INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: interleukin-4; cytokines; chronic bronchitis; COPD; smoking; lymphocytes

Chronic productive cough is the clinical hallmark of chronic bronchitis (CB) and, with increasing severity, mucus hypersecretion is associated with increasing morbidity, hospitalization, and mortality (1). Histologically, CB is usually associated with hypertrophy of bronchial mucus-secreting submucosal glands, and the increased size of the glands has been traditionally thought to account for the increased production of sputum (2). However, scores of the inflammation, reported to consist mainly of lymphocytes and plasma cells, in and around the mucus-secreting glands have shown a better concordance with mucus hypersecretion than gland size per se (3). The mechanisms for developing the morphological changes of enlargement and hypersecretion by airway submucosal glands are still unclear. Although interleukin (IL)-4 is known to be implicated in the pathogenesis of asthma and airway wall eosinophilia (4, 5), IL-4 is also known for its capacity to induce mucin production and secretion (6, 7). However, its potential role in the development of CB in humans is unknown.

Immunohistological analyses of endobronchial biopsies from smokers with CB with or without airflow obstruction have shown that compared with nonobstructive CB, when there is also airflow obstruction (i.e., CB + [chronic obstructive pulmonary disease] COPD) there are significantly higher numbers of (CD3⁺) T-lymphocytes and (CD68⁺) macrophages (8). Neutrophils, mast cells, B-lymphocytes, and eosinophils are not increased significantly and plasma cells have not been quantified due to lack of an immunohistological marker specific for them. The predominant T cell subset in the airway wall of large airways of subjects with CB + COPD is the (CD8⁺) T-cytotoxic/suppressor phenotype and there is a trend to an increase in the CB group without airflow obstruction (9, 10). There is a significant inverse relationship of CD8⁺ cell number and FEV₁ % of predicted in both large and small airways (9, 11) and this cell phenotype has become of central interest in the generation of smoke-induced CB and COPD.

Comparing CB and COPD with asthma, the distinct predominance of the respective CD8 and CD4 T cell subsets and the relative recruitment of neutrophils and eosinophils, together with other distinctive features of these two conditions, led us to the hypothesis that the pattern of regulatory cytokines would also be different (12). Accordingly, we initially predicted there would be a lack of expression of the regulatory cytokines IL-4 and IL-5 in CB + COPD. However, Mueller and colleagues, in an immunohistological study of bronchial biopsies, reported an increase in the number of cells containing IL-4 protein in the airway mucosa of patients with CB as compared with a group of normal healthy controls (13). In addition, the last study reported a lack of cells containing IL-5 protein in CB. Detection of IL-5 by immunohistological staining has, however, proved difficult. We wished, therefore, to apply in situ hybridization to detect the mRNA for these two regulatory cytokines and we wanted to examine both the subepithelium (i.e., mucosa) and submucosal mucus-secreting glands. As endobronchial biopsy samples contain little, if any, gland we obtained tissues freshly at surgical resection for tumor from carefully characterized groups of smokers. Of the smokers who fulfilled the criteria for CB, some had and others did not have airflow obstruction (i.e., COPD) as defined by internationally recognized guidelines. We predicted that if there was gene expression for IL-4 or IL-5 that it would most likely be expressed by CD8 T cells of the type 2 CD8 cytotoxic phenotype (14, 15).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Three groups of smokers (> 20 packs/yr) coming to surgery for lung tumor were included whose overall age range was 50-80 yr. Group I asymptomatic smokers (AS, n = 11)were nonatopic, nonasthmatic, male and female patients with normal lung function. Group II chronic bronchitis (CB, n = 11)were patients with chronic cough and sputum as defined by the Medical Research Council (MRC) criteria for CB (16) but with normal lung function (i.e., FEV₁ > 80% of predicted; FEV₁/FVC > 70%). Group IIICB and coexisting COPD (CB + COPD, n = 10)were CB patients with chronic persistent irreversible airflow obstruction with FEV₁ < 80% (i.e., range 43-77%) of predicted as well as FEV₁/FVC < 70% (i.e., range 51-68%) with cough and sputum satisfying the British Thoracic Society recommendations for the diagnosis of COPD (17) and MRC criteria for CB (16). These patients each had < 10% reversibility or < 200 ml increase of FEV₁ in response to inhaled ₂-agonist. Atopic patients or patients with a history of wheezing, asthma, allergic rhinitis, or any allergic disease were excluded from the study. None of the patients studied had had a respiratory infection within the preceding month. Similarly, none had received antibiotics or inhaled or oral corticosteroids within the preceding month.

