Increased Bronchial Density of CD25+Foxp3+ Regulatory T Cells in Occupational Asthma: Relationship to Current Smoking


Correspondence to: T. S. Halstensen MD, PhD, Laboratory for mucosal immunology (LMI), IOB, University of Oslo, Pb. 1052 Blindern, 0316 Oslo, Norway. E-mail:


To identify activated T cell subset in the asthmatic bronchia, we developed a triple-colour immunohistofluorescence labelling technique on cryo-section to discriminate activated CD4+CD25+ T cells, (effector T cells) from Foxp3+ regulatory T cells (Treg). Additional coexpression of activation and proliferation markers was also examined in situ. Bronchial biopsies were taken from 20 aluminium potroom workers (12 smokers) with asthma (>12% reversibility), 15 non-asthmatic potroom workers (7 smokers) and 10 non-smoking, non-exposed controls. Non-smoking asthmatics had significantly higher subepithelial density of both Tregs, effector T cells, activated (HLA-DR+) CD8+ and activated CD4+ T cells. Moreover, both Tregs, effector T cells and CD8+ T cells proliferated in the non-smoking asthmatics, only. Although smoking asthmatics had no asthma-associated increase in bronchial T cell, both had a significantly increase in effector T cell to Treg ratios. The significantly increased bronchial density of Tregs, effector T cells, proliferative T cells and activated CD8+ T cells in non-smoking asthmatics clearly showed that both the effector T cells and the inhibitory Treg system were activated in asthma.


The bronchial inflammation in various asthma subgroups may share features responsible for the reversible bronchial obstruction. Thus, examination in occupational asthma may reveal immunological phenomenon applicable to other asthma subtypes. Potroom workers in the aluminium industry develop an asthma [1], with similar bronchial pathophysiology as non-atopic asthma [2, 3], including bronchial basement membrane thickening and increased subepithelial CD4+ T cells density [4]. These asthma-associated CD4+ T cells often co-express CD25+ in atopic [5], non-atopic [6] and occupational asthma [7] and therefore assumed to represented asthma-promoting Th2 cells. However, the CD4+ CD25+ T cell subset is known to include regulatory T cells (Tregs), identified by nuclear co-expression of the forkhead/winged-helix transcription factor Foxp3 [8]. These Tregs are important in maintaining immunological homoeostasis [9, 10] and may inhibit both autoimmune and allergic diseases through suppression of Th1 as well as Th2 type of immune responses [11]. Some reports have suggested that asthma is associated with decreased number or function of Tregs [11, 12], based on findings in peripheral blood [11, 13-15] and bronchoalveolar lavage (BAL) [13]. However, there are no reports on bronchial Tregs in situ in asthma.

The aim of this study was to identify which T cell subtypes that in particular were activated in both non-smoking and smoking subjects with asthma, with particular emphasis on the relatively increased bronchial density of CD8/CD4 T cell ratio previously observed in smoking asthmatics [4]. Because activation markers are expressed on several different cell types, we used a novel triple-colour immunohistofluorescent labelling technique to phenotypically discriminate the regulatory T cell subsets (CD4+CD25+Foxp3+) from the Foxp3negative CD25+ T cell subset (assumed to be activated effector T cells). We also examined whether any of the bronchial T cell subsets expressed the activation marker human leucocyte antigen (HLA)-DR or if any T cell subset proliferated in situ.



