Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Over the last few years, compelling evidence has established an important physiologic role for apoptotic cell death in maintaining optimal cell numbers in multicellular organisms (1). Proteins such as Bcl-2 or Fas and its ligand, Fas ligand (FasL), have been shown to regulate this process (2, 3).

Bcl-2 is an intracellular protein that inhibits apoptotic death induced by a variety of stimuli in different cell types (2). FasL and Fas are transmembrane proteins, belonging to the tumor necrosis factor- and its receptor families, respectively (3). The activation of Fas receptor by its natural ligand or by specific antibodies (Ab) induces apoptosis (3). A dysregulation of the balance between cell death and cell growth or between Bcl-2 and Fas or FasL expression has been involved in the pathogenesis of a number of diseases, such as cancer; some autoimmune, hematological, or neurodegenerative disorders; and acquired immune deficiency syndrome (4).

Although apoptosis is claimed to be fundamentally significant in the resolution of the inflammation response (1), a defect in cell death associated with an altered expression of different pro- and antiapoptotic factors has never been demonstrated in asthma. This syndrome is characterized by a chronic inflammation of the bronchial submucosa, in which eosinophils and T lymphocytes play a critical role (5). Indeed, T cells of the Th2-like phenotype orchestrate this inflammation via the release of cytokines, which influence several aspects of the eosinophil biology, including proliferation and differentiation, chemoattraction, survival, and activation (5). Eosinophils contribute to the pathogenesis of asthma and to the development and maintenance of airway inflammation and tissue injury by releasing proinflammatory cytokines, cytotoxic cationic proteins, oxygen metabolites, and lipid mediators (7).

Epithelial damage is also a characteristic feature of asthma and contributes directly to its severity (8, 9). This phenomenon, which is claimed to be the consequence of the effects of mediators released from activated inflammatory cells, is followed by repair mechanisms with proliferation of basal cells after local injury (10). Growth factors secreted by different cell types, including mononuclear phagocytes, fibroblasts, and the epithelial cells themselves, can stimulate proliferation in an autocrine or paracrine manner (10).

Steroids have a prominent place in the treatment of asthma (11), as they reduce airway inflammation by different mechanisms, including the inhibition of the synthesis of proinflammatory cytokines from various cell types, particularly T lymphocytes (12). It is noteworthy that these drugs promote directly the in vitro apoptotic death of human eosinophils (13, 14) and of activated T lymphocytes (15).

In the present study, the expression of some molecules involved in apoptotic cell death, such as Bcl-2, Fas, and FasL, was examined by immunohistochemistry in bronchial biopsies from control individuals and from intrinsic and extrinsic asthmatic patients, treated or not treated with inhaled or oral steroids. In parallel, terminal transferase-mediated deoxyuridyltriphosphate nick-end labeling (TUNEL) technique and immunohistochemical detection of proliferating cell nuclear antigen (PCNA) were used to determine the presence of apoptotic and proliferating cells, respectively.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Asthmatic and Control Subjects

Eighteen asthmatic patients fulfilling the criteria of the International Consensus Report on Diagnosis and Management of Asthma (16) were recruited. All of them exhibited a history of dyspnea and wheeze episodes and demonstrated a reversible airflow limitation characterized by a 20% improvement in forced expiratory volume in 1 s (FEV₁) measurement after the inhalation of 200 µg of albuterol. The atopic status of these patients was defined by positive skin-prick tests to extracts of common aeroallergens and evaluation of serum-specific immunoglobulin (Ig)E by enzyme-linked immunosorbent assay (ELISA) (Behring Hoechst, Reuil Malmaison, France). All patients were nonsmokers, had normal sinus and chest X-ray films, and had no history suggestive of a viral upper respiratory tract infection within the 4 wk preceding the study. The 18 asthmatics were distributed into two groups, according to their treatment at the time of the study. The first group consisted of nine subjects who received intermittent inhaled short acting 2 agonists, taken as needed, and were further defined as untreated asthmatics. The second group included nine asthmatic patients who were treated with inhaled glucocorticosteroids associated with oral steroids in two cases (Table 1, Patients 2 and 9). These nine patients were defined as steroid-treated asthmatics. In the case of oral corticotherapy, the treatment of Patients 2 and 9 (Table 1) had been started 4 mo and 6 wk, respectively, before the biopsy removal. The clinical severity of asthma was assessed according to the Aas scoring system (17), which grades asthma severity on a scale of 1 to 5. Thus, a score of 1 defines very mild asthma and a score of 5 indicates an incapacitating disease requiring treatment.

