Combined inhibition of nitric oxide and vasodilating prostaglandins abolishes forearm vasodilatation to systemic hypoxia in healthy humans

Arterial catheterization

A 20 gauge, 7.5 cm catheter was placed in the brachial artery of the non-dominant arm under aseptic conditions after administration of local anaesthesia (2% lidocaine) for local infusion of study drugs. The catheter was connected to a pressure transducer which allowed for the constant measurement of mean arterial pressure (MAP). The line was continuously flushed with heparinized saline (Crecelius et al. 2010; Kirby et al. 2010). Arterial blood samples were obtained at specific time points for analysis of blood gases (Fig. 1).

Open in a separate window

Figure 1

Timeline and experimental protocol

A, Overall experimental timeline. Subjects were instrumented with a brachial artery catheter and received continuous phentolamine and propranolol to block α- and β-adrenoceptors (respectively) before they underwent a total of 3 hypoxia trials. Each trial was separated by 30 min of rest. B, Trials 1 and 2 were designed to determine the individual contributions of NO and PGs to hypoxic vasodilatation. These trials began with a 3 min baseline, followed by 2 min transition to the target (∼85%). After 5 min at steady-state hypoxia, l-NAME (NOS inhibitor) or ketorolac (COX inhibitor) were infused for an additional 5 min while the subject remained at ∼ 85%. One group received l-NAME first, the other received ketorolac first, and maintenance doses of the respective inhibitors were continuously infused for the remainder of the study. Trial 3 was designed to determine the hypoxic vasodilator response from rest to steady-state hypoxia after combined blockade of NOS and COX, and therefore no additional drug was given during this trial. This trial consisted of the baseline, transition and 5 min steady-state period. Arterial blood samples (↓) were taken at the end of baseline, steady-state hypoxia and drug infusion for measures of , pH and haemoglobin O₂ saturation.

Systemic isocapnic hypoxia

We employed the use of a self-regulating partial rebreathe system developed by Banzett et al. (2000) to isolate the effects of hypoxia. This system allows for constant alveolar fresh air ventilation independent of changes in breathing frequency or tidal volume (Banzett et al. 2000; Weisbrod et al. 2001; Dinenno et al. 2003). Additionally, this system allows for clamping of end-tidal carbon dioxide levels despite large changes in minute ventilation in response to hypoxia. Oxygen (O₂) levels were manipulated by mixing nitrogen with air via a medical gas blender. The level of O₂ was titrated down to achieve a steady arterial O₂ saturation of ∼85% as assessed by pulse oximetry of the earlobe. Subjects breathed through a scuba mouthpiece with a nose-clip to prevent any nasal breathing. Gas concentrations were monitored at the mouthpiece by an anaesthesia monitor (Cardiocap/5, Datex-Ohmeda, Louisville, CO, USA) that was also used to determine heart rate (HR; 3-lead ECG). Ventilation was measured via a pneumotachograph (model VMM-2a, Interface Associates, Laguna Niguel, CA, USA). Brachial artery blood samples were analysed with a clinical blood gas analyzer (Siemens Rapid Point 400 Series Automatic Blood Gas System, Los Angeles, CA, USA) for , pH and haemoglobin O₂ saturation.

Forearm blood flow and vascular conductance

A 4 MHz pulsed Doppler probe (model 500V, Multigon Industries, Mt Vernon, NY, USA) was used to measure brachial artery mean blood velocity (MBV) proximal to the catheter insertion site. The probe was securely fixed over the brachial artery with an insonation angle of 45 deg as previously described (Dinenno & Joyner, 2003, 2004). A linear 7.0 MHz echo Doppler ultrasound probe (Hewlett-Packard Sonos 4500, Andover, MA, USA) was placed immediately proximal to the velocity probe for measurement of brachial artery diameter. Forearm blood flow (FBF) was calculated as:

where FBF is in ml min⁻¹, MBV is in cm s⁻¹, brachial artery diameter is in cm, and 60 is used to convert from ml s⁻¹ to ml min⁻¹. To take into account any changes in blood pressure, forearm vascular conductance (FVC) was calculated as (FBF/MAP) × 100, and expressed in millilitres per minute per 100 mmHg.

