Contribution of Endothelial Nitric Oxide to Blood Pressure in Humans

Nitric oxide is one of the most widely studied substances in biology. It is produced from L-arginine by NO synthase (NOS) and modulates vascular tone, platelet activation, and neural function, among other actions. NO is produced by 3 isoforms of NOS. An inducible form is found mostly in macrophages (NOS2). Two constitutively expressed isoforms are expressed, one in epithelial and neural cells (NOS1), and another in endothelial cells, platelets, and myocardial cells (NOS3). There is great interest in defining the contribution of endogenous NO to blood pressure modulation. Endothelial cells tonically produce NO (eNO), which lowers basal blood pressure by producing vasodilation. Neurally derived NO (nNO) modulates blood pressure primarily through its interaction with the autonomic nervous system, mainly through inhibition of central sympathetic outflow.^1,2 There is controversy about the relative contribution of eNO and nNO in the regulation of blood pressure. Mice made deficient of NOS3 have a significant increase in baseline blood pressure,³ whereas mice lacking NOS1 do not.⁴ In humans, the contribution of the sympathetic nervous system to the increase in blood pressure produced by systemic administration of NOS inhibitors has been difficult to assess. In some studies, the increase in blood pressure produced by systemic administration of the NOS inhibitor N^G-monomethyl-l-arginine (l-NMMA) results in a reflex decrease in muscle sympathetic nerve activity⁵ that is similar to that produced by an equipressor dose of phenylephrine, suggesting that NO does not tonically restrain central sympathetic outflow in humans.⁶ On the other hand, in other studies, NOS inhibition with N^G-nitro-l-arginine methylester resulted in a larger increase in blood pressure that was partially reversed with the α-adrenergic antagonist phentolamine, most likely reflecting a tonic inhibition of sympathetic tone by nNOS.⁷

The relative contribution of endothelial-derived NO to blood pressure, therefore, remains difficult to assess. In humans, this could be achieved by specific inhibition of eNOS. Unfortunately, selective eNOS inhibitors are not available for use in humans; currently available inhibitors act on both eNOS and nNOS. An alternative approach has been to study an isolated vascular bed, assuming that the vascular effects of NO-dependent vasodilators reflect eNO function. The forearm is commonly used, and forearm blood flow is monitored in response to intrabrachial infusions of endothelium-dependent and -independent vasodilators and to nonselective NOS inhibitors. There are, however, limitations to this approach. Commonly used “endothelium-dependent vasodilators” can also induce vasodilation through mechanisms independent of NO, including endothelium-derived hyperpolarizing factor.^8,9 Also, this approach studies only one vascular bed with limited influence to overall blood pressure levels, and it is not certain that this approach selectively examines eNO function. nNOS is also expressed in skeletal muscle cells and has been suggested to be the source of the NO-mediated inhibition of sympathetic vasoconstriction in contracting muscle.¹⁰ nNOS is also expressed in presynaptic noradrenergic nerve terminals, and its inhibition could theoretically increase NE release and contribute to forearm vasoconstriction.

An alternative approach has been to measure flow-mediated dilation, usually by monitoring the brachial artery diameter in response to reactive hyperemia.¹¹ This response, however, seems to be influenced by sympathetic activity.^12,13 Therefore, it is difficult to estimate how important eNO is to overall blood pressure regulation from the use of these conventional methods.

The goal of this study was to develop an experimental approach that will allow us to preferentially examine the role of endothelial NO on blood pressure regulation. We used blockade of autonomic ganglia with the N_N-nicotinic receptor antagonist trimethaphan to eliminate the restraining effect of the baroreflex, thus allowing for the full expression of the effect of NOS blockade on blood pressure. Furthermore, the effects of nNO become irrelevant, because they depend on its interaction with the autonomic nervous system, which is no longer operative (Figure 1). Under these experimental conditions, the increase in blood pressure induced by systemic NOS inhibition can be assigned to preferential inhibition of tonic eNO.

