Effect of 1-year smoking cessation on airway inflammation in COPD and asymptomatic smokers

Subjects

Subjects were recruited from the pulmonary outpatient clinic of the University Medical Centre Groningen (Groningen, the Netherlands) and by advertisements in local newspapers from September 1998 to December 2000. A total of 63 participants were included in the study. COPD patients were recruited according to the European Respiratory Society (ERS) criteria 17. At the analysis, the participants were re-categorised according to the more recent American Thoracic Society (ATS)/ERS guidelines 18. This resulted in three groups of individuals. 1) COPD. This group was categorised, in accordance with the guidelines, as COPD being present when FEV₁/forced vital capacity (FVC) post-bronchodilator was <0.7. The severity of COPD was based on FEV₁ % predicted post-bronchodilator; e.g. ≥80% was mild COPD, ≥50–80% was moderate COPD, ≥30–50% was severe COPD and <30% was very severe COPD. All COPD patients included had chronic respiratory symptoms, i.e. chronic cough and sputum production for at least 3 months for 2 successive yrs. 2) Smokers with chronic bronchitis. Smokers with chronic respiratory symptoms and without airflow limitation, i.e. at risk of developing COPD according to the ATS/ERS criteria were included. 3) Asymptomatic smokers. This group consisted of individuals without chronic respiratory symptoms or airway obstruction, and an FEV₁ ≥85% pred; they also belong to the at-risk group. Since the authors felt that symptomatic and asymptomatic smokers with normal lung function are different groups of smokers, they were analysed separately.

All participants met the following criteria: aged 45–75 yrs, >10 pack-yrs smoking, actual smoking ≥10 cigarettes·day⁻¹, reversibility to salbutamol <9% of the predicted FEV₁, no use of inhaled or oral corticosteroids in the previous 6 months, no signs of atopy (no positive skin-prick test for 10 common aeroallergens and serum total immunoglobulin (Ig)E <200 IU), and no respiratory tract infections 1 month prior to the study. During the study patients only used long-acting or short-acting β₂-agonist or ipratropium on a regular basis; inhaled corticosteroids were not used. Only in cases of an exacerbation was a short course of oral corticosteroids allowed.

The local medical ethics committee (University Medical Centre Groningen, Groningen, the Netherlands) approved the study protocol and all subjects gave their written informed consent.

Study design

Before entering the 1-yr smoking cessation behavioral programme all subjects underwent lung function tests, blood gas analysis, sputum induction and bronchoscopy. Lung function tests and sputum induction were repeated after 2, 6 and 12 months smoking cessation. Bronchoscopy was repeated in patients who successfully quit smoking for 1 yr. Before each measurement, subjects were asked not to use long or short-acting β₂-agonists and/or ipratropium at least 12 h before the test. They did not suffer from a respiratory tract infection nor used oral corticosteroids in the month prior to any of the measurements.

Smoking cessation programme

The smoking cessation behavioural programme consisted of an intensive group-orientated course for 3 months, followed by five meetings throughout the rest of the year. If necessary, nicotine replacements were administered during the first 3 months; no bupropion or antidepressants were prescribed. Measuring cotinine levels in urine verified smoking cessation before, then 2, 6 and 12 months after smoking cessation. A quitter was defined as someone who refrained from smoking for a minimum of 1 yr, with negative cotinine levels (i.e. <25 ng·mL⁻¹) at 2, 6 and 12 months after smoking cessation.

Lung function

Lung functions (FEV₁, FEV₁/vital capacity) were measured using dry wedge spirometry (Masterscope, Jaeger, Breda, the Netherlands) according to standardised guidelines 19. Bronchodilator responsiveness to 400 μg salbutamol was measured and post-bronchodilator values were used. Specific airway conductance (_sG_aw), total lung capacity (TLC) and residual volume (RV) were measured by body plethysmography (Masterscope, Jaeger). Transfer factor of the lung carbon monoxide (T_L,CO) was assessed by the single breath method using standard equipment (Masterscope, Jaeger).

