- Split View
-
Views
-
Cite
Cite
Chris Carlsten, Meaghan J. MacNutt, Zhihong Zhang, Francesco Sava, Mandy M. Pui, Anti-Oxidant N-Acetylcysteine Diminishes Diesel Exhaust-Induced Increased Airway Responsiveness in Person with Airway Hyper-Reactivity, Toxicological Sciences, Volume 139, Issue 2, June 2014, Pages 479–487, https://doi.org/10.1093/toxsci/kfu040
- Share Icon Share
Abstract
Background: Inhalation of diesel exhaust (DE) at moderate concentrations causes increased airway responsiveness in asthmatics and increased airway resistance in both healthy and asthmatic subjects, but the effect of baseline airway responsiveness and anti-oxidant supplementation on this dynamic is unknown. Objectives: We aimed to determine if changes in airway responsiveness due to DE are attenuated by thiol anti-oxidant supplementation, particularly in those with underlying airway hyper-responsiveness. Methods: Participants took N-acetylcysteine (600 mg) or placebo capsules three times daily for 6 days. On the last of these 6 days, participants were exposed for 2 h to either filtered air (FA) or DE (300 μg/m3 of particulate matter smaller than 2.5 microns). Twenty-six non-smokers were studied under each of three experimental conditions (filtered air with placebo, diesel exhaust with placebo, and diesel exhaust with N-acetylcysteine) using a randomized, double-blind, crossover design, with a 2-week washout between conditions. Methacholine challenge was performed pre-exposure (baseline airway responsiveness) and post-exposure (effect of exposure). Results: Anti-oxidant supplementation reduced baseline airway responsiveness in hyper-responsive individuals by 20% (p = 0.001). In hyper-responsive individuals, airway responsiveness increased 42% following DE compared with FA (p = 0.03) and this increase was abrogated with anti-oxidant supplementation (diesel exhaust with N-acetylcysteine vs. filtered air with placebo, p = 0.85). Conclusions: Anti-oxidant (N-acetylcysteine) supplementation protects against increased airway responsiveness associated with DE inhalation and reduces need for supplement bronchodilators in those with baseline airway hyper-responsiveness. Individuals with variants in genes of oxidative stress metabolism when exposed to DE are protected from increases in airway responsiveness if taking anti-oxidant supplementation.
ABBREVIATIONS
- ATS
American Thoracic Society
- DE
diesel exhaust
- DEN
diesel exhaust with N-acetylcysteine
- DEP
diesel exhaust with placebo
- DRS
dose-response slope
- FA
filtered air
- FAP
filtered air with placebo
- FEV1
forced expiratory volume at 1 second
- PM
particulate matter
- SABA
short-acting β-2 agonists
Asthma is a disease associated with considerable health consequences (Mannino et al., 2002) whose prevalence has increased sharply over the past 40 years on a global level (Braman, 2006). Evidence links combustion-derived particulate matter (PM) to asthma symptoms and exacerbations (Atkinson et al., 2001). In asthmatic human subjects, diesel exhaust (DE) has been strongly associated with acute worsening of lung function (McCreanor et al., 2007) and increased airway resistance (Stenfors et al., 2004). DE is associated with airway hyper-responsiveness in both mice (Ohta et al., 1999) and in asthmatic human subjects (Nordenhall et al., 2001).
Constituents in DE have the potential to produce reactive oxygen species (Xia et al., 2004). Oxidative stress in response to DE has been demonstrated in vitro (Hashimoto et al., 2000; Koike et al., 2004; Li et al., 2002) and may be reversible with the anti-oxidant N-acetylcysteine (Koike et al., 2004). An endogenous anti-oxidant response to DE has also been shown in human bronchial epithelial cells in vitro (Li et al., 2002; Xiao et al., 2003) and a controlled human exposure model has shown increases in glutathione in bronchial wash at 6 h and in bronchoalveolar lavage at 18 h following exposure to DE (Behndig et al., 20062006). Finally, in asthmatics exposed to DE, there is a drop in the pH of exhaled breath condensate (McCreanor et al., 2007), suggestive of oxidative stress.
Deficient oxidant homeostasis or anti-oxidant reserves have been linked to asthma prevalence and severity (Mak and Chan-Yeung, 2006) and anti-oxidants attenuate airway responsiveness (Chang and Crapo, 2002). However, the direct effects of anti-oxidants on airway responsiveness have been investigated primarily using only animal models (Dong et al., 2005; Katsumata et al., 1990). A mouse model provided evidence that airway responsiveness was increased with exposure to DE particles and that this effect was attenuated by pre-treatment with a non-specific inhibitor of nitric oxide synthase (Lim et al., 1998) though this is not an anti-oxidant per se. Anti-oxidant supplementation attenuates declines in lung function caused by subacute ozone exposure (Romieu et al., 2002). However, it is not known whether this is true for airway responsiveness specifically and for pollutants other than ozone; to our knowledge, an effect of anti-oxidant supplementation on acute changes in DE-related airway function has not been demonstrated in humans (Tashakkor et al., 2011). Thus, we aimed to determine if DE-associated airway responsiveness is attenuated by thiol anti-oxidant supplementation administered prior to exposure, hypothesizing that this would be the case in those with underlying airway hyper-responsiveness.
