Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The accumulation of inflammatory cells in the bronchial tissue during allergic reactions is T lymphocyte-dependent (1). Specifically, T cells from the Th2 subtype secrete cytokines that induce the maturation, migration, and accumulation of effector cells, particularly eosinophils. Ovalbumin (OVA)-immunized mice challenged locally with OVA respond with an influx of eosinophils, lymphocytes, and, to a lesser extent, other inflammatory cells into their airways (2, 3). These mice may display antigen-dependent functional respiratory disturbances similar to those of asthma, such as bronchopulmonary hyperreactivity (BHR). Nevertheless, most of the protocols so far described in mice require multiple immunizations and/or challenges with antigen in order to achieve significant BHR (2, 4). Although multiple provocations during long periods of time may simulate the events of asthma, the mechanisms of inflammatory-cell recruitment are probably better understood under a single antigenic challenge. For this reason, we investigated inflammatory responses and BHR following a single antigenic challenge in the hyperimmunoglobulin (Ig)E BP2 mice. In previous studies, these mice were found to express BHR after four antigenic challenges (5) under conditions that left BALB/c mice normoresponsive. In the presently described single OVA challenge protocol, nonanesthesized BP2 mice became hyperreactive to inhaled methacholine. We followed the kinetics of recruitment of lymphocytes, eosinophils, mast cells, and non-mast cell IgE-bearing cells to the lungs, as well as the major epithelial modifications in the bronchioles, under both light and electron microscopy in BP2 and BALB/c mice. To study the role of inflammatory cells for BHR, we depleted T cells with antibodies and took advantage of the granulocytopenic effects of vinblastine (6, 7) to correlate granulocyte (eosinophil) recruitment and epithelial modifications. Our results show that the recruitment of eosinophils and lymphocytes into the peribronchiolar and perivascular tissues within hours after a single antigen instillation into the murine airways is accompanied by that of IgE-bearing polymorphonuclear leukocytes (PMN), most likely basophils, and by a BHR that starts by 1 h after challenge. Epithelial cells of the bronchi and bronchioles undergo a mucous-cell metaplasia that starts within the first 6 h and persists for more than 11 d after challenge. Altogether, inflammatory-cell infiltration, mucous-cell metaplasia, and BHR were suppressed when CD3⁺ lymphocytes were depleted, and were partially inhibited by anti-CD4 antibodies. Finally, vinblastine abrogated granulocyte recruitment into the lungs, but lymphocyte numbers and mucous-cell metaplasia remained unaffected. Mucus secretion is thus independent of the presence of granulocytes, but correlates with the presence of lymphocytes.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals, Immunizations, and OVA-Challenge Protocols

BP2 and BALB/c mice aged 6 to 8 wk (Centre d'Elevage R. Janvier, Le Genest Saint-Isle, France) were immunized subcutaneously at Day 0 and Day 7 with 100 µg OVA (Immunobiologicals, Lisle, IL) in the presence of 1.6 mg alum (Merck, Darmstadt, Germany) and diluted in a 0.4 ml saline solution. Intranasal OVA challenge was performed 1 wk after the second immunization, namely, at Day 14, with 10 µg OVA diluted in 50 µl of alum-free saline solution or with saline solution as a control under ether anesthesia. OVA challenge was performed once in one group and four times (twice a day) in another group. Six to 12 mice were taken at 1, 3 , 6, 24, 48, and 72 h, and also at 6 and 11 d for the evaluation of pulmonary functions, and for histologic examination of the lungs. For electron microscopy studies of cells from the bronchoalveolar lavage fluid (BALF), mice were challenged four times to yield an optimum number of eosinophils (5).

Evaluation of BHR

Unrestrained, conscious mice were placed in a plethysmographic chamber (Buxco Electronics, Sharon, CT), and respiratory parameters were measured before and after an aerosol of methacholine (Sigma-Aldrich, Stonheim, Germany) delivered for 20 s at 3 × 10² M in the aerosolator. The resistance was expressed as enhanced pause (Penh), calculated as: {(Expiratory time)/40% of (Relaxation time)  1} × {(Peak expiratory flow)/(Peak inspiratory flow)} × 0.67, according to the manufacturer's recommendations. Every 20 s an average value of Penh was recorded. For the graphic representation, each value was expressed for every minute. Accordingly, each point represents the average of three values.

Sampling for Cytology and Histology

Mice were deeply anesthetized with urethane (intraperitoneally, 15 mg/10 g body wt), and the abdominal cavity was opened. Blood samples for serum were taken when needed; otherwise, animals were bled and killed by cutting of the large abdominal vessels. BALF was collected by cannulating the trachea and washing the lung with a total of 4 ml saline solution (8 × 0.5 ml each). After cell counting, cytospin was prepared with a standard apparatus. Lungs were then inflated with 1 ml of 50% Optimum Cutting Temperature compound (Sakura Finetek, Torrance, CA) solution through the cannulated trachea. The left lobe was placed in a tube and immediately frozen in liquid nitrogen, and the remaining right lobes were fixed in 10% formaldehyde.