Sputum induction and processing

Sputum was induced by inhalation of hypertonic saline and processed as described previously 20. Sputum samples containing >80% of squamous cells were excluded from analysis because of poor cytospin quality. Sputum cytospin slides were stained with May-Grünwald-Giemsa for differential cell counts, which were performed by one observer. Data on reproducibility have previously been reported 21. IL-8 concentration was measured using ELISA (CLB, Amsterdam, the Netherlands) and eosinophil cationic protein (ECP) concentration was detected using a fluor-enzyme immunoassay (ImmunoCAP ECP, Pharmacia, Uppsala, Sweden).

Bronchoscopy and biopsy processing

Bronchial biopsies were taken from the subcarinae of the right-middle or -lower lobe as described previously 22. Two biopsies were directly snap-frozen and four biopsies were embedded in paraffin.

From the paraffin biopsies serial sections were cut to a thickness of 4 μm and stored at room temperature. Selection of morphological optimal tissue was based on a haematoxylin and eosin (HE) stained slide. Tissue slides were deparaffinised with xylene (15 min) and dehydrated before staining. Immunohistochemical staining was performed with monoclonal antibodies against: CD3 (A0452; DAKO, Copenhagen, Denmark), CD4 (CD4-368; Novocastra, Newcastle upon Tyne, UK), CD8 (M7103; DAKO), B-cells (CD20, L26, M0755; DAKO), mast cell tryptase (AA1, M7052; DAKO), neutrophil elastase (NP57, M0752; DAKO), macrophages (CD68, M0814; DAKO), secreted form of eosinophilic cationic protein EG2 (Pharmacia Diagnostics, Sweden), granzyme B (Mon7029; Monosan, the Netherlands) and T-cell intracellular antigen-1 (Tia-1; PN IM 2550; Immunotech, France). Negative controls were obtained by omission of the primary antibody. Slides were pre-treated with 1mM EDTA buffer pH8 (CD4, CD8), 10 mM citrate buffer pH6 (granzyme B) and 0.1 mM tris-HCL buffer pH9.0 (CD20) in the microwave for 8, 8 or 30 min respectively or with 1% protease for 30 min at room temperature (CD68, NP57, AA1, EG2). CD3 slides were incubated overnight at 80°C with tris-HCL buffer pH9. Slides stained with Tia-1 were not pre-treated. All staining except CD4 was performed automatically using the Dako Autostainer (DAKO). CD4 was done manually. Pre-incubation with 0.3% hydrogen peroxide was used to block endogenous peroxidase.

The dilutions used were: CD3 1:100; CD4 1:25; CD8 1:100; CD68 1:50; EG2 1:200; NE 1:200; AA1 1:100; CD20 1:400; granzyme B 1:5; Tia-1 1:500. The LSAB+(K0690, DAKO) detection system was used except for CD4 where the Envision detection system was used (K5007; DAKO). Either the 3-Amino-9-Ethyl Carbazole (AEC; K3469; DAKO) or for CD4, Nova Red (SK4800; Vector, USA) were used as a chromogen (substrate) giving a reddish-brown reaction product. Haematoxylin was used as a counterstain.

The frozen biopsies were used to investigate expression of vascular adhesion molecules. Serial sections were cut with a thickness of 4 μm and stored at -20°C until used. Selection of morphological optimal tissue was based on HE stained slides. Before immunostaining the sections were dried and fixed in acetone for 10 min and washed in PBS. Double immunostaining was performed to visualise all vessels using a monoclonal antibody against CD31 (MON1142; Monosan), which is present on all endothelial cells in combination with monoclonal antibodies against the adhesion molecules: intercellular adhesion molecule (ICAM)-1 (anti-CD54; Novocastra), vascular cell adhesion molecule (VCAM; CD106; Genzyme, Cambridge, MA, USA), P-selectin (CD62P; BD Biosciences, San Jose, CA, USA), and E-selectin (CD62E; Genzyme). The concentrations used were: ICAM 1:800, VCAM 1:1000, P-selectin 1:2500 and E-selectin 1:700. Labelling of anti-CD31 was performed by isotype specific biotinylated goat anti-mouse Ig and subsequently by streptavidin conjugated to alkaline phosphatase, providing a blue reaction product using Fast Blue BB salt as a chromogen. Labelling of the adhesion antibodies was performed by isotype-specific goat anti-mouse antibody conjugated to peroxidase, using AEC as a reagent, giving a reddish brown reaction.