MATERIALS AND METHODS
The ethical review boards of the University of British Columbia and the Vancouver Coastal Health Research Institute approved study procedures. Written informed consent was obtained from each participant. The study was registered at clinicaltrials.gov (trial no. NCT01699204).
Study design
Participants took N-acetylcysteine (N; 600 mg) or placebo (P) capsules three times daily for 6 days. This is a dose that is used standardly for treatment of idiopathic pulmonary fibrosis and is generally well tolerated. On the last of these 6 days, participants were exposed to either filtered air (FA) or DE. Each participant was studied under each of three experimental conditions (filtered air with placebo, FAP; diesel exhaust with placebo, DEP; and diesel exhaust with N-acetylcysteine, DEN) using a balanced, randomized, double-blind, crossover design, with a minimum of 2-week washout between conditions. Symptoms, lung function, and airway responsiveness (represented by the dose-response slope (DRS) to methacholine, as noted below) were assessed and blood and urine were collected at several time points from 18 h before, to 30 h after, the onset of each environmental exposure (see Fig. 1).
Participants and study procedures
Twenty-six young (19–46 years old), generally healthy, non-smoking males and females were recruited to participate. Sample size was based on two-sided tests with an alpha level of 0.05 and power of 90% using prior data for DE-associated change in PC20 (Nordenhall et al., 2001) and inclusive of the standard deviation in short-term serial methacholine challenges. Atopy was assessed at screening by skin prick testing for common aeroallergens. Presence of GSTM1 (null), GSTP1 (Ile105Val; rs1695), and NFκB (ins/del; rs28362491) variants was assessed by PCR-Restriction Fragment Length Polymorphism using DNA isolated from blood.
Participants did not use dietary supplements containing Vitamins A, C, or E, and each participant's diet was held constant for each exposure day. Participants were asked to withhold caffeine for 4 h, short-acting β-2 agonists (SABAs) for 8 h, and long-acting β-2 antagonists for 48 h before any testing procedures. A validated common cold questionnaire (Jackson et al., 1958) was used to confirm that participants were free of viral infection for at least 4 weeks before each testing session. Compliance with the protocol's non-smoking requirement was assessed by urinary cotinine (Cotinine Urine Micro-Plate EIA, OraSure Technologies, Bethlehem, PA) in a randomly selected sample of participants.
Forced expiratory volume in 1 second (FEV1) was determined by standard spirometry performed according to American Thoracic Society (ATS) guidelines (1995). Eye, nose, throat, chest, head, and non-specific symptoms were evaluated by a brief questionnaire (see Supplementary fig. 1). Cardiorespiratory function (heart and respiratory rate, blood pressure, and oxyhaemoglobin saturation) was assessed using standard clinical methodology. Airway responsiveness was evaluated by methacholine challenge using the 2-min tidal breathing method (Crapo et al., 2000). Starting at 0.5 mg/ml, methacholine concentrations were increased in doubling doses to a maximum of 16 mg/ml. Airway responsiveness is described by the dose-response slope (DRS = log[(% fall in FEV1/cumulative dose methacholine) + 1]), which allowed determination of a numerical value even when a FEV1 dropped less than 20% following inhalation of the final dose of methacholine (de Meer et al., 2005). Baseline hyper-responsiveness was defined as DRS ≤ 0.78 (equivalent to PC20 of 8 mg/ml) at pre-FAP and/or pre-DEP. A sample DRS calculation is provided in Figure 2; the rationale for the use of DRS is described further within the Discussion section. For the purposes of this manuscript's results, DRS is interchangeable with airway responsiveness.
Blood was collected by venipuncture into EDTA tubes and differential complete blood counts were completed in the Vancouver General Hospital hematology laboratory using an automated analyzer (Sysmex XE-200i, TOA Medical Electronics Co., Kobe, Japan). Sputum was induced using an aerosol of inhaled hypertonic saline. Prior to sputum induction, inhaled salbutamol (200 mg) was given to inhibit possible airway constriction. The Fisoneb ultrasonic nebulizer (Clement Clarke International Ltd) with an output of 0.87 ml/min and an aerodynamic mass median diameter of 5.58 mm was used to deliver saline. Concentrations of saline at 3.00, 4.00, and 5.00% (each for 7 min) were given by inhalation through a mouthpiece without a valve or noseclip. Sputum was analyzed within 2 h of sample collection. The sample of sputum was transferred to a Petri dish and all visible sputum plugs were selected, placed in a pre-weighed conical tube, and weighed. Freshly prepared dithiothreitol (dilution of one part to nine parts distilled water) was added in a volume (ml) equal to four times the weight of sputum (mg) in order to homogenize the sample. The sample was vortexed for 15 s and then rocked using a bench rocker for 15 min. An equal volume of Dulbecco's phosphate buffered saline was added and the sample was then filtered to remove PM. The resultant samples were each centrifuged at 5006G for 10 min and the supernatant was aspirated into aliquots and stored at −80°C until use. The remaining pellet after centrifugation was resuspended in phosphate buffered saline and the cell count was determined as follows. The number of cells per milliliter of processed sputum was calculated. The viability of cells was evaluated using the trypan blue exclusion method. The cell suspension was adjusted to 1×106 cells/ml, and two cytospins were prepared with 50 µl of the cell suspension. These slides were air-dried, fixed, and stained with Wright's stain. At least 400 non-squamous cells were counted and a differential cell count was made. The percentage assigned to each cell type represents the proportion of each cell type within those 400 cells.
