Nitric oxide (NO) is generated from l-arginine by the enzyme NO synthase (NOS), of which constitutive (cNOS) and inducible (iNOS) isoforms have been described functionally.¹ Both are present in the airways.² cNOS (endothelial NOS (eNOS) and neuronal NOS (nNOS)) release small amounts of NO (in the picomolar range) within seconds in response to increases in cytosolic Ca²⁺. Several agonists, including bradykinin, activate cNOS upon receptor stimulation.³ In contrast, the expression of iNOS is induced by proinflammatory cytokines, a process that results in a protracted and Ca²⁺ independent release of high levels of NO (in the nanomolar range)⁴ which is inhibited by corticosteroids.^5,6 NO formed by cNOS leads to cGMP dependent airway smooth muscle relaxation, whereas NO released by iNOS is associated with proinflammatory effects.^7,8 Both enzyme isoforms are likely to play a role in asthma.^7,8 NO is detectable in the exhaled air of healthy volunteers⁹ and is increased in patients with atopic asthma.¹⁰ It has recently been shown that the increased levels of NO in the exhaled air of asthmatic patients are related to airway eosinophilic inflammation¹¹ and to the increased expression of iNOS.⁶ Exhaled NO (FE_NO) is inhibited by treatment with inhaled corticosteroids⁵ and NOS inhibitors,¹² suggesting that it is sensitive to steroids and produced by the enzymatic NOS pathway.

In most sensitised asthmatics, allergen exposure induces an early (EAR) and late (LAR) asthmatic response. The EAR, with early onset (10 minutes after exposure) and short duration (1–2 hours), is followed by the LAR (3–10 hours) which is characterised by bronchoconstriction associated with airway cellular inflammation.¹³ Interestingly, it has been shown that the LAR is associated with increased levels of NO in the exhaled air,¹⁴ suggesting that FE_NO may reflect allergic inflammation in asthmatic airways.

Bradykinin is an autacoid involved in the pathophysiology of asthma.¹⁵ Bradykinin induced hyperresponsiveness is increased by NOS inhibitors in asthma,¹⁶ suggesting that bradykinin releases Ca²⁺ dependent cNOS derived NO in the airways.¹⁶ Conversely, inhalation of bradykinin,¹⁷ as well as of prostaglandin E₂ (PGE₂),¹⁸ reduces FE_NO in stable asthmatics, which suggests that bradykinin has a role in modulating FE_NO in the absence of disease activity. However, it is unknown whether bradykinin can regulate increased FE_NO levels during late phase inflammation induced by allergen in asthma and whether it can modulate FE_NO via NOS inhibition.

We measured the effect of bradykinin inhalation on NO levels in the exhaled air of asthmatic patients, both at baseline and after experimental allergen exposure in the absence or presence of the NOS inhibitor l-NMMA. This study is part of a larger project for which the study design, methods, and results for other outcome parameters have already been published.¹⁶

METHODS

Study design

The screening period for the selection criteria has been described previously.¹⁶ The study had a randomised, placebo controlled, crossover design and consisted of three periods (diluent/placebo/bradykinin (Dil/P/BK), allergen/placebo/bradykinin (All/P/BK), and allergen/l-NMMA/bradykinin (All/L/BK)) of five study days each, separated by a washout interval of 2–4 weeks. During each study period the participants pretreated with placebo (Dil/P/BK and All/P/BK periods) or l-NMMA (All/L/BK period) underwent a bradykinin challenge before (day 3) and 48 hours after (day 10) the diluent (Dil/P/BK period) or allergen challenge (All/P/BK and All/L/BK periods, day 8; fig 1⇓). On day 9 the participants underwent spirometric tests and FE_NO monitoring (fig 1⇓). On day 1 of each study period a control PC₂₀ histamine measurement was performed.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Study design. Dil=diluent; All=allergen; P=placebo; BK=bradykinin; L=N^G-monomethyl-l-arginine.

FE_NO was measured (2 minutes after spirometry) at baseline, after placebo or l-NMMA, and at the end of bradykinin (after 5 minutes and every 10 minutes up to 1 hour) and allergen/diluent challenge (every 10 minutes in the first hour, every 30 minutes in the second hour, every 60 minutes up to 10 hours, and at 24 hours), when a fall in forced expiratory volume in 1 second (FEV₁) of ⩾20% was achieved or after three consecutive diluent inhalations.