For each immunohistochemical staining, except for granzyme B and vascular adhesions molecules, sections from two bronchial biopsies were quantified by computer-assisted image analysis at ×200 (Qwin; Leica Microsystems Imaging Solutions Ltd, Cambridge, UK). Automated image analysis to quantify immunopositivity used the next algorithm: first the intensity of the positive area (cells) was appointed in each biopsy by the observer, followed by the intensity of the total area of the biopsy based on the red-green-blue colour model. After this, all images of the biopsy were analysed; excluded were epithelium, submucosal glands, airway smooth muscle and damaged tissue. Afterwards, the algorithm analysed the positive area and the measurable area of the biopsy leading to the percentage-positive area per biopsy. A positive area was ≥11.8 μm², to exclude false-positive areas. The minimal measurable area was 0.4 mm². The mean percentage-positive area of two biopsies was used. Measurements were performed in a blinded way by one observer. Part of the biopsies were studied by both observers to determine inter-observer reproducibility, using Bland and Altman 23 analysis, showing good agreement between the two observers (mean difference 0.01±0.07). Intra-observer reproducibility of the cellular parameters showed a mean difference ranging from 0.008 for EG2 to 0.09 for CD8 measurements. In addition, 95% of the differences of repeated measurements were between ±2 sd of the mean of differences and thus the reproducibility of inflammatory cells and markers was satisfactory in this respect.

Quantification of granzyme B was performed using morphometric measurements with a light microscope (Olympus BX40, Olympus, Japan) connected to a camera linked to a computerised image system (Research Assistant v4; RVC, Soest, the Netherlands) to measure the total measurable area (excluding epithelium, submucosal glands, airway smooth muscle tissue and damaged tissue). The number of granzyme B-cells was counted in each biopsy using a light microscope, at ×200, in a blinded manner by one observer. The mean of two bronchial biopsies was used to calculate the number of positive cells per mm⁻².

The vascular adhesion molecules were counted using a light microscope at ×400. All vessels in one section were scored positive or negative for immunopositivity for an adhesion molecule and expressed as a percentage of the total number CD31 positive vessels. A minimum of 20 vessels were counted.

Data analysis

Values of p<0.05 were considered statistically significant. Clinical data were expressed in means±sd, inflammatory variables were expressed in medians (minimum–maximum). Differences between smokers with COPD, smokers with chronic bronchitis and asymptomatic smokers at baseline were analysed using the Kruskall-Wallis test, a nonparametric equivalent to one-way ANOVA. When the Kruskall-Wallis test was significant the Mann-Whitney U-test was used to analyse the observed differences between the three groups. In the group of subjects who successfully managed to quit smoking, differences in sputum inflammation before and after 2, 6 and 12 months, smoking cessation were analysed using the Friedman's nonparametric two-way ANOVA, a nonparametric test to investigate multiple paired time points. If the Friedman test was significant, the Wilcoxon signed-rank test was used to define which time-points were different. Differences in airway inflammation in bronchial biopsies before and after 12 months smoking cessation were analysed using the Wilcoxon signed-rank test. Since the authors wanted to compare airway inflammation in sputum and bronchial biopsies, the differences in sputum airway inflammation before and after 12 months smoking cessation were also analysed separately using the Wilcoxon-signed rank test. The Mann-Whitney U-test was used to compare baseline characteristics between the group of patients who quit smoking for 1 yr and the group of patients who did not.

Patient characteristics

A total of 28 smokers with COPD (13 mild, eight moderate and seven severe), 10 smokers with chronic bronchitis and normal lung function and 25 asymptomatic smokers with normal lung function, were included in the 1-yr smoking cessation programme. A total of 12 COPD patients (six mild, four moderate and two severe), four smokers with chronic bronchitis and 16 asymptomatic smokers successfully quit smoking for 1 yr. All urinary cotinine levels confirmed smoking cessation at 2, 6 and 12 months. None of the participants used nicotine replacements. Patient characteristics are presented in table 1⇓.

Successful quitters with COPD had slightly lower percentage of sputum lymphocytes at baseline and a higher number of granzyme B+ cells and higher percentage of ICAM-positive vessels than nonquitters with COPD (0.7 versus 1.1, 0.18 versus 0, and 86 versus 70%, respectively). The asymptomatic smokers who successfully quit were younger and the percentage of eosinophils in sputum was higher at baseline (49 versus 52 yrs and 1.1 versus 0.4%, respectively) than in the asymptomatic nonquitters. The effects of smoking cessation on airway inflammation in smokers with chronic bronchitis could not be evaluated since only four smokers successfully quit smoking for a 1-yr period.