Exposures
Participants were exposed to FA or DE for 2 hr in the Air Pollution Exposure Laboratory (Birger et al., 2011) at the University of British Columbia. DE was generated on site by a 6.0 kW diesel generator that was operated under discrete cyclic loads to simulate on-road diesel emissions. After two dilution steps, DE was delivered into the exposure chamber where the real-time concentration of particles smaller than 2.5 μm (PM2.5) was monitored using a nephelometer and tapered element oscillating microbalance and maintained nominally at 300 μg/m3, with particle mass median aerodynamic diameter of 102.5 nm (standard deviation = 11.4). Participants rode a bicycle ergometer for 15 min each hour at a pre-determined workload that elicited a minute ventilation of 15 l/min/m2, but otherwise were seated at rest in the chamber. In our experience, blinding to exposures is effective (Carlsten et al., 2013).
Data analysis
DRS was, a priori, designated the primary study endpoint. Effects of condition, atopic status, baseline airway hyper-responsiveness, order, sex, and genotype on study endpoints were evaluated using repeated-measures ANOVA. Given the potential for inherent variability in baseline airway responsiveness to alter DRS values and thus obscure experimental effects, and for more direct comparison with previous work by Nordenhall and colleagues (Nordenhall et al., 2001), DRS data collected post-FAP, DEP, and DEN were used to analyze the effect of exposure on airway responsiveness. Statistical analyses were completed using R version 2.11.1 (R Development Core Team, 2010) with significance defined as p < 0.05 and borderline significance defined as p ≤ 0.1. All data are presented as mean ± standard deviation. Times are given relative to the onset of FA or DE exposure. Our statistical approach was performed in consultation with the Statistical Consulting and Research Laboratory at the University of British Columbia, as acknowledged.
RESULTS
Summary characteristics of study participants can be found in Table 1 (detailed in Supplementary table 1). Eight individuals had normal baseline airway responsiveness and 18 were hyper-responsive. Fourteen of 18 hyper-responsive individuals had been previously diagnosed with asthma. Two of the subjects previously doctor-diagnosed with asthma were not hyper-responsive. Eight doctor-diagnosed asthmatics were not prescribed any asthmatic medications. Eight hyper-responsive individuals were prescribed β-2 agonists (with or without inhaled corticosteroids) but only four reported regular use of β-2 agonists, and these were held according to the ethically approved protocol, as noted above. Because it specifically recruited asthmatics, the study was enriched for those with variants in genes known to confer risk for asthmatic pathways (Gilliland et al., 2004). Cotinine was negative (less than the 500 ng/ml calibrator) in all urine samples tested.
Description of Study Participants
Sex . | Age (year) . | BMI (kg/m2) . | FEV1 (% predicted) . | Atopy . | Asthma diagnosis . | Hyper-responsive . | NFκB variant . | GSTP1 variant . | GSTM1 variant . |
---|---|---|---|---|---|---|---|---|---|
50% female | Mean = 29 ± 8 | Mean = 24 ± 3 | Mean = 94 ± 13 | 62% | 62% | 69% | 81% | 65% | 69% |
Sex . | Age (year) . | BMI (kg/m2) . | FEV1 (% predicted) . | Atopy . | Asthma diagnosis . | Hyper-responsive . | NFκB variant . | GSTP1 variant . | GSTM1 variant . |
---|---|---|---|---|---|---|---|---|---|
50% female | Mean = 29 ± 8 | Mean = 24 ± 3 | Mean = 94 ± 13 | 62% | 62% | 69% | 81% | 65% | 69% |
Sex . | Age (year) . | BMI (kg/m2) . | FEV1 (% predicted) . | Atopy . | Asthma diagnosis . | Hyper-responsive . | NFκB variant . | GSTP1 variant . | GSTM1 variant . |
---|---|---|---|---|---|---|---|---|---|
50% female | Mean = 29 ± 8 | Mean = 24 ± 3 | Mean = 94 ± 13 | 62% | 62% | 69% | 81% | 65% | 69% |
Sex . | Age (year) . | BMI (kg/m2) . | FEV1 (% predicted) . | Atopy . | Asthma diagnosis . | Hyper-responsive . | NFκB variant . | GSTP1 variant . | GSTM1 variant . |
---|---|---|---|---|---|---|---|---|---|
50% female | Mean = 29 ± 8 | Mean = 24 ± 3 | Mean = 94 ± 13 | 62% | 62% | 69% | 81% | 65% | 69% |
Environmental exposures are described in Table 2. For DE exposures, PM2.5 was tightly regulated at the target of 300 μg/m3. The concentration of each component was significantly higher in both DEP and DEN compared with FAP but there were no significant differences between DEP and DEN.