A group of nine healthy volunteers with no known diseases and taking no medication was also studied. All of them were nonsmokers and had no history of allergy or asthma. Skin-prick tests were negative and total serum IgE values were below 10 kU/liter.

Each patient signed an informed consent statement, and the protocol was approved by the ethical committee of the Centre Hospitalier Régional Universitaire de Lille (CP 93/07; Lille, France).

Control and asthmatic characteristics are summarized in Table 1.

Fiberoptic Bronchoscopy and Bronchial Biopsy Processing

The FEV₁ was measured before the fiberoptic bronchoscopy as a safety precaution, and the investigations were carried out only in patients with FEV₁ values of more than 70%. The bronchoscopy was performed according to the guidelines outlined by the American Thoracic Society (18). The subjects were premedicated with 5 mg preservative-free salbutamol via an inspiron nebulizer before bronchoscopy. Next, the larynx and upper airways were anesthetized with 2% xylocain spray. The oxygen saturation was monitored with a digital oximeter. The bronchoscope used was an Olympus FT20 model (Olympus Optical, Tokyo, Japan). Four endobronchial mucosal biopsies were taken between the carena and the right lower lobes using alligator forceps.

Bronchial biopsies were mounted over cork disks, covered by an optimum cutting temperature compound (BDH, Poole, UK) and snap-frozen in isopentane, cooled by liquid nitrogen. The frozen blocks were kept at 80°C prior to use. Five-micrometer sections were cut in a cryostat, and two consecutive sections were collected on glass slides previously coated with acetone containing 5% of 3-amino-propyltriethoxy silane (Sigma, St. Quentin Fallavier, France). Sections were then dehydrated in acetone for 10 min at room temperature, wrapped in a plastic film, and kept at 20°C until use.

Immunohistochemistry

The Abs used and their corresponding dilutions or concentrations were as follows: mouse IgG1 antihuman eosinophil cationic protein, clone EG2, at 10 µg/ml (Pharmacia, Uppsala, Sweden); mouse IgG1 antihuman cluster of differentiation (CD)3, clone UCHT-1; and mouse IgG1 antihuman CD4, clone Q4120, both at 10 µg/ml (Sigma); mouse IgG1 antihuman Bcl-2, clone 124, at 6 µg/ml; mouse IgG2a anti-PCNA, clone PC10, at 12 µg/ml; rabbit Ig antihuman CD3, at 1/50; irrelevant mouse IgG1, clone DAK-GO1, at 10 and 50 µg/ml; irrelevant mouse IgG2a, clone DAK-GO5, at 12 µg/ml; goat antimouse Ig and alkaline phosphatase-mouse antialkaline phosphatase (APAAP) monoclonal antibody (mAb), both at 1/25 (Dako, Trappes, France); mouse IgG1 antihuman Fas, clone UB2, at 50 µg/ ml (Immugenex, Los Angeles, CA); rabbit Ig antihuman, mouse and rat FasL, sc-834, and rabbit Ig antihuman Fas, sc-714, both at 2 µg/ml (Santa Cruz Biotechnologies, Santa Cruz, CA); irrelevant rabbit IgG, at 2 µg/ml (Biogenesis, Poole, UK); fluorescein isothiocyanate (FITC)-conjugated Ab to mouse IgG1, at 1/150 (Southern Biotechnology Associates, Birmingham, AL); rhodamine-labeled Ab to rabbit Ig, at 1/100 (Immunotech, Marseille, France).

Cryostat sections were incubated for 30 min or 1 h at 20°C or 37°C, in the presence of the specific murine mAbs, or of their corresponding isotypes. Goat antimouse Ig and mouse APAAP were next used. All Abs were diluted in Tris buffer, pH 7.6, supplemented with 1% bovine serum albumin (BSA). Substrate solution, consisting of 0.5 mM naphtol AS-XM phosphate, 2% dimethylformamide, 100 mM Tris (pH 8.2), 1.3 mM levamisole, and 1 mg/ml Fast Red TR salt (all from Sigma), was applied for 30 min at 20°C. A light nuclear hematoxylin counterstaining was performed next, and slides were mounted using IMMU-MOUNT (Shandon, Pittsburgh, PA). Slides were washed twice for 10 min at 20°C in Tris buffer (pH 7.6) between each step. For rabbit Ig anti-FasL or control rabbit Ig, the alkaline phosphatase LSAB⁺ kit (Dako) was used as the revelation system. Positive cells appeared with a blue-counterstained nucleus surrounded by red staining.