Twenty potroom workers with asthma (eight non-smokers and 12 smokers (median 13 pack-years)), 15 healthy potroom workers (eight non-smokers and seven smokers (median 17 pack-years)) and 10 healthy non-exposed non-smoking controls were included (Table 1) as detailed elsewhere [3]. Briefly, symptoms of airway obstruction (dyspnoea, wheezing and cough) and reversible airway obstruction defined as >12% increase in forced expiratory volume in onesecond (FEV1) after inhalation of a β2-agonist were documented in all asthmatics. Ex-smokers (who had ceased smoking >1 year previously) and never-smokers were classified as non-smokers. Two of the asthmatics and three of the healthy workers were ex-smokers. Median smoking load was 13 pack-years for the asthmatics and 17 pack-years for the healthy workers. Half of the asthmatics had been relocated to non-polluted working environments. None of the participants had a history of allergy, familiar asthma or childhood asthma, and their total IgE-levels were within the normal range. The controls had no symptoms from upper or lower airways, except for one smoking healthy potroom worker. By inclusion, he had no respiratory symptoms, normal blood samples and denied any respiratory infection the last weeks. However, as he had a particular high proliferative T cell index, he admitted having some cough and sore throat several days prior to inclusion. Thus, he was probably recovering from a respiratory tract infection which could explain the particular strong proliferative activity.

Table 1. Characteristics of the study population
 Asthmatic workersHealthy workersNonexposed controls
Non-smokers (= 8)Smokers (= 12)Nonsmokers (= 8)Smokers (= 7)Non-smokers (= 10)
  1. Data are presented as median (range). FEV1: forced expiratory volume in one-second;% pred:% of predicted value; FVC: forced vital capacity.

Age (years)











FEV1% pred






















Inhaled corticosteroids13000
Inhaled β2-agonist46000

Spirometry and flexible bronchoscopy were performed as previously described [3], and bronchial biopsies were snap frozen. The study was approved by the Regional Ethics Committee, and informed written consent was obtained from all subjects.

Multicolour immunohistofluorescent labelling

Mucosal T cell subsets (CD4 and CD8) were examined by double- or triple-colour immunohistofluorescent labelling on bronchial cryo-sections, using antibodies to CD3 (Diatec AS, Oslo, Norway, clone RIV9; Monosan, Am Uden, the Netherlands; rabbit antiserum to a CD3ε-peptide; DAKO A/S, Glostrup, Denmark) in various combinations with mAb against HLA-DR (clone L123, IgG2a; Becton Dickinson, San Jose, CA, USA), Ki-67 (Zymed Invitrogen, Carlsbad, CA, USA); CD4 (clone MT310, DAKO); CD8 (clone DK25, DAKO or clone 4B11, SeroTech Ltd. Oxford UK); CD25 (clone ACT-1, DAKO) or Foxp3 (ab450, Abcam plc, Cambridge, UK), followed by appropriate combinations of peroxidate-, biotin-, (Southern Biotechnology, Birgmingham, AL, USA), Alexa-488 or -594 conjugated mouse IgG-subclass-specific goat antisera, (Molecular Probes, Eugene, OR, USA) and either thyramide signal amplification (TSA)-coumarin (PerkinElmer Life Science, Boston, MA, USA) or 4',6-diamino-2-phenylindole (DAPI, Molecular Probes). Methodological negative controls included sections incubated with non-immune mouse Ig and rabbit serum in similar concentrations.

Quantification of leucocytes

All slides were coded and analysed using a Zeiss Axioplane2 microscope (Carl Zeiss, Oberkochen, Germany) at 630× magnification. Positively stained cells were counted in a 114 μm subepithelial zone beneath the reticular basement membrane, excluding cells in submucosal glands and vessels. A total of 88.430 CD3+ T cells (median 1.423 cells per subject, range 690–5.461) were individually examined for Foxp3 and CD25 expression in triple-labelling, covering an area of median 1.8 mm2 per subject (range 1–4 mm2), corresponding to median 13 mm basement membrane length (range 6–28 mm). Similarly, a total of 16.718 CD3+ T cells (median 282 cells per subject, range 103–1.664) were individually examined for Ki-67 expression, covering an area of median 0.9 mm2 per subject (range 0.3–2.4 mm2), corresponding to median 6.6 mm basement membrane length (range 2.4–17.0 mm) as recommended [16].