• Download figure
• Download PowerPoint

Methods

Subjects

We studied a total of 25 subjects. In protocol 1, we studied 17 healthy nonsmoking normotensive subjects (9 women and 8 men; 32±8.8 years; 25±3.9 kg/m² body mass index [BMI]) on 2 separate days. In protocol 2, we compared 10 healthy nonsmoking normotensive subjects with no family history of hypertension (6 women and 4 men; 30±3.1 years; 25±2.0 kg/m² BMI), with 6 normotensive heavy smokers (≥2 packs a day) but otherwise healthy subjects (2 women and 4 men; 31±2.5 years; 25±1.1 kg/m² BMI). Some healthy controls from the first protocol also participated in the second protocol. Participants were recruited from the Vanderbilt University General Clinical Research Center volunteer database. Subjects abstained from all drugs, including caffeine and nicotine, for ≥72 hours before testing. All of the subjects underwent a thorough clinical examination, ECG, and admission urinalysis and blood work. Written informed consent was obtained before study entry. All of the studies were approved by the Vanderbilt University Institutional Review Board.

Study Design

Protocol 1: To Determine the Contribution of eNO to Blood Pressure Regulation

Seventeen healthy volunteers were studied on 2 separate study days randomly assigned ≥1 week apart using a crossover design. One day was designed to assess the effect of NOS blockade with the autonomic nervous system intact (intact study day) and the other one with the autonomic nervous system temporarily blocked with trimethaphan (blocked study day).

Four days before any study, volunteers were on a diet free of food containing methylxanthines. The volunteers were admitted to the Elliot V. Newman Clinical Research Center at Vanderbilt University Medical Center the day that testing was performed.

The studies were conducted in the morning with the subject in the recumbent position ≥8 hours after their last meal. Heart rate was determined with continuous ECG monitoring, blood pressure through the volume clamp method (Finapres 2300; Ohmeda), and also automated brachial cuff pressure with standard sphygmomanometry (Dinamap). Two intravenous lines were placed in different arms. Three infusion ports were connected to the catheter placed in a large antecubital vein in the left arm, one for trimethaphan infusion, the second for infusion of phenylephrine, and the third for L-NMMA. In the other arm, 1 heparin lock was placed to assess cardiovascular responses to phenylephrine before and during trimethaphan to ensure complete autonomic blockade.

After a stable baseline was reached, phenylephrine boluses were started with 25 μg. The dose of phenylephrine was increased every 3 minutes until an increase in systolic blood pressure (SBP) of ≥20 mm Hg was achieved. The changes in SBP were used to calculate baseline dose–response curves to phenylephrine. After this, N_N-cholinergic receptors were blocked by continuous infusion of trimethaphan (Cambridge Pharmaceuticals) at 4 μg/min. We have shown previously that this dose induces complete autonomic blockade.¹⁴ During ganglionic blockade, phenylephrine boluses were repeated with adjustment of the dose to account for the loss of baroreflex buffering, starting at a dose of 2.5 μg. The dose of phenylephrine was increased every 3 minutes until an increase in SBP of ≥20 mm Hg was achieved to calculate the dose–response curve to phenylephrine during autonomic blockade. Blood pressure was then restored by infusing phenylephrine at individually titrated doses, starting with 0.05 μg/kg per minute. l-NMMA was then infused at 2 different doses for 15 minutes each (250 and 500 μg/kg per minute) or until SBP reached 150 mm Hg. On the “intact” day, saline was infused instead of trimethaphan and phenylephrine.