RESULTS

Exhaled NO

NO was detectable in the exhaled air of all asthmatic subjects. Baseline FE_NO values in the Dil/P/BK period (day 3: 37 (11) ppb; day 8: 38 (12) ppb), All/P/BK period (day 3: 35 (9) ppb; day 8: 33 (8) ppb), and All/L/BK period (day 3: 37 (10) ppb; day 8: 31 (8) ppb) were not different (MANOVA, p=0.48).

There was no change in FE_NO up to 48 hours after diluent challenge (day 8), whereas allergen challenge on day 8 reduced FE_NO in the first hour after the end of the challenge in the All/P/BK and All/L/BK periods (fig 2⇓), as shown by the significant differences between AUC_0–1 for FE_NO in the All/P/BK period (19.9 ppb/h (95% CI 9.2 to 30.7)), the All/L/BK period (17.9 ppb/h (95% CI 8.5 to 27.4)), and the Dil/P/BK period (38.1 ppb/h (95% CI 9.6 to 66.6); p=0.04). FE_NO was increased in the LAR after allergen challenge in the All/P/BK and All/L/BK periods (fig 2⇓). AUC_1–48 for FE_NO on day 8 was significantly different in the All/P/BK (3282 ppb/h (95% CI 1693 to 4870)), All/L/BK (2579 ppb/h (95% CI 1467 to 3690)), and Dil/P/BK periods (1781 ppb/h (95% CI 471 to 3091); p=0.008). The differences in total AUC_0–48/48 for FE_NO, interpreted as the mean FE_NO, on day 8 in the All/P/BK (68.8 ppb (95% CI 35 to 102)), All/L/BK (54.1 ppb (95% CI 31 to 77)), and Dil/P/BK periods (37.8 ppb (95% CI 10 to 65)) were significant (p=0.009). The differences in all the AUCs for FE_NO between diluent and the two allergen challenges are shown in table 2⇓.

View this table:

• View inline
• View popup

Table 2

Mean (95% CI) differences in AUC_0–1, AUC_1–48, and total AUC/48 for FE_NO between diluent (Dil/P/BK) and allergen (All/P/BK and All/L/BK) challenges

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Time course of the change in exhaled nitric oxide levels after diluent (circles) and allergen (allergen/placebo/bradykinin: squares; allergen/l-NMMA/bradykinin: rhombus) challenges in 10 atopic patients with mild asthma. Mean (SE) values are shown.

In the Dil/P/BK period FE_NO values were significantly lower after bradykinin than after placebo (fig 3⇓) on day 3 (difference 23.5 ppb (95% CI 6.2 to 40.9); p=0.01) and day 10 (difference 22.5 ppb (95% CI 7.3 to 37.7); p=0.008). Both FE_NO values after bradykinin in the Dil/P/BK period were not significantly different from the corresponding FE_NO value on the pre-allergen day in the All/P/BK period (p=0.15). Forty eight hours after allergen challenge in the All/P/BK period bradykinin completely suppressed the increased FE_NO levels (fig 3⇓), as shown by the reduction in the allergen induced increase in FE_NO (difference from placebo: 54.3 ppb (95% CI 23.8 to 84.8); p=0.003) to the baseline level of the pre-allergen day (p=0.51). The difference between the reductions in FE_NO induced by bradykinin from the respective placebo 48 hours after allergen in the All/P/BK period and after diluent in the Dil/P/BK period was statistically significant (31.8 ppb (95% CI 7.1 to 56.5); p=0.017).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Effect of bradykinin challenge on exhaled nitric oxide in (A) the diluent/placebo/bradykinin period, (B) the allergen/placebo/bradykinin period, and (C) the allergen/l-NMMA/bradykinin period. Exhaled nitric oxide levels were measured before pre-treatment (open columns), after pretreatment (with placebo or l-NMMA: shaded columns), and 5 minutes after bradykinin challenge (filled columns) on days 3 and 10 of each study period. On day 8 diluent (diluent/placebo/bradykinin) and allergen (allergen/placebo/bradykinin and allergen/l-NMMA/bradykinin) challenges were performed. Mean (SE) values are shown. * p<0.05 v respective placebo or l-NMMA.

In the Dil/P/BK and All/P/BK periods, neither placebo nor l-NMMA (0.9% saline) changed baseline FE_NO on any study days (fig 3⇑). In the All/L/BK period the differences between the level of FE_NO after l-NMMA and placebo before and after allergen were significant (All/L/BK: 10.85 ppb (95% CI 1.3 to 20.4), p=0.029; All/P/BK: 36.2 ppb (95% CI 5.5 to 66.9), p=0.026).