Our goal was to obtain accurate beat-by-beat measurement of MBV for the following reasons: (a) in order to capture any transient drop (compensatory response) to local inhibition of NO and PGs (Schrage et al. 2004) and (b) to make sure we were capturing the drug's peak effect so as to quantify the data accurately. This precluded us from obtaining diameter measurements during the hypoxia trials. However, pilot data from our laboratory indicated no change in brachial artery diameter in response to mild-to-moderate acute isocapnic hypoxia in healthy subjects (data not shown), and our previous studies indicate no change in brachial artery diameter during infusion of NOS and COX inhibitors (Schrage et al. 2004). As a result, we chose to only perform diameter measurements prior to each of the three hypoxia trials and these values were used to calculate FBF and FVC.

To restrict our haemodynamic measures to the forearm circulation, a small blood pressure cuff was placed around the wrist and inflated above systolic pressure to occlude blood flow to the hand circulation during all hypoxia trials. Additionally, a fan was directed toward the experimental forearm to minimize the contribution of skin blood flow to forearm haemodynamics.

Regional sympathoadrenal blockade

Phentolamine mesylate (CIBA Pharmaceutical, Summit, NJ, USA) was administered intra-arterially prior to the experimental trials to inhibit the effects of α-adrenoceptor stimulation in the forearm (200 μg min⁻¹ for 5 min; total loading dose equalled 1000 μg). Maintenance doses were administered at 50 μg min⁻¹ for the duration of the study to ensure continuous blockade. These doses are twice as great as those previously documented to effectively block α-adrenoceptor (Eklund & Kaijser, 1976; Dietz et al. 1997; Halliwill et al. 1997). Propranolol hydrochloride (SoloPak Laboratories, Elk Grove Village, IL, USA) was administered intra-arterially prior to experimental trials to inhibit the effects of β₂-adrenoceptor stimulation in the forearm (200 μg min⁻¹ for 5 min; total loading dose equalled 1000 μg). Maintenance doses were administered at 50 μg min⁻¹ for the duration of the study to ensure continuous blockade. This dose of propranolol was shown to inhibit the forearm vasodilatory response to the non-selective β-adrenoceptor agonist isoproteronol (Johnsson, 1967), as well as attenuate the observed decrease in vascular resistance in the resting forearm during contralateral isometric handgrip exercise (Eklund & Kaijser, 1976).

Regional NO synthase (NOS) inhibition

N^G-Nitro-l-arginine methyl ester (l-NAME; Aerbio/Clinalfa, Darmstadt, Germany) was administered intra-arterially at 5 mg min⁻¹ for 5 min to inhibit NOS and the production of NO (total loading dose of 25 mg). Maintenance doses were administered at 1.25 mg min⁻¹ for the duration of the study to ensure continuous blockade. This dose of l-NAME has been previously shown to significantly reduce basal tone as well as block the vasodilatory effects of acetylcholine (Dinenno & Joyner, 2003; Schrage et al. 2004), consistent with effective NOS inhibition.

Regional cyclooxygenase (COX) inhibition

Ketorolac (trade name Toradol, Abbott Laboratories, Abbott Park, IL, USA) was administered intra-arterially at 600 μg min⁻¹ for 5 min to inhibit COX and the production of PGs (total loading dose of 3 mg). Maintenance doses were administered at 150 μg min⁻¹ for the duration of the study to ensure continuous blockade. Previously, the efficacy of this dose of ketorolac has been demonstrated by measurements of arterial and deep venous PGF_1α (a stable breakdown product of PGs) showing a decrease in circulating levels after 3 min of handgrip exercise compared to the exercise control condition (Dinenno & Joyner, 2004). Further, this dose of ketorolac caused a transient but consistent reduction in FBF during exercise hyperaemia indicating the drug was effective at inhibiting COX (Schrage et al. 2004).

