INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

NITRIC OXIDE (NO) is produced by the enzyme NO synthase (NOS) in many cells of the respiratory tract and plays an important role in host defense, regulation of ciliary activity and pulmonary blood flow, and possibly the matching of perfusion with ventilation in the lung (15). NO is a unique respiratory tract product, since it is gaseous and is produced by the epithelial cells lining the conducting airways and the nose (1, 8, 24, 30, 31, 34). The function of NO in the nose is not known, but it may play a role in host defense, and the nose may act as a source of pulmonary NO during nasal inspiration (25).

NO produced by NOS in airway walls may be excreted into the lumen in the gas phase, or it may diffuse submucosally, where it will combine avidly with heme (17) and other reactants (9). The exchange dynamics of NO are therefore very different from those of well-studied biological gases, such as CO₂, which is excreted from the alveolar region of the lung. An understanding of these dynamics is fundamental to the development of our knowledge of the functions of NO in the respiratory tract.

Several features of these dynamics have been described. It is known that during breath holding, with no luminal gas flow, NO accumulates in the nose and in the bronchi (21), indicating that the rate of NO production exceeds the rate of submucosal diffusion at both sites, although accumulation of NO is more rapid in the nose (3, 21). When gas flows in the nose and bronchi, NO excretion on the luminal side increases (5, 11, 36). This feature of NO exchange dynamics has been suggested to occur as luminal gas flow leads to a lower luminal NO concentration ([NO]), in turn altering the equilibrium between luminal NO excretion and submucosal diffusion (36). There is no experimental evidence to support or refute this hypothesis. An alternative, and perhaps more intriguing, hypothesis would be that luminal gas flow directly affects NOS activity in epithelial cells, in an analogous fashion to luminal blood flow increasing NOS activity in vascular endothelial cells (28).

Pulmonary NO production is also known to be dependent on inspired oxygen tension (6, 32), presumably because oxygen is an NOS substrate (6, 22), and a recent study has suggested a similar relationship in the nose (13). Although some of the features of respiratory tract NO exchange dynamics have been studied in the lung, this approach is hampered by the complexity of pulmonary anatomy and physiology (40). Because the nasal cavity is of relatively fixed volume and is easily isolated and accessed, we sought to further define the exchange dynamics of NO in healthy human subjects in this organ as a first step in more accurately describing NO exchange in the bronchial tree. The specific aims of this study were to 1) determine whether nasal NO output is a positive function of luminal gas flow rate, 2) develop a mathematical model for use in the nose to determine the diffusing capacity for NO (D_NO) and the mucosal [NO] [wall concentration (C_w)], 3) apply that model to determine D_NO and C_w in healthy human subjects, 4) determine D_NO and C_w at different luminal gas flow rates to define the mechanism of aim 1 (above) and 5) determine the relationship between luminal oxygen tension and nasal NO output.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Apparatus. A schematic for the apparatus used in the experiments is presented in Fig. 1. Gas was available from two cylinders, each fitted with a two-stage regulator; the gases were blended using digital mass flow-controlled valves (model 5850S, Brooks Instruments). Operating pressure was 1 bar. These valves allow the delivery of flows of 0-1,000 and 0-2,000 ml/min, respectively, with tolerance of <1% of maximum flow. For experiment 1, a different valve, which delivered a flow rate of 0.5-5 l/min (Rotork Analysis), was used. After it was blended, the gas mix was delivered via polytetrafluoroethylene tubing to the right nostril via a tight-fitting nasal olive. The [NO] in this gas is referred to as [NO]_in. Another tight-fitting nasal olive was applied to the left nostril and connected via a T piece to the sampling port of the chemiluminescence analyzer, where [NO]_out was determined.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Apparatus used in the experimental series. NO, nitric oxide.

Chemiluminescent analyzer. Concentrations of NO were detected using chemiluminescence (Logan 2000, Rochester, Kent, UK) in accordance with European Respiratory Society and American Thoracic Society guidelines (19, 35). The analyzer has a sensitivity of 0.3 ppb and is able to analyze at 25 Hz for up to 80 s continuously and in real time. Concurrent analysis of CO₂ concentration is available, and the analyzer has a response time of <2 s. The sampling flow rate used during these experiments was ~4 ml/s. The analyzer was calibrated daily using a known standard (99 ppb NO in nitrogen) and zeroed before each experimental run.

