Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

A marked increase in the numbers of goblet cells in airway epithelium occurs in asthma and chronic bronchitis (1, 2). The resulting overproduction of mucin contributes to airway obstruction (3) and makes it difficult for patients to clear their chests. We have used a murine model of atopic asthma to investigate various aspects of goblet cell hyperplasia (GCH), with the aim of eventually defining its role in the disease and its response to therapy.

In previously published work we showed that a marked and extensive GCH develops in the airways of ovalbumin (OA)-sensitized mice following repeated intratracheal challenge with the allergen (4). This dose-related increase in goblet cell numbers is referred to here as hyperplasia (of existing goblet cells), although the speed with which the phenotype change occurs and the small number of mitotic figures seen in the epithelium suggest that some degree of bronchial epithelial metaplasia (of another cell type) may also be involved. The change takes place particularly in the larger airways, but also in smaller and terminal bronchioles. While airways from naive or sham-challenged mice contain only the occasional goblet cell, lung sections from OA-challenged mice show that about 35% of smaller airways and 100% of larger airways contain goblet cells as more than 50% (and in some cases almost 100%) of the total airway epithelial cells. Other published models of pulmonary inflammation in mice have also shown increased mucus production following airway challenge with allergen (5).

Here, we define the duration of GCH by following the time course of its spontaneous resolution in the absence of further exposures to allergen and present results that illustrate the relatively poor ability of lipopolysaccharide (LPS) to induce GCH when compared with that of allergen. We also describe the effects of a subclinical infection with respiratory syncytial virus (RSV) on the airway goblet cells and report the ability of daily treatment with dexamethasone to both suppress the development of GCH and to accelerate the restoration of a normal epithelial phenotype in airways with an established GCH. We offer a working hypothesis to explain a possible mechanism for induction of GCH in airway epithelium.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

This research complied with national legislation and with company policy on the care and use of animals.

Sensitization

Sensitization and challenge of male Balb-C mice (Charles River UK Ltd, Margate, Kent, UK) were performed as previously reported (4). Briefly, mice (16-18 g, 4-5 wk old) were immunized, using an adjuvant-free protocol (8, 9), by the intraperitoneal injection of 10 µg ovalbumin (grade V, Sigma No. A-5503; Sigma Chemical Company, Poole, Dorset, UK) in 0.1 ml endotoxin-free saline (sodium chloride for injection BP, 0.9% weight/volume [wt/ vol], Evans Medical Ltd, Langhurst, Horsham, West Sussex, UK) on each of seven alternate days.

For the study with respiratory syncytial virus infection, mice were selected on the basis of their allergic status. Only animals with a serum immunoglobin E (IgE) titer of at least 80% of that of a pooled high-titer laboratory standard were used. Ovalbumin-specific serum IgE levels at approximately 40 d after the first intraperitoneal injection of OA were measured with an enzyme-linked immunosorbent assay (ELISA) employing antimouse IgE antibody from clone EM95.3.

Intratracheal Instillation

We have continued to use instillation of allergen solution directly into the trachea because it avoids contact of the upper airways (nasopharynx) with allergen and because of its ability to deliver larger amounts of allergen in a more consistent way than can be delivered by inhaled aerosol techniques. This ability probably accounts for the marked histologic changes seen in our model and the rapidity of their induction when compared with effects seen in models employing inhaled aerosol challenges. In humans, the cellular response following segmental challenge of the lung with allergen solution more closely approximates the findings of spontaneous exacerbations of asthma than does the cellular response after aerosol challenge, and for this reason the segmental technique has been proposed as a better model for the study of human asthma (10).

On or after Day 40, when the mice were approximately 25-26 g in weight, challenge of sensitized mice (normally in groups of five) was performed according to the standard protocol (challenge on 3 d, each 3 d apart) adopted by us in our initial study (4), using anesthesia with 0.2-ml Saffan (alphaxolone, 0.9% wt/vol + alphadalone, 0.3% wt/vol; Vet Drug Ltd, Dunnington, York, UK) intraperitoneally and the nonsurgical intratracheal instillation of 20 µg OA in 10 µl endotoxin-free saline by the dosing technique of Ho and Furst (11). This quantity of OA induces a submaximal increase in goblet cell numbers. We have shown that challenges with 1.25, 5, 20, or 80 µg OA induce dose- related increases in goblet cell numbers in the airway epithelium (unpublished results). Sham-challenged mice received 10 µl of endotoxin-free saline only.

The effectiveness of intratracheal instillation for the distribution of solutions within the airways was illustrated using microfocal radiography. Lungs were imaged with a Pantak HF200M microfocal X-ray unit (Astrophysics Research Ltd, Windsor, UK) at an exposure of 50 kV, 0.5 mA, 1-min duration, onto 3M type S mammography X-ray film. This microfocal X-ray unit produces magnified X rays at high resolution (12). The lungs were held between two perspex plates during radiography and were positioned in the X-ray beam by use of a motorized specimen table and video imaging system. Radiographs of pairs of lungs following intratracheal instillations of 10 µl of a 115% wt/vol suspension of radio-opaque barium sulphate contrast medium (Micropaque HD; Nicholas Laboratories Ltd, Slough, UK) in saline showed widespread distribution of the liquid to the periphery of both lungs by 20 min (Figure 1). One mouse was used for each time point (5, 10, 15, and 20 min).

View larger version (75K): [in this window] [in a new window]   Figure 1.   Microfocal radiograph showing the progressive distribution of radio-opaque contrast medium (115% wt/vol suspension of barium sulfate in saline) to the peripheral airways of individual 25-g male Balb-C mice at 5, 10, 15, and 20 min after intratracheal instillation of 10 µl.

