Funded by King's College London School of Medicine.
C. Corrigan, King's College London, Department of Asthma, Allergy and Respiratory Science, 5th Floor Tower Wing, Guy's Hospital, Great Maze Pond, London SE1 9RT, UK. E-mail: email@example.com
Bronchial mucosal CD8+ cells are implicated in chronic obstructive pulmonary disease (COPD) pathogenesis, but there are few data on their functional properties. We have developed a novel technique to outgrow these cells from COPD patients in sufficient numbers to examine effector functions. Endobronchial biopsies from 15 COPD smokers and 12 ex-smokers, 11 control smokers and 10 non-smokers were cultured with anti-CD3/interleukin (IL)-2 ± IL-15. Outgrown CD3+ T cells were characterized in terms of phenotype (expression of CD4, 8, 25, 28, 69 and 56), cytotoxicity and expression of COPD-related cytokines. Compared with IL-2 alone, additional IL-15 increased the yield and viability of biopsy-derived CD3+ T cells (12–16-day culture without restimulation) without alteration of CD4+/CD8+ ratios or expression of accessory/activation molecules. Biopsy-derived T cells, principally CD8+/CD56+ cells, exhibited statistically significantly greater cytotoxic activity in current or ex-smokers with COPD compared with controls (P < 0·01). Elevated percentages of CD8+ T cells expressed interferon (IFN)-γ, tumour necrosis factor (TNF)-α and IL-13 (P < 0·01) in current COPD smokers compared with all comparison groups. It is possible to perform functional studies on bronchial mucosal T cells in COPD. We demonstrate increased CD8+CD56+ T cell cytotoxic activity and expression of remodelling cytokines in smokers who develop COPD.
Immunostaining of endobronchial biopsies and resected lung tissue [1–6] has implicated mucosal CD8+ T cells in the pathogenesis of chronic obstructive pulmonary disease (COPD), in the sense that their numbers correlate in some studies with the degree of emphysema  and airflow limitation [4–6]. Animal models of COPD also implicate cytokines, particularly interferon (IFN)-γ, interleukin (IL-13 and tumour necrosis factor (TNF)-α in tissue remodelling . Despite endless speculation about the properties of these CD8+ T cells [8,9], hard data are lacking because available technologies limit functional analysis while they remain embedded in the tissue.
One approach to this problem is to expand T cells from the target tissues. IL-2 has been employed to outgrow T cells from resected lung tissue [10,11] and endobronchial biopsies from human asthmatics , but these studies involved lengthy propagation and repeated stimulation of the T cells. It has since been discovered that IL-15, which shares receptor components with IL-2, additionally promotes growth and survival of memory T cells, as well as CD56+ natural killer (NK) and NK T cells [13,14].
Our aim in this study was to develop a technique for outgrowth of T cells from the bronchial mucosa of smokers with COPD and appropriate controls in sufficient numbers to perform functional studies using a single, pan-T cell physiological stimulus (anti-CD3 monoclonal antibody) while avoiding prolonged culture and repetitive stimulation. We hypothesized that outgrowth in the presence of IL-15, compared with IL-2 alone, improves yield, viability and expansion of T cells without changing their phenotypes.
As an illustration of the general applicability of this technique, we investigated functions of bronchial mucosal T cells speculated to be relevant to COPD pathogenesis. Specifically, we hypothesized first that bronchial mucosal T cells in smokers who develop COPD display elevated cytotoxic potential compared with those who do not, and that this potential is particularly enriched in cells of the CD8+CD56+ phenotype. Secondly, we hypothesized that bronchial mucosal T cells are a potential source of cytokines implicated in causing airways remodelling and destruction in COPD, and that their potential for production of these cytokines is elevated in smokers who develop COPD compared with those who do not.
We studied 15 patients with COPD who currently smoked (smokers with COPD) and 12 subjects with COPD who had not smoked for at least 6 months (ex-smokers with COPD). The definition of COPD followed the Global Initiative for Chronic Obstructive Lung Diseases (GOLD) guidelines , in the sense that all patients had post-bronchodilator forced expiratory volume in 1 s/forced vital capacity (FEV1/FVC) ratio ≤ 0·7 and FEV1 < 70% of the predicted value. All patients had a smoking history of at least 15 pack-years and had not used oral corticosteroids or suffered disease exacerbations in the 4 weeks prior to the study. The majority of the COPD subjects was treated with short or long-acting β2-agonist, while some were taking inhaled corticosteroids and two were taking a combined corticosteroid/long-acting β2-agonist preparation (Table 1). The control groups comprised 11 asymptomatic smokers (control smokers) and 10 non-smokers (control non-smokers). All the control subjects had FEV1/FVC ratio ≥ 0·7 and FEV1 > 90% of the predicted value, and were clinically free from respiratory infection in the 4 weeks prior to the study (Table 1). Subjects provided written informed consent to participate in the study, which was approved by the Local Research Ethics Committee of King's College Hospital, and all procedures were performed in agreement with the Helsinki Declaration of 1975 (1983 revision).