Ambient Conditions During 2-h Exposures for Filtered Air with Placebo (FAP), Diesel Exhaust with Placebo (DEP), and Diesel Exhaust with Anti-oxidant (DEN) Conditions
. | FAP . | DEP . | DEN . |
---|---|---|---|
TVOC (ppb) | 103 ± 126 | 2010 ± 1211 | 1947 ± 1091 |
CO (ppm) | 2 ± 1 | 15 ± 10 | 13 ± 8 |
CO2 (ppm) | 849 ± 119 | 2352 ± 322 | 2551 ± 557 |
NO (ppb) | 59 ± 116 | 8026 ± 3057 | 8781 ± 3865 |
NO2 (ppb) | 26 ±11 | 385 ± 403 | 378 ± 390 |
NOx (ppb) | 86 ± 121 | 8411 ± 3178 | 9159 ± 4034 |
PM2.5 (μg/m3) | 11 ± 8 | 299 ± 20 | 303 ± 12 |
. | FAP . | DEP . | DEN . |
---|---|---|---|
TVOC (ppb) | 103 ± 126 | 2010 ± 1211 | 1947 ± 1091 |
CO (ppm) | 2 ± 1 | 15 ± 10 | 13 ± 8 |
CO2 (ppm) | 849 ± 119 | 2352 ± 322 | 2551 ± 557 |
NO (ppb) | 59 ± 116 | 8026 ± 3057 | 8781 ± 3865 |
NO2 (ppb) | 26 ±11 | 385 ± 403 | 378 ± 390 |
NOx (ppb) | 86 ± 121 | 8411 ± 3178 | 9159 ± 4034 |
PM2.5 (μg/m3) | 11 ± 8 | 299 ± 20 | 303 ± 12 |
Note. Data represent group mean ± standard deviation for 26 subjects receiving all three conditions in random order. Concentrations of total volatile organic compounds (TVOC), carbon monoxide (CO), carbon dioxide (CO2), nitric oxide (NO), nitric dioxide, total nitrogen oxides (Knox et al.1997), and particulate matter smaller than 2.5 μm (PM2.5) are given.
. | FAP . | DEP . | DEN . |
---|---|---|---|
TVOC (ppb) | 103 ± 126 | 2010 ± 1211 | 1947 ± 1091 |
CO (ppm) | 2 ± 1 | 15 ± 10 | 13 ± 8 |
CO2 (ppm) | 849 ± 119 | 2352 ± 322 | 2551 ± 557 |
NO (ppb) | 59 ± 116 | 8026 ± 3057 | 8781 ± 3865 |
NO2 (ppb) | 26 ±11 | 385 ± 403 | 378 ± 390 |
NOx (ppb) | 86 ± 121 | 8411 ± 3178 | 9159 ± 4034 |
PM2.5 (μg/m3) | 11 ± 8 | 299 ± 20 | 303 ± 12 |
. | FAP . | DEP . | DEN . |
---|---|---|---|
TVOC (ppb) | 103 ± 126 | 2010 ± 1211 | 1947 ± 1091 |
CO (ppm) | 2 ± 1 | 15 ± 10 | 13 ± 8 |
CO2 (ppm) | 849 ± 119 | 2352 ± 322 | 2551 ± 557 |
NO (ppb) | 59 ± 116 | 8026 ± 3057 | 8781 ± 3865 |
NO2 (ppb) | 26 ±11 | 385 ± 403 | 378 ± 390 |
NOx (ppb) | 86 ± 121 | 8411 ± 3178 | 9159 ± 4034 |
PM2.5 (μg/m3) | 11 ± 8 | 299 ± 20 | 303 ± 12 |
Note. Data represent group mean ± standard deviation for 26 subjects receiving all three conditions in random order. Concentrations of total volatile organic compounds (TVOC), carbon monoxide (CO), carbon dioxide (CO2), nitric oxide (NO), nitric dioxide, total nitrogen oxides (Knox et al.1997), and particulate matter smaller than 2.5 μm (PM2.5) are given.
There was no effect of the order of conditions or sex on any endpoint. FEV1 did not change over time in any condition and was not affected by baseline airway responsiveness or genotype. All FEV1 data are given in Supplementary table S1.
FAP caused no change in symptoms. With DEP, across all subjects, total symptom score was elevated from 1.4 ± 2.1 at 0 h to 4.7 ± 8.1 at +2 h (p < 0.001) and at +6 h had returned to baseline levels. There were no significant differences in symptom scores between DEP and DEN. Across all time points, symptom scores were higher in hyper-responsive compared with normally responsive individuals (p < 0.001) but the effects of DEP on symptoms were similar for both groups. Symptom scores are detailed in Supplementary table S2.
Although there was no effect of anti-oxidant supplementation on baseline DRS (measured prior to exposure to DE or FA) in normally responsive individuals, baseline DRS was reduced in hyper-responsive individuals by an average of 20% following 5 days of anti-oxidant supplementation compared with that following 5 days of placebo supplements (p = 0.001; Fig. 3). This was accompanied by a mean 58% reduction (p = 0.046) in self-reported SABA use in those individuals who reported regular use at baseline.