Processing of Tissue

The tissues obtained at surgical resection for cancer were obtained through the departments of respiratory diseases at Ghent University Hospital (Belgium) and at the University of Turin (Italy). Grossly normal airways far from tumor and not blocked by tumor were selected to include lobar and segmental bronchi (i.e., airways with cartilage and mucus-secreting glands). These were freshly fixed in 10% buffered formaldehyde and processed to paraffin wax. Serial sections 5 µm thick were first cut and stained with hematoxylin-eosin (H&E) in order to visualize the morphology and conduct counts of the total number of inflammatory cells. A monoclonal antibody (mAb) directed against the cell surface epitope of CD8⁺ cells was applied to the sections and the cells were identified by immunohistochemistry. IL-4 and IL-5 cRNA were used for in situ hybridization (ISH) procedures applied to tissue sections.

In double labeling experiments, cells were immunostained for CD8⁺ cells as well as being hybridized for IL-4 or IL-5 or tumor necrosis factor (TNF)- mRNA.

Immunohistochemistry

The avidin-biotinylated peroxidase complex (ABC) technique was used to identify CD8⁺ cells as previously described but with minor modification (18). Briefly, following microwave antigen retrieval (1 mM EDTA at pH 8.0; 5 min), the sections were incubated with 0.3% hydrogen peroxide (H₂O₂) in distilled water and then immunostained for 60 min with appropriately diluted (1 in 50 in phosphate-buffered saline [PBS]) monoclonal antibodies to CD8⁺ cells (DAKO, Cambridge, UK, M7103). Sections were washed and incubated for a further 30 min with biotinylated rabbit antimouse immunoglobulins (DAKO E0464) diluted 1 in 300. Sections were then washed and incubated for 50 min with avidin complexed with biotinylated peroxidase (DAKO K0355). After a further wash, bound peroxidase was detected as a dark brown product following 10 min incubation with 0.03% H₂O₂ and 0.5 mg/ml diaminobenzidine (DAB). The slides were counterstained with Harris' hematoxylin to provide morphological detail, then mounted in DPX. We used an irrelevant antibody: mouse immunoglobulin G (IgG)₁ (MOPC21 Sigma, Dorset, UK, M7894), as a replacement for the primary mAb as a negative control procedure.

Preparation of cRNA Probes

Digoxigenin (Dig)-labeled and ³⁵S-labeled antisense and sense complementary ribonucleic acid (cRNA) probes were generated from complementary deoxyribonucleic acid (cDNA) according to a well-tried and published method (19). One microgram linearized DNA, 2 µl (of 10× concentration) Dig- or ³⁵S-NTP: adenosine, guanosine, cytidine, and Dig- or ³⁵S-uridine triphosphate (ATP, GTP, CTP, and Dig- or ³⁵S-UTP) labeling mixture was added to 2 µl (of 10× concentration) transcription buffer followed by 2 µl SP6 or T7 RNA polymerase made up to a total of 20 µl with RNase-free double-distilled water to produce either antisense or sense probes according to the orientation of each particular probe. After incubation for 2 h at 37° C, 20 U DNase I, RNase free was added and incubated for 15 min at 37° C. Ethylenediamine tetraacetic acid (EDTA), 2 µl, 0.2 M, pH 8.0, was added to stop the reaction and precipitate the probe with cold ethanol. The sample was centrifuged and dried under vacuum. To check transcription efficiency and quantify the probe, 1 µl Dig-RNA was taken for electrophoresis in 1.2% agarose gel and 1 µl ³⁵S-RNA in scintillation was counted, after which it was stored at 80° C. The sizes (base pairs), vectors, and sources of the cDNA probes were as follows: IL-4 (318 bp; PGEM-1), IL-5 (412 bp; PGEM-IV; both from Glaxo Wellcome Biomedical Research, Geneva, Switzerland), and TNF- (500 bp; PGEM-3Z; R & D Systems Europe Ltd, Oxon, UK).