Formaldehyde-fixed samples were embedded in paraffin for standard histologic examination. Longitudinal and transversal sections of the major intrapulmonary bronchioles of 4-µm thickness were prepared and stained with hematoxylin and eosin and with periodic acid-Schiff and Alcian blue (PAS.Ab) reaction to study mucus-containing cells. PAS stains neutral mucin, and Ab stains acidic mucin. The thicknesses of the bronchial epithelium and the smooth muscle of the transversely sectioned preparations were measured using a computerized image analyzer with Optimas software (Bioscan, Inc., Edmons, WA). In addition to the lung, secondary lymphoid organs, spleen, and mesenteric lymph nodes were also examined histologically when needed.

Uniform 6-µm cryostat sections containing longitudinal sections of the major bronchiole were prepared on Superfrost-plus slides (Menzel-Glasser, Braunschweig, Germany), rolled on cellophane paper, and stored frozen at 20°C until further use. Slides with cytospin preparations were fixed and stored in the same manner.

Histochemical Staining

Eosinophils were selectively stained for eosinophil peroxidase (EPO) on cryostat sections as described (8). Briefly, acetone-fixed sections were incubated for 10 min with a phosphate buffer, pH 7.4, containing diaminobenzidine 10 mg/13.3 ml, sodium cyanide, and hydrogen peroxide; washed; counterstained with Harris hematoxylin; and mounted with aquamount (Gurr; BDH Laboratories, Poole, UK). All other chemicals were from Sigma-Aldrich. Mast cells were stained and identified using a histochemical chloroacetate esterase stain for nonspecific esterase (9). Briefly, acetone-fixed sections were incubated with a solution containing the substrate naphthol AS-D acetate (5 mg dissolved in 0.5 ml dimethylformamide), 25 ml distilled water, 25 ml Tris buffer (pH 7.1), and 30 mg of Fast blue RR salt (all from Sigma-Aldrich).

Immunohistochemistry

CD4⁺ and CD8⁺ lymphocytes and IgE-bearing cells were identified using immunohistochemical techniques. Anti-CD4 and anti-CD8 antibodies were purchased from Caltag (San Francisco, CA). Anti-IgE antibody EM95-3 (10) was kindly provided by Dr. Y. Chvatchko (Serono Pharmaceutical Research Institute, Geneva, Switzerland). Briefly, after covering the nonspecific site with 10% rabbit serum, sections were incubated for 1 h with the first antibody, a rat antimouse antibody (all antibodies were of rat origin), followed by washing in Tris-saline buffer (pH 7.6) and then incubation for 45 min with a second biotinylated rabbit antirat antibody (Dako A/S, Glostrup, Denmark). Washing was followed by another 45-min incubation with alkaline phosphatase-labeled streptavidin (Dako A/S), followed by another washing. The reaction was revealed with an alkaline phosphatase substrate solution containing naphtol AS-MX phosphate, Fast red, and levamisole in Tris buffer (pH 8.2). Unless otherwise stated, all chemicals were from Sigma-Aldrich. Double immunohistochemical staining for IgE and for B lymphocytes or for neutrophils was performed consecutively using the same method. This was done to make a distinction between IgE-bearing PMN, neutrophils, and IgE-secreting and/or -bearing B lymphocytes. Antineutrophil and anti-B lymphocyte B220 antibodies were kindly provided by Dr. Geneviève Milon (Unité d'Immunophysiologie Cellulaire, Institut Pasteur, Paris, France).

Electron Microscopy

Lungs from anesthetized mice were perfused first with phosphate-buffered saline (PBS) solution to remove blood, followed by a fixative of 1% glutaraldehyde in 0.1 M cacodylate buffer at pH 7.4 (11). Lungs were slowly instilled intratracheally with the same fixative and then removed. The major intrapulmonary bronchiole was dissected out into four cubes of 1 mm each for each mouse and was left in a fixative containing 2.5% glutaraldehyde overnight. The next day, tissues were washed, postfixed in 2% osmium tetraoxide, and processed as described (11).

In another group of mice, BALF and blood eosinophils from OVA-challenged mice were isolated by Percoll centrifugation as described (12). Because saline-challenged mice did not contain sufficient eosinophil numbers to serve as controls, blood eosinophils were collected from transgenic mice for interleukin (IL)-5, which do not become hyperreactive after a single antigenic provocation (13). Eosinophils were resuspended in PBS solution, washed, recentrifuged, and fixed in glutaraldehyde at 2.5% for 1 h at 4°C. After rinsing in phosphate buffer, they were centrifuged and resuspended with two drops of 8% bovine serum albumin and agglutinated by a brief incubation in two drops of 25% glutaraldehyde as described (11). The pellets were washed and postfixed in 2% osmium tetraoxide for 1 h. They were then processed as described for fixed tissues (11).

Th2 Cytokine Measurements

IL-4 and IL-5 were evaluated in BALF and in serum at different times after challenge. IL-4 was measured with an enzyme-linked immunosorbent assay technique (Valbiotech, Paris, France). Briefly, 96-well plates were coated with 2 µg/ml of rat antimouse IL-4 (BVD4-1D11; Pharmingen, San Diego, CA), to which were added dilutions of recombinant (rc)IL-4 standard (7 to 1,000 pg/ml) or of the sample, followed by a biotinylated rat anti-IL-4 antibody (BVD6-24G2; Pharmingen) at 0.5 mg/ml. The reaction was revealed using a substrate containing o-phenylenediamine dihydrochloride in phosphate buffer and read at 490 nm.