Table 1. Characteristics of study subjects.
Subjects with COPD
P ≤ 0·05,
P ≤ 0·001 compared with the control groups. P < 0·05 was regarded as significant. Data are expressed as the median and range. Smoking status: (+) current smoker, (EX) ex-smoker, (−) lifelong non-smoker. Differences between groups were analysed by one-way analysis of variance and the Newman–Keuls multiple comparison test. COPD: chronic obstructive pulmonary disease; M: male; F: female; FEV1: forced expiratory volume in 1 s; FVC: forced vital capacity (post-bronchodilator in COPD patients).
Three endobronchial biopsies from the second- and third-generation bronchi of the right lower lobe were obtained at fibreoptic bronchoscopy (Olympus BFT 20 bronchoscope; Olympus Corp., London, UK) using FB-35C-1 2·8-mm channel forceps (Olympus Corp). Subjects were sedated with intravenous alfentanyl (100–500 µg; Janssen-Cilag, High Wycombe, UK). Local anaesthesia was achieved with topical lignocaine (4%, 2%) (Astra Zeneca, Luton, UK). Capillary blood oxygen saturation was monitored with a pulse oximeter. The procedure was well tolerated by all subjects. Sixty ml peripheral venous blood was collected into 10% sodium citrate just prior to the bronchoscopy.
Response of lung-derived T cells to IL-15
In preliminary experiments to determine the optimal culture time and IL-15 concentration required to outgrow T cells from the airways, three bronchial biopsies from each of four healthy non-smokers were pooled, teased into small pieces in 1·2 ml of complete medium (see below) and the resulting suspension divided approximately equally between six wells of a 24-well plate using an Eppendorf pipettor with sterile, disposable tips. Complete medium (2 ml, see composition below) containing 1 × 106/ml autologous irradiated (4000 RAD) peripheral blood mononuclear cells (PBMCs) as feeder cells, 1 µg/ml soluble anti-CD3 monoclonal antibody (mAb) (Pharmingen, Oxford, UK), 2·5 µg/ml amphotericin (Sigma-Aldrich, Poole, UK), a previously optimized concentration of 50 U/ml IL-2 (Peprotech, London, UK)  and increasing concentrations of IL-15 (Peprotech) (0, 0·1, 1 and 10 ng/ml) were added to four of the culture wells. Medium containing IL-2 (50 U/ml) and IL-15 (10 ng/ml) but no feeders was added to the fifth well, and medium alone without IL-2 or IL-15 was added to the sixth.
Biopsy tissue was cultured for up to 20 days (37°C, 5% CO2) with periodic topping-up of medium containing the appropriate concentrations of IL-2 and IL-15. T cell numbers and viability were measured on days 8, 12 and 16 after resuspending the cells in each of the six wells and performing a total cell count on a 20 µl aliquot using a haemocytometer and analysing a further 100 µl aliquot by flow cytometry. In brief, cells were adjusted to a maximum of 106/ml in 0·5 ml phosphate-buffered saline (PBS)/1% fetal calf serum (FCS) and stained with peridinin–chlorophyll–protein complex (Per-CP)-conjugated anti-CD3 and allophycocyanin (APC)-conjugated anti-CD8 or anti-CD4 (Becton Dickinson, Oxford, UK), then resuspended in 0·5 ml calcium binding buffer (Becton Dickinson). Propidium iodide (10 µl of 50 µg/ml; Sigma-Aldrich, Poole, UK) containing 0·1% sodium citrate (Sigma Aldrich) and 5 µl annexin V-fluorescein isothiocyanate (FITC) (Becton Dickinson) was added and the cells incubated for 15 min at room temperature in the dark. A further 400 µl 1× calcium binding buffer were added and the cells analysed by flow cytometry within 1 h. Total CD3+ T cell counts were estimated by multiplying the percentages of viable CD3+ cells by the total cell counts obtained from the haemocytometer.