BAL and Histologic Analysis of the Lung

These were performed as previously described (4). Briefly, groups of mice were killed by an intraperitoneal injection of 0.1 ml pentobarbitone sodium, 200 mg/ml (Euthatal; Rhone Merieux, Harlow, Essex, UK). Bronchoalveolar lavage (BAL) was performed by introducing 1 ml phosphate-buffered saline (PBS) into the lungs via a tracheal cannula, withdrawing this fluid into a test tube on ice, and repeating the procedure four more times. This procedure has been found to recover all of the intraluminal cells that can be recovered. Further, 1-ml washes merely serve to dilute the BAL cell suspension. The fluid (approximately 4.5 ml) was centrifuged (300 × g for 6 min) and the cells resuspended in 0.5 ml PBS. Total cells were counted using an improved Neubauer hemocytometer chamber, and an air-dried slide preparation made of each sample (Cytospin 3; Shandon Scientific, Runcorn, Cheshire, UK). These were stained with May-Grunwald-Giemsa stain, and differential counts of at least 200 cells were made according to standard morphologic criteria. The numbers of cells recovered per mouse were then expressed as the means and SEM for each treatment group. We regard the expression of recovered cell numbers as "per mouse" as appropriate since expressing them as "cells per ml BAL fluid" does not indicate efficient recovery of cells, but merely the extent of their dilution.

The lungs were fixed by slow in situ inflation with 1 ml 10% phosphate-buffered neutral formalin (BNF), pH 7.0, via the tracheal cannula and, following immediate removal from the thorax, immersion in BNF for a minimum period of 24 h. After fixation of the lung tissue and processing to paraffin wax, sections (3-4 µm thick) were cut longitudinally through the left lung (one lobe) and right lung (cranial, medial, accessory, and caudal lobes) so as to include all lobes (sometimes with the exception of the right accessory lobe). Sections were stained with hematoxylin and eosin for general morphology or Alcian Blue-Periodic Acid Schiff (ABPAS), with -amylase predigestion to remove glycogen, for the detection of acid and neutral mucins and identification of goblet cells. A histopathologist who was unaware of the BAL results analyzed one representative section per lung (30-60 airways per section) for GCH, using a semiquantitative, nonlinear, five-point scoring system related to the total numbers of goblet cells per section (severity score). Only those cells that were stained purple/magenta with ABPAS and were morphologically typical of epithelial goblet cells were counted. An incidence score that described the extent of affected airways within a section was produced by calculating the reciprocal of the proportion of airways in which goblet cells were present. A negative control slide of mouse lung was referred to periodically throughout the analysis to ensure that consistency of scoring was maintained. By multiplying the incidence score by the severity score, a combined score was produced that described the degree of hyperplasia present. We consider this method of analysis to be an acceptable alternative to other methods of assessing goblet cell numbers, especially with regard to the expression of a phenotypic change throughout an entire lung.

Scoring for inflammation was done using a five-point scoring system to describe numbers of individual cell types in hematoxylin and eosin-stained sections of lung tissue.

Treatment with LPS

LPS (endotoxin from Escherichia coli serotype 055:B5, Sigma No. L-6529) was dissolved in endotoxin-free saline and administered intratracheally at doses of 1, 10, or 100 µg, in a volume of 10 µl using the above technique. Single instillations (groups of three mice) or triple instillations (groups of six mice) were given, with triple exposures being 3 d apart. The mice were killed and examined 1 d after the last exposure to LPS.

Infection with RSV

The Balb-C strain of mice is among the most susceptible to RSV infection (13). The peak titer of RSV recovered from the lungs of these mice is at 4-6 d after intranasal inoculation, while lung lesions are maximal at Days 6-7 (14) and resolve by Day 9 or 10 (15).

The human A2 strain of RSV (originally supplied by Professor Peter Openshaw, St. Mary's Hospital Medical School, London) was grown and assayed for infectivity in a human nasopharyngeal cell line (Hep2c). The severity of RSV-induced illness in Balb-C mice is related directly to the inoculum dose. Mice inoculated with 10⁷ focal-forming units (FFU) display anorexia, cachexia, ruffled fur, pneumonia, and death by Day 7 postinfection; those receiving 5 × 10⁶ FFU exhibit milder signs of disease; and those receiving 2 × 10⁶ FFU show no signs of illness or weight loss (16). Our intention was to mimic a subclinical RSV infection in a human subject with asthma because in many cases of exacerbation, viral infection is low grade, with the presence of virus in samples being detectable only by the use of sensitive techniques such as polymerase chain reaction (PCR) (17). We used an inoculum titer of 1.5 × 10⁵ FFU per mouse. This titer is used in a well-established in-house model and caused no signs of illness (weight loss, changes in appearance or behavior) to occur. Paired groups of naive (nonsensitized, nonchallenged), sensitized (OA-sensitized, nonchallenged), sham-challenged (OA-sensitized, saline-challenged), or OA-challenged (OA-sensitized, OA-challenged) mice were given intranasal saline only, or RSV, on the day of the first allergen (OA) challenge of the standard mouse asthma model protocol. Using a well-characterized model of RSV infection, mice (groups of five) were anesthetized (for no longer than 10-15 s) by inhalation of ether, and 50 µl of the virus diluted in endotoxin-free saline was instilled intranasally. Control mice were sham-inoculated with 50 µl saline. The mice were usually killed and examined 24 h after the third challenge with OA, a time which coincided with the reported peak effects of RSV given by this method on lung pathology (14). Figure 2 shows a time line of the protocol used in the RSV studies. In one study, groups of mice were killed at 0, 12, 24, 48, 72, and 96 h after the third challenge in order to follow the time course of cell recruitment to the airways. The presence of live replicating virus in the lungs of animals from all infected groups was confirmed by positive results in an assay (18) of viral infectivity in homogenized lung tissue taken on the concluding day of the study. Lungs from sham-inoculated control mice were all negative for RSV.