IL-5 was measured using the immunometric assay as described (5). Briefly, 96-well plates were coated with 10 µg/ml of rat antimouse IL-5 (TRFK-4) to which were added an rcIL-5 standard (7.8 to 1,000 pg/ml) or the sample, followed by an AchE-labeled rat antimouse IL-5 antibody (TRFK-5) at 10 Ellman U/ml. Absorbance was read at 405 nm. The lower limit of detection of both assays was approximately 5 pg/ml of the sample. Both TRFK-4 and TRFK-5 were purified from ascitic fluids (cloned hybridomas provided by Dr. P. Minoprio, Institut Pasteur).

In Vivo Treatments

Hamster antimouse CD3 monoclonal antibody (mAb) 145-2C11 (14) (kindly provided by Dr. L. Majlessi, Unité d'Immunophysiologie Moléculaire, Institut Pasteur) or hamster IgG as a control was given once intravenously at a dose of 200 µg per animal. The next day mice were challenged intranasally with OVA. In another group of animals, OVA challenge was performed 1 wk after the treatment with anti-CD3. Administration of whole anti-CD3 antibody causes high morbidity and mortality from immunodepression and from excessive secretion of IL-2 and interferon- in mice (15). For this reason our mice were kept in rigid isolation in specific pathogen-free conditions to prevent secondary infection.

To deplete CD4 lymphocytes, rat antimouse CD4 mAb GK 1.5 (kindly provided by Dr. G. Bordenave, Unité d'Immunophysiologie Moléculaire, Institut Pasteur) or rat IgG as a control was given at a dose of 300 µg/mouse/d during 3 consecutive days, a week before challenge, insuring a maximum depletion of CD4 cells. In a preliminary test, one injection of 1 mg of anti-CD4 antibody failed to deplete lymphocytes and to inhibit eosinophil infiltration in BP2 mice, but depleted CD4 lymphocytes and inhibited eosinophil infiltration into the lungs in BALB/c mice (data not shown). We did not identify the cause of the relative resistance of BP2 mice to anti-CD4-dependent depletion. Depletion or absence of depletion of lymphocytes was verified in the spleen and mesenteric lymph nodes by fluorescence-activated cell sorter (FACS) analysis using labeled antibodies as described (16), and by histologic examination of the lymphoid organs.

To study whether granulocyte depletion might affect inflammatory-cell recruitment to the lungs and mucus production by bronchial epithelial cells and BHR, vinblastine (Lilly France, Saint Cloud, France) was administered intravenously at 5 mg/kg 72 h before the antigenic challenge. Because mucus production by epithelial cells is best quantified histologically 72 h after antigenic provocation, evaluation was performed at this time point.

Statistics

All results are presented as means ± SEM. Significance levels were calculated using the nonparametric Mann-Whitney U test, and P < 0.05 was taken as significant difference between data (17).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Eosinophil and Lymphocyte Recruitment Is More Intense in BP2 Than in BALB/c Mice after a Single OVA Challenge

Saline-challenged control mice did not show eosinophil recruitment in the airways. In OVA-challenged mice, the kinetics of eosinophil recruitment to the airways were similar in BP2 and BALB/c mice, but the intensity differed: BP2 mice recruited > 2-fold more cells (measured at 24 h after challenge) than BALB/c mice (Figures 1a, 1c, 2a, and 2b). These differences were particularly clear after a single provocation, but persisted after four challenges (Figures 1b and 1d).

View larger version (30K): [in this window] [in a new window]   Figure 1.   Time-dependent inflammatory cell recruitment and epithelial mucus secretion in BP2 (a and b) and BALB/c (c and d) mice after one (a and c) or four (b and d) intranasal OVA challenges. Mice challenged with saline as antigen control did not show inflammatory cell infiltration or mucus secretion. Eosinophils (open circles) and CD4+ (closed triangles) and CD8+ (open triangles) lymphocytes were counted as cells per millimeter length in the major intrapulmonary bronchi. IgE+ PMN (open squares) were counted as cells per square millimeter of the lung section. Mucus secretion (closed circles) was evaluated as the proportion of mucus-containing cells over total epithelial cells of the major intrapulmonary bronchi. Standard errors (not shown for simplicity) were below 12%.

Three hours after challenge, eosinophils marginated onto the endothelial surface. By 6 h, they started to infiltrate the peribronchial and perivascular connective tissues. Their numbers increased by 24 h, to peak by 48 to 72 h (Figure 1a), a time when focal eosinophilic alveolitis was also seen. Peribronchial and perivascular infiltration increased and alveolar eosinophilia decreased with time. Tissue eosinophilia decreased slightly but still persisted at Day 11 (not shown). Four OVA challenges increased the numbers of eosinophils and T lymphocytes infiltrating the lung and prolonged the persistence of IgE-bearing PMN numbers. As observed using the EPO stain, no extracellular eosinophil granules were present at any time.

Saline-challenged control mice did not show lymphocyte infiltration in the airways. T lymphocytes infiltrated the peribronchial and perivascular tissues as early as 6 h after challenge and their numbers increased with time (Figure 1a). At 24 h and onward, B lymphocytes were the most numerous cells recruited, followed by CD4⁺ T lymphocytes (not shown). Lymphoid follicles containing germinal centers were clearly noted from 6 h on after challenge, particularly in BP2 mice. The intensity of lymphocyte recruitment increased when mice were challenged four times.