The effect of IL-2 and IL-15 on the proliferation of airways-derived T cells was assessed by tritiated thymidine incorporation at day 12. In brief, cells from further aliquots of the cell suspension from each well were adjusted to 1 × 106/ml viable CD3+ T cells using complete medium and 100 µl of suspension transferred in triplicate to wells of 96-well flat-bottomed microtitre plates, then pulsed with tritiated thymidine (0·5 µCi/well) and cultured for 24 h. The T cells were harvested onto glass fibre self-aligning filters. Cell division was assessed by tritium incorporation, measured using a Canberra Packard matrix 96-beta counter and expressed as counts per minute (cpm).
Outgrowth of T cells from bronchial biopsies
T cells were outgrown from bronchial biopsies using a novel technique as follows. Following retrieval, three endobronchial biopsies were immersed immediately in 5 ml complete medium [RPMI-1640 with 2 mM glutamine, 1% sodium pyruvate, 1% non-essential amino acids, 1:200 10 µg/ml gentamicin (Sigma-Aldrich) and 10% human serum (PPA Laboratories, Austria)], then transferred in the laboratory to 1 ml complete medium in a small, sterile Petri dish, under a sterile hood, teased into small pieces using sharp sterile forceps and scissors and the resulting suspension divided approximately equally between five wells of a 24-well plate using an Eppendorf pipettor with sterile, disposable tips. To each of four of these wells was added complete medium (2 ml) containing 1 µg/ml anti-CD3 mAb, 1 × 106/ml irradiated (4000 RAD) autologous PBMCs, 2·5 µg/ml amphotericin and 50 U/ml IL-2 with or without 10 ng/ml IL-15 (duplicate wells for each condition). A separate fifth well containing biopsy fragments without irradiated PBMCs or supplementary cytokines and two separate wells containing irradiated feeders with the anti-CD3 stimulus and IL-2 50 U/ml with or without IL-15 10 ng/ml were set up as controls.
Biopsy fragments were cultured for 12 days (37°C, 5% CO2) and fresh medium with cytokines (but no further anti-CD3) added on days 4, 7 and 10. On day 12 cells were harvested, the cells from duplicate wells were pooled and biopsy debris removed by filtering through a 70-µm sieve. Biopsy material was divided approximately between wells, and so it was not possible to determine the initial numbers of T cells in the starting cultures. Final CD3+ T cell numbers for Table 2 were computed by multiplying the percentages of viable CD3+ cells in the cultures obtained by flow cytometry by the total cell counts in the wells determined using a haemocytometer. In Table 3 the results are expressed as percentages of the total expanded CD3+ T cell population.
Table 2. Numbers of lung-derived CD3+ T cells outgrown from bronchial biopsies.
Anti-CD3 stimulation with:
Numbers of outgrown cells
P ≤ 0·01,
P ≤ 0·001 compared with lung-derived T cells cultured with interleukin (IL)-2 alone. P < 0·05 was regarded as significant. Teased bronchial biopsies from smokers with chronic obstructive pulmonary disease (COPD) (COPD SM), ex-smokers with COPD (COPD-EX), control smokers (Con SM) and non-smokers (Con NS) were cultured with anti-CD3 monoclonal antibody (mAb) supplemented with 50 U/ml IL-2 either alone or with 10 ng/ml IL-15 for 12 days with irradiated feeder cells. Pooled total cell numbers were determined with a haemocytometer and the total numbers of CD3+ T cells computed by multiplying this total by the percentages of viable CD3+ cells as measured by flow cytometry. Differences between groups were analysed by one-way analysis of variance and Newman–Keuls multiple comparison test. s.e.m.: standard error of the mean; no.: numbers of samples in which detectable T cell outgrowth was achieved.
Table 3. The effect of interleukin (IL)-2 and IL-15 on the percentages of anti-CD3 stimulated lung-derived T cells expressing phenotypic markers.
P ≤ 0·001 compared to all other groups,
P ≤ 0·001 compared with control non-smokers. Bronchial biopsies from smokers with chronic obstructive pulmonary disease (COPD) (COPD SM), ex-smokers with COPD (COPD EX), control smokers (Con SM) and non-smokers (Con NS) were stimulated with anti-CD3 supplemented with 50 U/ml IL-2 either alone or with 10 ng/ml IL-15 for 12 days with irradiated feeder cells. The data show the percentages [mean, standard error of the mean (s.e.m.)] of outgrown CD3+ T cells expressing CD4 and CD8 and the percentages of CD4+ and CD8+ T cells co-expressing phenotypic markers. Differences were compared with one-way analysis of variance and Newman–Keuls multiple comparison test. Allowing for multiple comparisons, P < 0·002 was regarded as significant.