To identify the phenotype of Fas- and FasL-positive cells, double-immunofluorescence stainings were performed. Abs were diluted in phosphate-buffered saline (PBS), pH 7.4, supplemented with 1% BSA and slides were washed in PBS at 20°C in the dark. Slides were sequentially incubated for 30 min or 1 h at 37°C in the dark with the anti-Fas or anti-FasL polyclonal Ab, followed by the rhodamine- labeled Ab to rabbit Ig, by the EG2 or anti-CD3 mAb, and finally by the FITC-conjugated Ab to mouse IgG1. Slides were mounted using IMMU-MOUNT containing 100 mg/ ml DABCO (Sigma). This protocol was also carried out for identifying the colocalization of CD3 and Bcl-2, using the anti-CD3 polyclonal Ab prior to the anti-Bcl-2 mAb. To verify whether eosinophils express Bcl-2, enzymatic detection of the peroxidase, which is contained in high levels in eosinophils (19), was combined with the immunohistochemical staining of Bcl-2. After Bcl-2 immunostaining and washing twice in Tris buffer (pH 7.6), sections were incubated in acetate buffer (pH 5.0) containing 100 µg/ml 3-amino-9-ethylcarbazole (Sigma), 5% dimethylformamide, and 0.015% hydrogen peroxide for 20 min at 20°C. Bcl-2-positive cells were detected by the APAAP technique, as described previously, except that Fast Red TR salt in the substrate solution was replaced by Fast Blue BB salt (Sigma) and no hematoxylin counterstaining was performed. Eosinophils were identified by an orange color and Bcl-2⁺ cells were detected by blue staining. Double-positive cells stained for both colors.

In Situ Apoptosis Detection

In situ apoptosis was performed using the APOPDETEK cell death assay system (Enzo Diagnostic, Farmingdale, NY), based on the TUNEL technique (20). This technique allows the labeling of the 3'OH-free DNA ends, which are generated during DNA fragmentation, a phenomenon specific to apoptosis (20). The instructions of the manufacturer were modified, as 25 µl of the label reagent and 1.5 µl of terminal transferase were used and the substrate was replaced by naphtol phosphate/Fast Blue solution, prepared as described previously. Slides were incubated for 20 min in the presence of the buffer containing the enzyme and dUTP-biotin, next with the solution containing alkaline phosphatase-conjugated streptavidin, and finally for 30 min with the substrate solution. Apoptotic cells appeared with a blue nucleus, whereas negative nuclei were pink, due to red nuclear counterstaining.

In some sections, in situ apoptosis and immunohistochemistry, using the EG2 or the anti-CD3 mAb, were combined. No counterstaining was performed. Double-positive cells appeared with a purple nucleus surrounded by red staining.

Quantitation

Slides were randomly coded and counted by two observers in a blind fashion. Counts from both observers, which were compared using the Wilcoxon test for nonparametric paired values, were not statistically different. A minimum of two nonserial sections from the same biopsy or sections from two different biopsies were analyzed and the mean values calculated.

The numbers of EG2⁺, CD3⁺, CD4⁺, Bcl-2⁺, Fas⁺, FasL⁺, and PCNA⁺ cells and apoptotic cells were determined in the bronchial submucosa, excluding glands, in a zone 125 µm deep, as defined by a calibrated squared eyepiece graticule, along the length of the basement membrane. Results represent the number of positive cells per unit length (1 mm) of basement membrane. In the case of double-stainings, values are expressed as percentage of eosinophils or of CD3⁺ T lymphocytes expressing Bcl-2, Fas, or FasL, or undergoing apoptosis, and as percentage of Bcl-2⁺, Fas⁺, FasL⁺ cells, or apoptotic cells that were eosinophils or CD3⁺ T lymphocytes.

The intensity of the expression of Fas, FasL, and PCNA in the epithelium was evaluated with a semiquantitative scoring system, as described previously (21, 22). Scores 0, 1, 2, or 3 defined the absence of staining, or the presence of a staining of weak, moderate, or high intensity, respectively.