Estimation of the Relative Contribution of eNO and nNO to Blood Pressure Regulation

The increase in blood pressure induced by systemic NO inhibition in the presence of intact sympathetic tone is the resultant of 3 components: (1) blockade of eNO synthesis resulting in vasoconstriction; (2) blockade of nNO synthesis resulting in an increase in sympathetic tone; and (3) the presence of baroreflexes that restrain the increase in blood pressure that would otherwise result from the first 2 components. These interactions are shown in the following equations for illustration purposes only:

From our experiment we can determine the value of the left side of the equation (“↑BP”) by measuring the increase in blood pressure produced by 250 μg/kg per minute IV of l-NMMA in autonomically intact subjects (+eNO +nNO−baroreflex restraint). We can also measure the increase in blood pressure produced by the same dose in autonomically blocked subjects (“eNO inhibition” in the equation). Baroreflex restraint can be individually estimated in each subject by taking into account the potentiation of the pressor response to phenylephrine during autonomic blockade compared with baseline¹⁵ (baroreflex restraint). We can, therefore, resolve the unknown component of the equation (“nNO inhibition”) to estimate the contribution of nNO to blood pressure regulation (see Results for further explanation).

Protocol 2: To Compare the Importance of eNO on Blood Pressure Regulation in Normal Individuals and Smokers

A parallel group design was used to compare the effects of selective eNOS inhibition between control subjects and smokers. The inclusion criteria for controls were refined to include only normotensive subjects with no hypertensive genetic background (normotensive with no hypertensive parents), or heavy smokers (≥2 packs per day), a group known to have endothelial dysfunction. Subjects were excluded if they met criteria for stage 1 hypertension (SBP >140 or DBP ≥90 mm Hg) or were on any medication. Some volunteers from protocol 1 were also included in this study in the normotensive subjects with no hypertensive genetic background group. Instrumentation and pharmacological testing were identical to the “blocked” day of protocol 1.

Data Acquisition

The surface ECG was amplified, and no additional filters were applied. ECG, blood pressure, and impedance recordings were digitized with 14-bit resolution and 500 Hz sample frequency and recorded using the WINDAQ data acquisition system (DATAQ). Data were analyzed offline using a customized program for data analysis (DIANA, Dr André Diedrich, Vanderbilt University, Nashville, TN) written in PV-Wave (VNI).

Heart Rate and Blood Pressure Variability

A robust QRS detection algorithm, modified from Pan and Tompkins,¹⁶ was used to generate beat-to-beat values. The nonequidistant event time series of RR intervals and blood pressure values were interpolated, low-pass filtered (cutoff: 2 Hz), and resampled at 4 Hz. The estimation of the power spectral density was done by the Welch method, which is a fast Fourier transform–based algorithm. Data segments of 128 s recorded at the end of baseline and were used for spectral analysis. Linear trends were removed, and power spectral density was estimated with the fast Fourier transform–based Welch algorithm using 3 segments of 256 data points with 50% overlapping and Hanning window.¹⁷ The Hanning window was applied to previous estimation of the power spectral density. The power in the frequency ranges for very low frequencies (0.003 to <0.04 Hz), low frequencies (0.04 to <0.15 Hz), and high frequencies (0.15 to <0.40 Hz) were calculated for each interval in according to task force recommendations.¹⁸

Statistical Analysis

Unless otherwise noted, data are presented as mean±SEM. A preliminary comparison of outcomes by study days was performed by Wilcoxon signed-rank test, and comparisons of demographics and outcomes by smoking status were performed using Mann–Whitney U test.

Random-effects models were used to take into account a correlation between measurements taken over time within a subject. The null hypotheses were that there was no difference in outcomes in the response to systemic infusion of l-NMMA between the intact and blocked study days for protocol 1 or between nonsmokers and smokers for protocol 2. The primary end point for both protocols was SBP achieved during interventions. More specific hypothesis tests were performed to answer to the following questions: (1) how the overall dose effects of l-NMMA would differ between the intact and blocked study days for protocol 1 and between nonsmokers and smokers for protocol 2; and (2) if any difference was found between 2 intervention days or smoking status as a function of dose, at what doses the difference has arisen. The major outcomes of interests were systolic blood pressure (SBP) and diastolic blood pressure (DBP). Two random-effects models were built to answer to the above 2 questions for each protocol, model 1 for question 1 and model 2 for question 2. The blocked day indicator (smoking status for protocol 2) and dose (as a continuous variable in model 1 and as indicator variable in model 2) were the main effects, and the interaction between them was also analyzed. To adjust for the potential confounding factors and carryover effects for protocol 1, the random-effects models included the order of intervention, period indicator (first or second period), BMI, sex, and the baseline (for model 2) as covariates. Models for protocol 2 included only pulse pressure, which was found to be significantly associated with smoking status, as covariate. None of the other baseline parameters were included in the model as covariates, considering the limited sample size and that no evidence of association between the outcomes and sex and BMI were seen in final analysis.