Experimental protocol

Prior to the initiation of the experimental trials involving systemic hypoxia, combined blockade of α- and β-adrenoceptors (via phentolamine and propranolol) was performed to eliminate sympathoadrenal influences on the vasculature, and thus isolate local influences on vascular tone. Maintenance doses of phentolamine and propranolol were infused for the remainder of the study to ensure continuous blockade.

An initial rest period of 3 min was given to establish baseline values while the subject breathed on the mouthpiece open to room air. For all trials the level of inspired O₂ was then titrated to achieve a of ∼85% as established by pulse oximetry on the earlobe. This transition from baseline to the target level of hypoxia typically occurred in 2 min. The subject was then maintained at ∼85% for 5 min to establish steady-state (SS) levels of , , as well as forearm haemodynamics during hypoxia. At this time point (for Trials 1 and 2), the subject was given either l-NAME (n = 10) or ketorolac (n = 10) for a total of 5 min while remaining at ∼85%. A maintenance dose of the infused drug was given for the remainder of the study. After a 30 min rest period, we repeated a second hypoxia trial (identical to Trial 1) where the subject received the other pharmacological inhibitor. Drug order was randomized and counterbalanced across all subjects, and maintenance doses of both drugs were given for the remainder of the study. After an additional rest period of 30 min following Trial 2, the subject underwent a third bout of hypoxia while receiving maintenance doses of all drugs. Figure 1 summarizes the experimental protocol in further detail. Trial 1 was designed to establish the normal hypoxic vasodilator response (rest to SS hypoxia) as well as determine the independent effects of either NO or PG inhibition on vascular tone during SS hypoxia. Trial 2 was designed to determine the influence of single NO or PG inhibition on the vasodilator response from rest to SS hypoxia, as well as the interactive effects of these two vasodilating substances on vascular tone during SS hypoxia. Finally, Trial 3 was designed to determine whether combined NO and PG inhibition further reduces the vasodilatation to acute systemic hypoxia (rest to SS hypoxia).

FBF and FVC responses to α- and β-adrenoceptor blockade

Combined α- and β-adrenoceptor blockade significantly increased basal FVC in the experimental forearm of both the l-NAME 1st and Ketorolac 1st groups by ∼90% (P < 0.05) prior to any hypoxia trials. Table 1 summarizes the baseline forearm haemodynamic responses in both groups before and after sympathoadrenal blockade, as well at each key time point during the experimental trials.

Effects of acute systemic hypoxia on blood gases, ventilation and systemic haemodynamics

There were no significant differences between the l-NAME 1st and the Ketorolac 1st group in any blood gas variables, ventilation, or systemic haemodynamics at rest or during hypoxia. In general, decreased while ventilation and HR increased significantly compared with baseline for all trials. MAP tended to increase during hypoxia, but this was not significant except for the l-NAME 1st group in Trial 3. Table 2 further summarizes these responses to acute systemic hypoxia.