Nonisotopic in Situ Hybridization (ISH)

The sections were deparaffinized, rehydrated, and washed in PBS then incubated with 5 µg/ml proteinase K in 100 mM Tris-HCl pH 8.0, 50 mM EDTA for 30 min at 37° C to permeabilize the cells. After postfixation in 4% paraformaldehyde in PBS, hybridization buffer (2× Denhardt's solution, 50 µg/ml salmon sperm DNA, 100 µg/ml yeast tRNA, 50% formamide) containing 50 ng-100 ng/ml Dig- labeled cRNA probe was added and incubated at 42° C overnight. Sense probes were used as the most appropriate negative control. Slides were washed in solution of 2 × sodium chloride and trisodium citrate (SSC) at 37° C and incubated in 20 µg/ml RNase A in 2 × SSC. Then posthybridization washes were carried out with two changes of 2 × SSC at 37° C. The remaining probe hybridized with the mRNA of interest detected with anti-Dig antibody conjugated with alkaline phosphatase (Rote Diagnostics Limited, Sussex, UK). Detection was carried out by addition of 5-bromo-4-chloro-3-indolyl phosphate/ nitroblue tetrazolium (NBT/BCIP) substrate (Roche Diagnostics Limited, Sussex, UK) and incubation for 30-60 min. The counterstain used was nuclear fast red.

Isotopic in Situ and Immunohistochemistry Double Staining

Isotopic in situ and immunohistochemical techniques were used in sequence to investigate whether CD8⁺ cells contained mRNA for IL-4 or IL-5 or TNF-. The TNF- probe was used as a positive control. ³⁵S in situ hybridization was used to detect cytoplasmic mRNA for IL-4, IL-5, and TNF- and alkaline phosphatase-antialkaline phosphatase (APAAP) immunostaining was used to detect the membrane epitope identifying CD8⁺ cells.

The hybridization procedure was carried out as a first sequence according to the method previously described (20). Briefly, 8-µm-thick sections were dewaxed, rehydrated, and digested by 1 µg/ml proteinase K in 100 mM Tris-HCl pH 8.0, 50 mM EDTA for 15 min at 37° C. Following postfixation in 4% paraformaldehyde in PBS, the sections were immersed in 0.25% acetic anhydride, 0.1 M triethanolamine. Then 60 µl containing nine volumes of hybridization buffer and one volume of ³⁵S-labeled IL-4, IL-5, and TNF- sense or antisense probes to give 1.5-3 × 10⁶ counts per minute per section was applied, respectively. Sections were hybridized overnight at 42° C. Posthybridization washes were carried out with two changes of 2 × SSC and 10 mM dithiothreitol at 37° C. Unhybridized single-stranded RNA was selectively removed with 20 µg/ml RNase at 37° C. Sections were then immersed in Tris-buffered saline (TBS) pH 7.4 followed by incubation for 60 min with mAb to CD8⁺ T cells diluted 1:50 in TBS containing 20% normal human serum. Sections were washed and incubated for a further 60 min with unconjugated rabbit antimouse immunoglobulin (DAKO Z0259) diluted 1 in 25 in TBS containing 20% of normal human serum. Sections were then washed and incubated for 30 min with APAAP mouse mAb (DAKO D0651) diluted 1 in 50 with TBS containing 20% of normal human serum. After a further wash, bound alkaline phosphatase was detected as a red product following 20 min incubation with Naphthol AS-MX phosphate and 1 mg/ml New Fuschin. The slides were air dried and dipped in K5 emulsion (Ilford Scientific Products, Cheshire, UK) in dark, stored at 4° C in light-tight boxes, and exposed to the radioisotope for 14 d. The autoradiograms were developed in Kodak D-19 developer (Jemmerton Ltd, London, UK) and fixed in Ilford Hypam rapid fixer (Ilford Scientific Products) and counterstained and then mounted in Aqua perm mounting medium. We used an irrelevant antibody: mouse IgG₁ (MOPC21 Sigma, Dorset, UK, M7894) for the primary layer as negative control procedures.

Quantification and Statistics

Areas of epithelium, interstitial areas in the subepithelium (a zone between the external edge of the reticular basement membrane and the inner aspect of the bronchial muscle), areas of submucosal mucus-secreting gland (including interacinar interstitium, secretory acini, and ducts), and the remainder of the airway wall (excluding cartilage) of each entire section were assessed using Apple Macintosh computer Image 1.5 software (Apple Mac, Cupertino, CA). IL-4 and IL-5 mRNA⁺ cells and CD8 immunopositive cells were counted using a Leitz Dialux 20 light microscope (Leitz Wetzlar, West Germany) at ×200 magnification using an eyepiece graticule divided into 100 squares in order to ensure no overlap of fields of view. Slides were coded to avoid observer bias. To test the consistency of quantification and the inherent variation of repeated counts, one section was selected and counted five times over the period of the entire study. The sections, which had been hybridized with the sense (negative) probe, were examined first in order to detect and be aware of the background signal. Following the observation of increased numbers of IL-4 positive cells in the CB group as compared with the CB + COPD group we wished also to determine the total numbers of inflammatory cells in these two groups. We used the H&E-stained sections to count the total number of inflammatory cells infiltrating the inner wall, in the subepithelial zone and in the interstitium between gland acini.