The numbers of apoptotic cells per 100 total epithelial cells and that of Bcl-2⁺ cells per 100 epithelial cells of the basal layer were also determined.

Results are expressed as means ± SEM. The number of positive cells and the scores for most of the mAbs used showed a skew distribution. Therefore, the nonparametric Mann-Whitney U test for unpaired values was used to determine significance among groups. Correlation coefficients (r') were calculated using the nonparametric Spearman's rank-order method. P values of  0.05 were considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Cellular and In Situ Apoptosis Analysis in the Bronchial Submucosa

Immunohistochemical analysis revealed a significant increase in the number of eosinophils in the bronchial submucosa from steroid-untreated asthmatics, compared with that of controls (Table 2). No significant changes in CD3⁺ or CD4⁺ T lymphocyte counts were observed (Table 2). Treatment with steroids decreased eosinophilic infiltrate without modifying the number of T cells (Table 2).

The number of cells expressing Fas, FasL, or Bcl-2 was then determined (representative stainings are illustrated in Figure 1). Both control and asthmatic patients, treated or not treated with steroids, showed comparable numbers of Bcl-2⁺, Fas⁺, or FasL⁺ cells in their bronchial submucosa (Table 2).

View larger version (133K): [in this window] [in a new window]   Figure 1.   Identification of Bcl-2- (A), Fas- (B), and FasL (C)-positive cells (immunohistochemistry and APAAP), and of apoptotic nuclei (D, TUNEL technique) in bronchial biopsies from asthmatic patients. A, B, and C: Positive cells stain red, whereas cellular nuclei are counterstained in blue. Positive cells are indicated by the arrows in B and C. D: Apoptotic nuclei, shown in the mucosa and the epithelium, stain blue, whereas nonapoptotic nuclei are pink. Scale bar = approximately 32 µm.

In situ apoptosis and proliferating cell detection were performed next by the TUNEL technique and immunohistochemistry using the anti-PCNA mAb, respectively. The number of apoptotic cells, as identified by an intense nuclear staining (Figure 1D), was significantly decreased in the submucosa from steroid-untreated asthmatics, compared with that of control subjects (Table 2). In parallel, a trend toward an elevation in the number of apoptotic cells was noted in steroid-treated as opposed to untreated asthmatics. The high degree of variability in their counts, however, precluded the results from achieving statistical significance (Table 2).

Most of the biopsies failed to show PCNA-expressing cells in the bronchial submucosa and, when present, only rare positive cells were identified. The number of proliferating cells was similar in the three groups of patients (data not shown).

Next, cell distribution and apoptotic marker expression in the bronchial mucosa of intrinsic and extrinsic asthmatics, either steroid-treated or untreated, were compared. The numbers of CD4⁺ and Bcl-2⁺ cells were statistically increased in the submucosa of intrinsic (n = 4), compared with extrinsic (n = 5) steroid-untreated asthmatic patients, as 58.9 ± 19.4 and 24.3 ± 10.0 CD4⁺ cells (P = 0.050), and 29.9 ± 14.0 and 5.4 ± 3.3 Bcl-2⁺ cells (P = 0.046), respectively, were observed. No additional statistical differences were found (data not shown).

To determine whether eosinophils and T lymphocytes express Bcl-2, Fas, or FasL, double-immunofluorescence stainings were performed on six individuals from each group. Results are summarized in Table 3 and illustrated in Figure 2. No differences in the percentage of eosinophils and T lymphocytes bearing Bcl-2, Fas, or FasL were observed when control subjects and steroid-untreated and -treated asthmatics were compared (Table 3). When data from all patients were averaged, approximately 20% of eosinophils were shown to express Fas, 4% were FasL⁺, and only 2% of them were Bcl-2⁺. In contrast, nearly 30% of T lymphocytes expressed Bcl-2, whereas approximately 10% stained for Fas and 5% were FasL⁺.

View larger version (37K): [in this window] [in a new window]   Figure 2.   Immunofluorescent double-stained cells in bronchial sections from asthmatic patients. Sections were incubated with the following couples of Abs: mouse EG2/rabbit anti-Fas (A and B), mouse EG2/ rabbit anti-FasL (C and D), rabbit anti-CD3/mouse anti-Bcl-2 (E and F), mouse anti-CD3/rabbit anti-Fas (G and H), mouse anti-CD3/rabbit anti-FasL (I and J). FITC-conjugated antimouse Ig and rhodamine-labeled antirabbit Ig were used as revelation systems. The same microscopy field was photographed twice, first with a filter specific for FITC (A, C, F, G, and I) and next with one for rhodamine (B, D, E, H, and J) using exposition times of 7.5 to 30 s. Double-positive cells are indicated by the arrows, except in G and H, where a double-stained cell cluster, surrounded by four double-positive cells, is shown. A and B: Scale bars = approximately 21 µm; C to J: Scale bars = approximately 19 µm.