View larger version (15K): [in this window] [in a new window]   Figure 2.   Time line of the protocol for infection of OA-sensitized mice with RSV on the same day as the first of three challenges with OA. The reported peaks for viral titer in the lungs and RSV-induced changes in lung tissue are from Reference 14.

Systemic Dosing Protocols

Figure 3 shows a time line to illustrate the two dosing protocols for treatment with dexamethasone. Protocol 1 was designed to study the effects of a treatment on the development of changes in the lungs during triple exposures to allergen, and involved dosing on each of the 8 d of the challenge period, beginning 1 d before the first challenge. Protocol 2 was designed to study the effects of a treatment on the lungs once GCH had been established by triple exposures to allergen, and involved dosing on each of the 7 d following the third challenge, during which no further exposure to allergen occurred.

View larger version (21K): [in this window] [in a new window]   Figure 3.   Protocols for the systemic dosing of mice. Protocol 1 involved daily dosing during the 8-d period of allergen challenge, beginning 1 d before the first challenge. Protocol 2 involved daily dosing on the seventh day following the challenges, beginning on the day after the third allergen challenge.

Treatment with Dexamethasone

Dexamethasone-21-phosphate, di-sodium salt (Sigma No. D-1159) was dissolved in endotoxin-free saline and administered daily by intraperitoneal injection in a volume of 0.1 ml by Protocol 1 (groups of five mice) or Protocol 2 (groups of six mice) (Figure 3). Doses of dexamethasone (0.1 or 1 mg/kg/d) are expressed in terms of the base.

Statistical Analysis

The statistical analysis performed for each data set is described in the appropriate figure legend. For analyses of GCH, the proportion of small airways showing goblet cells in each group (reciprocal of proportion = R: incidence) was normalized by arcsine transformation and analyzed by one-way analysis of variance followed by Dunnett's multiple comparison test to examine whether the mean values in the treated groups were significantly different from controls. The average goblet cell scores (S: severity) within affected airways and the R × S (incidence × severity) values were analyzed nonparametrically using one-way Dunn's test, based on Kruskal-Wallis rank sums (19), to test whether the median values for each treatment group were different from a control group.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Duration and Spontaneous Resolution of GCH after Three Challenges with OA

Previously reported studies with this model showed that the degree of GCH at 11 d after a third intratracheal challenge with 20 µg OA remained as great as at 1 d after the challenge, despite numbers of eosinophils and lymphocytes (both greatly elevated at 24-72 h after the challenge) in the airway lumen having fallen to low levels (4). We have now monitored the spontaneous resolution of this feature from Day 14, and the changes in cell numbers in the airway lumen in the absence of further exposures to the allergen. Following the standard protocol for our model, paired groups of five mice (sham-sensitized and sham-challenged with saline, or OA-sensitized and challenged with 20 µg OA) were killed at weekly intervals, starting at Day 14 after the third OA challenge and continuing until Day 50. Eosinophil numbers in the airway lumen (i.e., recovered by BAL) were very small at Day 14 (mean/SEM = 0.895/0.37 × 10⁴ per mouse), and fell to zero by Days 35- 40 (Figure 4a). Lymphocyte numbers in the lumen fell to a plateau at Day 35 but were still elevated at Day 50 (mean/ SEM = 0.24/0.10 × 10⁴ per mouse) compared with those in sham-challenged controls (0.03/0.016 × 10⁴) (Figure 4b). Lymphocyte numbers in the tissues of the lung (as demonstrated by histology) also were still elevated at Day 50 when compared with the control group (results not shown). These lymphocytes may serve to maintain a surveillance role as memory cells within and near the airway. There were no significant differences between OA-challenged and sham-challenged groups at any time point in the numbers of macrophages or neutrophils in the airway lumen. The decline in the score for goblet cell numbers in the airway epithelium followed a similar pattern, falling rapidly until about Day 30, and reaching control levels around Day 50 (Figure 4c). This may indicate that cellular inflammation plays a direct role in the maintenance of GCH or that it is a parallel phenomenon whose resolution follows approximately the same time course as that of GCH.

View larger version (31K): [in this window] [in a new window]   Figure 4.   The time course of the spontaneous resolution of luminal eosinophil (a), luminal lymphocyte (b), and epithelial goblet cell (c) numbers in the airways of mice sham-sensitized and challenged with saline (open symbols), or sensitized and challenged with OA (closed symbols). Statistical significance: *P < 0.05, ***P < 0.001, Student's t test (a and b); data for the nonparametric goblet cell scores (c) were analyzed using Dunn's multiple comparison test (all possible comparisons) based on Kruskal-Wallis rank sums. Median/range values for (incidence × severity) scores were: sham 14 d (0 /0-0), 21 d (0 /0-4.08), 28 d (0 /0- 2.94), 35 d (0 /0-1.89), 42 d (0 /0-0), 50 d (0 /0-0); OA 14 d (27.04/ 0- 54.84), 21 d (16.245/1.56-42.62), 28 d (6.12 /1.15-45.62), 35 d (4.48/ 1.28-6.90), 42 d (4.655/0-8.11), 50 d (1.89 /1.04-4.00). No significant differences were found (P < 0.05).