Mann–Whitney U-test was used to compare groups. Subtyping the CD25+ T cells as being Foxp3+ Tregs or Foxp3-negative, putative effector T cells revealed often <10 positive cells per smoking subject, making individual statistics difficult to justify. Group-based statistics were therefore applied by adding all counted CD25+Foxp3+ CD3+ T cells; all CD25+ Foxp3negative CD3+ T cells and the total number of CD25+ CD3+ T cells within each group in a Chi-Square test to examine any group-based differences. All statistics were performed with and without the convalescent non-asthmatic worker, but his participation did not change the significance of the results. His data are, nevertheless, identified in all scatter diagrams.


Increased bronchial density of CD25+ T cells in non-smoking asthmatics

Non-smoking asthmatics had significantly higher subepithelial CD25+CD3+ T cell density (median 31, range 13–99 cells/mm2) compared with smoking asthmatics (median 9, range 3–53 cells/mm2), non-smoking potroom workers (median 7, range 4–59 cells/mm2) or non-exposed non-smoking controls (median 8, range 2–16 cells/mm2) (Figs 1, 2). Almost all were CD8 negative and thus CD4+ (medians 90–100%; data not shown) as we previously showed that <2% of the T cells were CD4 and CD8 double negative (TCR γ/δ+ T cells [4]).

Figure 1.

(A–D): Triple-colour immunohistofluorescent staining for CD3 (a, green), Foxp3 (b, blue nuclear staining) and CD25 (c, red). The triple, CD25+Foxp3+CD3+ labelled cells (D) represent putative regulatory T cells and appear yellow (red and green fluorescence) with blue nuclear Foxp3-positivity. Blue nuclear Foxp3-positivity was also noted in several CD25 negative green CD3+ T cells (D). CD25-positivity on CD3+ T cells which did not express Foxp3 (CD25+Foxp3negCD3+) is yellow without blue nuclear positivity and represents putative effector T cells (d). Figures e-h is triple-colour immunohistofluorescent staining for CD3 (E, blue), Foxp3 (F, green) and Ki-67 (G, red nuclear staining) visualized. A proliferating (Ki-67+) Foxp3+CD3+ T cell (arrow) is indicated in H. Note several proliferating (Ki-67+) epithelial cells. Multicolour immunohistofluorescent staining of bronchial cryo-sections from a non-smoking asthmatic patient. Single colour images were captured with a MicroMax CCD digital camera system and the AnalySIS Soft Imaging system.

Figure 2.

Individual subepithelial cell density (cells/mm2) of a) CD25+Foxp3+ T cells (regulatory T cells) and b) CD25+Foxp3neg T cells (effector T cells) in non-smoking (image_n/sji12035-gra-0001.png) and smoking (image_n/sji12035-gra-0002.png) asthmatic workers, healthy workers and non-exposed controls. The bars represent median values. image_n/sji12035-gra-0003.png: treatment with inhaled corticosteroids; image_n/sji12035-gra-0004.png ex-smokers; image_n/sji12035-gra-0005.png convalescent subject; image_n/sji12035-gra-0006.png: other individuals.

The CD25+ T cells were predominantly Foxp3+ Tregs

Triple-colour immunohistofluorescent staining for CD4, Foxp3 and CD3 revealed that 99% of the Foxp3+ T cells co-stained for CD4, in all groups. The majority of the CD25+ T cells (CD3+) were Foxp3+ and therefore phenotypically CD4+ Tregs, both in the non-smoking asthmatics (median 68%), smoking asthmatics (51%), smoking (67%) and non-smoking (86%) healthy potroom workers and non-exposed controls (88%). Non-smoking asthmatics had a significantly increased density of both bronchial CD25+Foxp3+ Tregs (Figs 1, 2A) and CD25+ Foxp3negative, putative CD4+ effector T cells (Fig. 2B).

