Nitrite Regulates Hypoxic Vasodilation via Myoglobin–Dependent Nitric Oxide Generation

Assessment of in vivo hemodynamics

Mice were anesthetized by intraperitoneal (i.p.) injection of ketamine (45 mg kg⁻¹) and xylazin (Rompun, 10 mg kg⁻¹). A tracheal tube was inserted and mechanical ventilation initiated according to the body weight (volume controlled ventilation, Inspira, Harvard Apparatus, March-Hugstetten, Germany). Ventilation was controlled by end-tidal capnography (Hugo-Sachs, March-Hugstetten, Germany). 1.2 vol% isoflurane was added as anesthetic to O₂ and nitrogen. A pressure volume catheter (Millar Instruments) was inserted via the right carotid artery into the thoracic aorta and heparin (70 IU kg⁻¹) was injected i.p.. Heart rate, systolic (P_sys) and diastolic blood pressure (P_dias) were recorded beat-to-beat.

To assess basal hypoxic vasodilation, hypoxia was induced by ventilation with 10% O₂/90% N₂ and the effects on hemodynamics were recorded continuously without the administration of exogenous nitrite. We measured heart functions through a pressure volume catheter placed in the left ventricle (1.4 F, Millar Instruments, Seeheim-Ober Beerbach, Germany).

To investigate the effects on blood pressure of increased nitrite levels by adding exogenous nitrite under steady-state hypoxia, we first injected the NOS inhibitor L-NIO (100 mg kg⁻¹) i.p. to inhibit endogenous NO^˙ synthesis via NOS and unmask the genuine effects of exogenous nitrite.⁴ After 30 min, we injected nitrite (1.67 μmol kg⁻¹ or 16.7 μmol kg⁻¹) intravenously (i.v.) and measured hemodynamics for the following 20 min. For the hypoxic experiments, the ventilation gas mixture was changed to 10% O₂/90% N₂ after the injection of NOS inhibitor. Hypoxic ventilation was conducted for 30 min to establish steady state conditions before nitrite was injected.

Final blood gas composition was determined (Siemens Blood Gas Analyzer 865, Eschborn, Germany) in all in vivo experiments (steady-state hypoxia with exogenous nitrite and basal hypoxic vasodilation) and only those animals with a pO₂ of 36±6 mmHg (standard deviation [SD]) and with normal CO₂ and pH were included.

Supplemental Methodology

For a detailed explanation of the methods applied to determine Mb gene regulation, to demonstrate Mb mRNA and protein in situ, to investigate vascular Mb nitrite-reductase activity, to analyze cGMP formation, to measure nitrite-derived NO^˙ and nitrite, and for aortic ring bioassay and the assessment of down-stream signaling pathways, please see the Method section of the online-only Data Supplement.

Statistical analysis

Descriptive statistics such as mean±SD or mean±SEM (standard error of the mean) were used to summarize continuous variables. The data were analyzed by Student's t-test (two groups) and ANOVA (multiple groups) with post hoc Bonferroni adjustment using statistical software Prism 5.0 (GraphPad) and Sigma Plot 11 (SigmaPlot). Two-way ANOVA (for [i] mouse species and [ii] time) with pairwise comparisons (Holm-Sidak procedure) was used for comparisons on in vivo basal hypoxic vasodilation. P-value < 0.05 was considered to be statistically significant. However, given the smaller sample size of some approaches (n=3), P-values should not be over-interpreted.

Mb is expressed in vascular smooth muscle and reduces nitrite to NO^˙

The expression of Mb transcript in mouse aorta tissue was confirmed by two methods. First, we used reverse transcription-PCR, checked by sequencing, to demonstrate the presence of Mb RNA transcripts in mouse vascular tissue (data not shown). Second, we used RNA-in situ hybridization of an anti-sense riboprobe directed against the murine Mb transcript to identify the smooth muscle cells of the media as the source of Mb mRNA (Figure 1A). The corresponding sense riboprobe was inactive (Figure 1 of the online-only Data Supplement) and no signal was obtained in Mb^−/− mice (Figure 1B). The presence of Mb protein was confirmed using both immunohistology, using a custom-made polyclonal antibody directed against the C-terminus of mouse Mb (Figure 1C and D; negative controls in Figure 1 of the online-only Data Supplement) and Western blotting (Figure 1E). Figures 1C–E shows that aortas from Mb^+/+ mice contain Mb, namely in all smooth muscle cells; vessel from Mb^−/− mice did not exhibit a signal.