Table 2

Effects of hypoxia on blood gases, ventilation and systemic haemodynamics

		Trial 1			Trial 2			Trial 3
		Baseline	SS	Drug	Baseline	SS	Drug	Baseline	SS
HB Sat (%)	LK	96 ± 1	84 ± 1*	83 ± 1*	95 + 1	83 ± 1*	82 ± 1*	95 ± 1	82 ± 1*
	KL	96 ± 1	84 ± 1*	84 ± 1*	96 ± 1	84 ± 1*	84 ± 1*	95 ± 1	84 ± 1*
(%)	LK	98 ± 1	84 ± 1*	83 ± 1*	98 ± 1	82 ± 1*	83 ± 1*	98 + 1	83 ± 1*
	KL	98 ± 1	82 ± 1*	82 ± 1*	98 ± 1	83 ± 1*	83 ± 1*	98 ± 1	83 ± 1*
(mmHg)	LK	88 ± 3	48 ± 1*	47 ± 2*	84 ± 3	48 ± 1*	48 ± 1*	84 ± 2	48 ± 1*
	KL	88 ± 3	48 ± 1*	48 ± 2*	89 ± 3	49 ± 1*	50 ± 2*	88 ± 3	49 ± 1*
(mmHg)	LK	38 ± 1	38 ± 1	39 ± 1	38 ± 1	38 ± 1	38 ± 1	39 ± 1	38 ± 1
	KL	39 ± 1	38 ± 1	39 ± 1	38 ± 1	38 ± 1	39 ± 1	38 ± 1	39 ± 1
pH	LK	7.44 ± .01	7.44 ± .01	7.43 ± .01	7.42 ± .01	7.43 ± .01	7.42 ± .01	7.42 ± .01	7.42 ± .01
	KL	7.46 ± .02	7.46 ± .02	7.46 ± .02	7.45 ± .02	7.46 ± .02	7.45 ± .02†	7.44 ± .01	7.45 ± .02
Minute vent. (l min−1)	LK	7.2 ± 1.2	11.5 ± 1.9*	10.0 ± 1.8*	6.7 ± 0.8	14.0 ± 2.2*	10.0 ± 1.3*†	6.8 ± 0.9	14.7 ± 1.8*
	KL	7.3 ± 0.8	13.6 ± 2.5*	12.7 ± 2.3*	7.3 ± 0.9	10.6 ± 2.2*	10.2 ± 1.7*	7.0 ± 1.0	12.5 ± 1.7*
HR (beats min−1)	LK	57 ± 3	68 ± 4*	65 ± 4	52 ± 2	66 ± 3*	63 ± 3*	52 ± 2	65 ± 3*
	KL	55 ± 1	67 ± 3*	65 ± 3*	52 ± 1	66 ± 3*	63 ± 3*	50 ± 2	64 ± 3*
MAP (mmHg)	LK	95 ± 2	98 ± 2	98 ± 3	103 ± 3	104 ± 4	105 ± 3	102 ± 4	107 ± 4*
	KL	96 ± 2	98 ± 3	96 ± 3	97 ± 3	98 ± 3	100 ± 3	101 ± 2	102 ± 3

Open in a separate window

HB Sat, haemoglobin saturation; , arterial oxygen saturation; , partial pressure of O₂; HR, heart rate; , end-tidal CO₂; vent., ventilation; MAP, mean arterial pressure; *P < 0.05 vs. baseline; †P < 0.05 vs. SS hypoxia; LK, l-NAME first, ketorolac second; KL, ketorolac first, l-NAME second.

FVC responses to acute systemic hypoxia: l-NAME 1st group

In Trial 1, SS hypoxia while under local sympathoadrenal blockade significantly increased FVC from baseline in the experimental forearm prior to the infusion of l-NAME (ΔFVC = 21 ± 3%; P < 0.05 vs. baseline; Fig. 2A). After 5 min of SS hypoxia, inhibition of NOS via l-NAME significantly reduced FVC by −33 ± 3% (P < 0.05 vs. zero; Fig. 3A). In Trial 2 with continuous NOS inhibition, the increase in FVC from baseline to SS hypoxia was similar to that in Trial 1 (20 ± 3%; Fig. 2A). After 5 min of SS hypoxia, inhibition of PGs via ketorolac significantly reduced FVC by −15 ± 3% (P < 0.05 vs. zero; Fig. 3B). In Trial 3 with combined inhibition of NOS and COX, the increase in FVC from baseline to SS hypoxia was abolished (3 ± 3%; NS vs. zero; Fig. 2A). Absolute levels of forearm haemodynamics are presented for each trial in Table 1.

Open in a separate window

Figure 2

Forearm vasodilator responses to systemic hypoxia

Percentage changes in vascular conductance from baseline to steady-state hypoxia for l-NAME 1st group (A) and Ketorolac 1st group (B). Black bars indicate the Control condition (Trial 1), grey bars indicate the single inhibition trial (Trial 2), and open bars indicate the combined inhibition trial (Trial 3). *P < 0.05 vs. zero; †P < 0.05 vs. control and single inhibition (Trial 1 and 2, respectively).