Mean pre-exposure DRS was 440% higher in hyper-responsive compared with normally responsive individuals (p = 0.0002). The effect of condition on DRS among all subjects was of borderline statistical significance (p = 0.10). However, according to our a priori hypothesis, separate analyses for normally responsive and hyper-responsive individuals were performed (Fig. 4); consistent with our hypothesis, there was an effect of condition for hyper-responsive (p = 0.04) but not normally responsive (p = 0.66) individuals. In neither subset of individuals was DRS affected by atopic status. In hyper-responsive individuals, AR was 20 ± 64% higher (p = 0.03) after DEP (DRS = 1.4 ± 0.7) than FAP (DRS = 1.1 ± 0.6), but was not different from FAP following DE exposure with anti-oxidant supplementation (DEN; AR = 1.1 ± 0.7; p = 0.43 for DEN vs. FAP).
Genotype effects on DRS changes by exposure are shown in Figure 5. For all three genes (GSTM1, GSTP1, and NFκB), under the variant condition but not the wild-type condition, anti-oxidant supplementation to DEP significantly decreased DRS.
Post-exposure sputum cell counts (Table 3) showed high variability, as expected. Considering all subjects, there were no significant differences attributable to DEP 30 h after exposure. However, in hyper-responsive individuals, sputum% bronchial epithelial cells increased following DEP relative to FAP (p = 0.06), and this effect was attenuated by DEN. In normally responsive individuals, there was a decrease in neutrophils with DEP relative to FAP whereas in hyper-responsive individuals there was a trend for increased neutrophils with DEP and for DEN to attenuate this effect. Overall, interpretation of the sputum cell differences is challenging due to the noted high variability.
Sputum Cell Differentials at 30 h Post-exposure
. | . | Bronchial epithelial cells . | Macrophages . | Lymphocytes . | Neutrophils . | Eosinophils . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|
. | . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . |
All (n = 11) | FAP | 31 ± 25 | 6 ± 5 | 83 ± 79 | 19 ± 10 | 4 ± 6 | 1 ± 1 | 605 ± 676 | 73 ± 18 | 6 ± 8 | 0.02 ± 0.03 |
DEP | 61 ± 80 | 5 ± 3 | 195 ± 287 | 22 ± 22 | 4 ± 10 | 0 ± 0 | 1059 ± 1797 | 72 ± 24 | 3 ± 5 | 0.00 ± 0.00 | |
DEN | 33 ± 25 | 5 ± 3 | 135 ± 130** | 24 ± 20 | 2 ± 3 | 0 ± 0 | 521 ± 484 | 71 ± 21 | 3 ± 4 | 0.00 ± 0.01 | |
Normally responsive (n = 5) | FAP | 40 ± 25 | 5 ± 4 | 79 ± 67 | 10 ± 8 | 5 ± 4 | 1 ± 1 | 974 ± 842 | 84 ± 10 | 7 ± 7 | 0.01 ± 0.01 |
DEP | 31 ± 20* | 5 ± 4 | 79 ± 59 | 15 ± 11 | 1 ± 2 | 0 ± 0 | 595 ± 501** | 80 ± 14 | 2 ± 3 | 0.00 ± 0.01 | |
DEN | 24 ± 12 | 4 ± 2 | 93 ± 61 | 18 ± 15* | 1 ± 1 | 0 ± 0 | 555 ± 439* | 77 ± 15 | 4 ± 5 | 0.01 ± 0.01 | |
Hyper-responsive (n = 6) | FAP | 24 ± 24 | 6 ± 4 | 86 ± 87 | 27 ± 18 | 4 ± 7 | 1 ± 1 | 297 ± 309 | 64 ± 20 | 5 ± 9 | 0.02 ± 0.05 |
DEP | 85 ± 104* | 6 ± 2 | 291 ± 371 | 28 ± 28 | 7 ± 13 | 0 ± 0 | 1447 ± 2420 | 66 ± 30 | 4 ± 7 | 0.00 ± 0.00 | |
DEN | 41 ± 31 | 6 ± 3 | 170 ± 166** | 28 ± 23 | 3 ± 3 | 0 ± 0 | 492 ± 559* | 65 ± 26 | 2 ± 2 | 0.00 ± 0.00 |
. | . | Bronchial epithelial cells . | Macrophages . | Lymphocytes . | Neutrophils . | Eosinophils . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|
. | . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . |
All (n = 11) | FAP | 31 ± 25 | 6 ± 5 | 83 ± 79 | 19 ± 10 | 4 ± 6 | 1 ± 1 | 605 ± 676 | 73 ± 18 | 6 ± 8 | 0.02 ± 0.03 |
DEP | 61 ± 80 | 5 ± 3 | 195 ± 287 | 22 ± 22 | 4 ± 10 | 0 ± 0 | 1059 ± 1797 | 72 ± 24 | 3 ± 5 | 0.00 ± 0.00 | |
DEN | 33 ± 25 | 5 ± 3 | 135 ± 130** | 24 ± 20 | 2 ± 3 | 0 ± 0 | 521 ± 484 | 71 ± 21 | 3 ± 4 | 0.00 ± 0.01 | |
Normally responsive (n = 5) | FAP | 40 ± 25 | 5 ± 4 | 79 ± 67 | 10 ± 8 | 5 ± 4 | 1 ± 1 | 974 ± 842 | 84 ± 10 | 7 ± 7 | 0.01 ± 0.01 |
DEP | 31 ± 20* | 5 ± 4 | 79 ± 59 | 15 ± 11 | 1 ± 2 | 0 ± 0 | 595 ± 501** | 80 ± 14 | 2 ± 3 | 0.00 ± 0.01 | |