We then examined whether vascular tissue can generate NO^˙ from nitrite under hypoxia and whether this was dependent on the presence of deoxyMb. Homogenates of mouse aortas were added to a PBS solution (pH 5.5) containing nitrite (100 μM, final) under anaerobic conditions, and the sustained release of NO^˙ was monitored by headspace gas-phase chemiluminescence. Figure 1F shows that the rate of NO^˙ release from aortas of Mb^−/− mice was substantially lower than for Mb^+/+ mice. Figure 1G shows that the inhibition of the XOR, by simultaneous addition of its inhibitors allopurinol (100 μM) and diphenyliodonium (DPI, 10 μM) and the inhibition of the electron transfer from ubiquinol to bc1-complex in the mitochondrial respiratory chain by incubation with 300 μM myxothiazol had no effect in Mb^+/+ and Mb^−/− tissue, consistent with a negligible contribution in previous observations.^26,33,34 Preincubation with 50 mM ferricyanide to oxidize all cellular heme proteins significantly reduced the rate of NO^˙ generation in Mb^+/+. By contrast, no significant NO^˙ release from nitrite from aortic tissues was detected under normoxia (data not shown). Figure 1H shows that addition of 20 μM Hb did not significantly affect the rate of NO^˙ generation in both Mb^+/+ and Mb^−/− tissue. However, we observed a tendency to greater rates of NO^˙ release in the Mb^−/− group and therefore a statistical analysis of the Mb^−/− approaches was conducted. This revealed a small but significant increase of nitrite reduction in Mb^−/− tissue after incubation with Hb which is consistent with our previously reported observations under the same experimental conditions.²⁷

Regulation of physiological in vivo vasodilation under hypoxia depends on reduction of endogenous nitrite via Mb

In light of relatively high levels of nitrite in the vessel walls compared to other tissues and blood³⁵ and the demonstrated nitrite-reductase activity of vascular Mb, we investigated acute in vivo responses to hypoxia using Mb^+/+ and Mb^−/− mice, which have comparable levels of nitrite (Figure 2A of the online-only Data Supplement). Anaesthetized mice were intubated and mechanically ventilated (scheme in Figure 2A). Hypoxia was then induced and the immediate effects on hemodynamics were recorded. In Mb^+/+ mice, we observed a reduction in systolic (P_sys) and diastolic blood pressure (P_dias) within 2–4 min of hypoxia to a lower plateau after 5–10 min. In Mb^−/− mice, the acute blood pressure response was reduced by up to 54% (Figure 2B and C) and instead we saw a continuing but gradual decline in pressure.

Nitrite-induced hypoxic vasodilation depends on NO^˙/sGC/cGMP signaling pathway

The in vivo differences between Mb^+/+ and Mb^−/− mice were consistent with significantly higher levels of nitroso species in aortic tissue relating to higher rates of NO^˙ production in Mb^+/+ mice (Figure 3A; comparable normoxic values and RNNO levels in Figure 2B–D of the online-only Data Supplement). Given the smaller sample size of these approaches, P-values should not be over-interpreted. However, this was further evidenced by higher plasma cGMP levels and in aortic tissue of Mb^−/− mice (Figure 3B; relevant normoxic baseline values Figure 2E and F of the online-only Data Supplement). These cGMP levels were quantified by competitive enzyme immunoassay in excised thoracic aortas incubated under hypoxia with 10 μM nitrite for 10 min.³⁶ This suggests that nitrite-dependent formation of NO^˙ via deoxyMb significantly elevates the cellular signaling molecule, cGMP. Using an ex vivo bioassay with phenylephrine-constricted Mb^+/+ and Mb^−/− aortic rings, we then assessed the effect of the NO^˙ scavenger cPTIO and the sGC inhibitor ODQ upon vasorelaxation under 10 μM nitrite. In Mb^+/+ vessels both inhibitors significantly but not totally³³ inhibited the nitrite-dependent vasodilatory response (Figure 3C). To quantify the production of NO^˙ we used EPR spectroscopy on single excised aortas that were pre-incubated with 100 μM nitrite under hypoxia (1% O₂, pH 7.4). A spin trap was generated anoxically from ferrous sulfate and diethyldithiocarbamate (0.2 mM Fe-(DETC)₂) and added to trap any NO^˙. The resulting NO-Fe-(DETC)₂ EPR signal was used to quantify the amount of NO^˙ generated over a 10 min period. Figure 3D shows that the signal obtained from Mb^+/+ mice was distinctive from that of Mb^−/− mice. Quantification of the NO-Fe-(DETC)₂ EPR signal (Figure 3E) revealed a ~50% decrease of NO^˙ formation in Mb^−/− aortas relative to vessels from Mb^+/+ mice. Taken together these data suggest that the nitrite-induced, Mb-dependent, in vivo vasodilation directly depends upon the canonical NO^˙/sGC/cGMP pathway located in the vessel wall.