Isolation of nasal cavity. The nasal cavity was effectively isolated through voluntary elevation of the soft palate during breath holding. The application of tight-fitting olives to the nares completed the isolation procedure. Soft palate closure was confirmed through the exclusion of CO₂, and the adequacy of fit of the olives was confirmed by the rapid attainment of a stable plateau for NO (19, 35). Deliberate relaxation of the soft palate led to immediate release of CO₂ into the sample, and deliberate dislodgement of the nasal olives led to immediate deterioration in the stability of [NO]_out. These parameters were therefore carefully monitored to ensure effective isolation of the nasal cavity. Some subjects were unable to satisfactorily complete this maneuver and were excluded from further study. Subjects were asked to maintain breath holding and soft palate closure for 35 s. Equilibration occurred during the first 10-15 s of breath hold. [NO] for each experimental run was calculated from the mean of four measurements taken at 15, 20, 25, and 30 s after the commencement of breath hold.

Experiment 1: effect of luminal gas flow rate on nasal NO output. Six healthy, nonsmoking subjects (25-35 yr old, 3 men and 3 women), all of whom were able to complete the nasal isolation maneuver, were recruited. Inasmuch as nasal NO output increases within 3 wk of common cold infection (27), subjects who had had an upper respiratory tract infection in the previous 6 wk were excluded from further study, as were those taking any form of medication. Only one of the cylinders was used for this experiment, and gas blending was not necessary (Fig. 1). Medical air (21% oxygen-79% nitrogen, <1 ppb NO) was infused into the right nostril at 10 different gas flow rates in random order (0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, and 5 l/min), and the gas escaping from the left nostril was analyzed for NO ([NO]_out). Each experimental run was completed six times for each subject at each of the 10 flow rates. Nasal NO output (nl/min) was calculated from NO_out (ppb) × gas flow rate (l/min), according the recommendations of the American Thoracic Society (35).

Experiment 2: effect of luminal NO concentration on [NO]_out. Eight healthy, nonsmoking subjects (20-31 yr old, 5 men and 3 women) were recruited. Again, none had had an upper respiratory tract infection for the previous 6 wk or were taking any medication, and all were able to satisfactorily complete the nasal cavity isolation maneuver. NO (49 ppm) in nitrogen and medical air (21% oxygen) were used. The gases were blended to produce air with nominal [NO] of 0, 500, 1,000, and 2,000 ppb at each of two total flow rates (1 and 2 l/min). Because the flow rate for the NO in nitrogen gas was very low compared with the flow rate of the medical air, the final oxygen concentration varied little from 21%. At the lowest [NO], the mass flow-controlled valve was operating near the limit of its tolerance, so before each experimental run, the actual [NO] delivered ([NO]_in) was measured over 30 s by direct sampling from the right nasal olive and verified to be stable before introduction of the gas mix into the right nostril. The eight different flow-[NO]_in combinations were applied to the right nostril of each subject, each for three experimental runs, in random order. For each subject, NO_out was plotted against NO_in.

(1)

1

1

(2)

(3)

Experiment 3: effect of luminal oxygen concentration on nasal NO output. Oxygen and nitrogen were blended to give gas mixes with final oxygen tensions of 0, 2.4, 4.9, 7.4, 9.9, 14.9, 20.9, 29.7, 49.8, 69.7, 89.9, and 100% at a flow rate of 2 l/min, each with [NO]_in < 1 ppb. Nine subjects (24-36 yr old, 4 men and 5 women) were recruited for the study. All fulfilled the entry criteria described for experiments 1 and 2. [NO]_out was measured during three experimental runs at each of the 12 different oxygen tensions, again in random order, for each subject. The nasal NO output (nl/min) was calculated according to American Thoracic Society guidelines, as for experiment 1, from [NO]_out (ppb) × gas flow rate (l/min).