The proportion of Bcl-2-, Fas-, or FasL-positive cells that were eosinophils or CD3⁺ T lymphocytes was then examined. A high proportion of Bcl-2⁺ cells were identified as T cells (Table 3), a finding paralleling positive correlations between the number of CD3⁺ or CD4⁺ T lymphocytes and that of Bcl-2⁺ cells in the bronchial submucosa from controls and steroid-untreated or -treated asthmatics (CD3⁺ versus Bcl-2⁺: r' = 0.722, P < 0.001, n = 27; CD4⁺ versus Bcl-2⁺: r' = 0.602, P = 0.002, n = 27). No differences in the proportion of Bcl-2-, Fas-, or FasL-positive cells that were eosinophils or T lymphocytes were observed when control subjects and steroid-untreated and -treated asthmatics were compared (Table 3).

To distinguish apoptotic eosinophils from T lymphocytes, TUNEL technique was combined with immunohistochemistry. A slight, although not significant, increase in the proportion of apoptotic eosinophils was observed in the submucosa of steroid-treated asthmatics, compared with that of untreated asthmatics or control subjects (Table 3). In contrast, the proportions of apoptotic T lymphocytes were similar in the three groups of individuals (Table 3).

Finally, the individual numbers of eosinophils and the FEV₁ values negatively correlated (r' = 0.442, P = 0.026, n = 27). No additional correlations were found when EG2⁺, CD3⁺, CD4⁺, Bcl-2⁺, Fas⁺, FasL⁺, or apoptotic cells in the submucosa were related to each other or to FEV₁ values or Aas scores.

Bcl-2, Fas, FasL, and PCNA Expression and In Situ Apoptosis in the Bronchial Epithelium

Bcl-2, Fas, FasL, and PCNA were detected in the bronchial epithelium of the majority of the patients, irrespective of their groups (Figure 3). No cell positivity was detected when bronchial sections were incubated with the isotype controls (Figure 3).

View larger version (100K): [in this window] [in a new window]   Figure 3.   Immunohistochemical identification of Bcl-2, Fas, FasL, and PCNA in bronchial epithelium from steroid-treated asthmatic patients. Serial cryostat sections were incubated with irrelevant isotype controls (A, C, E, and G) or with their corresponding specific anti-Bcl-2 (B), anti-Fas (D), anti-FasL (F), or anti-PCNA (H) Abs, followed by APAAP. Positive cells stain red. Contrary to Fas, FasL, and PCNA, which are expressed uniformly by all the epithelial cells, Bcl-2 staining is concentrated on the basal epithelial cell layer. Scale bar = approximately 25 µm.

Contrary to Fas, FasL, and PCNA, which were present in all epithelial cells, Bcl-2 expression was focal and restricted to the basal epithelial cell layer (Figure 3). Because of this different cell distribution, the number of Bcl-2-positive cells per 100 epithelial cells of the basal layer was determined, whereas the presence of Fas, FasL, and PCNA was quantitated in terms of intensity and expressed as a scoring system. Fas expression was significantly enhanced in the bronchial epithelium of steroid-untreated asthmatics, compared with that of control subjects (Figure 4B). This expression was further increased, although not significantly, in the steroid-treated group (Figure 4B). In contrast, Bcl-2⁺ cell counts, FasL, and PCNA scores were significantly augmented in the epithelium of steroid-treated asthmatics exclusively (Figures 4A, 4C, and 4D).

View larger version (23K): [in this window] [in a new window]   Figure 4.   Expression of Bcl-2 (A), Fas (B), FasL (C), and PCNA (D) in the bronchial epithelium from control individuals (C), and from asthmatic patients treated (G) or not treated (A) with steroids. A: The number of Bcl-2+ cells per 100 epithelial cells from the basal layer was evaluated. B, C, and D: The intensity of the Fas, FasL, and PCNA stainings was expressed as a scoring sytem, in which scores 0, 1, 2, and 3 correspond to the absence of staining, or the presence of a red staining of weak, moderate, or high intensity, respectively. Results are expressed as the means ± SEM from nine control subjects, nine untreated asthmatics, and nine steroid-treated asthmatics. When differences were statistically significant, P values are indicated.