Open in a separate window

Figure 3

Effects of NO and PG inhibition on steady-state forearm vascular tone

Percentage changes in vascular conductance during steady-state hypoxia to l-NAME (A) and Ketorolac (B) when given independently (1st; grey bars) vs. with prior NOS or COX inhibition (2nd; open bars). *P < 0.05 vs. steady-state hypoxia, †P < 0.05 vs. independent COX inhibition (KET 1st) Trial.

FVC responses to acute systemic hypoxia: Ketorolac 1st group

In Trial 1, SS hypoxia while under local sympathoadrenal blockade significantly increased FVC from baseline in the experimental forearm prior to the infusion of ketorolac (ΔFVC = 24 ± 5%; P < 0.05 vs. baseline; Fig. 2B). After 5 min of SS hypoxia, inhibition of COX via ketorolac did not significantly reduce FVC (−4 ± 3%; P = 0.35 vs. zero; Fig. 3B). In Trial 2 with continuous COX inhibition, the increase in FVC from baseline to SS hypoxia was similar to that in Trial 1 (27 ± 4%; Fig. 2B). After 5 min of SS hypoxia, inhibition of NOS via l-NAME significantly reduced FVC by −36 ± 7% (P < 0.05 vs. zero; Fig. 3A). In Trial 3 with combined inhibition of COX and NOS, the increase in FVC from baseline to SS hypoxia was abolished (3 ± 2%; NS vs. zero; Fig. 2B). Thus, regardless of drug order, combined blockade of NO and PGs abolished the vasodilatory response in both groups to hypoxia in Trial 3. Absolute levels of forearm haemodynamics are presented for each Trial in Table 1.

Contributions of NO to the hypoxic vasodilator response in humans: rest to SS hypoxia

It has previously been demonstrated that acute NOS inhibition via intra-arterial infusion of N^G-monomethyl-l-arginine (l-NMMA) attenuates the hypoxic vasodilatation observed in the forearm circulation of healthy humans (Blitzer et al. 1996; Casey et al. 2010). Interestingly, Weisbrod et al. (2001) demonstrated that acute β₂-adrenoceptor blockade (via propranolol) reduced the vasodilatory response by ∼50% during a similar level of hypoxia as used in the current study, yet this response was not further reduced with the addition of the NOS inhibitor l-NMMA. This latter observation led to the hypothesis that local β₂-adrenoceptor stimulation via circulating adrenaline is the primary mechanism for elevated NO synthesis during hypoxia in humans (Weisbrod et al. 2001). In the present study under conditions of sympathoadrenal blockade (via phentolamine and propranolol), infusion of the NOS inhibitor l-NAME resulted in a significant reduction in resting FVC. However, despite these baseline changes in FVC, the vasodilatory response from rest to SS hypoxia was similar to that in control conditions (∼25% increase in FVC in both trials), consistent with the results from Weisbrod et al. (2001). Although these data could be interpreted that NO synthesis during systemic hypoxia is downstream of β₂-adrenoceptor stimulation, an alternative explanation might be that NO does play a role in the vasodilatory response from rest to hypoxia, but the synthesis of another substance was augmented under these conditions and compensated for NOS inhibition thereby evoking ‘normal’ vasodilatation (see below).

Contributions of PGs to the hypoxic vasodilator response in humans: rest to SS hypoxia