The data for single staining are expressed as number of positive cells/mm² tissue for four zones: (1) epithelium, (2) a subepithelium, (3) that containing mucus-secreting gland acini, and (4) the remainder of the airway wall. The data for double staining for CD8/IL-4, CD8/ IL-5, and CD8/TNF- were expressed as percentage of double stained cells of the total positive cells stained for CD8 or each of cytokines in the subepithelial zone. The coefficient of variation (CV = SD/mean × 100) was used to express the error of repeat counts in a single section of a patient with CB. The errors for repeat counts for the numbers of IL-4⁺ cells in the subepithelium and submucosal glands by one observer (J.Z.) were 5.6% and 4.3%, respectively. As the data were nonnormally distributed, the Mann-Whitney U test was applied to compare differences amoung the three patient groups. The Spearman rank correlation test was used to determine the correlation between the number of CD8⁺ cells and IL-4 or IL-5 mRNA⁺ cells and for the relationship between the numbers of IL-4⁺ cells and total inflammatory cells. A p value < 0.05 was accepted as indicating a significant difference for both the Mann-Whitney U test and Spearman rank correlation.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The characteristics of the asymptomatic smokers and those with CB or CB + COPD are shown in Table 1. The three groups of smokers were similar with regard to age and smoking history. By comparison with the AS and CB groups, the CB + COPD group had a significantly lower value for FEV₁ % of predicted (p < 0.0001 for both groups) and FEV₁/FVC (p < 0.0001 for both groups).

IL-4 and IL-5 mRNA Expression

Inflammatory cells were located throughout the airway wall, in the subepithelium and also in the submucosa, in the interstitium between the secretory acini of the mucus-secreting glands. Compared with the AS and CB + COPD groups, there appeared to be greater numbers of inflammatory cells associated with the submucosal glands of the CB group and this was confirmed by cell counts. The counts expressed as number of inflammatory cells/mm² tissue were in the subepithelium (median and range): CB = 634 (275-948) versus COPD = 206 (97-605) (p < 0.002) and gland: CB = 410 (181-729) versus COPD = 151 (72-246) (p < 0.001). Most of the increased inflammatory cell infiltrate in CB appeared to be plasma cells, identified by their typical nuclear morphology and abundant basophilic cytoplasm, but neutrophils were also present.

Gene expression. mRNA positivity was detected by in situ hybridization as blue/black cytoplasmic staining. In the CB group, there was an abundance of strongly IL-4 mRNA positive mononuclear cells located in the subepithelium (see Figure 1A) and particularly in the interstitial tissue between the mucus-secreting gland acini (Figure 1B) and ducts. The serous demilunes of the glands were also mildly positive. IL-4 mRNA was also localized to the surface epithelium (see Figure 1A), particularly associated with basal cells and areas of squamous metaplasia.

View larger version (116K): [in this window] [in a new window]   View larger version (134K): [in this window] [in a new window]   View larger version (140K): [in this window] [in a new window]   View larger version (127K): [in this window] [in a new window]   View larger version (127K): [in this window] [in a new window]   Figure 1.   Nonisotopic in situ hybridization of a tissue section of a bronchial resection from a patient with CB: (A) the antisense probe shows strong cytoplasmic expression of IL-4 mRNA (dark blue) in mononuclear inflammatory cells located in a subepithelial zone and moderate intensity of IL-4 gene expression in some of the epithelial cells also (arrowhead). (B) IL-4 gene expression in the interstitial tissue located between the submucosal gland acini. (C ) There is relatively weak expression of IL-5 in mononuclear cells in the subepithelial zone and (D) in the interstitial tissue between gland acini; (E ) there was no positivity using the IL-4 sense probe. Original magnification: ×230.