In electron microscopic preparations, numerous eosinophils were observed in the submucosa of the airways. Nevertheless, in both strains of mice no eosinophils with morphologic aspect of activation were seen, and neither rarefaction of the crystalloid or of the matrix of the granules (a morphologic counterpart of activation) nor free granules were observed in the submucosa or in the epithelium (not shown). Eosinophils freshly isolated from the BALF and from blood did not show altered granules or other modifications, such as the increase in the number of lipid bodies described for human eosinophils from allergic lesions (Figure 3c). In one sample from an OVA-challenged BP2 mouse, free eosinophils were noted in the bronchial lumen next to the surface of the epithelium, displaying characteristics of degeneration such as swollen and disrupted mitochondria and disintegrating granules coming out of the cell (Figure 3d).

View larger version (173K): [in this window] [in a new window]   Figure 3.   Electron micrographs of tissues and of cells isolated from the BALF of OVA-challenged mice and their controls. (a) Bronchiolar epithelium of saline-challenged control mouse showing alternating ciliated cells (Ci) and nonciliated or Clara cells (C). Bar = 6 µm. (b) Bronchiolar epithelium from a mouse 48 h after OVA challenge showing alternating ciliated (Ci) and nonciliated (C) cells containing mucus granules (white arrow) in their cytoplasm. Bar = 6 µm. (c) Percoll-isolated eosinophils from the BALF of OVA-challenged BP2 mice show a normal structure. Mature granules (arrow) and immature granules (arrowhead) are shown. Note the absence of piecemeal degranulation and lipid bodies. Bar = 2 µm. (d) Morphology of eosinophils in situ in the bronchiolar lumen. Degenerating eosinophils (E) in the bronchiolar lumen with disintegrating granules (arrows), mitochondrial ballooning, and a neutrophil (N). This was prepared from a cluster of eosinophils such as those observed in the bronchiolar lumen (Figure 2a, arrowhead). Bar = 3.5 µm. (e) Mast cell from saline-challenged control mouse with dense and normally appearing granules. Bar = 2 µm. ( f ) Mast cell from OVA-challenged mouse with rarefied granules, a morphologic aspect of degranulation. Bar = 2 µm.

View larger version (116K): [in this window] [in a new window]   Figure 2.   Characterization of inflammatory cells and epithelial changes in the airways of OVA-challenged mice. (a and b) OVA-induced recruitment of eosinophils (brown) into the bronchiolar wall following a single OVA challenge is more intense in BP2 (a) than in BALB/c (b) mice. Note the presence of the cluster of eosinophils in the bronchiolar lumen in (a) (arrowhead). Histochemical staining for EPO on frozen lung sections was revealed with diaminobenzidine substrate. Bar = 160 µm. (c) Localization and distribution of mast cells (blue-stained cells) in the mouse lung is limited to the first 2 to 3 mm of the major intrapulmonary bronchiole. Staining of nonspecific esterase for proteases on frozen lung section was revealed with fast blue substrate. Bar = 160 µm. (d) OVA-induced recruitment of IgE-positive PMN (red-stained cells). Slides of frozen lung sections were stained with the indirect labeled-antibody staining method, revealed with fast red and counterstained with hematoxylin. Bar = 30 µm. (e) IgE-positive PMN (red-stained cells) in cytospin preparation of the BALF from OVA-challenged BP2 mice. To show that these cells are not B lymphocytes, double staining for B220, a B lymphocyte marker (blue-stained cells), was performed. Slides were stained with the indirect labeled-antibody staining method, revealed with fast red (IgE) and fast blue (B220), and counterstained lightly with Metanil Yellow. Bar = 30 µm. ( f to k) Time-dependent changes in bronchiolar epithelium of mice after a single OVA challenge, showing a saline-challenged control ( f ), and 6 (g), 24 (h), 48 (i), and 72 h ( j and k) after OVA challenge. PAS-positive mucus granules developed in the apical pole of the cytoplasm (arrow) and increased with time. Mucus was also excreted to the lumen (arrowhead in j). White arrows in (k) show eosinophils infiltrating the submucosa. PAS reaction (purple) counterstained with hematoxylin. Bars: f to j, 10 µm; k, 40 µm. (l to n) Vinblastine abrogates eosinophil (arrowhead) recruitment into the lung but not mucous-cell metaplasia (arrow). (l) Saline-treated and saline-challenged control mice. (m) Saline-treated and OVA-challenged mice with eosinophil infiltrating the bronchiolar wall (arrowhead). (n) Vinblastine-treated and OVA-challenged mice. Combined EPO (red-brown) and PAS.Ab reaction (blue) on frozen lung sections without counterstaining. EPO was revealed with diaminobenzidine substrate. Bar = 160 µm.