CD3+ T cells were isolated using microbeads in accordance with the manufacturer's instructions. In brief, biopsy-derived cells were washed and resuspended in 80 µl of buffer (PBS/2% FCS) per 107 total cells. Twenty µl of anti-CD3 microbeads (Miltenyi Biotech, Surrey, UK) were added per 107 total cells and the mixture incubated for 15 min at 4°C. The cells/beads were washed, resuspended in 500 µl buffer and trapped on an MS magnetic column (Invitrogen, Paisley, UK). Purity was consistently >90%, as defined by flow cytometric analysis of CD3 expression.
CD4+ and CD8+ T cell isolation
CD4+ and CD8+ T cells were isolated by positive and negative selection from biopsy-derived expanded cells using Dynabeads® (Invitrogen), in accordance with the manufacturer's instructions. Briefly, T cells were resuspended at 0·5–1·0 × 107 total cells/ml in RPMI-1640 with 2% FCS. Anti-CD4 Dynabeads® were added at an approximately 5:1 bead : cell ratio and the mixture incubated for 30 min at 4°C on a rolling rocker. Captured CD4+ T cells were isolated and washed using a magnetic particle concentrator (MPC-1, Dynal; Invitrogen). Captured CD4+ T cells were released from the Dynabeads® by bead detachment with Detachabeads® (Invitrogen). CD8+ T cells were collected in the effluent. Positively selected CD4+ and negatively selected CD8+ T cells were consistently >98% pure, as assessed by flow cytometric analysis.
Isolation of CD56+ and CD56- T cells
CD56+ and CD56- subsets of CD4+ and CD8+ T cells were isolated using microbeads (Miltenyi Biotech) as above. In brief, purified CD4+ and CD8+ T cells were washed and resuspended in 80 µl of buffer (PBS/2% FCS) per 107 total cells. Twenty µl of anti-CD56 microbeads were added per 107 total cells and the mixture incubated for 15 min at 4°C. The cells/beads were washed, resuspended in 500 µl buffer and passed through an MS magnetic column where the CD56+ T cells were trapped and the CD56- T cells collected in the effluent. Purity was consistently >95%, as defined by flow cytometric analysis of CD3 expression.
Cytotoxic assay (time-resolved fluorometry)
T cell cytotoxic function was assessed using a cytotoxic assay kit (Perkin Elmer, Bucks, UK) and time-resolved fluorometry, as described previously . In brief, K562 target cells (European Collection of Cell Cultures, Salisbury, UK) (1 × 106/ml) were labelled with 10 µl of a fluorescent enhancing ligand [2,2′:6′,2″-terpyridine-6,6″-dicarboxylate (BATDA), bis(acetoxymethyl) 2,2′:6′2″-terpyridine-6,6″-dicarboxylate] for 25 min at 37°C in 2–4 ml culture medium. Labelled target cells (5 × 103) were incubated in a total volume of 200 µl with effector cells (CD4+ or CD8+ T cells and their CD56+/− subsets) in a humidified 5% CO2 atmosphere at 37°C (E/T ratio = 4:1, 8:1, 16:1, 32:1) in complete medium in 96-well V-bottomed microtitre plates. After incubation (4 h) the cells were pelleted by centrifugation (5 min at 500 g) and 20 µl supernatant from each well transferred into the wells of flat-bottomed 96-well plates; 180 µl of Europium solution (Eu) were added to each well. Following 15 min incubation with shaking at room temperature, the fluorescence of the Europium tartryl diazide (EuTDA) chelates formed was measured in a time-resolved fluorometer. The percentage-specific release of the label was calculated as: (experimental release − spontaneous release/maximum release − spontaneous release) × 100%. Maximum and spontaneous release was determined as the TDA release from target cells, respectively, lysed with detergent and cultured with medium.
Surface and intracellular staining for flow cytometry
Lung-derived T cells were analysed by flow cytometry. Fluorochrome-conjugated monoclonal antibodies were used to identify T cell phenotypes: fluorescein isothiocyanate (FITC)-conjugated anti-CD4, anti-αβ T cell receptor (TCR), phycoerythrin (PE)-conjugated anti-CD28, anti-CD56, anti-CD25 anti-CD69, anti-γδ TCR, PerCP-conjugated anti-CD3 and APC-conjugated anti-CD8 (Becton Dickinson). Staining solutions routinely contained 1% FCS. Data were acquired on a fluorescence activated cell sorter (FACSCalibur) cytometer and analysed using CellQuest software (Becton Dickinson). Propidium iodide staining dead cells were routinely gated out, and indeed dead cells were clearly distinguishable from live cells on forward-scatter criteria.