Quantification. Counts of IL-4 and IL-5 mRNA positive cells was made in the three groups subjects. These data are expressed as the number of positive cells/mm² of (1) subepithelium, (2) submucosal gland, and (3) the remainder of the inner airway wall. The data are shown as median and range in Figures 2 and 3. In each of the AS, CB, and CB + COPD groups, IL-4 and IL-5 mRNA⁺ cells were located mainly in the glandular and subepithelial zones rather than the remainder of the airway wall. There were significantly more IL-4 mRNA⁺ than IL-5 mRNA⁺ cells in both subepithelial and glandular compartments and this difference was so in each of the three subject groups (p < 0.01). There were significantly higher numbers of IL-4 mRNA⁺ cells in CB patients than in the asymptomatic and COPD groups in both subepithelial (p < 0.05) and glandular compartments (p < 0.01) (see Figure 2). In the CB group, the number of IL-4 mRNA⁺ and IL-5 mRNA⁺ cells associated with submucosal glands was greater than that of the subepithelium (p < 0.05) (Figures 2 and 3). The numbers of cells expressing IL-4 showed a strong correlation with the total numbers of inflammatory cells counted in both the subepithelial zone and gland areas in CB (r = 0.60, p < 0.01 and r = 0.70, p < 0.02, respectively) (Figure 4). There was also a significant correlation between IL-4 mRNA⁺ cells and the numbers expressing IL-5 in each of the three subject groups (r > 0.66, p < 0.01). However, although there were similar trends between the numbers of cells expressing IL-4 and IL-5, there were no significant between-group differences in respect to IL-5 (Figure 3).

View larger version (18K): [in this window] [in a new window]   Figure 2.   Counts for IL-4 mRNA positive cells in distinct zones of the airway wall of asymptomatic (AS) smokers, and smokers with chronic bronchitis (CB) and CB with airflow obstruction (CB + COPD). The results are expressed as the number of positive cells/mm2 subepithelial zone: individual patient values and medians for the groups. The Mann-Whitney U test was used to compare the differences between distinct tissue zones and subject groups. G = gland; S = subepithelial zone; R = remainder of the airway wall.

View larger version (15K): [in this window] [in a new window]   Figure 3.   Counts for IL-5 mRNA positive cells in distinct zones of the airway wall of AS smokers and those with CB and CB + COPD. The results are expressed as the number of positive cells/mm2 subepithelial zone: individual patient values and medians for the groups. The Mann-Whitney U test was used for the comparison of differences between distinct zones and groups. G = gland; S = subepithelial zone; R = remainder of the airway wall.

View larger version (15K): [in this window] [in a new window]   Figure 4.   Correlations between the numbers of IL-4 mRNA positive cells and total number of inflammatory cells/mm2 of subepithelium and submucosal glands in both the CB and CB + COPD groups (Spearman rank correlation; n = 21).

CD8⁺ Cells and Their Association with IL-4 or IL-5 mRNA Expression

CD8⁺ inflammatory cells were particularly abundant in the surface epithelium and in the subepithelial zone immediately adjacent to the surface epithelium (Figure 5). By contrast, they were relatively sparse in submucosal glands. The data for the counts of CD8⁺ T cells are expressed as number of positive cells/mm² of (1) epithelium, (2) subepithelial zone, (3) gland, and (4) the remainder of the airway wall and are shown in Figure 6 as median and range.

View larger version (106K): [in this window] [in a new window]   Figure 5.   Light microscopic image of immunostained bronchial mucosa from a patient with CB. CD8+ cells (stained brown with the ABC technique) were particularly abundant in the surface epithelium and in a subepithelial zone adjacent to the surface. Original magnification: ×230.

View larger version (18K): [in this window] [in a new window]   Figure 6.   Counts for CD8+ cells in distinct zones of the airway wall of AS patients and those with CB and CB + COPD. The results are expressed as the number of immunopositive cells/mm2 subepithelial zone and show individiual and median values. The Mann-Whitney U test was used to compare differences between distinct zones and groups. E = epithelial area; G = gland; S = subepithelial zone; R = remainder of the airway wall.

We predicted that because CD8⁺ cells are frequently found in smokers, the IL-4 and IL-5 mRNA would be colocalized to CD8⁺ T cells of the Tc2-like phenotype. However the predominant locations of the CD8⁺ cells and of IL-4 and IL-5 mRNA⁺ cells were distinct (see Table 2). Furthermore, when we looked for associations by correlation between CD8⁺ cells and IL-4 or IL-5 mRNA⁺ cells in the three groups, none was found. In the CB group, r and p values for IL-4 and CD8 cells in the subepithelial zone and gland area were r = 0.43, p = 0.10 and r = 0.12, p = 0.99, respectively. The values for IL-5 and CD8 in the subepithelial zone and gland area were r = 0.09, p = 0.20 and r = 0.01, p = 0.10, respectively.