The majority of the epithelia expressing PCNA (score > 1) concomitantly showed a positivity for Bcl-2, Fas, and FasL, whereas epithelia that failed to express PCNA were either Fas⁺/FasL or Fas/FasL⁺ or Fas/FasL (data not shown). Accordingly, the individual Bcl-2⁺ cell counts or the intensities of the expression of Fas and FasL in the bronchial epithelium from all patients were significantly related to PCNA score (Figure 5). Furthermore, the expressions of Bcl-2, Fas, and FasL positively correlated with each other (Bcl-2⁺ versus Fas⁺: r' = 0.516, P = 0.008, n = 27; Bcl-2⁺ versus FasL⁺: r' = 0.699, P < 0.001, n = 27; and Fas⁺ versus FasL⁺: r' = 0.517, P = 0.008, n = 27).

View larger version (10K): [in this window] [in a new window]   Figure 5.   Correlations between individual numbers of Bcl-2+ cells (A), intensities of Fas (B) and FasL (C) expression and that of PCNA in the bronchial epithelium from control healthy individuals (open circles) and from steroid-untreated (closed circles) and -treated (open squares) asthmatic individuals. The number of Bcl-2+ cells was evaluated per 100 epithelial cells of the basal layer, and the intensity of the Fas, FasL, or PCNA labeling was quantified using a scoring system as defined in the legend of Figure 4. Correlation coefficients (r') and P values are indicated.

In situ apoptosis analysis, as assessed by the TUNEL technique, disclosed similar numbers of apoptotic cells in the epithelium of control subjects and of steroid-untreated and -treated asthmatics (3.8 ± 0.9, 2.1 ± 0.5, and 2.4 ± 0.6 apoptotic cells, respectively, n = 9, differences not statistically significant).

Next, we compared intrinsic and extrinsic asthmatics, either steroid-untreated or -treated, in terms of their ability to express Bcl-2, Fas, FasL, and PCNA in the bronchial epithelium. FasL and PCNA scores were slightly but statistically increased in the epithelium from intrinsic, compared with extrinsic, steroid-untreated asthmatics. Accordingly, the FasL scores for intrinsic (n = 4) and extrinsic (n = 5) asthmatics were of 1.5 ± 0.2 and 0.7 ± 0.3 for FasL (P = 0.049), and those of PCNA were of 0.8 ± 0.1 and 0.1 ± 0.0 (P = 0.012), respectively. No additional statistical differences were found when these groups of patients were compared (data not shown).

Finally, Bcl-2, Fas, FasL, and PCNA stainings, as well as the number of apoptotic cells in the epithelium, failed to correlate significantly with the individual FEV₁ values or Aas scores (data not shown).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

As reported previously, bronchial biopsies from asthmatics showed a steroid-sensitive increase in the numbers of eosinophils, compared with those from controls (12, 23- 25). This increase, however, was moderate and was unaccompanied by a rise in either CD3⁺ or CD4⁺ T cell counts, a phenomenon that may be related to the fact that the steroid-untreated patients included in the present study were mild asthmatics. When extrinsic and intrinsic asthmatics were considered separately, however, the number of CD4⁺ T lymphocytes was augmented in patients with intrinsic asthma, compared with extrinsic asthmatics and control subjects.

A diminished number of apoptotic cells, particularly eosinophils, was observed in the submucosa of steroid- untreated asthmatics, compared with that of controls. This suggests that a defect in eosinophil apoptosis, probably secondary to an increased expression of certain cytokines responsible for their survival (26, 27), may be involved in the development and the maintenance of bronchial inflammation in asthmatics. Steroid administration augmented the proportion of apoptotic eosinophils, as recently shown in the sputum of asthmatics (28). This phenomenon may be a consequence of a direct effect of steroids on eosinophil apoptosis (13, 14) and/or of the inhibition of the synthesis of interleukin-5 and granulocyte macrophage colony-stimulating factor, two cytokines that enhance eosinophil survival (12, 29).