For intracellular staining, the cells were first stained with cell surface Per-CP-conjugated anti-CD3 and APC-conjugated anti-CD8 (Becton Dickinson), washed and fixed in 4% formaldehyde/PBS and permeabilized in 0·5% saponin/1% BSA/PBS for 30 min at room temperature in the dark. The cells were then incubated with anti-cytokine monoclonal antibodies FITC-conjugated anti-IFN-γ and PE-conjugated anti-TNF-α and anti-IL-13 (Becton Dickinson)] or isotype-matched controls for 30 min at 4°C, fixed in 1% paraformaldehyde (PFA) in PBS and acquired immediately by flow cytometry. CD3+ T cells were gated, with CD8+ T cells stained positively and the remainder of CD3+ T cells defined as CD4+ T cells.
Group data were expressed as means ± standard error of the mean (s.e.m.) when distributed parametrically and as individual values and median when distributed non-parametrically. Parametric distribution of data was assessed using the Kolmogorov–Smirnov (KS) test. The Bartlett's test was used to test for equal variance. The data in Tables 1–3 were analysed by one-way analysis of variance and the Newman–Keuls multiple comparison test. For Tables 1 and 2P < 0·05 was taken as significant; in Table 3, because of the multiple comparisons, P < 0·002 was taken as significant. The analysis in Fig. 1 was by Student's t-test. Differences in cytotoxic potential of CD3+ T cells between the four groups (Fig. 2) were analysed using one-way analysis of variance and the Newman–Keuls multiple comparison test. Group variance of intracellular cytokine immunoreactivity between the four groups was analysed by the non-parametric Kruskal–Wallis test, with differences between the four groups compared further using Dunn's multiple comparison test. For the data in all three figures a P < 0·05 was accepted as significant.
Effect of IL-15 on T cell outgrowth from bronchial biopsies
In preliminary experiments to determine the optimal culture time and IL-15 concentrations required to expand mucosal T cells in culture, biopsy tissue from four healthy non-smoking subjects was cultured with irradiated autologous peripheral blood feeder cells, 1 µg/ml soluble anti-CD3 mAb and a previously optimized concentration of 50 U/ml IL-2 in the absence or presence of increasing concentrations of IL-15 (Fig. 1a). In the absence of IL-15, there was no increase in CD3+ T cell numbers (Fig. 1a) or thymidine incorporation (Fig. 1b) compared to cells cultured in medium alone. With IL-2 and IL-15 CD3+ T cell numbers increased with time, most markedly in the presence of 10 ng/ml IL-15, where statistically significant expansion (Fig. 1a) and proliferation (Fig. 1b) of CD3+ T cells was observed by day 12, the former plateauing between days 16 and 20 (not shown). IL-15 also inhibited apoptosis (typically >70% of T cells did not stain with annexin V or propidium iodide after 12 days of culture, compared with <20% of cells cultured with IL-2 alone). No growth was observed in the absence of feeders even with IL-2 and IL-15. Expanded CD3+ T cells comprised of both CD4+ and CD8+ populations, although a small percentage expressed both markers (Fig. 1c). In view of these data, a culture period of 12 days and a concentration of IL-15 of 10 ng/ml were employed in all subsequent experiments.
Numbers and phenotype of T cells outgrown from bronchial biopsies
Table 1 shows the characteristics of individuals within the four study groups. Table 2 shows the numbers of CD3+ T cells grown out from teased bronchial biopsies in the presence of feeders, anti-CD3 and IL-2 with or without IL-15. In accordance with the preliminary data above, CD3+ T cell outgrowth was statistically enhanced significantly only in the presence of IL-15. In a minority of subjects (eight of 48, all current smokers, and six with COPD), no outgrowth was detected even in the presence of IL-15 (in which case all data pertaining to these samples were omitted from analysis).