Double labeling was used to identify and count the CD8/IL-4 or CD8/IL-5 double-labeled cells that occurred. Double labeling of CD8/TNF- was used as a positive control for cells expressing both signal: 34% of TNF- mRNA⁺ cells were also CD8⁺ and 15% of CD8⁺ cells were TNF- mRNA⁺ (Figure 7). By contrast of 1328 IL-4 mRNA⁺ and 1404 CD8⁺ cells counted, none was double labeled (Figure 8). And of 727 IL-5⁺ and 1569 CD8⁺ cells, none was double labeled. We conclude that cells other than CD8⁺ T-cells produce IL-4 and IL-5 in smokers with CB.

View larger version (132K): [in this window] [in a new window]   Figure 7.   A double-labeling experiment to demonstrate the result of immunostaining for CD8+ cells by the APAAP technique using new fuschin (red) combined with isotopic in situ hybridization for TNF- mRNA shown as dense clusters of silver grains in some cells colocalized to CD8+ cells (arrow) in a case from the CB group. About 15% of CD8+ cells expressed mRNA for TNF-. Original magnification: ×230.

View larger version (133K): [in this window] [in a new window]   Figure 8.   A double-labeling experiment in a CB patient showing the result of immunostaining for CD8+ cells by the APAAP technique (red) combined with isotopic in situ hybridization for IL-4 mRNA. None of the CD8+ cells expressed mRNA for IL-4. Original magnification: ×230.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The main findings of the present study are the evidence for strong IL-4 gene expression by relatively large numbers of inflammatory cells infiltrating the bronchial subepithelium and submucosal mucus-secreting glands of smokers with CB. There is a relative reduction of IL-4 expression and of the total numbers of inflammatory cells when there is coexisting airflow obstruction (CB + COPD). The IL-4 or IL-5 gene expression present is not expressed by cells of the CD8⁺ phenotype present in these tissue compartments.

In support of our findings of IL-4 gene expression, there is one previous immunohistological study, of IL-4 in bronchial biopsies taken from subjects with chronic bronchitis (without airflow obstruction), in which the authors use a semiquantitative score to demonstrate an increase of IL-4 immunopositive cells in CB by comparison with normal healthy nonsmokers (13). Our finding of an increased IL-4 mRNA expression in our CB group confirms and extends these results by a quantitative comparison with CB + COPD and asymptomatic groups and by examination of submucosal glands not accessible to study by examination of endobronchial biopsies. Our examination of resection material has allowed us to count and compare IL-4 positive cells in both the subepithelium and glandular zone deep within the airway wall.

What then is the potential role(s) of IL-4 in CB? The critical role of IL-4 and other Th2-type cytokines in the development of allergic inflammation and asthma is, of course, well recognized (5). IL-4 induces an immunoglobulin isotype switch to IgE in B cells: this is important to the allergen-specific priming of mast cells involved in the immediate reaction to allergen exposure (4). IL-4 up-regulates the IgE receptor on mast cells (21) and B-lymphocytes and mononuclear phagocytic cells (22, 23). IL-4 is associated with directly stimulating the increased expression of vascular cell adhesion molecule-1 (VCAM-1) on the surface of endothelial cells (24) and intercellular adhesion molecule-1 (ICAM-1) on bronchial epithelial cells (25), selectively promoting the migration of eosinophils from blood vessels to the airway wall epithelium and lumen.

Apart from the regulatory role and direct action of IL-4 on effector cells in allergic inflammation, this cytokine appears to have actions of relevance to CB and COPD also. IL-4 induces mucous glycoprotein synthesis and results in the accumulation of mucus in nonciliated epithelial cells; it also promotes goblet cell hyperplasia and the secretion of airway mucus (6, 7). The presence of relatively large numbers of inflammatory cells expressing this cytokine, particularly in the subepithelium and in the submucosal glands of our symptomatic, smokers with bronchitis would be in keeping with a prosecretory effect of such inflammation and support the earlier conclusions of Mullen and coworkers (3). We speculate that the persistence of IL-4 promotes both the hypertrophy and hypersecretion of bronchial submucosal glands and the goblet cell hyperplasia characteristic of CB.

