Hypoxia, red blood cells, and nitrite regulate NO-dependent hypoxic vasodilation

Abstract

Local vasodilation in response to hypoxia is a fundamental physiologic response ensuring oxygen delivery to tissues under metabolic stress. Recent studies identify a role for the red blood cell (RBC), with hemoglobin the hypoxic sensor. Herein, we investigate the mechanisms regulating this process and explore the relative roles of adenosine triphosphate, S-nitrosohemoglobin, and nitrite as effectors. We provide evidence that hypoxic RBCs mediate vasodilation by reducing nitrite to nitric oxide (NO) and ATP release. NO dependence for nitrite-mediated vasodilation was evidenced by NO gas formation, stimulation of cGMP production, and inhibition of mitochondrial respiration in a process sensitive to the NO scavenger C-PTIO. The nitrite reductase activity of hemoglobin is modulated by heme deoxygenation and heme redox potential, with maximal activity observed at 50% hemoglobin oxygenation (P(50)). Concomitantly, vasodilation is initiated at the P(50), suggesting that oxygen sensing by hemoglobin is mechanistically linked to nitrite reduction and stimulation of vasodilation. Mutation of the conserved beta93cys residue decreases the heme redox potential (ie, decreases E(1/2)), an effect that increases nitrite reductase activity and vasodilation at any given hemoglobin saturation. These data support a function for RBC hemoglobin as an allosterically and redox-regulated nitrite reductase whose "enzyme activity" couples hypoxia to increased NO-dependent blood flow.

Figure 1.

Hypoxia, RBCs, and nitrite stimulate vasodilation. Nitrite-dependent vasodilation alone (▪) or in the presence of RBCs (0.3% HCT) (•) at 60 mmHg (A), 25 mmHg (B), and 15 mmHg (C) oxygen. Using a P₅₀ value of 35 mmHg for rat RBCs and a Hill coefficient of 2.5,, at 15 mmHg, 25 mmHg, and 60 mmHg oxygen, Hb saturation corresponded to 11%, 30%, and 82%, respectively, encompassing physiologic oxygen saturations. Data are mean ± SEM (n = 3). *P < .05; #P < .01 relative to the corresponding nitrite dose without RBCs. (D) EC₂₀ (nitrite concentration that stimulated 20% dilation) values (determined from 3 independent experiments) of nitrite (▪) and nitrite + 0.3% HCT rat RBCs (□) at 15 mmHg, 25 mmHg, and 60 mmHg oxygen. *P < .01 relative to corresponding controls (n = 3).

Figure 2.

Nitrite metabolism by deoxygenated RBCs stimulates NO-dependent vasodilation and inhibition of mitochondrial respiration. (A) Representative vessel tension traces showing rat RBCs (0.3% HCT) and nitrite-dependent vasodilation at 25 mmHg oxygen in the presence (black) and absence (gray) of 200 μM C-PTIO. Arrows indicate the times and concentrations of nitrite added. (B) Changes in cGMP in rat thoracic aorta treated as indicated for 10 minutes at 25 mmHg oxygen (▪) or 600 mmHg oxygen (□). Final concentrations were nitrite 3 μM, rat RBCs 0.3% HCT, and C-PTIO 200 μM. *P < .01 relative to control and nitrite alone; *P < .05 relative to RBCs alone (n = 3). (inset) To estimate the amount of NO produced by RBC/nitrite, a calibration curve was determined for increased vessel cGMP in response to increasing concentrations of the NO donor, DeaNonoate, at 25 mmHg oxygen. In a concentration-dependent manner, NO increased cGMP, which was inhibited by C-PTIO (○). Specifically, 100 nM DeaNonoate increased cGMP from 526.3 ± 46 pmol/g to 5602.6 ± 1340 pmol/g in the absence of C-PTIO and to 653 ± 97.9 nM in the presence of C-PTIO (values represent mean ± SEM). Using this curve, an estimated 20 nM NO was produced by RBCs (0.3% HCT) and nitrite (3 μM) over 10 minutes. (C) Representative traces of oxygen concentration as a function of time for mitochondria in state 3 respiration (black line) in the presence of 20 μM nitrite (gray line), 0.3% HCT RBCs (dashed line), or RBCs and nitrite (dotted line) are shown. Respiratory substrates were added to oxygen electrode chamber containing mitochondria and RBCs. The arrow indicates the time of nitrite addition. (D) Respiration rate measured at 30 μM oxygen. Inhibition of respiration by RBCs + nitrite was completely reversed by C-PTIO, consistent with NO formation. C-PTIO had no effect alone (not shown), and neither did RBCs alone. *P < .001 relative to RBCs alone. #P < .002 relative to RBCs + nitrite + PTIO (n = 3).

Figure 3.