After pretreatment with l-NMMA, bradykinin significantly reduced FE_NO on pre-allergen (difference from l-NMMA: 14.1 ppb (95% CI 6.8 to 21.4); p=0.002) and post-allergen (difference from l-NMMA: 29.7 ppb (95% CI 16 to 43.5); p=0.0008) days (fig 3⇑). The differences between the level of FE_NO after placebo plus bradykinin in the All/P/BK period and after l-NMMA plus bradykinin in the All/L/BK period on pre- and post-allergen days (1.96 ppb (95% CI −2.1 to 6), p=0.3; and 9.4 ppb (95% CI −0.5 to 19.5), p=0.06, respectively) were not statistically significant. The difference between the baseline values of FE_NO (48 hours after allergen) on the post-allergen day (day 10) in the All/P/BK and All/L/BK periods was 15.8 ppb (95% CI −5 to 36.4); p=0.11), which explains the slight difference between FE_NO after placebo plus bradykinin in the All/P/BK period and after l-NMMA plus bradykinin in the All/L/BK period on the post-allergen day.

On all bradykinin treatment days the recovery time (up to 60 minutes) was characterised by a complete spontaneous recovery in FEV₁ values to baseline levels, whereas FE_NO levels remained significantly lower than baseline (p<0.05, fig 4⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Time course of the change in exhaled nitric oxide (rhombus) and forced expiratory volume in one second (FEV₁; squares) during the recovery period after diluent in (A) the diluent/placebo/bradykinin period and after allergen in (B) the allergen/placebo/bradykinin period and (C) the allergen/l-NMMA/bradykinin period on the day of bradykinin challenge (day 10) in 10 atopic patients with mild asthma. Mean (SE) values are shown.

DISCUSSION

This study shows, for the first time, that the allergen induced increase in FE_NO in atopic asthma is suppressed by bradykinin, and that the NOS inhibitor l-NMMA significantly reduces pre- and post-allergen FE_NO levels and does not potentiate the bradykinin induced reduction in FE_NO. Taken together, our results indicate that bradykinin, via NOS inhibition, modulates the prolonged increase in FE_NO induced by allergen exposure. Our findings suggest that bradykinin inhibits excessive NO production resulting from an overexpression of the NOS pathway in asthmatic airways during an exacerbation.

The bradykinin induced fall in FE_NO levels in stable asthmatics has recently been reported by Kharitonov et al.¹⁷. We have shown that bradykinin suppresses increased levels of FE_NO 48 hours after allergen challenge, which suggests that it has a role in modulating exacerbations of allergic inflammation. Furthermore, the finding that FE_NO levels were still significantly reduced 60 minutes after the end of the pre-allergen and post-allergen bradykinin challenge, when FEV₁ had completely recovered, suggests that the rapid suppression of FE_NO by bradykinin has a relatively long duration and is not dependent on airway calibre. One possible explanation for the action of bradykinin is vasodilation¹⁵ with potentially enhanced NO trapping by haemoglobin²³ and/or altered diffusion of NO to the airway lumen due to mucosal swelling and gland secretion¹⁵ with subsequent airway obstruction; however, the short term effect of vasodilation and the reduction in FE_NO even after recovery of the obstruction limit the validity of this hypothesis.

In this study we also examined whether bradykinin negatively modulates FE_NO through inhibition of NOS. l-NMMA has previously been found to reduce FE_NO levels significantly in stable asthma,¹² so we studied the effect of l-NMMA on the bradykinin induced reduction in FE_NO. The lack of significant potentiation by l-NMMA on the bradykinin induced reduction in FE_NO suggests that bradykinin and l-NMMA share a common pathway in inhibiting FE_NO production in the airways of patients with asthma. It has recently been found that prostaglandins also reduce FE_NO in asthmatic subjects.¹⁸ Bradykinin stimulates PGE₂ release from airway epithelial cells, either by the constitutive isoform of cyclo-oxygenase enzyme (COX-1) or by the inducible isoform (COX-2),²⁴ and PGE₂ negatively modulates the induction of iNOS expression at the transcriptional level.²⁵ It is therefore possible that the reduction in FE_NO induced by bradykinin is a result of the inhibition of iNOS expression through the release of PGE₂. However, the rapid reduction in FE_NO observed after bradykinin inhalation does not favour such a delayed mechanism, but suggests a possible direct or indirect influence of bradykinin on NOS at the post-transcriptional level. It has previously been shown that high levels of NO can downregulate NOS pathways by negative feedback,²⁶ and also that augmented production of FE_NO in the lower respiratory tract of eNOS deficient mice is associated with pulmonary iNOS overexpression, indicating a counter-regulation between the two NOS isoforms.²⁷ Furthermore, a recent study in human mesangial cells showed a negative post-transcriptional regulation of iNOS by the NO/cGMP pathway.²⁸ On the basis of the present results we believe that bradykinin is not only an activator of the Ca²⁺ dependent cNOS/cGMP pathway which modulates airway hyperresponsiveness,²⁹ but also may downregulate excessive NO release, probably derived by iNOS, through negative feedback in allergic asthma.