IgE-Bearing PMN Are Recruited into the Lungs within the First Hours after OVA Challenge

In immunized saline-challenged controls, faint IgE staining was observed only on a few mast cells and on lymphocytes, and very strong staining was noted on PMN circulating in the capillaries. In OVA-challenged mice, the intensity of mast-cell staining and numbers of IgE-bearing PMN increased (Figure 2d), whereas lymphocytes were still very faintly stained, even though their number increased. The first cells to infiltrate the lungs 3 to 6 h after challenge were IgE-bearing PMN. From double immunohistochemical or combined immunohistochemical and histochemical stainings, these IgE-bearing PMN were negative for EPO, for the antineutrophil antibody NIMP-R14, for the B-lymphocyte marker B220 (Figure 2e), and for nonspecific esterase (18), as seen both in histologic preparations and in cytospin preparations of BALF. This, and the morphology of the cells in the cytospin (Figure 2e), allowed us to conclude that these IgE-bearing PMN are basophils. The kinetics of infiltration of these IgE-bearing PMN after a single OVA challenge are presented in Figure 1. The absence of a significant increase in the recruitment of IgE-bearing PMN after a single OVA challenge in BALB/c mice, which are not turned hyperreactive using the single antigenic provocation protocol, is noteworthy. The number of IgE-bearing PMN decreased rapidly to basal values by 72 h in BP2 mice challenged once (Figure 1a), but after four OVA challenges persisted for longer (Figure 1b).

No Variation in Mast Cell Numbers after OVA Challenge in Mice

Mast cells were scarce and localized on the first 2 to 3 mm of the longitudinally sectioned major intrapulmonary bronchioles of mice (Figure 2c). Slightly more mast cells were seen in the lungs of BALB/c than in those of BP2 mice (5.1 ± 3.8/section of the left lobe for BP2, and 6.5 ± 3.5 for BALB/c, P < 0.02). In general, mast cells were more concentrated in the connective tissue between the large vessels and the major bronchiole at the level of the hilus of each lobe, and were rarely seen deep in the lung parenchyma. No significant differences in mast cell numbers were noted between OVA-challenged mice and their respective saline-challenged control mice.

Only a few mast cells, with piecemeal degranulation, were seen in the electron microscopic preparations of bronchial wall in both BP2 and BALB/c mice challenged with OVA (Figure 3f).

Epithelial Lesions after OVA Challenge Consist Essentially of Mucous-Cell Metaplasia

OVA-challenged BP2 and BALB/c mice showed a significant increase in the height of the bronchial epithelial lining and in the thickness of the smooth muscle at 72 h after challenge (Figure 4), when compared with saline-challenged controls. The other major epithelial change observed was the appearance of PAS.Ab-positive mucus- secreting cells in the epithelium. Approximately 50% of the cells of the major bronchioles of mice are Clara cells, very few of which showed PAS.Ab-positive secretory granules in the apical part of the cytoplasm. In OVA-challenged mice, secretory granules formed by both acidic (Alcian blue-stainable) and neutral (PAS-stainable) mucus started to appear in some Clara cells by 6 h. At 24 h, the number of granules increased substantially, and by 48 to 72 h after challenge almost all Clara cells (i.e., 50% of the total epithelial cell numbers) were transformed into mucus-secreting cells (Figures 2f to 2k). This was confirmed in electron microscopic preparations of the bronchiolar epithelium (Figure 3b). When individual cells were analyzed, the mucus granules started to appear in the apical pole of the cells and gradually increased in numbers in the mid- and basal parts. Mucus was also secreted in the bronchial lumen (Figure 2j). Saline-challenged and nonimmunized OVA-challenged control mice did not show stainable mucus in the bronchiolar epithelium. No thickening of the basal membrane was observed even in mice repeatedly challenged for several days (data not shown).

View larger version (21K): [in this window] [in a new window]   Figure 4.   Box plot for the height of the epithelium (a) and cross-sectional width (b) of smooth muscle of the major bronchiole of BP2 mice measured 72 h after one intranasal OVA challenge. The left lung lobe was sectioned transversally at around 4 mm from the entry of the bronchi. Measurement was performed at high-power magnification using an image analyzer. *P < 0.05.

IL-4 Is Detected Only in the BALF

IL-4 and IL-5 were detected in the BALF only at 24 h after challenge. IL-5 levels in the serum increased starting 6 h after challenge (Figure 5).

View larger version (21K): [in this window] [in a new window]   Figure 5.   IL-4 and IL-5 levels in the serum (open columns) and in the BALF (closed columns) in BP2 mice at different time points after a single intranasal OVA challenge.

BP2 Mice Develop OVA-Induced BHR after a Single Challenge

BP2 mice studied 1 h after the OVA challenge displayed BHR in response to methacholine inhalation, which was over at 3 h, rose at 6 h, and persisted for up to 6 d (Figure 6). Similarly challenged BALB/c mice showed no BHR, except for a small rise at 72 h (data not shown).

View larger version (17K): [in this window] [in a new window]   Figure 6.   Time-dependent BHR in BP2 mice. An average value of the Penh was calculated every minute before and after the methacholine aerosol inhalation. BP2 mice showed early and late antigen reponses similar to those described in human asthma. Open squares: immunized mice challenged with saline solution; closed squares: immunized mice challenged with OVA.

Inflammatory-Cell Recruitment and Mucous-Cell Metaplasia Are Independent of Each Other in Experimental Allergic Inflammation

We observed no mortality in mice receiving anti-CD3 antibody, but they showed roughened coats and decreased weight. T lymphocytes from mice treated with anti-CD3 were totally, and with the anti-CD4 were partially, depleted in spleens (FACS data, not shown). The anti-CD3 antibody suppressed the OVA-dependent recruitment of eosinophils, lymphocytes, and IgE-bearing PMN (Figure 7), as well as mucus production and BHR in response to methacholine (Figure 8a). When the animals were challenged 1 wk after anti-CD3 treatment, infiltration of inflammatory cells was only partly recovered (Figure 7e), and inhibition of BHR persisted (Figure 8b).