Limited Ability of Intratracheally Instilled LPS to Induce GCH

When a single intratracheal instillation of bacterial LPS endotoxin (0, 1, 10, or 100 µg) was given to sham-sensitized or OA-sensitized mice, a dose-related infiltration of neutrophils into the airway lumen took place when examined after 24 h (Figure 5a). There were no eosinophils in the lumen, and the numbers of lymphocytes and macrophages were not increased after exposure to LPS (results not shown). There were no statistically significant differences between responses in sham- or OA-sensitized mice. Despite an intense neutrophilic inflammation in the lung parenchyma, submucosa, and epithelium, and also massive transepithelial migration into the lumen of the airways, there was no increase in the number of airway goblet cells in these LPS-treated animals when compared with the saline-treated mice (virtually zero goblet cells in all groups).

View larger version (25K): [in this window] [in a new window]   Figure 5.   Effect of intratracheal instillation of LPS. Panel a shows the dose-related increase in neutrophil numbers in the airway lumen 24 h after a single instillation of LPS. There was no statistically significant difference between the response of sham-sensitized or OA-sensitized mice (t test). Panel b shows a similar response 24 h after a third instillation of LPS. Neutrophil numbers are no higher than after a single exposure. Panel c shows the slight increase in goblet cell numbers 24 h after a third exposure to LPS. Median/range values were sham/0 µg LPS: 0/0- 6.0; sham/1 µg LPS: 3.225/1.23- 11.36; sham/10 µg LPS: 4.66/1.28- 8.93; sham/100 µg LPS: 5.395/ 3.77-7.02; OA/0 µg LPS: 0/0-3.3; OA/1 µg LPS: 2.535/0-6.67; OA/ 10 µg LPS: 6.02/1.64-13.21. Analysis as for Figure 7b, to test whether the mean rank sum of each LPS-treated group was higher than that of the relevant control (zero LPS) group (*P < 0.05).

View larger version (32K): [in this window] [in a new window]   Figure 7.   Effects of chronic treatment with dexamethasone during developing GCH. Panel a dose- related suppression by dexamethasone (0.1 or 1 mg/kg/d, intraperitoneally by dosing protocol 1) of cellular infiltration into the airway lumen of OA-sensitized mice 1 d after a third OA challenge. Group A: sham challenges with saline (intratracheally) and saline treatment (intraperitoneally); group B: OA challenges (intratracheally) and saline treatment (intraperitoneally); group C: OA challenges (intratracheally) and 0.1 mg/kg/d dexamethasone treatment (intraperitoneally); group D: OA challenges (intratracheally) and 1 mg/kg/d dexamethasone treatment (intraperitoneally). The higher dose of dexamethasone (group D) has suppressed cell numbers to those in nonchallenged mice (group A). *P < 0.05, Student's t test. Panel b dose-related suppression by dexamethasone of the degree of GCH. Groups and treatments were those in 10a. The scores for goblet cell numbers and incidence in the airways were analyzed by the nonparametric Dunn's test, based on Kruskal-Wallis rank sums. One-way analyses were performed to test whether the mean rank of each treatment (groups C and D) was less than that of the positive control group (B). Median/range values were group A: 0/; group B: 76.92/21.62-127.28; group C: 28.28/14.14-72.0; group D: 8.34/0- 14.52 (**P < 0.01).

When the same doses of LPS were given as repeated intratracheal instillations on 3 d, each 3 d apart, and examined 24 h after the third instillation, there was again a dose-related influx of neutrophils into the airway lumen, but the influx after the third exposure was no greater than after a single exposure to LPS (Figure 5b). (This contrasts with the progressive increase in numbers of airway eosinophils in OA-sensitized mice, which is directly related to the number of intratracheal exposures to OA [4].) There were no eosinophils in the airways, but there was a dose-related increase in the number, and particularly the size, of macrophages in the lumen. Cachexia led to only two of the group of six sham-sensitized mice and none of the group of six OA-sensitized mice surviving three exposures to the top dose of LPS (100 µg). Figures 5c and 6A and 6B show that despite this repeated influx of large numbers of neutrophils into the airways and a marked activation of airway macrophages, multiple exposures to LPS caused only a relatively small, though dose-related, increase in goblet cell numbers. The score for goblet cell numbers induced by three exposures to 100 µg LPS (median score of 6.02) was small when compared with the massive increase achieved after three exposures to 20 µg OA (median score of 76.92: group B, Figure 7b). These results indicate that although the prolonged presence of severe nonallergic cellular inflammation in the airways of sham-sensitized or OA-sensitized mice was able to cause small increases in goblet cell numbers, it was not sufficient to reproduce the marked GCH seen after induction of allergic inflammation in the airways by exposure of sensitized mice to allergen during the same time scale.