Hypoxic vasodilation is independent of enzymatic NO^˙ production with normal responses in eNOS^−/− and iNOS^−/− mice

NO^˙ could be generated by the O₂-dependent eNOS or iNOS rather than by Mb. To explore the involvement of eNOS/iNOS in the initiation of in vivo hypoxic vasodilation we used eNOS^−/− and iNOS^−/− mice. eNOS^−/− mice and to a lesser extent iNOS^−/− mice displayed somewhat higher vascular pressures than the corresponding wild-type mice (genetic background: C57BL/6), but despite this, both displayed pronounced hypoxic vasodilation responses, which in relative terms were indistinguishable from their respective wild-type controls (Figure 3F and Figure 3 of the online-only Data Supplement). This is incompatible with a role for eNOS and iNOS for hypoxic vasodilation.

Effects of escalating nitrite doses on hypoxic vasodilation in vivo depend on the presence of Mb

To further demonstrate the role of nitrite as the principal source of NO^˙, we assessed the in vivo Mb-dependent effects on blood pressure under steady state hypoxia using different exogenous nitrite doses. Pharmacological concentrations of exogenous nitrite have been demonstrated to relax pre-constricted isolated arteries.^5,17,18 In our experimental protocol endogenous NO^˙ synthesis was inhibited pharmacologically using L-NIO⁴, and mice were exposed for 30 min to hypoxia (ventilation with 10% O₂ to reach a final paO₂ of 30 mmHg). Exogenous nitrite was then administered i.v. and hemodynamics were monitored (schema in Figure 4A). Nitrite at 16.7 μmol kg⁻¹ caused a significantly greater reduction in blood pressure in Mb^+/+ than in Mb^−/− mice (Figure 4B), and the same but smaller effects were evident at 1.67 μmol kg⁻¹ nitrite, which is a physiological dose (Figure 4C and absolute values and the effects of exogenous nitrite under normoxia as controls in Figure 4 and 5 of the online-only Data Supplement).

If NO^˙ was generated from applied nitrite via vascular Mb then we anticipated the formation of iron-nitrosylated Mb products (MbNO)^30,37 in aortic extracts - an indirect marker of nitrite reduction. Indeed, after hypoxic incubation of aortas with the isotope [¹⁵N]-labeled nitrite we detected the Mb[¹⁵N]NO product using EPR spectroscopy in Mb^+/+ but not Mb^−/− aortas (Figure 5A). The EPR spectrum was identical to that obtained for an authentic Mb[¹⁵N]NO solution which directly links the nitrite-dependent production of NO^˙ to the Mb protein itself and not to other heme proteins in smooth muscle, including cytoglobin³⁸, which should occur in Mb^−/− mice. Other potential nitrite reduction mechanisms^33,39,40 could also be negated by using ex vivo aortic ring preparations under hypoxic conditions. Thus Figure 5B shows that neither the specific inhibition of XOR or aldehyde oxidase nor the absence of eNOS by removal of the endothelium affected the nitrite-induced (10 μM) vasodilation or the impact of Mb.

Nitrite-driven hypoxic vasodilation via vascular Mb is not related to changes in cardiac output

Finally, hypoxic vasodilation might be linked to mechanisms of cardiac origin, either through reduced cardiac output or by transfer of cardiac NO^˙ to vascular smooth muscles. The former is unlikely since we found sustained or even enhanced heart functions during hypoxic challenge using a pressure volume catheter placed in the left ventricle (Figure 6A and Figure 6 of the online-only Data Supplement), perhaps due to a decrease in afterload. Also Figure 6B shows that the circulating NO^˙ pool represented by nitrite and nitroso compounds in the plasma was identical in both species under hypoxia and unaffected by absence of Mb, which argues against any influence on vasodilation of NO^˙ generated by the heart.

Role of heme globins

A nitrite reductase activity has been identified in a variety of different proteins ranging from XOR³⁶, cytochrome c oxidase⁴¹, cytochrome c⁴², to eNOS⁴⁰. A comparable activity was also described for heme globins, e.g. Hb and Mb.^4,19,26 and, neuroglobin has recently emerged as novel redox-regulated nitrite-reductase.²⁸ Since heme globins can also react with NO^˙ to form stable nitrosyl-heme complexes which may limit NO^˙ bioavailability, a role for heme globin driven NO^˙ signaling has been questioned. Despite a general belief that Mb expression and function is limited to cardiomyocytes and striated muscle cells, recent experimental studies confirmed the existence of Mb in the vasculature and in a wide range of non-muscle tissues of the hypoxia-tolerant carp.^31,43 Here, we have confirmed the presence of Mb transcript in smooth muscle layers of vessels and using EPR spectroscopy we have confirmed Mb protein expression according to its distinct spectrum.³⁷ No transcript signal was evident in endothelial cells as recently described for capillaries located in the central nervous system³¹ consistent with its presence in smooth muscle cells. We also showed that ablation of Mb in vascular tissue led to a marked decrease in reduction of nitrite to NO^˙, an observation that was independently verified by chemiluminescence and by EPR spectroscopy. By contrast, we failed to detect a role for XOR-mediated nitrite reduction³⁶, and the contribution of acidic disproportionation³⁷ was negligible. The local presence of Mb in the vessel wall and a relevant Mb-dependent NO^˙ generation implicates a role for this protein in the regulation of vascular functions under hypoxic conditions.