To the best of our knowledge, the contribution of PGs to hypoxic vasodilatation has not yet been explored in humans and, as such, this study marks the first attempt to understand their role under these conditions. There is evidence from isolated vessels of experimental animals to suggest that PGs are involved in hypoxic vasodilatation (Busse et al. 1984; Messina et al. 1992; Goodwill et al. 2008), and infusion of a COX inhibitor into the rat hindlimb attenuated the vasodilatation during systemic hypoxia (Ray et al. 2002). Additionally, Ray and colleagues (2002) demonstrated a unique interaction between PGs and NO synthesis during hypoxia, such that locally released adenosine acts on adenosine (A₁) receptors on the endothelium to increase PG synthesis, and the subsequent increase in intracellular cyclic AMP stimulates NO synthesis which in turn contributes to the hypoxic vasodilatation. In the present study, we found that individual blockade of PGs with ketorolac did not impair the vasodilatory response from baseline to SS hypoxia (∼25% increase in FVC in both trials). Further, the magnitude of reduction in forearm vascular tone during hypoxia with NOS inhibition was not impacted by prior inhibition of COX (∼35% in both conditions). Our results differ from these studies utilizing isolated vessels from animals and in the rat hindlimb circulation in that we found a non-obligatory role for PGs under normal conditions in mediating the hypoxic vasodilator response in humans, and that PG synthesis was not required to stimulate NO synthesis. It is at present unclear why this is the case, but species and/or methodological differences could contribute to the disparate observations.

Combined inhibition of NO and PGs abolishes hypoxic vasodilatation in humans

Many studies under a variety of experimental conditions and models have demonstrated a unique interplay between NO and PGs in the regulation of vascular tone (Sun et al. 1999; Ray et al. 2002; Schrage et al. 2004). As such, in the present study we determined whether these putative endothelium-dependent vasodilators act in a compensatory manner and evoke ‘normal’ vasodilatation when one pathway is inhibited. As discussed above, individual inhibition of NO or PGs did not impact that ability of the forearm vasculature to vasodilate from rest to SS hypoxia. In stark contrast, combined blockade of NO and PGs abolished the vasodilatory response to SS hypoxia seen in Trial 3 in both groups of subjects (Fig. 2). This was independent of drug order as there was a non-significant increase in FVC of ∼3% to hypoxia in both the l-NAME 1st and Ketorolac 1st groups. These data provide the first evidence that NO and PGs operate within a redundant system to regulate vascular tone under hypoxic conditions. Collectively, our data indicate that these two endothelium-dependent vasodilators act synergistically and together contribute largely to the forearm vasodilatation during systemic hypoxia in healthy humans.

Contribution of NO and PGs to vascular regulation during steady-state hypoxia

The present study was designed to not only allow us to determine the roles of NO and PGs in mediating hypoxic vasodilatation (from rest to SS hypoxia as discussed above), but also to determine how NO and PGs interact to regulate vascular tone during hypoxia. When l-NAME was infused to inhibit NOS during SS hypoxia, we found a significant decrease in FVC whether the drug was given prior to (Group 1) or subsequent to (Group 2) COX inhibition (–33%vs.–36%, respectively). Conversely, COX inhibition via ketorolac did not influence FVC during SS hypoxia under normal conditions (Group 2; ΔFVC ∼−4%); however it did significantly reduce FVC when performed subsequent to NOS inhibition (Group 1; ΔFVC ∼−15%). This observation that COX inhibition significantly reduced forearm vascular tone during hypoxia only after NOS inhibition is worthy of discussion. Given that the hypoxic vasodilator response was preserved in Trial 2 after l-NAME infusion (Group 1; rest to SS hypoxia), this could imply that PGs were augmented during NOS inhibition and as such contributed in a compensatory manner to maintain vasodilatory responses to hypoxia. Interestingly, the opposite was not true in the case of PG inhibition. The lack of an effect of independent COX inhibition on SS hypoxic FVC could indicate two possibilities: (a) that PGs are not contributing to hypoxic vasodilatation under normal conditions, or (b) some other vasodilator is capable of compensating immediately for the loss of PGs during the hypoxic stimulus. If there was a compensatory response, it does not appear to be the result of augmented NO synthesis as we found l-NAME had similar effects on SS hypoxic vascular tone whether given prior to or subsequent to COX inhibition.