Effects of Respiratory Infection with RSV When Superimposed on the Asthma Model

Paired groups of naive (nonsensitized, nonchallenged: groups A and B), sensitized (OA-sensitized, nonchallenged: groups C and D), sham-challenged (OA-sensitized, saline-challenged: groups E and F), or OA-challenged (OA-sensitized, OA-challenged: groups G and H) mice were given intranasal saline or RSV on the day of the first allergen (OA) challenge of the standard protocol of the mouse asthma model (Figure 2). RSV infection caused no statistically significant changes in the numbers of leukocytes (eosinophils, neutrophils, lymphocytes, monocytes, and basophils) in the blood of any of the groups at 1 d after the third OA challenge (results not shown). BAL of the mice at this time showed that RSV infection induced a small, though not statistically significant, increase in neutrophil numbers in the airways of mice that had not been challenged with OA (group pairs A/B and C/D), and an increased number of macrophages (which were much enlarged and vacuolated) in all groups (group pairs A/B, C/D, E/F, G/H) (Figure 8a). In group pairs E/F and G/H, RSV induced an increase in lymphocyte numbers in the airways, and a statistically significant increase in the score for lymphocyte numbers in the lung tissue of all infected groups (Figure 8c). These lymphocytes were mainly at perivascular and peribronchiolar sites. However, RSV neither induced eosinophil recruitment in naive (A/B), OA-sensitized (C/D), or sham-challenged (E/ F) groups nor enhanced it in those mice in which eosinophil recruitment was induced by OA challenge (G/H). Analysis of scores for goblet cell numbers in the airway epithelium showed that RSV infection neither induced a statistically significant GCH in nonchallenged groups nor enhanced the degree of GCH in mice in which the phenotypic change was induced by allergen challenge (group pair G/H) (Figure 8b). Figures 6C and 6D illustrate the lack of GCH-inducing activity of the lymphocytic infiltrate induced by RSV. However, mice in group H did demonstrate evidence of an enhanced discharge of mucin from goblet cells when compared with group G (Figure 9). This suggests that RSV infection induced a discharge of mucin in excess of that which follows exposure to OA in this model. It is therefore possible that, if their mucin content had been totally depleted by the virus, some goblet cells in group H (OA-challenged, RSV-positive) would have remained unstained by ABPAS. In this case, they would not have been recognized as goblet cells, and an artificially low score may have resulted.

View larger version (28K): [in this window] [in a new window]   View larger version (23K): [in this window] [in a new window]   View larger version (44K): [in this window] [in a new window]   Figure 8.   Panel a shows cell numbers recovered by BAL at 24 h after a third intratracheal challenge with OA in groups of mice infected intranasally with saline only or RSV. Groups A and B: naive mice neither sensitized nor challenged; groups C and D: OA-sensitized, nonchallenged mice; groups E and F: OA-sensitized mice challenged with saline; groups G and H: OA-sensitized mice challenged with OA. RSV infection neither induced eosinophil recruitment nor enhanced it when induced by allergen, but did induce an increase in the numbers of macrophages and, particularly in OA-challenged mice, lymphocytes. Panel b shows scores for goblet cell numbers in the airway epithelium of mice (groups as in Panel a). RSV infection neither induced GCH nor enhanced it when induced by allergen. The reduction in the score in group H compared with group G was not statistically significant. Panel c shows the scores for lymphocyte numbers in the lung tissues of these animals. All RSV-infected groups (B, D, F, and H) show statistically significant increases in the numbers of lymphocytes when compared with their paired sham-infected control groups (A, C, E, and G). Most of these lymphocytes were at perivascular and peribronchiolar sites (*P < 0.05, **P < 0.01).

View larger version (151K): [in this window] [in a new window]   View larger version (143K): [in this window] [in a new window]   View larger version (168K): [in this window] [in a new window]   View larger version (156K): [in this window] [in a new window]   Figure 6.   Features of nonallergic inflammation. Lung sections stained with hematoxylin and eosin for general morphology (A and C) and with ABPAS for mucin (B and D). Following three intratracheal instillations of 100 µg LPS (each 3 d apart), a predominantly neutrophilic infiltration was induced (A), but with little GCH (B). Infection of the airways with RSV, given on the day of the first of three sham challenges, induced a predominantly lymphocytic infiltrate with some neutrophils (C) but no GCH (D). All tissues were taken 1 d after a third challenge. Bars = 20 µm.

View larger version (114K): [in this window] [in a new window]   View larger version (120K): [in this window] [in a new window]   Figure 9.   ABPAS-stained sections of lungs from OA-sensitized mice show that goblet cells from OA-challenged, RSV-negative mice (group G) have discharged little of their magenta-colored mucus contents (A), while goblet cells from OA-challenged, RSV-positive mice (group H) appear to contain empty vacuoles and to have discharged much of their mucin contents (B). Mucus can be seen clinging to the apical surface of the epithelial cells in this section (arrow). Groups as in Figure 8. Bar = 20 µm.

In order to study the influence of infection with RSV on the time course of cell recruitment after a third OA challenge, groups of mice were killed at 12, 24, 48, 72, and 96 h after the challenge. Figure 10 shows that eosinophil and neutrophil numbers in RSV-negative or RSV-positive mice were similar at all time points. However, lymphocyte and macrophage numbers were higher in RSV-infected mice until 72 h, although these differences were not statistically significant. Figure 11 shows that the only statistical significance (P < 0.05) in the differences between goblet cell scores in infected and noninfected mice was at 0 h, indicating that after a second OA challenge, the degree of GCH was greater in RSV-infected mice, but that a third exposure to OA brought the number in noninfected mice up to the maximal achievable number seen in RSV-infected animals. This subclinical degree of RSV infection was therefore unable to increase the maximal degree of GCH induced by three exposures to allergen but able to accelerate the rate of its development at earlier stages.

View larger version (36K): [in this window] [in a new window]   Figure 10.   Cell numbers in BAL fluid from RSV-negative (open symbols) or RSV-positive (closed symbols) mice at various time points following a third OA challenge in OA-sensitized mice. Lymphocyte and macrophage numbers are higher in RSV-infected animals, although these differences are not statistically significant (t test).

View larger version (16K): [in this window] [in a new window]   Figure 11.   Goblet cell scores from RSV-negative (open symbols) or RSV-positive (closed symbols) mice at various time points following a third OA challenge in OA-sensitized mice. The only statistical significance is at 0 h (median score/range values were: noninfected: 47.06 /41.80-49.12; RSV-infected: 73.83/53.58-84.00), indicating that after a second OA challenge, the degree of GCH was greater in RSV-infected mice, but that a third exposure to OA brought the number in noninfected mice up to the maximal achievable number in RSV-infected animals. Respiratory syncytial virus infection may therefore accelerate the rate of development of GCH in its early stages.

