via nitrite reduction, including myoglobin, cytoglobin, neuroglobin and globin X.^7–9 This signaling has been shown to modulate hypoxic vasodilation, platelet activation, and mitochondrial respiration (recent platelet inhibition studies summarized¹⁰). Both models propose allosteric-coupled hemoglobin deoxygenation and NO delivery and both models have central mechanisms that are testable.

The current study by Bailey and colleagues in this issue of Circulation¹¹ revisits the measurement of NO metabolites in the cerebral and leg arterial and venous circulation, in the plasma and red blood cells, at rest, during hypoxia, and during two levels of graded aerobic exercise. These physiological studies are well designed and assess gradients across two systemic circulatory beds, with 5-fold increases in blood flow during exercise. However, the study suffers from a number of limitations that are recognized by the investigators, such as a sample size with only 7 volunteer subjects and a lack of measureable increases in blood flow in the cerebral and leg circulation during the hypoxic challenge (no evident hypoxic vasodilation). These limitations notwithstanding, and accepting the caveat that gradient analyses do not confirm mechanism, we can evaluate the results of this work and determine how well the data conform to predictions based on the two red cell-NO delivery hypotheses.

In brief summary, Bailey and colleagues observe significant gradients of nitrite with arterial levels greater than venous levels during rest. Following systemic hypoxia the arterial nitrite levels decrease to the level observed in the venous circulation. In contrast, they do not observe significant arterial-venous (A-V) gradients of SNO-Hb; while not significant, the SNO-Hb levels increase in the venous circulation (V>A) under all conditions tested. During exercise the arterial nitrite levels increase and the (A-V) gradients are preserved, with significant increases in the calculated nitrite consumption (using Fick estimates of consumption: blood flow multiplied times the A-V nitrite gradient). There is a very significant increase in red cell NO metabolite accumulation from A-V during hypoxia and especially exercise, with the dominant and only statistically significant species formed being iron-nitrosylated hemoglobin (NO bound to the heme of hemoglobin; Fe⁺²-NO).

The nitrite-reductase hypothesis was grounded in observations of A-V gradients in nitrite, with nitrite consumption from artery-to-vein (A>V), and a reaction in the red cell between nitrite and deoxyhemoglobin to form NO.⁵ The NO would then bind to viscinal deoxyhemoglobins to form iron-nitrosylated-hemoglobin, and to a lesser extent SNO-Hb (equation 1 and 2).

$${{\text{NO}}_2}^{-}(\text{nitrite})+{\text{Fe}}^{+2}(\text{deoxyhemoglobin})\to \text{NO}+{\text{Fe}}^{+3}(\text{methemoglobin})$$ (eq. 1)

$$\text{NO}(\text{nitric}\text{oxide})+{\text{Fe}}^{+2}(\text{deoxyhemoglobin})\to {\text{Fe}}^{+2}-\text{NO}(\text{iron}-\text{nitrosyl}-\text{hemoglobin})$$ (eq. 2)

In the current study they observe robust A-V plasma nitrite gradients across the brain at rest (380–286 nM) and across leg at rest (380–320 nM). The arterial levels of nitrite drop with hypoxia, as the hypothesis would predict, and exercise results in a significant increase in both arterial nitrite production and consumption across both vascular beds (the brain: 404–295 nM and leg: 404–324 nM). Shear-activation of conduit vessel eNOS will increase NO and nitrite production, consistent with higher arterial levels, and maintained A-V gradients indicate that nitrite is consumed during the metabolic stress of exercise. Note that the gradient is observed in the setting of a 5-fold increase in blood flow, indicating large increases in nitrite consumption, according to the Fick principle, similar to prior published studies of nitrite metabolism during forearm exercise.¹² Finally, red cell iron-nitrosylated hemoglobin (Fe⁺²-NO) increases dramatically in all of these conditions, with Fe⁺²-NO representing the largest, and only significant contributer to the total observed red cell NO species formed. Fe⁺²-NO levels increased significantly with hypoxia from artery to vein, with additional increases with exercise, again consistent with the nitrite reductase hypothesis (equations 1 and 2). These observations are similar to those observed with nitrite infusions in humans, with consumption of nitrite across the circulation and metabolism to form venous Fe⁺²-NO and, to a lesser extent SNO, in the red cell.⁵

The SNO-Hb hypothesis was grounded on observations of A-V gradients in SNO-Hb in the red cell, with allosterically mediated release from A-V, resulting in a drop in SNO-Hb in the red cell and a rise in plasma S-nitrosothiols in the venous circulation.⁴ The results here are the opposite, with no gradients in observed in plasma or red blood SNO-Hb under any conditions. The red cell SNO-Hb measures are contrary to the hypothesis, with non-significant increases in the venous circulation rather than a decrease consistent with S-nitrosothiol delivery. There are non-significant drops in plasma SNO in the brain circulation, but these should go up if S-nitrosothiols are released from SNO-Hb during deoxygenation. However, the rise is not close to significant with high standard deviations (note A-V levels with exercise of 34 ± 31 vs. 41 ± 20; n=7), and not consistent across all measured conditions and circulatory systems.

In conclusion, these measurements are consistent with consumption of nitrite in plasma from artery-to-vein, with enhanced nitrite consumption during exercise. Nitrite consumption is coupled to Fe⁺²-NO formation in the red blood cell from artery-to-vein at rest, during hypoxemia and especially during exercise. The nitrite consumption gradients and Fe⁺²-NO formation gradients are the most statistically robust observations in these studies and conform to equation 1 and 2. These gradients reproduce the original observations in the forearm circulation, during exercise and nitrite infusions, that informed the nitrite reductase hypothesis.^{5, 12} Recent studies that other globins exhibit nitrite-reductase signaling activity, confirmed with genetic knock-down approaches, provides more direct evidence in support of this hypothesis.^7–9