View larger version (17K): [in this window] [in a new window]   Figure 7.   Effect of CD3+ lymphocyte depletion on inflammatory cell recruitment into the airways and on epithelial mucus secretion in OVA-challenged BP2 mice. (a) Immunized mice treated with nonspecific rat antimouse IgG and challenged with saline 24 h later. (b) Immunized mice depleted of CD3+ lymphocytes and challenged with saline 24 h later. (c) Immunized mice treated with rat IgG and challenged with OVA 24 h later. (d) Immunized mice depleted of CD3+ lymphocytes and challenged with OVA 24 h later. (e) Immunized mice depleted of CD3+ lymphocytes and challenged with OVA 1 wk later. Because the distribution of the different cell types was variable in the lung tissue, cell counts appear as number of cells per millimeter length of epithelium for eosinophils and CD4+ and CD8+ lymphocytes; number of cells per square millimeter for IgE-bearing cells; and percentage of mucus-containing cells over total epithelial cells of the major intrapulmonary bronchioli. *P < 0.05.

View larger version (30K): [in this window] [in a new window]   Figure 8.   Effects of CD3+ (a and b) and CD4+ (c and d) lymphocyte depletion on BHR. (a) Results of BHR reading 5 d after anti-CD3 treatment and 72 h after challenges. (b) BHR reading 10 d after anti-CD3 treatment and 24 h after challenges. (c and d) BHR readings 24 and 72 h, respectively, after challenge. Open squares: immunized mice challenged with saline solution; closed squares: immunized mice challenged with OVA; closed circles: immunized and challenged mice that received the depleting antibodies.

The anti-CD4 antibody decreased significantly but did not suppress eosinophil recruitment and mucus production (Figure 9). In this case, the control group treated with multiple injections of rat IgG also showed a reduced number of eosinophils, thus complicating the interpretation. BHR was inhibited by anti-CD4 at 72 h after challenge but not before (Figures 8c and 8d). The infiltration of lymphocytes and the formation of lymphoid follicles were completely inhibited by the anti-CD3 but not by the anti-CD4 antibody (data not shown).

View larger version (38K): [in this window] [in a new window]   Figure 9.   Effect of anti-CD4 (aCD4) antibody treatment on eosinophil recruitment and epithelial mucus production in BP2 mice. (a) Mice received one injection of 1 mg anti-CD4 antibody and parameters were measured 72 h after OVA challenge. (b) Mice received 300 µg of anti-CD4 antibody per day for 3 d and parameters were measured 24 h after OVA challenge. (c) Mice received 300 µg of anti-CD4 antibody per day for 3 d and parameters were measured 72 h after OVA challenge. *P < 0.05.

The T-lymphocyte areas of secondary lymphoid organs (spleen and lymph nodes) were depleted of lymphocytes by the anti-CD3 but less so by the anti-CD4 antibody. In the latter case, only focal apoptotic lymphocyte aggregates were seen. Lymphoid organs and the bone marrow of these mice examined 1 wk after anti-CD3 administration showed an intense granulocyte proliferation, particularly of neutrophils, as if the bone marrow was recovering from severe depression.

Vinblastine, given at 72 h before challenge, completely suppressed the lung recruitment of eosinophils and of IgE-bearing PMN. It also reduced the number of intravascular neutrophils detected in lung sections to below basal levels (Figure 10). By contrast, vinblastine failed to affect the number of CD4 lymphocytes infiltrating the submucosa. Most importantly, it failed to modify the antigen-induced mucus-cell metaplasia (Figure 3n). Despite complete abrogation of granulocyte recruitment in the lungs, the proportion of mucus-containing cells remained equal to that in the untreated, OVA-challenged BP2 mice.

View larger version (40K): [in this window] [in a new window]   Figure 10.   Effects of vinblastine on OVA-induced recruitment of inflammatory cells and mucous-cell metaplasia in the lungs of mice. Open column: immunized mice challenged intranasally with saline solution; striped columns: immunized mice challenged intranasally with OVA; closed columns: mice that received 5 mg/kg of vinblastine 72 h before the intranasal OVA challenge. Results were read 72 h after OVA challenge. (a) Eosinophils, counted as number of cells per millimeter length of the major bronchiole. (b) IgE-bearing PMN, counted as number of cells per square millimeter of lung section. (c) Neutrophils, counted as number of cells per square millimeter of lung section. (d) CD4+ lymphocytes, counted as number of cells per millimeter length of the major bronchiole. (e) Mucus-containing epithelial cells in percentage of total epithelial cells of the major bronchiole. *P < 0.05.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this study we demonstrate that single antigenic provocations in BP2 mice induce BHR, which lasts up to 6 d, in parallel to eosinophil and lymphocyte recruitment into the airways. As compared with BALB/c mice, BP2 mice showed a more intense infiltration of eosinophils, lymphocytes, and IgE-bearing PMN in lungs. Eosinophils were located more frequently in the epithelium of BP2 as compared with BALB/c mice, but no associated epithelial sloughing was observed even in multichallenged mice, contrary to what is observed in asthmatic human patients (19). BHR was already noted in BP2 mice 1 h after challenge with OVA, well before the arrival of eosinophils to the lungs. This may relate to the immediate response to antigen, which is attributable to the IgE-dependent release of inflammatory mediators and was confirmed by the identification of partially degranulated mast cells. BHR was over in 3 h, returned at 6 h to last for 6 d, and normalized later (Figure 6). By contrast, BALB/c mice showed a very mild and nonsignificant BHR at 72 h after challenge (not shown). The dual response on BP2 mice recalls the similar kinetics reported for asthmatic human patients (20).