Statistics. For experiments 1 and 3, data are presented as means and 95% confidence intervals. An overall mean NO output has been calculated from the data for all subjects in these two experiments. For experiment 2, differences in D_NO and C_w at the two luminal flow rates were assessed for statistical significance using a paired t-test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Most subjects found the breath-holding maneuver easy to perform. The measurement of [NO]_out was straightforward and reproducible using this method, with the rapid attainment of a steady state even at the lowest flow (0.5 l/min). The individual coefficients of variation ranged from 1.82% to 4.99%. The mean nasal NO output for experiments 1 and 3 was 334 nl/min.

Experiment 1: effect of luminal gas flow rate on nasal NO output. [NO]_out was >100 ppb at the lowest flow (0.5 l/min) but declined as the flow of gas into the right nostril increased (Fig. 2A). However, as shown in Fig. 2B, when dilutional effects due to the higher gas flows were taken into account, the nasal output for NO (nl/min) increased with increasing gas flow. There was considerable between-subject variation in nasal NO output. At 2 l/min, for instance, the nasal NO output varied from 144 nl/min in one subject to 273 nl/min in another.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   A: concentration of NO at the left nostril ([NO]out) with infusion of medical air at flow rates of 0.5-5 l/min. ppb, Parts/billion. B: nasal NO output with infusion of medical air at flow rates of 0.5-5 l/min. Values are means ± 95% confidence intervals (n = 6).

Experiment 2: effect of luminal NO concentration on [NO]_out. When nonzero [NO] were introduced into the isolated nasal cavity, [NO]_out was a linear function of [NO]_in in all subjects (Fig. 3). Because this function was found to be universally linear, we were able to apply the model developed by Silkoff et al. (37) to determine nasal D_NO and nasal C_w at each flow rate for each subject (Table 1) using Eq. 3. The mean nasal D_NO was 1.80 × 10³ and 1.86 × 10³ nl · s¹ · ppb¹ when luminal flow was 1 and 2 l/min, respectively (P = 0.78). The mean nasal mucosal [NO] (C_w) declined with increasing luminal gas flow from 4,935 ppb at 1 l/min to 2,511 ppb at 2 l/min (P = 0.01). There was wide intersubject variation in C_w, and although all values were lower at higher flow, the difference in C_w between the two flows also showed wide variation (Table 1).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   [NO]out (ppb) (y-axes) with nominal NO concentration delivered to the right nostril ([NO]in) of 0, 500, 1,000, and 2,000 ppb (x-axes) at flow rates of 1 (A) and 2 l/min (B, n = 8). The linear relationship in each case allows the application of the mathematical model and determination of diffusing capacity of NO and mucosal NO concentration.

Experiment 3: effect of luminal oxygen concentration on nasal NO output. Figure 4 displays the effect of alterations in luminal oxygen tension on nasal NO output. Output decreased with decreasing oxygen tension, although most of this effect was noticeable at oxygen tensions <10%.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   Effect of luminal oxygen tension on nasal NO output. Values are means ± 95% confidence intervals (n = 9).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A number of conclusions regarding the nasal exchange dynamics of NO in healthy adults can be drawn from the results of this study. With increasing luminal gas flow rate, [NO] emitted from the nose declines. However, when the NO output (in nl/min) is calculated from the emitted concentration × gas flow rate, it is found to increase with increasing luminal gas flow rate, an effect previously described (5, 11). Nasal NO output ranged between ~100 and 500 nl/min in the three experiments, with an overall average of 334 nl/min. When nonzero [NO] are infused into the isolated nasal system at a particular flow rate, D_NO and C_w can be easily calculated using the mathematical model. D_NO does not vary with a doubling of luminal gas flow rate, but C_w declines roughly by one-half. Nasal NO output declined with decreasing luminal oxygen tension, an effect especially apparent at oxygen tensions <10%.