IL-4 also stimulates T- and B-lymphocytes, plasma cell differentiation, and mast cell growth. Interestingly in stable CB, the airway wall infiltration of inflammatory cells is composed predominantly of lymphocytes and plasma cells, with only occasional neutrophils and eosinophils (26, 27). Thus, IL-4 may be involved in B-lymphocyte cell maturation and result in differentiation and a consequent increase of plasma cells in the airway mucosa, particularly in the submucosal glands of CB. Mast cell numbers are also reported to be increased in the airway epithelium of smokers with CB and COPD (13, 28). Our recent findings (unpublished) demonstrate a significant increase both in the number of subepithelial mast cells and of submucosal gland plasma cells in CB as compared with COPD and asymptomatic smokers.

In the absence of interferon (IFN) IL-4 has also been shown to have profibrotic effects by stimulating fibroblasts to proliferate and secrete collagen (29). This profibrotic role may be of relevance to the airway fibrosis reported, particularly in the smaller airways of smokers with COPD (30). IL-4 also promotes the differentiation of naive T helper (Th)-0 lymphocytes into Th2 lymphocytes that then secrete IL-4, IL-5, IL-9, and IL-13 (31). The last is also a powerful inducer of mucous cell metaplasia (32).

We demonstrate in smokers with CB that cells expressing IL-4 mRNA were about three times more frequent and showed stronger expression than those expressing IL-5 mRNA. Although the numbers of cells expressing IL-5 were fewer than IL-4, there were still significantly higher numbers of cells expressing IL-5 in the submucosal glands than in the subepithelium of subjects with CB. Our results also show that there is a significant positive correlation between the number of inflammatory cells expressing IL-4 and IL-5 mRNA in both the subepithelial zone and gland regions. However, the role of IL-5 in CB is unclear. In the context of a tissue eosinophilia, IL-5 has the direct effect of promoting the terminal differentiation of bone marrow eosinophils and their release from the bone marrow into the circulation where their numbers increase (33). As a regulatory molecule IL-5 could be expressed at moderate levels, inducing a moderate peripheral blood eosinophilia. This could take place without the observation of tissue eosinophil recruitment if there were a lack of eosinophil chemokine expression and the absence of eosinophil chemoattractant gradient. However during an exacerbation of CB, perhaps due to viral infection of the epithelium, there could be up-regulation of epithelial-derived eotaxin (34) or RANTES with the consequent rapid recruitment of tissue eosinophils from the blood. Such an exacerbation-associated tissue eosinophilia approaching the levels found in asthma has been reported in subjects with bronchitis (35) and we have recently reported RANTES gene up-regulation in these cases, during acute exacerbations of bronchitis (36).

Compared with CB, the significantly lower level of expression for IL-4 in patients with CB + COPD was unexpected. This is clearly associated with the relative reduction of the total numbers of inflammatory cells that we report in CB + COPD (r  0.6, p  0.02). Our counts of the total numbers of inflammatory cells in the inner wall showed there were significantly fewer inflammatory cells in the airways of the obstructed bronchitic group than the nonobstructed bronchitic group, both in the subepithelium and in the submucosal glands (p < 0.001). Although this may seem puzzling and contrary to the results of biopsy studies, there are similar and complementary previous and plausible explanations for the apparent disparity (37, 38). Linden and colleagues lavaged the lungs of smokers with chronic bronchitis (CB) and smokers with COPD and coexisting CB (i.e., CB + COPD): those in the obstructed group had significantly fewer inflammatory cells in both bronchial wash and bronchoalveolar lavage compared with the nonobstructed group (37). Linden and colleagues concluded that "airway obstruction may reflect a later stage of inflammation with connective tissue deposition, which may lead to decreased numbers of cells migrating into the airspace." In this conclusion, there are parallels to be drawn with interstitial lung disease. In our own previously published observations, we have found a relative reduction of IL-4 and IL-5 gene expression in the lung parenchyma of patients with interstitial pulmonary fibrosis (IPF) as compared with lung fibrosis associated with systemic sclerosis: the latter are generally seen and studied earlier in the course of their condition (20). There is also an immunohistological study of subjects who had severe irreversible airway obstruction and who had died (referred to as "fatal chronic bronchitis"). The group was compared with another in which there were patients with chronic cough and sputum but who had never complained of dyspnea and who had died of nonrespiratory cause ("incidental bronchitis") (38). The group with fatal obstructive bronchitis had significantly lower numbers of plasma cells and IgA positive cells associated with their airway submucosal glands and mucosa than the fatal group with "incidental bronchitis" at each of the three airway levels studied.