Suppression of the Development of GCH by Chronic Treatment with Dexamethasone

When OA-sensitized mice were treated with daily intraperitoneal injections of 0.1 or 1 mg/kg dexamethasone for 8 d, beginning 1 d before the first of three intratracheal challenges with OA (dosing Protocol 1), infiltration of all cell types into the airways (Figure 7a) and the numbers of airway goblet cells (Figure 7b) at 1 d after the third challenge were reduced in a dose-related manner. Dexamethasone, when given daily, was therefore able to inhibit the development of the hyperplastic/metaplastic goblet cell phenotype in the airway epithelium (Figures 12E and 12F). Previously published work suggested that the processes of cellular infiltration and induction of GCH appeared to be causally unrelated because acute dosing with the glucocorticoid, on only the 3 d of OA challenge, could completely suppress cell recruitment without suppressing the development of GCH (4).

View larger version (151K): [in this window] [in a new window]   View larger version (152K): [in this window] [in a new window]   View larger version (148K): [in this window] [in a new window]   View larger version (145K): [in this window] [in a new window]   View larger version (136K): [in this window] [in a new window]   View larger version (154K): [in this window] [in a new window]   Figure 12.   Features of allergic inflammation and their suppression by dexamethasone. Lung sections from OA-sensitized mice stained with hematoxylin and eosin for general morphology (A, C, and E) and with ABPAS for mucin (B, D, and F ). Sham-challenged lung showed no cellular infiltration or epithelial thickening (A), and no GCH (B). Allergen-challenged lung (triple intratracheal challenge with OA) showed intense, predominantly eosinophilic, inflammation and thickening of the epithelial layer (C), in addition to GCH (D). All these mice were sham-treated with saline, intraperitoneally, by dosing protocol 1. Lungs from OA-challenged mice treated with 1 mg/kg/d dexamethasone, intraperitoneally, by dosing protocol 1 showed that the steroid prevented cellular infiltration (E) and the development of GCH ( F  ). All tissues were taken 1 d after a third allergen challenge. Bars = 20 µm.

Reduction of Established GCH by Chronic Treatment with Dexamethasone

When OA-sensitized mice, in which a marked airway GCH had been established by triple challenge with OA, were treated with daily intraperitoneal injections of 0.1 or 1 mg/kg dexamethasone for 7 d (dosing Protocol 2), the severity of GCH at 8 d after the third OA challenge was reduced by the higher dose of dexamethasone from that in saline-treated controls toward that expected in normal mice (Figure 13). Although the severity score (Figure 13a) for GCH in affected airways was significantly reduced, there was no reduction in the incidence (number of airways showing this feature), and this meant that the reduction seen in the severity × incidence value was not statistically significant (Figure 13b). This was to be expected because the number of airways showing the phenotype change would be determined by the distribution of allergen during the challenges and would be unlikely to be influenced by postchallenge treatment. An accelerated return from the increased goblet-to-ciliated cell ratio (group A) toward a normal epithelial phenotype was therefore produced in this model by daily treatment with 1 mg/kg of dexamethasone (group C).

View larger version (30K): [in this window] [in a new window]   Figure 13.   Panel a shows suppression of the severity of established GCH by daily intraperitoneal treatment with 1 mg/kg dexamethasone in OA-sensitized and OA-challenged mice at 8 d after the third OA challenge (dosing protocol 2). While saline-treated mice still showed a marked goblet cell hyperplasia (group A), dexamethasone-treated mice (group C) had an epithelial phenotype closer to normal in terms of the severity of GCH. *P < 0.05. Panel b shows that the (incidence × severity) values, though reduced, were not statistically lower in group C than in group A because, despite a reduction in the goblet cell score (numbers in affected airways), the number of airways showing GCH was not reduced. This was to be expected because the number of airways displaying the phenotype change would be determined by the distribution of allergen during the challenges and would be unlikely to be influenced by postchallenge treatment. NS = not significant.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The development of a model in which airway epithelial GCH could be reproducibly induced by a local allergic reaction in the pulmonary airways of mice (4) enabled us to characterize this feature in terms of its induction (by allergic or nonallergic inflammatory mechanisms), duration, and resolution, and to demonstrate the response of this phenotypically altered epithelium to concurrent infection with RSV and to treatment with dexamethasone.

An understanding of the mechanisms for induction of GCH and other features of airway remodeling is crucial to the successful treatment of asthma because the disease involves more than bronchoconstriction. These other features involve changes in cell phenotypes and are less easily reversed than bronchoconstriction. Although the GCH- inducing capacity of agents such as sulfur dioxide, ozone, and tobacco smoke has been reported, mechanisms for the etiology and pathogenesis of GCH remain unclear (20). Even in relatively symptom-free periods, ventilation scans in asthmatic patients with normal chest radiographs reveal poor ventilation in several areas, the typical wheeze being generated by an airway that is almost completely occluded (21). Since 1882, Curshmann's spirals (corkscrew-shaped twists of condensed mucus) have been recognized as characteristic features of asthmatic sputum (22), the mucus in asthma usually being more viscid and tenacious than that in chronic obstructive pulmonary disease (23). There is evidence showing that in patients with asthma, but not in normal subjects or patients with chronic bronchitis, some mucin is retained in the goblet cells to form a continuous mass extending into the airway lumen (24).

Although it has been reported as absent in some patients with mild asthma (25), GCH has been described as the most discernible finding in airway epithelium of even newly diagnosed asthmatics (26) and can lead to a reduction in the caliber of the airway lumen due to epithelial thickening and increased mucin production. In severe cases, complete occlusion of the lumen by plugs of mucus and cell debris occurs (27). Under these circumstances (as with airway mucosal inflammation, edema, or vasodilatation) the luminal diameter of the airway is reduced, respiratory resistance is increased, and otherwise trivial bronchoconstriction can lead to dangerous narrowing of the airway irrespective of the reactivity of airway smooth muscle to bronchoconstricting stimuli (28).