Repeated OVA challenges (twice a day for 2 d) in both strains of mice increased the number of eosinophils, lymphocytes, and IgE-bearing PMN infiltrating the lungs, but the BALB/c mice remained nonhyperreactive. In other studies (2) including those of our laboratory (21), multiple antigen sensitizations and challenges in BALB/c mice succeeded in inducing BHR, suggesting that more eosinophils and IgE-bearing PMN are required in the lungs to render BALB/c mice hyperreactive. Cell numbers and the severity of the inflammatory infiltrate, and, as a consequence, the amount of mediators made available, may thus determine whether BHR will be induced.

Antigen-induced eosinophil infiltration in the lungs was completely inhibited in mice depleted of CD3⁺ lymphocytes and partially in those receiving the anti-CD4 antibody. Several studies demonstrated that CD4⁺ lymphocytes, and particularly those of Th2 subtype, contribute to the infiltration of eosinophils in the lungs by their secretion of eosinophilic cytokines (22). Selective depletion of CD4⁺ lymphocytes by the anti-CD4 antibody significantly decreased this infiltration (23). In our study, significant inhibition of eosinophilia and BHR was obtained after three repeated injections of the antibody, but the injection of rat IgG into control mice itself reduced eosinophilia when compared with saline-injected controls. By contrast, administration of a single dose of 200 µg anti-CD3 depleted lymphocytes and completely inhibited OVA-induced eosinophilia, IgE-bearing PMN, and BHR. The mechanism of inhibition of eosinophilia by anti-CD3 can be explained by the suppression of T lymphocyte-secreted eosinotactic cytokines. Indeed, MacLean and colleagaues (24) demonstrated that anti-CD3 injection into mice inhibited the production of eotaxin and suggested that this was the cause of inhibition of the antigen-induced eosinophilia.

In our studies, we confirmed our previous observation that the eosinophils infiltrating the airways of the OVA-challenged mice are not degranulated (5), contrary to what has been described for human allergic tissues (25). Furthermore, Percoll-collected eosinophils from BALF also were not degranulated. However, while confirming this observation in the tissues examined, we noted granule exocytosis from degenerating eosinophils in the bronchial lumen (Figure 3d). Recently, Persson and Erjefalt (26) suggested that eosinophil lysis and release of free granules may be more important than intact cell degranulation for the release of mediators. Even though we did not observe these alterations in the bronchial submucosa, the presence of degenerating eosinophils and of granule exocytosis in the bronchial lumen supports this hypothesis. The local microenvironment may favor degranulation because the mucus layer of the bronchial lumen is rich in immunoglobulins, especially IgA, that have been shown to degranulate eosinophils in vitro (27). Another observation that needs further investigation is the absence of visible lipid bodies either in tissue or in BALF eosinophils. The increase of the numbers and size of lipid bodies in eosinophils is considered a morphologic sign of activation and of the increased synthesis of lipid mediators (28, 29). In comparison with human eosinophils, the absence of lipid bodies in mice in this study implies a possible limitation in the synthesis of lipid mediators of inflammation.

IL-4 and IL-5 are important cytokines for allergy because they promote IgE isotype switch and eosinophilia, respectively (22). Both were detected in the BALF of OVA-challenged mice, and IL-5 was found in the serum as well. The fact that IL-4 was not found in the serum may result from an exclusive local production, for instance by the basophils (30). Staining of cell-bound IgE allowed us to identify mast cells as well as EPO-negative, thus noneosinophilic, non-B lymphocyte IgE-bearing PMN. After confirming their polymorphonuclear morphology in cytospin preparations of BALF from OVA-challenged mice, these cells were identified as basophils by exclusion criteria. As indicated above, these IgE-bearing PMN were recruited to the lungs shortly after the instillation of antigen and, accordingly, well before the massive arrival of eosinophils and lymphocytes. This may have important implications. The arrival of IgE-bearing PMN implies the potential release of cytokines such as IL-4 that play a role in late asthmatic responses. Several observations implicate IgE in the initiation of late responses (31, 32). We (unpublished observations) and others (31, 33) have demonstrated that treating mice with an anti-IgE antibody within the first 24 h before OVA challenge significantly decreases eosinophil recruitment to the lungs. We also demonstrated that this antibody abrogates the antigen-dependent IL-4 secretion into the BALF, highlighting the potential role of IgE in late reactions, through its action on high-affinity IgE receptor-bearing cells such as the IgE-bearing PMN observed in our study. However, recent investigations have shown that IgE-deficient mice are also capable of mounting early- and late-phase reactions to antigen, possibly via the alternative role of IgG₁ in absence of IgE (34).

The absence of modifications in mast-cell numbers after OVA challenge does not imply that they are not activated. Indeed, mast cells were found to have more stainable IgE after OVA challenge in both mouse strains, some of the cells undergoing piecemeal degranulation, as described (35). No increase in mast-cell number has been described in tissue from asthmatic human patients (36), but others have noted an increase in tissue (37) and in BALF (38). In our studies, no mast cells, but a significant number of IgE-bearing PMN, were seen in BALF from OVA- and not from saline-challenged control mice. This indicates that IgE-bearing PMN, notably basophils, may play an important role as a source of biologically active substances in mice, in addition to the scarce mast cells present in the mouse airways.