All of the tests were 2-tailed, and a P value of <0.05 was considered significant. Analyses were performed with the SPSS statistical software (SPSS version 14.0, SPSS Inc), Stata 9.1 (Stata Corp), and R (www.r-project.org).

Results

Demographics and baseline values for both protocols obtained during the screening visit are shown in Table 1. Smokers in protocol 2 were comparable with controls in age, height, weight, and BMI. They also tend to have lower seated SBP but higher DBP while sitting; hence, pulse pressure was found to be significantly lower in the smokers group (P=0.01 using Mann–Whitney U test). Smokers also had reduced baroreflex sensitivity at baseline, in accordance to previous studies.¹⁹

Effect of Autonomic Blockade on Resting Cardiovascular Parameters and Pressor Response to Phenylephrine

As expected, trimethaphan eliminated heart rate and blood pressure variability, as evidence by a significant decrease in the power spectra parameters (Table 2). There was only a modest decrease in mean arterial blood pressure, from 78±2.1 to 74±2.0 mm Hg (P<0.01), consistent with the low contribution of the sympathetic nervous system to blood pressure in the supine position. In contrast, heart rate increased from 58±2.0 to 84±2.0 bpm (P<0.001) consistent with parasympathetic withdrawal. Baroreflex sensitivity also was abolished, with a reduction in the baroreflex gain from 12±1.7 to 1.8±0.2 ms/mm Hg (P<0.001). In the absence of baroreflex buffering, there was a significant increase in the pressor response to phenylephrine, as evidenced by a decrease in the dose required to increase blood pressure by 20 mm Hg (205±19.6 μg during saline versus 22±1.8 μg during trimethaphan; P<0.005).

Contribution of eNO to Blood Pressure Regulation

For protocol 1, during the blocked day, the increases in both SBP and DBP induced by l-NMMA were higher compared with the intact day (Figure 2 and Table 3). There were also statistically significant linear positive dose effects on both SBP and DBP for the intact day (SBP: P<0.001; DBP: P<0.001) and for the blocked day (SBP: P<0.001; DBP: P<0.001), and the linear dose–response rate in the blocked day was increased 3-fold compared with that of the intact day (SBP, P<0.001; DBP, P<0.001). The mean responses during the blocked day at dose 250 were significantly greater (SBP mean: 20.4 mm Hg; 95% CI: 15.8 to 25.0; P<0.001; DBP mean: 11.2 mm Hg; 95% CI 7.8 to 14.7; P<0.001) compared with those of the intact day after controlling for sequence order, period, BMI, sex, and baseline blood pressure. Although the mean responses on the blocked day at dose 500 were also greater than those of the intact day, they were not statistically significant (SBP: P=0.376; DBP: P=0.389). This could be because of missing measurements at this dose for many subjects in the blocked day because of safety concerns. The missing measurements were informative in the sense that they would have been much larger values than the nonmissing measurements at dose 500, because they were not be able to be taken from subjects who already showed high blood pressures at dose 250, and, hence, the next dose could not be given to them to maintain the blood pressures within the safe limit of 150 mm Hg of SBP.