DEN | 24 ± 12 | 4 ± 2 | 93 ± 61 | 18 ± 15* | 1 ± 1 | 0 ± 0 | 555 ± 439* | 77 ± 15 | 4 ± 5 | 0.01 ± 0.01 | |
Hyper-responsive (n = 6) | FAP | 24 ± 24 | 6 ± 4 | 86 ± 87 | 27 ± 18 | 4 ± 7 | 1 ± 1 | 297 ± 309 | 64 ± 20 | 5 ± 9 | 0.02 ± 0.05 |
DEP | 85 ± 104* | 6 ± 2 | 291 ± 371 | 28 ± 28 | 7 ± 13 | 0 ± 0 | 1447 ± 2420 | 66 ± 30 | 4 ± 7 | 0.00 ± 0.00 | |
DEN | 41 ± 31 | 6 ± 3 | 170 ± 166** | 28 ± 23 | 3 ± 3 | 0 ± 0 | 492 ± 559* | 65 ± 26 | 2 ± 2 | 0.00 ± 0.00 |
Note. Cell counts are given for each cell type as both number of cells per milligram of sputum and as a cell differential. Data are presented as mean ± standard deviation. Statistical comparisons were made using paired t-tests. For each group of subjects, differences from FAP are indicated with *(p < 0.10) or **(p < 0.05).
. | . | Bronchial epithelial cells . | Macrophages . | Lymphocytes . | Neutrophils . | Eosinophils . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|
. | . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . |
All (n = 11) | FAP | 31 ± 25 | 6 ± 5 | 83 ± 79 | 19 ± 10 | 4 ± 6 | 1 ± 1 | 605 ± 676 | 73 ± 18 | 6 ± 8 | 0.02 ± 0.03 |
DEP | 61 ± 80 | 5 ± 3 | 195 ± 287 | 22 ± 22 | 4 ± 10 | 0 ± 0 | 1059 ± 1797 | 72 ± 24 | 3 ± 5 | 0.00 ± 0.00 | |
DEN | 33 ± 25 | 5 ± 3 | 135 ± 130** | 24 ± 20 | 2 ± 3 | 0 ± 0 | 521 ± 484 | 71 ± 21 | 3 ± 4 | 0.00 ± 0.01 | |
Normally responsive (n = 5) | FAP | 40 ± 25 | 5 ± 4 | 79 ± 67 | 10 ± 8 | 5 ± 4 | 1 ± 1 | 974 ± 842 | 84 ± 10 | 7 ± 7 | 0.01 ± 0.01 |
DEP | 31 ± 20* | 5 ± 4 | 79 ± 59 | 15 ± 11 | 1 ± 2 | 0 ± 0 | 595 ± 501** | 80 ± 14 | 2 ± 3 | 0.00 ± 0.01 | |
DEN | 24 ± 12 | 4 ± 2 | 93 ± 61 | 18 ± 15* | 1 ± 1 | 0 ± 0 | 555 ± 439* | 77 ± 15 | 4 ± 5 | 0.01 ± 0.01 | |
Hyper-responsive (n = 6) | FAP | 24 ± 24 | 6 ± 4 | 86 ± 87 | 27 ± 18 | 4 ± 7 | 1 ± 1 | 297 ± 309 | 64 ± 20 | 5 ± 9 | 0.02 ± 0.05 |
DEP | 85 ± 104* | 6 ± 2 | 291 ± 371 | 28 ± 28 | 7 ± 13 | 0 ± 0 | 1447 ± 2420 | 66 ± 30 | 4 ± 7 | 0.00 ± 0.00 | |
DEN | 41 ± 31 | 6 ± 3 | 170 ± 166** | 28 ± 23 | 3 ± 3 | 0 ± 0 | 492 ± 559* | 65 ± 26 | 2 ± 2 | 0.00 ± 0.00 |
. | . | Bronchial epithelial cells . | Macrophages . | Lymphocytes . | Neutrophils . | Eosinophils . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|
. | . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . | Number/mg . | % . |
All (n = 11) | FAP | 31 ± 25 | 6 ± 5 | 83 ± 79 | 19 ± 10 | 4 ± 6 | 1 ± 1 | 605 ± 676 | 73 ± 18 | 6 ± 8 | 0.02 ± 0.03 |
DEP | 61 ± 80 | 5 ± 3 | 195 ± 287 | 22 ± 22 | 4 ± 10 | 0 ± 0 | 1059 ± 1797 | 72 ± 24 | 3 ± 5 | 0.00 ± 0.00 | |
DEN | 33 ± 25 | 5 ± 3 | 135 ± 130** | 24 ± 20 | 2 ± 3 | 0 ± 0 | 521 ± 484 | 71 ± 21 | 3 ± 4 | 0.00 ± 0.01 | |
Normally responsive (n = 5) | FAP | 40 ± 25 | 5 ± 4 | 79 ± 67 | 10 ± 8 | 5 ± 4 | 1 ± 1 | 974 ± 842 | 84 ± 10 | 7 ± 7 | 0.01 ± 0.01 |
DEP | 31 ± 20* | 5 ± 4 | 79 ± 59 | 15 ± 11 | 1 ± 2 | 0 ± 0 | 595 ± 501** | 80 ± 14 | 2 ± 3 | 0.00 ± 0.01 | |
DEN | 24 ± 12 | 4 ± 2 | 93 ± 61 | 18 ± 15* | 1 ± 1 | 0 ± 0 | 555 ± 439* | 77 ± 15 | 4 ± 5 | 0.01 ± 0.01 | |
Hyper-responsive (n = 6) | FAP | 24 ± 24 | 6 ± 4 | 86 ± 87 | 27 ± 18 | 4 ± 7 | 1 ± 1 | 297 ± 309 | 64 ± 20 | 5 ± 9 | 0.02 ± 0.05 |
DEP | 85 ± 104* | 6 ± 2 | 291 ± 371 | 28 ± 28 | 7 ± 13 | 0 ± 0 | 1447 ± 2420 | 66 ± 30 | 4 ± 7 | 0.00 ± 0.00 | |
DEN | 41 ± 31 | 6 ± 3 | 170 ± 166** | 28 ± 23 | 3 ± 3 | 0 ± 0 | 492 ± 559* | 65 ± 26 | 2 ± 2 | 0.00 ± 0.00 |
Note. Cell counts are given for each cell type as both number of cells per milligram of sputum and as a cell differential. Data are presented as mean ± standard deviation. Statistical comparisons were made using paired t-tests. For each group of subjects, differences from FAP are indicated with *(p < 0.10) or **(p < 0.05).
DISCUSSION