At baseline, COPD patients were slightly older, had more pack-yrs smoking and lower lung function than asymptomatic smokers with normal lung function (table 1⇑). Sputum volume, total cells, percentage of sputum neutrophils and sputum IL-8 levels in COPD were significantly higher than in asymptomatic smokers (tables 2⇓ and 3⇓). Inflammatory markers in bronchial biopsies did not differ between the two groups at baseline (table 4⇓). COPD patients had less P-selectin positive vessels than asymptomatic smokers (table 4⇓).

Effect of smoking cessation on lung function

FEV₁ post-bronchodilator (% predicted) improved slightly after a 1-yr smoking cessation in COPD. No changes were found in FEV₁/FVC post-bronchodilator (%), RV/TLC, T_L,CO or _sG_aw (table 1⇑). Only T_L,CO improved significantly after a 1-yr smoking cessation in asymptomatic quitters (table 1⇑).

Effect of smoking cessation on airway inflammation in sputum

Part of these data have been presented in a previous article in which the effect of smoking cessation on airway hyperresponsiveness is described in relation to sputum inflammation in smokers with COPD 21.

All 12 quitters with COPD, and 15 of the 16 asymptomatic quitters, produced both at baseline and at 2, 6 and 12 months after smoking cessation sufficient sputum of good quality (tables 2⇑ and 3⇑).

In COPD, cell concentration, number of neutrophils, number of lymphocytes, IL-8 and ECP levels in sputum increased significantly after 12 months of smoking cessation (table 2⇑, fig. 1⇓). The percentages of eosinophils and macrophages tended to decrease after 12 months smoking cessation (p = 0.06 and 0.08, respectively).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1—

Effect of 1-yr smoking cessation on airway inflammation in sputum from chronic obstructive pulmonary disease patients (•) and asymptomatic smokers (○) with normal lung function. Cell conc.: cell concentration; epithelial cells: bronchial epithelial cells; IL: interleukin; ECP: eosinophilic cationic protein. *: p<0.05.

In asymptomatic quitters, number and percentage of macrophages, percentage eosinophils and IL-8 levels in sputum decreased significantly after a 12-month smoking cessation (table 3⇑, fig. 1⇑). The percentage of neutrophils increased significantly in asymptomatic smokers at 6 and 12 months after smoking cessation was introduced, whereas the number of neutrophils did not (table 3⇑).

The response to smoking cessation was significantly different between the two groups for cell concentration, the number of neutrophils, macrophages and epithelial cells and, IL-8 and ECP levels in sputum (table 2⇑, 3⇑ and fig. 2⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2—

Change (Δ) in sputum inflammation after 1-yr smoking cessation in chronic obstructive pulmonary disease patients (•) and asymptomatic smokers (○) with normal lung function. Cell conc: cell concentration; epithelial cells: bronchial epithelial cells; IL: interleukin; ECP: eosinophilic cationic protein. *: p<0.05.

Effect of smoking cessation on airway inflammation in bronchial biopsies

Data on bronchial biopsies are presented in table 4⇑ and figure 3⇓. Biopsies of one COPD patient and of two asymptomatic smokers were of insufficient quality and therefore excluded from analyses. In COPD, none of the inflammatory cell markers and vascular adhesion molecules showed a significant change after ceasing to smoke for a 12-month period (table 4⇑).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3—

Effect of 1-yr smoking cessation on airway inflammation in bronchial biopsies in chronic obstructive pulmonary disease patients (•) and asymptomatic smokers (○) with normal lung function. CD20: B-cells. *: p<0.05.

In asymptomatic quitters, mast cells (tryptase+cells) decreased significantly, whereas B-cells (CD20) increased significantly at 12 months smoking cessation (table 4⇑, fig. 3⇑). The CD3+ cells tended to increase after smoking cessation (p = 0.09). The response to smoking cessation was significantly different between the two groups for mast cells.

The current authors also re-categorised the COPD patients according to the COPD criteria of the Global Initiative for Chronic Obstructive Lung Disease 24, which did not change the results (data not shown).