In contrast, the majority of Foxp3+ T cells did not co-express CD25 and were therefore not classical CD4+ Tregs. Although the bronchial density of Foxp3+ CD25negative CD4+ T cells was highest in non-smoking asthmatics (median 45, range 23–95 cells/mm2), it was not significantly different from non-smoking healthy workers (median 33, range 16–56 cells/mm2) or controls (median 29, range 17–46 cells/mm2). The situation was actually reversed among the smokers as smoking healthy workers had a higher density (42, range 18–70 cells/mm2) than smoking asthmatics (median 16, range 8–60 cells/mm2).

Increased HLA-DR+ T cell density in asthmatics

Activated T cells may express HLA-DR. Non-smoking asthmatics had significantly higher bronchial density of HLA-DR+ T cells (median 37, range 0–104 cells/mm2) than smoking asthmatics (median 15, range 0–73 cells/mm2), non-smoking healthy workers (median 0, range 0–100 cells/mm2), smoking healthy workers (median 15, range 0–48 cells/mm2) and non-smoking controls (median 8, range 0–38 cells/mm2). Interestingly, although there were more T cells in the non-smoking asthmatics, the percentage of bronchial T cells that expressed HLA-DR were similar in non-smoking asthmatics (median 3.7%, range 0–9,3%) and smoking asthmatics (median 3.9%, range 0–12,7%). Triple-labelling revealed that HLA-DR was mainly expressed on CD8+ T cells (medians 79–93%). However, non-smoking asthmatics had increased bronchial densities of both HLA-DR+CD4+ and HLA-DR+CD8+ T cells compared with the controls (P < 0.05).

Proliferation of subepithelial T cells and Tregs and in asthma

Scattered proliferative (Ki-67+) T cells were almost exclusively observed in non-smoking asthmatics (median 14 cells/mm2, range 0–186) compared with smoking asthmatics (median 0, range 0–4 cells/mm2), healthy potroom workers (none observed) and non-smoking controls (median 0, range 0–16 cells/mm2). Triple-labelling revealed that both CD8+ and CD4+ T cells proliferated equally in the non-smoking asthmatics (~1:1 ratio).

Ten samples from nine subjects with high T cell proliferative response were selected for detailed triple-immunohistofluorescent labelling to examine whether Foxp3+ T cells proliferated in the bronchial mucosa (Fig. 1E–H). Distinct nuclear Ki-67+ labelling was observed on 24% of 121 subepithelial Foxp3+ T cells. Further triple-labelling revealed that 27% of 48 subepithelial Ki-67+Foxp3+ T cells were proliferative Tregs as they co-expressed CD25.

There were few proliferative T cells in the controls, except for the one convalescent smoking non-asthmatic potroom worker.

Few bronchial CD25+ T cells in smoking asthmatics

In contrast to the non-smoking asthmatics, there were no increased bronchial density of neither CD25+Foxp3+ Tregs (median 9, range 3–53 cells/mm2), CD25+Foxp3negative effector T cells (median 4, range 1–12 cells/mm2) or proliferating (Ki-67+) T cells (median 0, range 0–4 cells/mm2) compared with the smoking non-asthmatic workers.

Increased percentage of CD25+ T cells are effector T cells in asthma

Group-based data, adding all examined CD25+ T cells and their Foxp3 expression in patients and controls, revealed that an increased percentage of the CD25+ T cells (CD3+) were effector T cells in asthma as they did not co-express Foxp3 (Fig. 3). Consequently, the percentage of CD25+ T cells that were Tregs were reduced from 87% in non-smoking healthy workers to 73% in non-smoking potroom asthmatics, reducing the Treg to effector T cell ratio from 6.7 to 2.7, respectively. Similarly, the percentage of CD25+ T cells (CD3+) that co-expressed Foxp3 (Tregs) were reduced from 78% in the smoking healthy workers to 63% in smoking asthmatics, reducing Treg to effector T cell ratio from 3.5 in the smoking healthy workers to 1.7 in the smoking asthmatic workers. Thus, the only phenomenon that smoking and non-smoking asthmatics had in common was a decreased Treg to effector T cell ratio compared with the non-asthmatic potroom workers.

Figure 3.