The contribution of plasma cells to the total numbers of inflammatory cells has, in our opinion, been overlooked by all the immunohistological studies of biopsies and surgically resected tissues. Thus although the numbers of immunolabeled cells (i.e., CD8⁺ cells) is reported to increase in COPD as compared with CB, the numbers of plasma cells may well decrease and the reduction would go undetected by immunohistological analyses incapable of demonstrating this cell type. The presence and predominance of plasma cells have been previously reported in CB on several occasions by use of routine H&E staining (26).

Our recent data on counts of plasma cells support this and are to be published in a companion paper.

In our present submission we have examined surgically resected tissue from three patient groups all of whom are smokers: asymptomatic, nonobstructed chronic bronchitic, and obstructed chronic bronchitic. On first reading it might appear that we have failed to confirm our own previous observations of a rise in the CD8⁺ T cell population. However, our original observations (9) and those of Lams and colleagues (10) were made using analyses of endobronchial biopsies and not tissues obtained at resection. Also our original biopsy report compared CB and CB + COPD against normal nonsmoker control subjects: we found a significant increase in CD8⁺ cells only in the CB + COPD group albeit there was also a trend in the CB (bronchitic) group. Of course, in the present study of resected material we could not include a normal healthy nonsmoker group for comparison. We acknowledge that the biopsy study by Lams and coworkers does report an increase in the CD8⁺ cell phenotype in a CB + COPD group by comparison with normal healthy smokers: however the p value is relatively weak (= 0.03) and a CB group was not included for comparison. We would suggest that the more appropriate literature comparison would be with another study of surgically resected tissue in carefully defined groups of patients. In a comparison of surgically resected airways from a nonobstructed asymptomatic group, Saetta and colleagues demonstrate, in both subepithelial and glandular tissue compartments, no significant increases in the CD8⁺ cell population in a CB + COPD group (39). Their patient groups are comparable to ours and in that were the same as in our study, and there were no significant differences between groups in respect to the CD8⁺ phenotype.

The phenotype of cells expressing IL-4 and IL-5 mRNA and protein production in asthmatic airways has been studied previously. In studies of bronchial biopsies Bradding and coworkers have previously reported that IL-4 and IL-5 protein product were predominantly colocalized to mast cells and eosinophils, but not T cells in biopsies of subjects with asthma (40). Ying and coworkers demonstrated that IL-4 and IL-5 mRNA were produced by both CD4⁺ and CD8⁺ T cells, eosinophils, and mast cells in bronchial biopsies obtained from atopic and nonatopic (intrinsic) subjects with asthma (41). These authors showed that >50% of IL-4 and IL-5 mRNA⁺ was localized to CD4⁺ T cells and about 10% of IL-4 and IL-5 mRNA⁺ was localized to CD8⁺ T cells (41). In the present study, we have observed the full depth of airway wall of subjects with CB and demonstrated that neither IL-4 nor IL-5 was colocalized with the CD8⁺ T cell phenotype.

As with CD4 cells, CD8 T cell clones, according to cytokine production, can be classified into three subsets: Tc0 clones producing IL-4 in addition to IFN-, Tc1 clones producing IFN- and a variety of other cytokines but not IL-4, and Tc2 clones producing IL-4 but not IFN- (42). Due to the predominance of the CD8 phenotype in COPD reported by us previously and confirmed independently by several groups (9, 43, 44), we hypothesized herein that it would likely be Tc2 CD8⁺ cells producing these cytokines. However, the double-labeling technique and counts applied herein demonstrate that although TNF- gene expression could be localized to CD8⁺ cells, IL-4 and IL-5 could not. Our data do not, therefore, support the hypothesis. The contention that IL-4 is associated with an inflammatory cell contributing to the infiltrate is supported by the results of a correlation of IL-4 and the counts of all inflammatory cells. We are currently investigating further the cellular origin of IL-4 in the mucus-secreting glands of patients with CB: these data will form the focus of a subsequent report. Other candidate cells include plasma cells, CD4⁺ cells, present in relatively lower numbers, and also mast cells.

In conclusion, we have shown that IL-4 and IL-5 gene expression is not restricted to asthma. Like the recent reports of their expression in the lung parenchyma of fibrosing lung disease (20), they, and especially IL-4, are expressed in the submucosal bronchial glands and subepithelium of smokers with CB. Thus, although there are clinical and tissue-based differences reported between nonsmoking subjects with asthma, with highly reversible disease, and smokers with CB and "fixed" airflow obstruction, there are similarities and overlap also (12). We suggest that IL-4 may play an important role in the development of the mucus hypersecretion of CB, whereas the CD8⁺ phenotype may be of greater relevance to the development of the destructive lesions associated with COPD.