Recent advances in understanding the role of this heme globin in cardiac function were achieved by taking advantage of the Mb-knockout mouse. We were thus able to demonstrate the relevance of cardiac Mb in nitrite reduction along the physiological O₂ gradient thus restoring myocardial energy balance and yielding a much reduced infarct size^26,30. Here we showed that genetic ablation of Mb impaired the hypoxia-induced vasodilation response by up to 54% which points directly at Mb as a key component in the signal transduction mechanism necessary for hypoxic vasodilation. While these observations specifically assess a role for myoglobin which accounts for more than half of the vasodilatory response, it is likely that multiple overlapping enzymatic and non-enzymatic pathways for nitrite reduction are present in both vascular tissue and in the blood compartment to allow for the graded reduction of nitrite to NO^˙ at different oxygen tensions along physiological gradients.²⁴ These data therefore also highlight the potential for Hb and red blood cells, neuroglobin, cytoglobin and potentially undiscovered proteins to be involved in hypoxic nitrite reduction accounting for the remainder of the vasodilatory response. Further investigations are needed to compare each of their relative contributions.

Role of nitrite

Nitrite is not only the oxidation product of NO^˙ but also a key reservoir for NO^˙ in blood and cellular compartments. The largest component of the bodily nitrite provision derives from endogenous generation by NOS, while a smaller percentage is derived from nutritional sources.⁴⁴ Nitrite is present in the blood in nanomolar concentrations while tissue levels in heart, liver, kidney and particularly in the vasculature are comparably higher irrespective of the species investigated.^35,45,46 We and others have recently demonstrated that exogenous supplementation with pharmacological nitrite doses has implications for tissue protection under pathological conditions. Nitrite-related protection has been described for general disease states, e.g. the cardiopulmonary resuscitation syndrome⁴⁷ and for chronic and acute intervention regimens.^48,49 Organ-specific protection has also been described for ischemic conditions of the brain, the kidney, the liver, and heart.^30,49,50

To substantiate the role of endogenous nitrite as a key effector by its Mb-induced activation to NO^˙ we investigated the involvement of the NO^˙/sGC/cGMP signaling pathway. A timely vasodilatory response requires the activation of smooth muscle cell sGC.¹³ We showed significantly higher levels of cGMP when Mb is present, this being a definitive indication of sGC activation upon nitrite reduction. As expected, an inhibition of sGC or scavenging of NO^˙ did not abolish the vasodilatory effects completely. Furthermore, a specific role for NOS was excluded using eNOS^−/− and iNOS^−/− mice. In addition, our experiments with different exogenous nitrite doses further confirmed that nitrite is the principal source of NO^˙ via heme globin driven signaling as exogenous nitrite at high and low pharmacological concentrations caused a significantly greater reduction in blood pressure in Mb^+/+ than in Mb^−/− mice.

Taken together, these new observations demonstrate that endogenous nitrite and the heme globin Mb play an essential role in activating the hypoxic vasodilation response. Our principal finding was that a range of nitrite-induced responses under hypoxia on vasodilation and subsequently on blood pressure was substantially reduced or absent in mice lacking Mb. This relates specifically to NO^˙ generation from nitrite and the activation of the NO^˙/sGC/cGMP signaling pathway. Critically, we also showed that the reduced blood pressure, induced in vivo by hypoxia, was largely impaired under the absence of Mb as nitrite-reductase, while knockout of eNOS or iNOS had no such effect. We also showed in vivo that vasodilation can be induced under sustained hypoxia by application of a physiological dose of exogenous nitrite. That hypoxic vasodilation occurred in aortic rings stripped of endothelium indicates that the vascular mechanism is located in the remaining smooth muscle cells, which we show is the cyto-location of Mb transcripts and protein. However, it has to be considered that these experimental models are an approximation of relevant vascular beds. In summary, these data provide evidence for a physiological role for endogenous nitrite and for a heme-globin-related signaling mechanism in vivo.