Before the findings are discussed, it is prudent to identify possible sources of variation in this study. A brief consideration of nasal anatomy and physiology is worthwhile. Inasmuch as the nasal cavity is bony, it is of relatively fixed volume. However, because the submucosal area is richly perfused with a venous vascular network, volume can, and indeed does, change. In some people, this happens in a cyclical fashion over hours. As the venous network dilates, the lumen of the nasal cavity shrinks, and as the network constricts, the lumen increases in volume (7). This feature of nasal physiology can potentially influence NO physiology in two ways. First, an increase in luminal volume tends to reduce luminal [NO] through simple dilution, with the assumption that NO production remains constant. Second, a decrease in submucosal blood reduces the availability of heme, with which NO reacts avidly (14), and so potentially increases the amount of NO excreted into the nasal lumen. Although these opposite effects will tend to oppose one another, Giraud et al. (10) demonstrate that the effect of nasal volume is probably greater. This group applied the topical vasoconstrictor oxymetazoline to the nasal mucosa of human volunteers and found that luminal [NO] decreased, an effect of the increase in nasal volume, rather than an effect of constriction in the volume of the vascular bed.

To simplify matters in terms of NO physiology, however, this cyclical change in nasal volume tends to occur in a see-saw fashion, with the left- and right-sided nasal volumes changing in opposite directions, so that the overall nasal volume remains relatively constant (7). Qian et al. (29) demonstrated that the effect of the nasal cycle on nasal NO output is minimal, and two groups have shown that nasal NO output does not vary between nostrils in healthy subjects (27, 29). Although changes in luminal temperature have no effect on nasal NO output (10), deep inhalation and breath holding tend to decrease the nasal resistance (i.e., increase luminal volume) (38), although this is unlikely to lead to bias in our study, since all experiments were conducted with the same deep inspiration and breath-holding maneuver. Because the nasal cavity is in communication with the paranasal sinuses, it is possible that some NO detected at the nares originates in the sinuses. High [NO] values have been found during direct sampling of the paranasal sinuses (12, 23); however, Haight et al. (12), in a study specifically designed to determine the relative contributions of the nose and paranasal sinuses to the NO appearing at the nares, found that 88% was produced in the nasal cavity itself. For all the subjects in this study, a stable [NO]_out was obtained within 10 s. The rapidity of this response suggests a relatively simple NO-producing nasal cavity architecture, in keeping with the findings of Haight et al. Because, in the systemic vasculature, endothelial cells produce NO in response to shear stress (39), it is possible that changes in nasal pressure may alter NO production by nasal epithelial cells. However, changes in nasal pressure in our experiments are likely to be small, since the system is open, and previous work has demonstrated no effect of changes in airway pressure on lower respiratory tract NO output (2, 16).

The [NO] obtained at the left nostril and the calculated nasal NO outputs in our studies are in good agreement with those reported by other authors (5, 10, 18, 21). Imada et al. (18) found the average nasal NO output to be 323 nl/min in healthy men, and Giraud et al. (10) describe an average of 540 nl/min, whereas Kimberly et al. (21) found an average of 468 nl/min, compared with our overall average of 334 nl/min.

The relationship between nasal NO output and luminal gas flow rate has been studied systematically by others. Imada et al. (18) found a hyperbolic relationship between the [NO] collected in a bag after room air was pumped through the isolated nasal airway at flow rates between 0.5 and 2.0 l/min, leading them to conclude that nasal NO output is constant across different gas flows. It may have been that this range of flows was too narrow for detection of an effect of luminal gas flow on nasal NO output. On the other hand, Dubois et al. (5), using an experimental design similar to ours, demonstrate NO output as a positive function of luminal gas flow, but at a range of flows lower than those used in our study (8.2-347 ml/min), and Giraud et al. (11) also demonstrated a positive relationship, but at flows from 1 to 22 l/min.

We found that the infusion of nonzero [NO] into the isolated nasal system allowed easy determination of nasal D_NO and C_w at a constant luminal gas flow rate. The mean D_NO and C_w in the lung in a group of 10 healthy adults were 5.58 × 10³ nl · s¹ · ppb¹ and 137 ppb, respectively (37). We have found lower values for nasal D_NO in our group of eight healthy subjects (0.52-2.98 × 10³ nl · s¹ · ppb¹) but values for C_w that are much higher and vary considerably between individuals (1,236-8,947 ppb). These very high values for nasal C_w are not unexpected given the high [NO] in the nasal lumen (20), but the degree of interindividual variation raises the possibility of interindividual differences in NO-dependent nasal physiology.