• Download figure
• Download PowerPoint

Estimation of the Contribution of nNO to Blood Pressure Regulation

To estimate the relative contribution of eNO and nNO to blood pressure, we constructed log dose SBP responses to phenylephrine before and after autonomic blockade in each subject. The potentiation of the pressor effects of phenylephrine observed after autonomic blockade estimates the magnitude of baroreflex buffering¹⁵ and can be used to predict how much blood pressure should have increased if only the baroreflex was removed. An example is given in Figure 3. The measured increase in blood pressure induced by 250 μg/kg per minute of l-NMMA in the presence of an intact autonomic nervous system (eNO+nNO−baroreflex restraint, from the equation shown in the Methods) was 5 mm Hg in this subject (Figure 3, top). This increase is equivalent to that produced by 1.67 log dose phenylephrine. By extrapolation to the phenylephrine dose response curve during autonomic blockade, we can predict that l-NMMA should have produced an increase in blood pressure of 18.2 mm Hg if only the baroreflex restraint was removed (eNO+nNO). A predicted increase in blood pressure during autonomic blockade (eNO+nNO) greater than the measured increase in blood pressure produced by l-NMMA during autonomic blockade (selective eNO) would imply that nNO tonically contributes to blood pressure. However, when individual responses were averaged, we saw no differences between the measured and the predicted responses to l-NMMA during autonomic blockade (Figure 3, bottom), implying that nNO did not significantly contributed to the increase in blood pressure induced by l-NMMA in normal subjects.

• Download figure
• Download PowerPoint

Effect of eNO Inhibition in Smokers

For protocol 2, no differences were found in the decrease in blood pressure induced by trimethaphan in normal controls compared with smokers (Figure 4, left). Baseline blood pressures were not significantly different between nonsmokers and smokers, but SBP was significantly greater in nonsmokers after 250 μg/kg per minute of l-NMMA (P=0.02 by Mann–Whitney U test; Table 4). Analysis using the random-effects model showed a statistically significant linear positive dose effects of l-NMMA on both SBP and DBP for both nonsmokers (SBP: P<0.001; DBP: P<0.001) and smokers (SBP: P<0.001; DBP: P=0.001), but the linear dose–response rate for nonsmokers was increased ≈2-fold compared with that of smokers in SBP (P=0.007) but not in DBP (P=0.493) indicating a greater effect of l-NMMA on nonsmokers. The mean SBP of nonsmokers at dose 250 was increased compared with smokers (mean: 11.6 mm Hg; 95% CI: 0.3 to −23.4; P<0.056), but it was marginally nonsignificant. No difference in the means for DBP between nonsmokers and smokers was found. Statistical analysis was not performed with the higher dose of l-NMMA (500 μg/kg per minute), because it could be given only to 4 subjects in the nonsmoker groups. Blood pressure exceeded our safety limit in the remainder.

• Download figure
• Download PowerPoint

Discussion

We report that endogenous endothelial-derived NO tonically restrains blood pressure in normal subjects by ≥30 mm Hg (Figure 2). We based this conclusion on the increase in blood pressure achieved by systemic blockade of NOS with l-NMMA in the presence of the ganglion blocker trimethaphan. This approach allowed us to observe the full expression of the increase in blood pressure in the absence of the restraining effect of baroreflex mechanisms. Furthermore, the contribution of neurally derived NO to blood pressure regulation is greatly diminished in the presence of autonomic blockade, because it is mediated mostly by its interaction with the autonomic nervous system, which is no longer operative. Therefore, the increase in blood pressure with l-NMMA can be ascribed to preferential inhibition of eNO production. It should be noted that our findings reflect only partial inhibition of eNO, because during ganglionic blockade, we clearly did not reach a maximal dose effect of NOS inhibition. We deemed it unsafe to increase blood pressure >150 mm Hg in our subjects, and this limited the dose of l-NMMA during autonomic blockade. It is likely, therefore, that our results underestimate the real importance of eNO in restraining blood pressure in normal subjects.