Our a priori primary endpoint was post-exposure measures of airway responsiveness to emulate the seminal work of Nordenhall et al.(2001) and that of McCreanor et al.(2007) (though technically the latter study looked at airflow rather than airway responsiveness). From that perspective, our study corroborates prior evidence that airway responsiveness increases with DE and extends this by providing novel evidence that anti-oxidant supplementation can abrogate that deleterious effect in those with pre-existing airway hyper-responsiveness (Fig. 4); this supports the hypothesis that oxidative stress is responsible for DE-associated increases in airway responsiveness and complements the prior air pollution literature on another traffic-related pollutant, ozone (Romieu et al., 2002).
Particular strengths of the study design, which may have enhanced our ability to support our hypothesis, include the crossover design and associated repeated-measures analysis (which virtually eliminates patient-specific confounding variables); this point is easily unappreciated by those who may simply look at the visual display of the data (for example, Fig. 4) and do not appreciate the advantageous power corresponding to the subject statistically being included as a random effect variable. Another strength is the effort made to assess for potential cigarette smoking or viral syndromes that could affect the results.
However, there are notable caveats to our primary findings. First, we note that pre-treatment with anti-oxidant attenuated baseline airway responsiveness (Fig. 3). Thus, although supplemental anti-oxidant does appear to protect against the effect of DE on airway responsiveness, it may do so more via a prophylactic phenomenon than by directly interfering with active oxidative stress during the exposure period, consistent with the unpublished observations of Riedl and colleagues (Riedl and Diaz-Sanchez, 2005). In other words, our data suggest that 5 days of potent anti-oxidant supplementation can condition the airway to be inherently less responsive to non-specific stimuli such that it is “resistant” to the increase in responsiveness apparent upon exposure to DE. Accordingly, if changes in airway responsiveness pre- to post-exposure are quantified (“post-hoc”; data not shown), then we fail to find any DE-associated increase in airway hyperresponsiveness; this is consistent with the work of Behndig et al. (2011) at a lower concentration (100 μg/m3). Overall, the modest magnitude of DE's effect on airway responsiveness may reflect in part the bronchodilatory effects of nitric oxide, though our subjects’ exposure, approximately 7 ppm (Birger et al., 2011a), was considerably lower than those used clinically for bronchodilation (Group, 1997; Ratjen et al., 1999).
Furthermore, it remains unclear to what extent the effect of DE on airway responsiveness depends on detailed differences in exposure (for example, concentration and duration), participant characteristics (for example, asthma severity), or assessment of DRS (for example, timing and technique). In particular, one concern about the inclusion of baseline airway responsiveness data, in terms of assessing the effect of exposure, is that that any “baseline” that occurs prior to the proper onset of the study procedures is affected by extra-experimental factors beyond the control of the investigators. Although we were very careful to control extraneous factors “within” the time frame during which we worked directly with the subject (from 18 h before the exposure to 30 h post-exposure), we are significantly less able to control them in the period prior to the subject coming for pre-exposure testing. Intra-individual instability in the response to methacholine is well known (Chinn and Schouten, 2005) such that integration of repeated measures simply compounds the likelihood for error due to inherent variability. Thus, we believe that the post-exposure airway responsiveness is most indicative of the study procedures over which we have such advantageous control.