In line with this hypothesis, we have previously shown that iNOS (but not eNOS or nNOS) immunostaining is higher after allergen exposure in the epithelium of bronchial biopsy sections of subjects with asthma,¹⁶ which suggests two major findings on the basis of the present data: (1) iNOS may be responsible for the increase in FE_NO after allergen; (2) bradykinin should affect iNOS activity by reducing FE_NO in asthmatic patients and, in particular, by suppressing the increase in FE_NO after exposure to allergen.

In the present study, allergen inhalation led to two opposing changes in FE_NO levels in atopic mild asthmatics. Exhaled NO tended to decrease during the EAR and exhibited a prolonged increase during and after the LAR. These findings are consistent with previous data showing the effect of allergen on FE_NO levels in atopic asthmatic patients with LAR.¹⁴ In the early phase we found a reduction in FE_NO levels compared with baseline, which confirms previous observations.³⁰ In previous studies acute airways obstruction has been found to have a partial effect on FE_NO levels.^31,32 It is therefore possible that bronchoconstriction could, at least partially, lead to a fall in FE_NO levels during the early phase of the bronchial allergic reaction, possibly as a result of limiting NO diffusion or increased airflow velocity in constricted airways. We also postulate an indirect effect of allergen on FE_NO levels during the early phase reaction via the release of proinflammatory mediators such as kinins or prostaglandins.^17,18

Exhaled NO appeared to increase 9–48 hours after allergen inhalation, confirming a previous report.¹⁴ In the latter study the authors observed a decline in FE_NO by 27 hours. The longer duration of increased FE_NO in our study could be a reflection of a possibly higher level of allergic inflammation induced by allergen exposure. Several lines of evidence suggest that FE_NO reflects iNOS expression in asthma,^5,6 which recognises that FE_NO is a non-invasive marker of airway inflammation in asthma.³³ The delayed and long term increase in FE_NO during the LAR could be explained by the action of proinflammatory cytokines released from inflammatory cells recruited to the airways after allergen exposure, and of reactive oxygen species with subsequent upregulation of iNOS expression.^34,35 Furthermore, an additional explanation could be derived from the recent evidence that airway acidification converts nitrite (NO₂⁻) to NO gas in quantities consistent with those observed in expired air during an asthma exacerbation,³⁶ or after the LAR to allergen.

To avoid differences in FE_NO as a result of smoking or medication use, we selected a homogeneous group of non-smoking, atopic, mild intermittent asthmatic subjects who were not using any steroid medication before or during the experiments. It has recently been shown that repeated forced vital capacity (FVC) manoeuvres significantly reduce FE_NO levels in healthy and asthmatic subjects by 13% and 10%, respectively.³⁷ Even though our FE_NO measurements were performed 2 minutes after an FVC manoeuvre, we do not consider that it had any effect because no effect on the FE_NO measurement was seen following diluent.

NO released by iNOS is likely to exert proinflammatory activities such as induction of eosinophil chemotaxis and activation of the Th2 driven immune response,^8,11 as observed during the late phase response to inhaled allergen. The bradykinin induced suppression of raised FE_NO levels during the late phase suggests that bradykinin can modulate allergic inflammation exacerbations in the airways by the inhibition of proinflammatory NOS. On the basis of our study, a clinically relevant scenario could be outlined in patients with asthma where an acute inflammatory mediator (bradykinin) could have a modulatory role in limiting the effect of a detrimental long lasting inflammatory pathway (iNOS). The reinforcement of such endogenous protective mechanisms should be considered as a potential therapeutic target in asthma.