The extent to which increased airway mucus production affects FEV₁ lung function measurements is still unclear (29), although calculations show that mucin discharge is able to amplify the increase in resistance produced by a moderate degree of airway smooth muscle shortening with fatal consequences (30). Normal human airway epithelium has a ratio of approximately three to five ciliated cells to each goblet cell (31, 32) with a maximum reported ratio of 10:1 (33). In asthma the number of goblet cells can equal or exceed that of ciliated cells (26), this ratio reversal being especially marked in patients who die in status asthmaticus, where there is a 30-fold increase in the percentage area of goblet cells in epithelium when compared with the percentage in patients dying of nonasthma respiratory disease (34). Total obstruction of airways with mucus plugs was the most common lesion seen in a study of fatalities from asthma (35), and increased mucus production appears to be particularly implicated where death follows rapidly after an asthma attack (36).

The ratio of goblet cells to ciliated cells in the airway epithelium of normal Balb-C mice (< 1:100) is much less than in normal human airways. The increased ratio after multiple challenges with 20 µg OA in our model can exceed that seen in human asthma and, in this respect, our mouse model may more closely resemble human chronic bronchitis than human asthma. Our studies show that after three challenges with 20 µg OA, and in the absence of further exposures to allergen, established GCH resolved over a period of 50 d. The time course of resolution to normal was similar to that of the reduction in eosinophil and lymphocyte numbers in the airway lumen. Lymphocytes, however, did not completely disappear, possibly maintaining a surveillance role as memory cells within and close to the airway lumen. The similarity in time courses may indicate that cellular inflammation plays a direct role in the maintenance of GCH or that it is a parallel phenomenon dependent on the return to normal of some independent factor(s).

The failure of repeated high doses of LPS to reproduce the marked degree of GCH achieved with allergen suggests that induction of such an extensive phenotype change required allergic (T-helper 2 [T_H2]-driven) mechanisms. Severe nonallergic inflammation induced by deposition of LPS in the airways attracted large numbers of neutrophils and monocytes to the lumen, and led to a visibly profound state of activation in airway macrophages (as did infection with RSV) but, despite this, neither LPS nor RSV were able to induce the extent of GCH seen with allergen. Processes other than macrophage activation or release of macrophage mediators (interleukin [IL]-1, IL-6, IL-8, tumor necrosis factor alpha [TNF], platelet-activating factor [PAF], etc.) must provide the inducing trigger. Our observations may be partly explained by recently published studies which suggest that the T_H2 cytokine IL-4, whose production accompanies an allergic response, plays a central role in the induction of mucin glycoprotein gene expression and mucin release in mouse respiratory epithelium (37).

We found that infection of the lungs with RSV increased the number of macrophages in the airways and the number of lymphocytes at perivascular and peribronchiolar sites in the lung tissue. RSV did not, however, induce GCH in control mice that were not challenged with allergen or enhance the degree of allergen-induced GCH. In the airways of mice that showed a marked GCH induced by allergen challenge, primary subclinical infection with RSV was accompanied by an apparent discharge of mucin from the goblet cells. A similar effect has been reported in goblet cells in the bronchial epithelium of guinea pigs infected with parainfluenza 3 virus (38), a model in which mast cells and histamine appear to play an important role in both inflammatory cell recruitment (39, 40) and the airway responsiveness induced by it (41). Acute respiratory infections are the most frequent illnesses of the human host (42), and infection of the respiratory tract of normal individuals with virus can cause airway inflammation and hyperresponsiveness, altered -adrenergic function, production of virus-specific IgE, and damage to the respiratory epithelium (43). Infection, either overt or subclinical, with respiratory virus in patients with asthma constitutes the most common precipitating event of exacerbations of the disease (44, 45). Viral pathogens have been shown to be present in greater than 80% of asthma exacerbations in a study of 9- to 11-yr-old children (17), and in 44% of asthma exacerbations in adults (44).

It is likely that many factors contribute to the exacerbating action of viruses, including an increase in the eosinophil number in the airways of patients previously sensitized by contact with the virus. Acute and persistent enhancement of histamine release in response to segmental challenge with allergen is seen in allergic patients during infection with rhinovirus, together with an increased recruitment of eosinophils to the airways 48 h after challenge (46). There is some evidence in humans of a relationship between viral infections and the potentiation of allergic responses, including inflammation, in the airways (46). Infection with human rhinovirus induces increases in histamine responsiveness and increases in submucosal lymphocytes and epithelial eosinophils in the airways which, in patients with asthma, persists into convalescence (47). Viruses exert many other potent effects on cells of the lung and on inflammatory cells recruited to the lungs (43). One example is the ability of RSV to cause a sustained stimulation of PAF synthesis that parallels viral replication (48). This may be relevant because PAF stimulates airway mucin production and impairs mucociliary clearance (49).

RSV replicates best at 37°C and therefore preferentially infects respiratory epithelium of the lower airways. It induces the production of IL-8 by respiratory epithelial cells (50), and of TNF, IL-10, and IL-12 from peripheral blood mononuclear cells (51). RSV-induced IL-8 production by A549 human lung epithelial cells can be inhibited by the Na+-channel blocker, amiloride, or ribavirin, the only specific agent available for treatment of RSV infection (52). Both intercellular adhesion molecule-1 (ICAM-1) and very late antigen-4 (VLA-4) adhesion molecules are involved in RSV-induced inflammation in mouse airways (53).