RBC P₅₀ regulates nitrite-dependent vasodilation. (A) Representative traces of simultaneously measured vessel tension and oxygen tension as a function of time with RBCs alone (gray trace) or RBCs + nitrite (black trace). (B) Nitrite-dependent dilation starts at higher oxygen tensions in response to human RBCs (hRBCs), HBOC-201, or HbA and IHP. Final concentrations: nitrite (2 μM), human RBCs (0.3% HCT), HBOC-201 (25 μM), HbA (25 μM), and IHP (625 μM). †P < .03 relative to nitrite and P < .01 relative to hRBCs alone (n = 3). *P < .001 relative to nitrite alone. #P < .01 relative to Hb + IHP. Data represent mean ± SEM (n = 3-5). (C) Relationship of Hb P₅₀ and the oxygen tension at the initiation of vessel dilation in the presence of nitrite (2 μM). Data from rat RBCs, hRBCs, HBOC-201, and cell-free HbA (in which P₅₀ was modulated by the addition of varying concentrations of IHP, resulting in heme/IHP ratios ranging from 1:0.25 to 1:25) are shown. A linear relationship (y = 19.6 + 0.62x; r = 0.89; note intercept is not zero because of basal dilatory responses of vessels to hypoxia) between P₅₀ and oxygen tension at onset of relaxation was observed. (inset) Calculated values for Hb fractional saturation (Y) at the oxygen tension at which nitrite-dependent vasodilation was initiated. (D) Using the calculated rate constants for nitrite reactions with RBC deoxyHb, the predicted velocity or initial rate of nitrite reaction with deoxyHb under physiologic conditions (total Hb = 20 mM in heme; circulating nitrite concentrations = 500 nM19) and as a function of Hb fractional saturation are shown and indicate a maximal rate of reaction around the Hb P₅₀. Data were collected using RBCs from 3 different preparations. (inset) Representative traces showing NO formation (measured by NO chemiluminescence) in the head-space of the reaction between RBC (0.3% hc) and nitrite (500 μM) after 2 minutes in the open-flow respirometer (head-space volume and reaction solution volume 5 mL each with 100 mL/min gas flow). Traces show NO production after injection of 200 μL of [A] air alone or head-space from [B] nitrite in PBS at 28 mmHg oxygen or nitrite + RBC at [C] 0 mmHg (0% fractional saturation), [D] 28 mmHg oxygen (50% fractional saturation), and [E] 155 mmHg (100% fractional saturation).

Figure 4.

β93cys residue controls nitrite reductase activity of Hb. (A) Hb(C→A) potentiates nitrite-dependent relaxation. Final concentrations: nitrite (2 μM), Hb(C→A) (25 μM), and IHP (625 μM). *P < .001 relative to nitrite alone. #P < .01 relative to Hb(C→A) + IHP. Data represent mean ± SEM (n = 3-5). (B) Relationship between P₅₀ of HbA (▪) or Hb (C→A) () and the oxygen tension at the initiation of vessel dilation in the presence of nitrite (2 μM). Data for HbA are taken from Figure 3C. P₅₀ was modulated by the addition of varying amounts of IHP and was determined under identical buffer, pH, temperature, and PCO₂ conditions as for vessel dilation studies. A linear relationship between P₅₀ and oxygen tension at the onset of relaxation was observed for both Hbs (black line: y = 20.7 + 0.67x; r = 0.96; gray line: y = 11.6 + 1.48x; r = 0.94). Data represent mean ± SEM (n = 3-6). (C) Representative reaction traces for metHb formation from the reaction between nitrite and various deoxyHbs (experiments performed under anoxic conditions). (D) Relationship between Hb P₅₀ and rate of nitrite reduction. Shown are the best fit line (R 2 = 0.99) and the average redox potentials (E_1/2) from values in the literature.,,

Figure 5.

ATP and nitrite represent independent mediators of RBC vasodilation. (A) Representative vessel tension traces (at 25 mmHg oxygen) showing that human and rat RBCs induce contraction or have no vasoactive effect on rat thoracic aorta in the absence or presence, respectively, of L-NMMA. (B) Representative vessel tension traces and (C) percentages of relaxation stimulated by the addition of human RBCs (0.3% HCT) or ATP (1 μM) to rabbit thoracic aorta at 15 mmHg oxygen tension alone or after preincubation with 8 U/mL apyrase, 100 μM reactive blue-2 (RB-2), or 100 μM L-NMMA. To control for the addition of oxygen and oxygen off-loading from RBCs, an equivalent volume of ice-cold PBS saturated with 95% oxygen was added that also resulted in relaxation in an L-NMMA–inhibitable manner. Values are mean ± SEM (n = 3-6). P values are shown. NS = not significant. (D) Human RBCs (0.3% HCT) and nitrite (2 μM) were added to rabbit thoracic aorta incubated with L-NMMA (100 μM) to inhibit ATP-dependent relaxation. Dilation was initiated in a manner similar to that observed with rat vessels (Figure 3) at an oxygen tension of approximately 27 mmHg. Data represent mean ± SEM (n = 4-5). *P < .04 relative to RBCs alone. #P < .01 relative to nitrite alone.

Figure 6.

Proposed model for allosteric regulation of nitrite reduction by deoxyhemoglobin. The rate of nitrite reduction by deoxyhemoglobin is proportional to the product of the deoxyheme concentration and the rate constant, the latter of which is determined by heme reduction potential being greater and having a more negative E_1/2. Deoxyheme concentration and heme reduction potential/rate constant are maximal at low and high oxygen fractional saturations, respectively, resulting in a maximal initial rate of nitrite reduction at approximately 50% fractional saturation (ie, P₅₀). Note that the precise relationship between heme redox potential and fractional saturation is not known. In the model, a linear dependence is approximated based on observations that the rate constant for nitrite-reduction increases linearly with fractional saturation. The points shown on figure are theoretical and are presented to illustrate the concept that the product of rate constant and deoxyhemoglobin concentration determine the initial rate. Also illustrated are hemoglobin conformations (using the symmetry model; for the sake of simplicity, conformations adhering to the sequential model are not shown and R-T transition is denoted as occurring between first and second oxygen binding steps) that are populated as a function of fractional saturation. Conformations highlighted in the gray box denote the proposed intermediates populated at intermediate fractional saturations that have increased nitrite reductase activities and available deoxyheme binding sites to maximally reduce nitrite to NO and stimulate vasodilation.