We further recognize as a limitation that the magnitude of change in airway responsiveness is less than what is typically considered of clinical significance. We note however that our primary intention was not to provide data that would argue for or against the use of anti-oxidants clinically; rather we were motivated to directly test, in a carefully controlled in vivo human crossover investigation, the mechanistic role of oxidative stress in this setting. To this end, we opted a priori to use the DRS to most fully use the spectrum of airway responsiveness inherent to our study subjects. Although DRS is not commonly used in the clinical setting, it is advantageous given our motivation (Marcon et al., 2014) as it allows determination of a numerical value even when a FEV1 drops less than 20% following inhalation of the final dose of methacholine (de Meer et al., 2005), a circumstance common to our study setting; relying simply on the PC20 would effectively assign a categorical value (PC20 = 16) to several subjects for whom the DRS can more specifically quantify airway responsiveness and thus enhance the statistical analysis. The 42% increase in DRS, noted in hyper-responsive individuals upon inhalation of DE, is equivalent to at least a doubling of airway responsiveness in terms of the traditional metric PC20 (provocative concentration of methacholine leading to a 20% drop in FEV1), which—if long-term—would be equivalent to the effect of the common asthma controller salmeterol (Cheung et al., 1992).
Our evidence that baseline DRS in hyper-responsive individuals can be reduced with anti-oxidant supplementation is remarkable given the mixed evidence for benefit of supplementary anti-oxidants on asthmatic phenomena (Fogarty et al., 2003). That we were able to show benefit may reflect the relative potency of N-acetylcysteine and our particularly strict administration because 600 mg orally, three times daily, is known to raise lung glutathione levels by 50% after 5 days (Bridgeman et al., 1994). However, our study cannot address the possibility that a longer duration of treatment could have an even more pronounced effect on airway responsiveness. Regardless, this adds to the evidence, from our laboratory (Yamamoto et al., 2013) and others, that anti-oxidant supplementation can attenuate a variety of measurable pollution-related endpoints.
Another notable contribution from our work is the clear differentiation of DE's effect on DRS between those with and without baseline hyper-responsiveness (Fig. 4). In prior work, post-DE changes in airway responsiveness were not greater in asthmatic subjects than in non-asthmatic subjects (Stenfors et al., 2004). That we did find a differential response between the phenotypes may be due to our use of methacholine challenge (rather than airway resistance, as was used by Stenfors et al.(2004)).
Though the current study assessed symptoms at different time points, the effect of DE on symptoms in our Vancouver laboratory approximated that found using a similar exposure system in Seattle (Carlsten et al., 2013); interestingly, the anti-oxidant did not alleviate those symptoms. As another secondary outcome, we evaluated gene-exposure interactions in our study, given the somewhat surprising ability of other studies of similar size and contexts to show such interactions (Gilliland et al., 2004). Though we did not find an overall genotype-by-condition interaction on our primary outcome (likely due to lack of sufficient power, particularly in the absence of adjuvant exposures such as allergen; Gilliland et al., 2004), we did find that subjects with variance in genes with biological plausibility in this context benefited from anti-oxidant in terms of dose-response to methacholine.
We did note a DE-induced increase in another secondary outcome, bronchial epithelial cells, that was abrogated by anti-oxidant (borderline significance). Given the prior data from animal models consistent with our findings, this represents a notable extension of the evidence for damage of the airway lining due to particulate-rich combustion products (Bayram et al., 1998). Because other airway cell types were not significantly affected by DE, consistent with prior study of asthmatics (Nordenhall et al., 2001), this supports the hypothesis that secondary sub-epithelial phenomena may drive major portions of the pathophysiology due to inhaled air pollution (Mills et al., 1999).
CONCLUSION
In summary, our data suggest that anti-oxidant supplementation can abrogate the increase in airway responsiveness induced by inhalation of DE at moderate concentrations in a vulnerable population (those with baseline airway hyper-responsiveness). Those variant in genes of oxidative stress metabolism had less airway responsiveness in the face of DE if taking anti-oxidant supplementation. This protection likely occurs due to pre-exposure moderation of airway responsiveness by N-acetylcysteine. Airway hyper-responsiveness is an important clinical endpoint because it is closely associated with asthma severity and exacerbations (Juniper et al., 1981). Though this study was not designed to directly test the benefit of such an anti-oxidant in diverse settings, future studies may aim to replicate these results and extend them to more generalizable contexts.
FUNDING
Vancouver Coastal Health Research Institute; WorkSafeBC [RS2011-OG07]; British Columbia Lung Association.
We thank the study subjects and also following individuals for helpful discussions at various stages within the project's development: Drs Anders Blomberg, Mark FitzGerald, Peter Pare, Michael Brauer, Thomas Sandstrom, Annelie Behndig, and Guy Marks. We thank Rick White of the UBC Statistical Consulting and Research Laboratory for his review and approval of our statistical methods. Conflict of interest statement: None declared.
These authors contributed equally to this study.
Dr MacNutt currently at Quest University, but at UBC during this project.
The authors certify that all research involving human subjects was done under full compliance with all government policies and the Helsinki Declaration.
REFERENCES
Comments