Table 3 shows the percentages of CD3+ T cells outgrown from the biopsies expressing CD4 and CD8 and the percentages of these cells co-expressing a variety of phenotypic and activation markers, as measured by flow cytometry. CD8+ T cells predominated in the cells grown out of biopsies from COPD smokers, in stark contrast to those grown from all other groups examined where CD4+ T cells predominated. The percentages of CD4+ and CD8+ T cells expressing the accessory molecule CD28 and the activation markers CD25 and CD69 did not differ statistically significantly between the four study groups. Elevated percentages of CD8+, but not CD4+ T cells, expressed CD56 in smokers and ex-smokers with COPD and control smokers compared to control non-smokers. In all four groups, the vast majority (97% ± 0·8%) of the T cells expressed the αβ form of the T cell antigen receptor, the remainder expressing the γδ receptor (data not shown). As shown in Table 3, although additional IL-15 promoted outgrowth of T cells it did not alter statistically significantly the proportion of CD4+ to CD8+ T cells or the percentages of these cells expressing any of the activation markers studied.
Cytotoxic potential of T cells outgrown from bronchial biopsies
Expanded total CD3+ T cells as well as CD4+ CD56+, CD4+ CD56-, CD8+ CD56+ and CD8+ CD56- subsets were isolated from the outgrown cells from a proportion of the subjects in all four study groups as described in the Methods, then cultured at various effector/target (E/T) ratios with fluorescently labelled K562 target cells.
Cytotoxic activity of the total CD3+ T cell population and the subsets was measured as percentage lysis of K562 target cells, measured using time-resolved fluorometry (Fig. 2). CD3+ T cells from patients with COPD, smokers and ex-smokers, showed marked cytotoxic activity statistically significantly greater than that observed in the non-COPD groups at E/T ratios of 16 or above (Fig. 2a). CD8+ CD56+ T cells were largely responsible for this increased cytotoxic activity, which clearly delineated the four study groups (Fig. 2b). In contrast, cytotoxic activity of CD8+ CD56- cells was statistically significantly lower and similar in the four study groups. Cytotoxic activity of CD4+ T cells was detectable, but approximately three- to fourfold lower than that observed with CD8+ T cells. Again, cytolytic activity tended to be higher in the CD4+ CD56+ subset, although not statistically significantly so (Fig. 2c).
Potential for cytokine production by CD3+ T cells outgrown from bronchial biopsies
The capacity of the mucosal T cells to produce cytokines thought to be relevant to COPD pathogenesis was assessed using intracellular cytokine staining at day 12 (Fig. 3). Statistically significantly greater percentages of CD8+ T cells from smokers with COPD expressed IFN-γ, TNF-α and IL-13 immunoreactivity compared to non-COPD controls (P < 0·01 in each case). Percentages of CD8+ cells showing immunoreactivity for IFN-γ, but not TNF-α or IL-13, were also increased statistically significantly in COPD smokers compared to ex-smokers with COPD. By comparison, the median percentages of CD8+ T cells expressing these cytokines were lower and statistically equivalent in COPD ex-smokers and both non-COPD control groups. Only the median percentage of IL-13 immunoreactive CD8+ T cells was elevated statistically significantly in the COPD ex-smokers compared with the control groups (Fig. 3a, c and e). In all four subject groups 93 ± 4·9% (mean ± s.e.m.) of the CD8+ T cells expressing IFN-γ also expressed TNF-α immunoreactivity, while 88 ± 3·8% of the IL-13 immunoreactive cells also expressed IFN-γ (data not shown).
In direct contrast to the situation with CD8+ T cells, a statistically significantly lower median percentage of CD4+ T cells from smokers with COPD expressed IFN-γ immunoreactivity compared to the COPD ex-smokers (P < 0·05) and control smokers (P < 0·01) (Fig. 3b). A similar but statistically non-significant trend was observed with CD4+ T cells expressing TNF-α (Fig. 3d). The median percentages of CD4+ T cells expressing IL-13 immunoreactivity were very low but were nevertheless higher in the COPD groups, especially COPD ex-smokers, compared to the control smokers (P < 0·01) and non-smokers (P < 0·001) (Fig. 3f).
We describe a novel and probably widely applicable technique for outgrowing sufficient numbers of CD3+ T cells from bronchial biopsies to perform functional studies while preserving their phenotypes. The fact that the phenotypic profiles of T cells outgrown in IL-2 alone (where significant proliferation did not occur) and IL-2/IL-15 (where it did) were comparable provides evidence that T cell outgrowth was not selective (Table 3), although the study was insufficiently powered to detect subtle differences in expanded cell numbers, which might be important in some situations. In particular, depending upon the functional readout under study, it clearly might be important to evaluate possible changes in expression of T cell and co-stimulatory molecules such as CD27 and receptors for cytokines and chemokines which may influence T cell function, such as CCR7. The technique also avoids repeated passage and restimulation of the cells. Few γδ T cells were outgrown with this technique, which again we speculate reflects their low numbers in vivo, or alternatively they may be particularly susceptible to IL-15-induced apoptosis as has been reported with IL-2 .