There is little doubt that NO tonically restrains blood pressure in normal subjects, but the magnitude of this effect has been difficult to estimate. Previous studies using systemic inhibition of NO production have consistently found an increase in mean arterial pressure of ≈10%, even when l-NMMA was given at higher doses than the one used in the present study.²⁰ Our results indicate that the actual importance of eNO on restraining blood pressure can be gauged if the autonomic nervous system is blocked. Under these conditions, eNO inhibition produced at least a 25% increase in mean arterial blood pressure, for an increase in SBPof ≥27 mm Hg and mean arterial blood pressure of ≥21 mm Hg. An increase in mean arterial blood pressure of similar magnitude was reported by Halliwill et al²¹ in normal subjects infused with l-NMMA after α-adrenergic blockade with phentolamine. Therefore, eNO is arguably the most important metabolic regulator of blood pressure in normal subjects.

We have validated this approach by studying heavy smokers, which are widely accepted to have an impaired production of NO, based on impaired forearm vasodilatory response to acetylcholine,²² impaired flow-mediated dilation,²³ and impaired activity of vascular NOS3.²⁴ As expected, we found that the increase in blood pressure produced by l-NMMA in the presence of trimethaphan was significantly lower in smokers than in normal control subjects, reflecting an impaired contribution of eNO to their blood pressure. Despite the documented eNO deficiency in smokers, it is important to note that they had a normal blood pressure. This has been observed in other studies and implies that other mechanisms compensate for impaired eNO function to maintain blood pressure within a reference range.

There is controversy about the contribution of nNO to blood pressure regulation. We cannot directly measure the importance of nNO on blood pressure, but we can at least provide an estimate after individually taking into account the buffering capacity provided by the baroreflex (see Results for details). Using this approach we did not find a significant contribution of nNO to the increase in blood pressure induced by l-NMMA in normal subjects. This is in agreement with the finding that blood pressure is not increased in mice lacking nNOS.

We do not claim that we have achieved selective inhibition of the NOS3 enzyme over NOS1, merely that the contribution of nNO to blood pressure is minimized under the condition of our experiments, because it is primarily mediated by inhibition of the sympathetic nervous system. We cannot rule out the possibility that NO, synthesized by nNOS, can have a direct effect on vascular tone unrelated to autonomic interactions. Similarly, we cannot exclude a potential role of NO derived from inducible NOS (NOS2). The lack of corroboration of endothelial dysfunction in the smokers group is a limitation of the study that needs to be considered.

The paradigm developed in this study, the use of systemic inhibition of NOS in the presence of autonomic blockade, complements other techniques currently used to evaluate eNO dysfunction in disease states. Flow-mediated dilation is commonly used for this purpose, but measures mostly the role of eNO in conduit arteries and may be influenced by sympathetic tone.¹² Forearm blood flow responses to intrabrachial infusion of endothelial-dependent vasodilators are also used for this purpose. This approach provides important information about endothelial function,²⁵ and forearm responses correlate with those of other vascular beds, most notably the coronary circulation.²⁶ However, none of the vasodilators available act exclusively via NO mechanisms, and it is difficult to quantify from these results the effect that eNO may have to overall blood pressure levels.

Perspectives

We describe a novel approach to preferentially gauge the importance of eNO on blood pressure by measuring the effect of systemic inhibition of NOS during autonomic blockade, thus eliminating the confounding effect of baroreflex mechanisms, and diminishing the role of nNO on blood pressure, which depends on its interaction with the autonomic nervous system. Using this approach, we found that eNO is likely the most important metabolic regulator of blood pressure, tonically restraining it by ≈30 mm Hg or more. On the other hand, we did not find nNO to be as important in the regulation of blood pressure in normal subjects, with the caveat that this conclusion is based on an indirect assessment. Our approach was validated by the observation that the increase in blood pressure during selective eNO inhibition was impaired in smokers, a group known to have eNO deficiency. The paradigm presented here can be used to gauge the importance of eNO in other pathological conditions.

Footnotes