Percentage of effector T cells of CD25+ T cells in non-smoking (□) and smoking (image_n/sji12035-gra-0002.png) asthmatic workers (A), healthy workers (Hw) and non-exposed controls (C). Group-based data (Chi-Square test).

Intraepithelial Tregs

Intraepithelial Foxp3+ T cells were occasionally observed, and grouped data revealed that 24% of 354 intraepithelial Foxp3+ T cells co-expressed CD25. Further triple-labelling experiments on selected specimens revealed that nine of 13 proliferative (Ki-67+) Foxp3+ T cells (69%) co-expressed CD25.


The study revealed several novel phenotypic characteristics of the bronchial T cell infiltration in asthma. Although the asthma-associated increase in bronchial CD25+ CD4+ T cells may have included both activated, asthma-promoting T-effector cells [17], as well as Foxp3+ regulatory T cells, triple-labelling revealed that the majority of the CD4+ CD25+ T cells were Foxp3+ and therefore phenotypically Tregs. This may be surprising that the asthmatic inflammation contains an increased density of putative immune inhibitory Tregs, simultaneously as the patients have active asthma. However, this is similar to the increased percentage of BAL Tregs in stable adult asthma [18], but contrary to the decreased percentage in active childhood asthma [13]. However, childhood asthma is predominantly an atopic asthma subtype in which IgE-mediated mast cell degranulation is important, whereas the current asthma patients had non-atopic, presumably T cell-mediated asthma [17]. The increase in bronchial Tregs may, moreover, be counteracted by similar increase in effector T cells. This was reflected in the increased bronchial density of both Tregs and effector T cells which increased the effector T cell to Treg ratio 2,5 times, from 2,7 in controls to 6,7 in non-smoking asthmatics. Although the balance between these systems may control bronchial inflammation, their functional state may be more important than their densities. This was partly revealed in childhood asthma [13] where inhaled corticosteroid treatment restored the asthma-associated impairments in Treg inhibition. Perhaps this was a result of corticosteroid induced reduction in epithelial thymic stromal lymphopoietin (TSLP) expression, a cytokine which has been shown to inhibit Treg function in vitro [19]. Similar mechanisms would explain why the increased bronchial Treg density was unable to fully inhibit the bronchial inflammation in the non-smokers as they all had active disease (>12% reversible airflow obstruction).

Although Tregs have been shown to prevent experimental allergy and asthma [12], they inhibit Th1- better than Th2-type of immune responses [20]. Thus, their stronger Th1-inhibitory capacity may push the well-known Th1-Th2 balance towards an asthma-promoting Th2 type of inflammation as illustrated in the murine allergic airway model [20, 21].

The bronchial Tregs in the non-smoking asthma were presumably activated as a considerable fraction of them expressed the proliferating marker Ki-67 in situ. This may appear in contrast to how anergic Tregs appear in vitro, but both murine [22] and human Tregs proliferate in vivo [23]. Thus, the bronchial increase in proliferative Tregs in the non-smoking asthmatics illustrated that the Treg system was locally activated in asthma.

The majority of the bronchial Foxp3+ T cells were actually CD25 negative. Although there are some methodologically discrepancies [24, 25], newly activated T cells may temporally express Foxp3, without gaining classical inhibitory properties (25, 26). They are presumably in a transitional state to either differentiate into IL-17A producing Th-17 cells with or without concomitant Foxp3+ expressions [27] or to become functional Tregs depending on local cytokine stimulation [28]. Although the functional repertoire for these Foxp3+CD25negative T cells are currently unknown, they have been reported to inhibit T cell proliferation, but not IFN-γ production, in systemic lupus erythematous (SLE; [26]), and transfer tolerance in a murine allergen induced asthma model [29]. Together this suggested that the Foxp3+CD25negative T cells have immune regulatory properties in the asthmatic bronchi as well. However, the bronchial concentration of these cells in asthmatics was not significant different from the controls; it was even higher in smoking healthy workers than in smoking asthmatics. Thus, whatever functional role these Foxp3+CD25negative T cells may play in asthma immunopathology needs, therefore, to be established through further studies.