Immunohistochemical analysis demonstrated limited changes in the number of cells expressing Bcl-2, Fas, FasL, or PCNA in the bronchial submucosa of control subjects and extrinsic or intrinsic asthmatic patients. Treatment with either oral or inhaled steroids did not modify significantly Bcl-2⁺, Fas⁺, FasL⁺, or PCNA⁺ cell counts.

Using a double-immunofluorescence staining technique, we failed to disclose any difference in the proportions of eosinophils or T cells expressing Bcl-2, Fas, or FasL in the submucosa of control or asthmatic subjects, either steroid untreated or treated. Approximately 30% of T cells in the bronchial submucosa express Bcl-2, whereas only 2% of eosinophils bear this protein, which is consistent with in vitro data, including ours, that show constitutive expression of Bcl-2 in mature T cells, but not in eosinophils (30, 31).

The expression of the two proapoptotic molecules Fas and FasL in the bronchial submucosa was restricted to less than 10% of T cells. This finding is compatible with the observation that only activated T lymphocytes express Fas and FasL (32), and that limited numbers of T cells bear the activation marker CD25 in the bronchial submucosa from asthmatics (25).

In keeping with in vitro data showing constitutive expression of Fas on human eosinophils (13, 33, 34), we found that 20% of eosinophils in the bronchial tissue bear this protein. In contrast, only 4% of tissue eosinophils were FasL⁺, whereas in the peripheral blood they fail to express this antigen (35). This suggests that FasL might be induced on eosinophil surface as a consequence of local activation.

In addition to T cells and eosinophils, other cell types in the bronchial submucosa may express Bcl-2, Fas, and/or FasL or be apoptotic. These antigens may indeed be present on neutrophils, fibroblasts, activated B cells, mast cells, or macrophages, as previously demonstrated in vitro (30, 35).

In this study, we present evidence for the first time that bronchial epithelium may express concomitantly Bcl-2, Fas, FasL, and PCNA. Accordingly, immunohistochemical analysis disclosed that the basal epithelial cell layer, a zone that is claimed to be responsible for the proliferation and regeneration of the bronchial epithelium (10), expressed Bcl-2, extending a previous report (42). The number of Bcl-2⁺ cells was statistically correlated with the intensity of PCNA expression, which supports the hypothesis that most of the epithelial cells expressing Bcl-2 are located in the proliferating zones, as demonstrated in the intestinal lower crypt (42). However, in vitro experiments have shown that Bcl-2 acts mainly by prolonging epithelial cell survival, rather than by directly favoring their proliferation (43). Because Bcl-2 and PCNA were highly increased in the epithelium from steroid-treated asthmatics, we suggest that steroids may facilitate a process of epithelial proliferation and survival, and that Bcl-2 may be considered a marker of these phenomena.

Our results show that human bronchial epithelium expresses Fas and FasL, thus extending previous findings demonstrating the presence of Fas and FasL mRNA in the murine bronchial epithelium and lung tissue, respectively (44, 45).

The expression of FasL was markedly increased in the epithelium of steroid-treated asthmatics only, indicating that this tissue may acquire the ability to trigger the apoptotic death of different Fas-bearing cells, including eosinophils and T cells. This concept has already been proposed by Dayan and colleagues (46) in the case of Hashimoto's thyroiditis, in which crosslinking of Fas by FasL present on thyroid epithelial cells may favor the deletion by apoptosis of potentially autoaggressive T cells.

Contrary to Bcl-2, PCNA, and FasL, the expression of Fas was significantly increased in the epithelium of steroid-untreated asthmatics, possibly as a consequence of its activation secondary to the ongoing inflammatory process. The intensity of Fas expression, however, failed to correlate with apoptotic death, suggesting that the number of apoptotic cells in the epithelium might be underestimated because of their detachment in the bronchial lumen, or that Fas antigen might not be functional. A positive correlation between Fas and PCNA expression was noted, indicating that Fas activation might be involved in epithelium proliferation rather than in apoptotic death. This paradoxical hypothesis is reminiscent of a recent report showing that Fas triggering resulted in human fibroblast proliferation (36).

In conclusion, we demonstrated that steroids exert beneficial effects in asthmatic patients by facilitating the resolution of eosinophilic inflammation and by promoting epithelial cell survival and proliferation, possibly through the expression of Bcl-2 and PCNA. By inducing FasL expression, steroids may also render the bronchial epithelium prone to exert a cytotoxic activity toward Fas-expressing cells invading the bronchial submucosa.