The determination of D_NO and C_w in the bronchial tree is possible only at multiple flow rates and with the assumption that C_w remains constant with varying flow, since [NO]_in cannot be controlled (37). In the nose, we have demonstrated that C_w roughly halves with a doubling in luminal gas flow rate, suggesting that this assumption may be incorrect. In summary, with increasing luminal gas flow, the total nasal NO output increases, but the mucosal concentration declines. Taken together, these two findings support the hypothesis of Silkoff et al. (36) that changes in luminal gas flow induce a change in equilibrium between diffusion of NO to the submucosa and the lumen, such that an increase in gas flow favors diffusion toward the lumen and away from the submucosa. Because C_w declined with increasing luminal gas flow rate, it seems unlikely that the nasal epithelium acts in an analogous fashion to the vascular endothelium. That is, increases in luminal gas flow rate appear not to lead to an increase in NOS activity in nasal epithelial cells.

We have demonstrated that, especially at low oxygen tensions (<10%), nasal NO output declines markedly. This is in line with the findings of Haight et al. (13), who also demonstrate a reduction in nasal NO output with hypoxia, particularly at oxygen tensions <10%. The most obvious explanation for this finding, and one well pursued by Dweik et al. (6) in the lower respiratory tract, is that NO production declines, since oxygen is a substrate for NOS. This may well be the case, but other explanations need to be considered. A number of authors have examined the relationship between systemic hypoxia and nasal resistance. Maltais et al. (26) and Dinh et al. (4) demonstrated that nasal resistance decreases with progressive and transient systemic hypoxia in men, probably through a neural mechanism. However, a decrease in nasal resistance (that is, an increase in nasal luminal volume) would tend to lead to an increase in dwell time of gas infused at constant flow and a reduction in available submucosal heme and so, potentially, to an increase in NO output. In any case, significant systemic hypoxia is unlikely with a breath hold of 35 s in healthy adults. The effect of the local application of a hypoxic gas mixture on nasal resistance has been less well studied. Seelagy et al. (33) found no effect of changes in inspired oxygen tension on nasal resistance in decerebrate cats. In contrast, Syabbalo et al. (38), in a human study, describe a rapid 40% increase in nasal resistance with the application of a gas mixture containing 10% oxygen. They ascribe this effect to a local vasodilator action of hypoxia on the nasal vascular bed (38). This effect could lead to a reduction in nasal NO output through an increase in the availability of heme in the richly vascular submucosa for reaction with NO (14) or through a reduction in nasal volume, so reducing dwell time for a gas infused at a constant flow. At this stage, we can only speculate as to whether the effect of luminal oxygen tension on nasal NO output results from an effect on NOS activity or an effect on nasal resistance. Further studies with the determination of D_NO and C_w at very low oxygen tensions may help resolve this issue.

The ease with which C_w and D_NO may be determined in the nose means that this technique may be utilized to study, in more depth, many aspects of nasal physiology and pathophysiology. Because, in terms of NO physiology, there are similarities between the nasal and bronchial epithelium (both comprise ciliated, NOS-containing epithelial cells that demonstrate a positive relationship between luminal gas flow, luminal oxygen tension, and NO output), further studies in the nose may provide insight into NO exchange dynamics in the bronchial tree.

In summary, we have demonstrated that nasal NO output is a positive function of luminal gas flow rate, probably secondary to a shift in equilibrium resulting from dilutional changes in luminal [NO], rather than a direct effect of luminal gas flow on NOS activity. Nasal NO output declines with local hypoxia, as a direct effect of hypoxia on NOS activity or secondary to the vasodilatory effect of hypoxia on the nasal vascular bed. Infusing nonzero [NO] into the isolated nasal cavity allows the straightforward determination of C_w and D_NO, and further application of this technique may improve our understanding of respiratory tract NO physiology and pathophysiology.