This is apparently the first biopsy study to compare the bronchial inflammatory infiltrate in non-smoking with smoking asthmatics. We previously reported that smoking asthmatics did not have the asthma-associated increase in bronchial CD4+ T cells as non-smoking asthmatics [4]. Because smoking-induced chronic obstructive pulmonary disease (COPD) and fatal asthma have been associated with increased density of CD8+ T cells [30, 31], we hypothesized that asthma in smoking asthmatics could be associated with activated CD8+, rather than CD4+ T cells. However, there was no evidence for activation for any particular T cell subsets. This was somewhat surprising because these patients had similar reversibility as the non-smoking asthmatics. Tobacco smoke has been shown to modulate immune reactions [32-34] and inhibit bronchial influx of CD4+ T cells and other leucocyte subsets (for references see [4]). The current study supported the concept that CD8+ T cells are involved in asthma pathophysiology, but the evidences were only visible in the non-smoking asthmatics where the bronchial mucosa contained an increased density of activated (HLA-DR+) and proliferating (Ki-67+) CD8+ T cells. This in contrast to smoking asthmatics where there was no evidence for any particular role for CD8+ T cells as there was no increase in bronchial CD8+ T cells nor any increase in activated CD8+ T cells. Although these results challenge our understanding of how bronchial T cells are involved in asthma pathophysiology [17], the lack of asthma-associated increase in bronchial T cells may have a more simple explanation. Tobacco smoke tar contains the most active immune modulating substances [31-33] and is known to condensate on the surface of the larger bronchia. It may therefore influence the bronchial inflammation in locations where biopsies are normally taken. Asthmatic symptoms, however, are generated by smooth muscular contraction in the distal smaller bronchioles, which may be less influenced by tobacco smoke explaining why there was no asthma-associated increase in CD4+ T cells in the larger bronchia. Only further examination of peripheral lung tissue in smoking asthmatics may clarify whether asthma-associated bronchial inflammation is present in the smaller airways in smoking asthmatics as well.

Although none of the asthma-promoting components in the potroom fume has been identified, it may induce asthma-associated bronchial inflammation in all workers. This was supported by slightly increased reticular basement membrane thickness [3], increased subepithelial eosinophil density [3], and an increased bronchial density and percentage of subepithelial CD4+ effector T cells (CD25+, Foxp3negative) observed in all non-smoking asymptomatic workers compared with non-smoking controls (current study). Whether this eventually may develop into clinical relevant asthma is presumably a matter of genetic predisposition, time and degree of exposure.


There is an increased bronchial density of both CD4+CD25+ Foxp3+ regulatory T cells and an increased bronchial density of putative effector T cells in non-smoking asthmatics. Both T cell subsets proliferated in situ. Thus, both inhibitory Tregs and inflammatory T cells were activated in the bronchial mucosa of non-smoking asthmatics. Although the smoking asthmatics did not have similar increase in bronchial T cell subtypes, both asthma subgroups had a decrease in Treg to T-effector cell ratio, and thus an relatively increase in effector T cells over Tregs. Thus, the functional balance between these systems may modulate asthma immunopathology in both asthma groups.


The authors would like to thank all the workers for participating in the study. K. Vold, T. Skorve, D. Malterud, F. Myklebust and O.A. Hauge are all acknowledged for participating in clinical investigations and recruitment of workers, and Solveig Stig for excellent technical assistance. TS had the primary patient contact, clinical examination, performed the immunohistochemical evaluation and writing of the draft manuscript. ØB performed the bronchoscopy, biopsy and participated in the clinical evaluation, PAD had clinical responsibility for treating and selection of asthmatic potroom workers, JK participated in study design, clinical examination and writing of the manuscript. TSH designed the study, participated in immunohistofluorescence staining, data evaluation and writing of the manuscript. The authors state there are no known conflicts of interest or economical interest in the current study.