There was a striking predominance of CD8+ over CD4+ T cells in the populations of cells outgrown from the biopsies of current smokers with COPD compared with the other comparison groups, in line with the widely reported increased numbers of CD8+ T cells in the bronchial mucosa of COPD patients who smoke [1,4,5]. We did not observe increased percentages of CD8+ T cells in COPD ex-smokers, which contrasts with a previous study  based on counting of cells in bronchial biopsies, suggesting that T cell numbers are indistinguishable in COPD smokers and ex-smokers, although powering issues and duration of smoking cessation  may also have influenced these conclusions. Our data would suggest that elevated CD8+ T cell infiltration is not a consequence of smoking alone. One possibility is that environmental conditions within the bronchial mucosa of smokers who are predisposed to develop COPD render CD8+ T cells hyperresponsive to IL-15, which is known to be over-expressed in the bronchial mucosa of smokers . This could be explored in further studies.
Although cytotoxicity is a well-recognized property of T cells, it was fascinating to observe a striking enhancement of this activity in cells from patients who smoked and a further enhancement in those smokers who developed COPD, even after stopping smoking. Although these associations do not necessarily imply causation, it is tempting to speculate that cigarette smoking enhances T cell cytotoxicity in the bronchial mucosa and that patients who develop COPD are somehow particularly susceptible to this effect. In the present study the majority of the cytotoxic activity appeared to be associated with a CD8+CD56+ subset of T cells. This is consistent with the hypothesis that the expanded cells comprise at least a subset of NK T cells, a hypothesis supported further by the lack of dependence of the observed cytotoxic effect on MHC antigens (the K562 target cells do not express MHC antigens) and antigen priming, both of which detract from a mechanism involving classical MHC class I restricted CD8+ cytotoxic T cells. This observation is intriguing in view of the current interest in a possible role for NK T cells in COPD, although the evidence to support such a role is currently controversial. In further experiments it would be fascinating to characterize these cells more clearly and also to investigate possible cytotoxic activity of T cells expanded from biopsies against autologous structural cells which do express class I MHC antigens, such as epithelial cells or fibroblasts, expanded from the airways of the same patients. In animal models, cytotoxic CD8+ T cells clearly cause lung tissue destruction and remodelling , and furthermore are required for cigarette smoke-induced lung inflammation , making a very strong case for the involvement of CD8+ T cell cytotoxicity in COPD pathogenesis even if the relevant target cells cannot yet be defined precisely. Some of the cytolytic activity detected in the CD4+ T cell subset in the present study may reflect the fact that some of these cells also expressed CD8 (Fig. 1c) and may have subsumed some of the functionality of the CD8+ T cells. Our data extend a previous study  demonstrating elevated cytolytic activity of induced sputum CD8+ T cells in patients with COPD. As discussed, however, it is not clear from our data whether the expanded CD8+ T cells are partly conventional cytotoxic CD8+ T cells induced to express CD56 or partly a subset of NK T cells, which also respond to IL-15 , or a mixture of both.
Finally, our data clearly reveal enhanced potential for bronchial mucosal CD8+ T cells to produce cytokines (IFN-γ, TNF-α, IL-13) implicated in the pathogenesis of COPD, again most clearly in those patients who continue to smoke. Interestingly, the percentages of CD4+ T cells expressing these cytokines tended to be conversely lower in the smokers with COPD, suggesting that there is no overall bias to T helper type 1 (Th1) or Th2 T cell differentiation in this particular environment. All three of these cytokines have been shown to produce emphysema-like changes when expressed selectively in the lungs of transgenic animals , probably through distinct mechanisms. The likelihood that these changes are largely irreversible may explain the disappointing therapeutic efficacy of some anti-cytokine strategies, such as TNF-α blockade, in established COPD , and may indicate a need to apply such interventions earlier. Although the association of elevated IFN-γ, TNF-α and IL-13 immunoreactivity in the outgrown CD8+ T cells with COPD and smoking does not necessarily imply cause and effect, it is interesting to note that this effect disappeared in ex-smokers. Again, it is possible to speculate that reversible effects of smoking on T cell cytokine production may explain in part why smoking cessation slows the deterioration of lung function in COPD . There are few studies of the expression of these cytokines in situ within the human bronchial mucosa in COPD. Elevated expression of IL-13 was reported in the central airways of surgically resected specimens of smokers with chronic bronchitis  although, unsurprisingly, T cells were not the only source of these mediators. In a study in which bronchial epithelial T cells brushed from the lining of the large airways were stained ex vivo for cytokine immunoreactivity (following suitable stimulation) in groups of subjects similar to those in the present study , the median percentage of CD8+ T cells expressing TNF-α was elevated in COPD smokers compared with normal controls. No statistically significant reduction was seen in ex-smokers, although the groups were small. There was a similar trend in the case of IFN-γ; IL-13 was not measured.