Experimental considerations

In many human studies employing the use of pharmacological inhibitors to probe physiological control mechanisms, it is difficult to assess the effectiveness of the drugs used. One could argue that the reason we observed such a marked attenuation in the vasodilatory response in Trial 3 was due to incomplete α- and β-adrenoceptor blockade. We do not believe that this was the case for β-blockade with propranolol as its efficacy has been demonstrated previously (Johnsson, 1967), it has an extremely long half-life, and we infused maintenance doses throughout the entire protocol. Further, if β-adrenoceptors were not completely blocked we would have expected to observe more vasodilatation in Trial 3, which was clearly not the case. Additionally, we tested a bolus infusion of the α-blocker (phentolamine) at the end of the protocol (extending SS hypoxia by 3 min in Trial 3) in a subgroup of subjects and did not find any differences in FBF or FVC. Therefore, we believe that we were successful in maintaining sympathoadrenal blockade throughout the entire protocol, and as such, this should not influence the interpretation of our data. Finally, repeated trials of systemic hypoxia have been shown to evoke repeatable forearm vasodilatory responses (Weisbrod et al. 2001), and thus the abolished response in Trial 3 cannot be explained simply by repeated exposure to the hypoxic stimulus.

In our recent studies designed to understand how NO and PGs contribute to exercise hyperaemia in humans, we performed blockade of NOS and COX during the same SS exercise conditions (Schrage et al. 2004, 2007; Crecelius et al. 2010), whereas in the present study we performed single inhibition during SS hypoxia then allowed subjects to return to resting normoxia conditions. We chose this approach to allow us to determine how NO and/or PGs contribute to vascular tone during hypoxia, as well as whether either substance was obligatory to observe the ‘normal’ vasodilatation from rest to hypoxia. This was of particular interest in the present study because no studies to date have determine what role, if any, PGs play in hypoxic vasodilatation in humans. In our previous study during exercise in young humans (Schrage et al. 2004), we observed a transient reduction in SS FVC during 5 min of ketorolac infusions (with or without concurrent NOS inhibition), and thus if the same compensatory response existed during hypoxia we would have been able to observe this in the present study. An additional consideration with respect to the present study design involves the reduction in baseline vascular tone observed during NOS and combined NOS/COX inhibition. We do not believe that this limits the interpretation of our data as (1) it has previously been demonstrated that infusion of phenylephrine (an α₁-adrenoceptor agonist) to reduce resting vascular tone to the same extent as observed during NOS inhibition does not impact the vasodilatory response to hypoxia (Blitzer et al. 1996), and (2) this cannot explain our observations that combined NOS and COX inhibition abolished the hypoxic vasodilatation (as quantified by % FVC) in all subjects.

The finding that inhibition of COX reduced FVC during SS hypoxia after prior inhibition of NOS could indicate that the synthesis of PGs was augmented under these conditions in an attempt to compensate for the loss of NO. We can only speculate that this occurred, however, since we did not perform any direct measurements of PGs or PG metabolites in the blood and demonstrate they were correspondingly elevated under these conditions.

Perspectives

The topic of NO and PG interaction in the regulation of vascular tone has recently been of great interest. It appears that these two endothelium-dependent vasodilators, in combination, are vital in the normal local vasodilator responses to acute systemic hypoxia in humans. As such, we speculate that our findings might have significant implications for the regulation of blood flow and oxygen delivery in populations that demonstrate endothelial dysfunction. For example, chronic pathological conditions such as diabetes, obstructive sleep apnoea, congestive heart failure, and even healthy older adults demonstrate impaired endothelial function, which is often due to reductions in NO and perhaps prostacyclin bioavailability (Feletou & Vanhoutte, 2006). However, it is also possible that under these pathological conditions, other pathways or signals (endothelium-derived hyperpolarizing ‘factors’) might be augmented in an attempt to compensate for the reduction in NO and vasodilating PG bioavailability (Taddei et al. 1999). Further investigations will be needed in order to understand the specific roles of these endothelium-dependent vasodilators in the regulation of blood flow and oxygen delivery during hypoxia in these populations that exhibit endothelial dysfunction.

We would like to thank the subjects who volunteered to participate in this study. This research was supported by National Institutes of Health awards, AG022337, HL087952 and HL095573 (F.A.D.).