The key difference between BP2 and BALB/c mice after a single OVA challenge lies in the difference in their degrees of pulmonary eosinophilia and of recruitment of IgE-bearing PMN. By contrast, both strains developed mucus-cell metaplasia. Mucus hypersecretion characterizes allergic and nonallergic chronic bronchitis (39). In humans, this hypersecretion is due to goblet-cell hyperplasia and probably metaplastic changes of other epithelial cells (42). In this study, we demonstrate that mucus secretion in mice follows metaplastic changes of the nonciliated Clara cells. Normal mice have very few mucus-secreting cells in the tracheobronchial epithelium (43). Following OVA challenge, mucus granules started to form in the apical part of the cytoplasm of Clara cells and gradually increased in number to fill the entire supranuclear cytoplasm by 48 to 72 h. Antigen-dependent metaplastic changes of bronchiolar Clara cells in humans, therefore, may aggravate the mucus secretion and plug small airways, contributing further to bronchoconstriction.

Mucus secretion is favored and induced by several eosinophil and mast-cell products (44). Henderson and associates (45) have shown that leukotriene inhibitors suppress antigen-induced eosinophil infiltration and mucus secretion in the murine lung. In our study, we show that vinblastine treatment abrogates the recruitment of eosinophils and IgE-bearing PMN and significantly reduces that of neutrophils, but does not affect antigen-induced mucus-cell metaplasia of the bronchial epithelium. Therefore, mucous-cell metaplasia of the respiratory epithelium is not directly related to the presence of granulocytes and their products. By contrast, anti-CD3 antibody administered to mice before challenge not only depleted T lymphocytes but also completely inhibited antigen-induced eosinophilia, IgE-bearing PMN recruitment, and mucous-cell metaplasia. Considering these findings together, it can safely be concluded that mucus production is not related to the presence of granulocytes but rather to that of lymphocytes in this model because anti-CD3 antibodies, which inhibited both granulocyte and lymphocyte recruitment into the lungs, also inhibited mucous-cell metaplasia, whereas vinblastine inhibited only granulocyte infitration in the lungs. A recent study (46) showed that, contrary to previous statements (47), IL-4 is not directly implicated with mucus secretion because transfer of OVA-specific T lymphocytes from IL-4- deficient mice induced mucus production in recipient mice.

Depletion of CD4⁺ and CD3⁺ lymphocytes inhibited the development of BHR and mucous-cell metaplasia. Vinblastine also inhibited BHR development (unpublished observation), but it did not affect mucous-cell metaplasia. Hyperreactive BP2 mice and normorreactive BALB/c mice showed equivalent mucous-cell metaplasia. Together, these findings suggest that expression of BHR is independent from this epithelial modification, even though the amount of mucus secreted might contribute to the physical reduction of the bronchial lumen.

Vinblastine was used in our study to inhibit granulocyte recruitment; however, in some cases it has been used to treat the hypereosinophilic syndrome in humans (48, 49). The exact mechanism of inhibition of granulocyte recruitment by vinblastine is not precisely known. Vinblastine induces the formation of paracrystalline aggregates of tubulin, leading to microtubule depolymerization and consequently cell-cycle arrest (50). Inhibition of granulocytes probably results from mitosis arrest leading to the temporary absence of the replenishment of the peripheral granulocyte pool by the bone marrow (51). In addition, the depolymerization of intermediate filaments may compromise cell migration, as has been described for skin Langerhans cells (52). The absence of inhibition of mucus production by vinblastine further supplements the observation that mucus hyperproduction in this model occurs through metaplastic changes of Clara cells rather than through mucous-cell mitosis and hyperplasia.

Additionally, our study shows that the size, topography, and cell composition of the murine airway epithelium and the low number of mast cells resemble the histology of the lower airways of humans and of big mammals. Postmortem and biopsy studies from asthmatic human patients reveal epithelial sloughing, sub-basal fibrosis, and degranulating mast cells and eosinophils, a morphologic marker of activation (53). These lesions were not observed in murine lungs, despite the recruitment of similar inflammatory cells. Murine models of asthma reflect the events taking place in the lower human airways rather than those occurring in the larger bronchi.

In conclusion, we demonstrated that IgE-bearing PMN are recruited to the airways early during the allergic reaction and that their numbers are higher in hyperreactive BP2 mice than in nonhyperreactive BALB/c. This underlines the importance of IgE-stimulated cells in the development of late antigen reponses. Mucus hypersecretion is antigen-related and T-lymphocyte-related and was not affected even when eosinophil, IgE-bearing PMN, and neutrophil recruitment was abrogated by vinblastine treatment. Mucus hypersecretion occurred through metaplastic changes of Clara cells in the bronchi and bronchioles, which are not directly related to eosinophils or to their products, as has been suggested previously (54). The implication of mucus for the development of BHR in asthma is frequently stressed, but no direct study on its role is available. Murine models such as ours can contribute to the understanding of the functional mechanisms of BHR, of the mechanisms of epithelial lesions, and of the role of mucus, particularly in the lower airways.