Our data raise some interpretational questions. Although we have provided good evidence that outgrowth of T cells did not alter their phenotypes, a more systematic approach to this would have been to assess phenotypes before and after expansion. We experimented with this using collagenase treatment of the biopsies, but found that this caused some T cell death and loss of some surface markers, especially CD4. Nevertheless, this could be pursued further in subsequent studies. It is impossible to gauge how far outgrowth, albeit relatively short-term, of the T cells from the bronchial biopsies might have altered their functional properties, although the impressive differences in T cell cytotoxicity and cytokine expression observed between the clinical study groups attest clearly to the fact that the function of these cells was not altered at random and argue strongly that functional differences are indeed preserved. A more systematic comparison of cytokine expression by T cells in situ within the bronchial mucosa, for example by immunohistochemical staining or in situ hybridization of cytokine-specific mRNA and cytokine expression in T cells outgrown from biopsies from the same patients might clarify this in future studies, although these techniques all present their own technical challenges and might not always be expected to correspond. As shown in Table 2, we failed to outgrow significant numbers of T cells from 13 of 48 of the biopsies cultured with IL-2 alone probably because, as shown in Fig. 1, this stimulus was insufficient to cause T cell proliferation under the culture conditions employed. Even when cultured in IL-2 and IL-15, however, significant T cell outgrowth was not achieved from the biopsies of eight subjects. A few of these biopsies could have been of poor quality but not, we suspect, as many as eight sets. It is perhaps relevant that all the subjects in whom T cell expansion failed were current smokers, suggesting the possibility of an inhibitory effect of smoking on T cell expansion under these conditions (for example, a failure of the T cells in the environment of the bronchial mucosa of some smokers to exhibit a critical threshold for responsiveness to IL-2/IL-15). This could be explored in further studies. It is conceivable that these eight subjects had no T cells in their bronchial mucosa, but we consider this extremely unlikely: there has been one report  that T cells are relatively depleted in patients with severe COPD, but never absent. Inhaled corticosteroid therapy does not alter statistically significantly the numbers of mucosal CD4+ and CD8+ T cells in patients with COPD , except apparently with concomitant long-acting β2-agonist therapy . Failure to outgrow T cells did not correlate with taking inhaled corticosteroid therapy in the COPD patients, and outgrowth was achieved successfully in those two patients taking combined corticosteroid/long-acting β2-agonist. Consequently, we do not believe that inhaled therapy precludes T cell outgrowth, although we cannot be certain that it does not influence the functional properties of the outgrown cells. There were no statistically significant differences in any of the functional readouts shown in Figs 2 and 3 in the COPD patients taking and not taking inhaled corticosteroid, although the study was not designed to be powered to detect such differences. Whatever the cause of outgrowth failure, this may have biased the functional data, blunting or alternatively enhancing the degree of difference between the smoking and non-smoking subjects, because all data pertaining to biopsies from which T cells failed to outgrow were omitted from analysis.
Notwithstanding these reservations, we believe that this novel technique will allow functional analysis of T cells from bronchial biopsies from patients with COPD and perhaps other airways inflammatory diseases. We believe that it will be useful for elucidating further the functions of CD8+ T cells in COPD; for example, mechanisms and targets of cytotoxicity, molecular effects of cigarette smoke and the effects of potential new therapeutic agents.
BJO'C has received honoraria from GlaxoSmithKline for speaking engagements in the past year and is a consultant to Chiesi Pharmaceuticals and Cipla Pharmaceuticals. CJC has received honoraria for speaking engagements from GlaxoSmithKline, Novartis Pharmaceuticals, Teva Pharmaceuticals and Allergy Therapeutics in the past year and is a consultant for Meda Pharmaceuticals. He also has research contracts with Novartis, Artu Biologicals, ALK Abello and Amgen. The remaining authors have no disclosures to make.