A link between viral infection and the T_H2 cell phenotype in mice is suggested by work showing that CD4+ T_H2 immune responses (probably involving IL-4) to OA can switch lymphocytic choriomeningitis virus peptide-specific CD8+ T cells in the lung to IL-5 production, inducing an influx of eosinophils following challenge with virus peptide (54).

Recently published work (55) confirms earlier findings (56), which showed that when Balb-C mice are immunized with formalin-inactivated RSV, memory T cells with a predominantly T_H2-like cytokine profile are induced. Subsequent challenge of these mice with an RSV infection led to a pulmonary influx of eosinophils and CD4+ cells, a marked expression of mRNA for T_H2-type cytokines (IL-5, IL-10, and IL-13), and a decrease in the mRNA for the T_H1-type cytokine, IL-12 (55). However, if mice were immunized with live RSV, memory T cells with a predominantly T_H1-like cytokine profile were induced. In mice with a primary infection with RSV (as in the studies reported here), little IL-2, IL-4, or IL-5 mRNA was found, while interferon-gamma (IFN-), IL-10, and IL-12 mRNAs increased during the first 8 d. This suggests that the local cytokine response to primary RSV infection was of a mixed but predominantly T_H1-like phenotype. Since the development of eosinophilia requires the T_H2 cytokine, IL-5, these reports may explain why, in our present studies of mice with a primary infection with RSV, there was no induction by the virus of eosinophilia in control groups not challenged with allergen, nor enhancement by the virus of the eosinophilic response to allergen. Indeed, eosinophil numbers in the RSV-infected group (Figure 8b, group H) were slightly smaller than in the sham-infected mice (group G), perhaps because of the known suppressing effect of IFN- on allergic responses (57) and lung eosinophilia (58) in mice.

We have previously reported that while short-term treatment of the mice with dexamethasone on only the 3 d of allergen challenge can completely suppress cell recruitment into the airways, it does so without reducing the degree of GCH, suggesting that cell influx into the airway lumen may be an incidental, but not a causative, phenomenon (4). Here, we show that when treatment with dexamethasone was given daily throughout the 8-d challenge period, a dose-related reduction of goblet cell numbers was achieved. Dexamethasone also exerted a nonspecific inhibitory effect on cell recruitment in the present experiments, suppressing numbers of eosinophils, neutrophils, lymphocytes, and macrophages in the airway lumen to those of unchallenged control mice.

The severity of established GCH was also reduced by daily systemic treatment with 1 mg/kg dexamethasone, reproducing the clinical observation that long-term treatment of patients with asthma with an inhaled glucocorticoid can normalize the ratio of ciliated cells to goblet cells in the airways (26). Glucocorticoids have been shown to inhibit GCH induced in rats by tobacco smoke (59) or neutrophil products (60).

We propose the working hypothesis that persistent exudation of plasma proteins from postcapillary venules, possibly made permeable by the action of mast cell mediators such as histamine, serotonin, PAF, or tryptase, and the subsequent passage of this exudate between airway epithelial cells may be an initiating or contributory event in the induction of goblet cell hyperplasia/metaplasia in airway epithelium. Such exudation occurs readily even in intact epithelium and exposes epithelial cells to a range of proteins in concentrations not normally present outside the blood (61). The admixture of exudate proteins with airway mucus is known to increase its viscosity and hence its tenacious nature. Cross-linking by allergen of IgE molecules on the surfaces of mast cells in the airway epithelium would lead to immediate degranulation and the release of allergic mediators, including preformed IL-4 and IL-5, characteristic of a T_H2-driven allergic response. In a similar murine model, lung mast cell degranulation and plasma protein exudation into airways have been shown to occur within 1 h of aerosol challenge with OA (62). This hypothesis would explain the rapid appearance of GCH following an exposure to allergen in our model (at a time before eosinophils or lymphocytes migrate across the airway epithelium) (4) and the inhibitory action of glucocorticoids on GCH, as described here.

Glucocorticoids are potent inhibitors of plasma exudation from postcapillary venules at inflammatory sites (63), dexamethasone having been shown to inhibit increases in vascular permeability due to histamine or allergen in mice (64). It may also explain why nonallergic inflammation induced by LPS or RSV was unable to induce marked GCH. If mast cell products are implicated as being required for induction of GCH, the hypothesis would account for the apparent lack of GCH-inducing activity in macrophages as indicated by the inability of macrophages, activated by LPS or infection with RSV, to induce marked phenotypic change. It is likely that the nonallergic inflammation induced by LPS, or infection with RSV, does not involve mast cell degranulation or plasma protein exudation to any great extent. We are currently testing the above hypothesis.

A report that agents which stimulate an increase in intracellular cyclic adenosine monophosphate (cholera toxin, dibutyryl cAMP, or PGE₁) were able to induce airway GCH in mice (65) led Ahlstedt and colleagues (66) to suggest that the GCH seen in their murine model of T-cell-mediated delayed-type hypersensitivity (DH) was due to the production of E-type prostaglandins by monocytes. Their model showed that, in addition to increased monocytes, an increase in mast cell numbers occurred around the airways following the DH reaction (67), and it is possible that mediator release from these mast cells, followed by plasma protein exudation, may have contributed to induction of GCH.

The results presented here confirm and extend the relevance of our model as a representation of the human disease and indicate that allergic inflammatory mechanisms are required for the full induction of GCH in the mouse. The ability to establish GCH in airway epithelium has made possible the demonstration of mucin discharge as a possible pathologically relevant consequence of respiratory viral infection (70), and the suppressing effect of a glucocorticoid on both developing and established GCH (71). It is hoped that understanding of the disease and its treatment may eventually be improved by studies with such models.