INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sputum production is a common symptom in asthma, especially during asthma exacerbations (1), and a history of sputum production is independently associated with an accelerated rate of decline in FEV₁ (2). Furthermore, hypersecretion of mucus plays a central role in the pathogenesis of severe airway obstruction and asphyxiation in fatal attacks of asthma (3). The mechanisms governing hypersecretion of mucus in asthma are poorly understood. Mucins are the major constituents of airway mucus and the major determinants of its viscoelastic and adhesive properties (12, 13).

Airway mucins are produced mainly by goblet cells and submucosal gland cells. Nine different mucin genes have been cloned (12, 14), but it is not known whether hypersecretion of mucus in asthma is associated with activation and selective expression of specific mucin genes. In addition, the relative contribution of goblet cells and submucosal gland cells to the mucin component of airway mucus is uncertain and is likely to vary by airway level and in health and disease. Reid (15) estimated that the volume of glands in the airway mucosa was 40 times greater than the volume of goblet cells, but this calculation was based on several assumptions about the frequency and distribution of goblet cells and glands in the airway. Using quantitative morphometry to analyze mucins in the surface epithelium and submucosal glands of airways of macaque monkeys, Heidsiek and coworkers (16) reported twice as much stainable stored mucins in the goblet cells in the epithelium of the tracheobronchial airway as in the submucosal glands. This suggests that in primates, goblet cells, not mucous cells, might be the principal source of airway mucins. Thus, acute degranulation of goblet cells may represent an important mechanism of airway obstruction during asthma exacerbations, and chronic degranulation may represent a mechanism for chronic airway narrowing in more severe forms of asthma.

We hypothesized that mild and moderate asthma is associated with an increase in mucin stores in the epithelium secondary to either goblet cell hypertrophy or hyperplasia, and that secreted mucin in the airway lumen would be increased in proportion to these abnormalities. To investigate this hypothesis, we analyzed stored and secreted mucin and mucin gene expression in samples obtained by bronchoscopy and sputum induction from healthy subjects and subjects with asthma.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects and Protocol

Twelve healthy subjects and 13 subjects with asthma were studied (Table 1). The subjects with asthma were taking inhaled short-acting  agonists only as treatment of their asthma. All subjects signed consent forms approved by the Committee on Human Research at the University of California, San Francisco. The study involved three visits for spirometry, allergen skin testing, and methacholine challenge (Visit 1), as well as sputum induction (Visit 2, 48 h after Visit 1) and bronchoscopy (Visit 3, 7 d after Visit 2), and these tests were performed as previously described (17). Sputum induction was performed according a 12-min sputum induction protocol (18). During bronchoscopy, the bronchoscope was introduced orally and advanced to the right mainstem bronchus, where up to nine mucosal biopsies (six or seven for morphometry and immunohistochemistry and two or three for RNA extraction) were taken from the carinae of the upper lobe, middle lobe, and superior segment, using a spiked fenestrated forceps.

Measurement of Secreted Mucin in Induced Sputum

Induced sputum was homogenized by mixing with an equal volume of 0.1% dithiothreitol, as previously described (20). Secreted mucin was measured in supernatants of induced sputum by enzyme-linked immunosorbent assay (ELISA), as previously described (21). Briefly, Dynex (Chantilly, VA) Microtiter plates (Immulon 2HB) were coated with purified IgG 17Q2 (0.2 µg/well) in coating buffer (0.05 M sodium carbonate buffer, pH 9.6) and incubated at 37°C for 2 h under an airtight cover. After washing with a solution of phosphate-buffered saline (PBS)-Tween 20 (0.05%), various amounts of standard mucin in PBS-Tween 20 (0.05%) (0.5. 1, 2, 4, 8, and 16 ng of protein per well) and the samples were added to the wells. These reactions were carried out under the same airtight conditions at 37° C for 2 h. After washing with PBS-Tween 20 solution, plates were reacted with alkaline phosphatase-conjugated IgG 17Q2 (0.2 to 0.4 µg/well) in 0.2% albumin and PBS-Tween 20. Incubation was carried out at 37° C for 2 h. Finally, after washing with PBS-Tween 20, phosphate substrate solution ( p-nitrophenyl phosphate, disodium, at 1 mg/ml in 10% diethanolamine solution, pH 9.8) was added for color development. The reaction was stopped by the addition of 50 µl of 3 N NaOH to each well. Developed color in each well was read at a wavelength of 405 nm.

This double-sandwich ELISA method measures "general" airway mucin, that is, the 17Q2 antibody specifically stains the granules of the surface goblet cells and mucous gland cells (lesser intensity) and the high molecular weight mucous glycoproteins found in sputum. Therefore, this general airway mucin probably includes both MUC5B and MUC5AC, and other mucin gene products. The epitope characterization for the 17Q2 antibody has been published previously (21).

Morphometry Measurements

Biopsies were fixed in 4% paraformaldehyde, dehydrated in alcohol, and infiltrated with glycomethacrylate (Polysciences, Warrington, PA). A 2-µM section was cut from each block and stained with Alcian blue and periodic acid-Schiff reagent (Sigma, St. Louis, MO). A microscopic image of each biopsy at a final magnification of ×80 was captured by video camera (3CCD; Dage, Michigan City, IN) linked to a computer (Power Tower Pro 200; Power Computing, Austin, TX) and NIH Image software (version 1.60; NIH, Bethesda, MD). A design-based stereologic analysis was then performed with NIH Image, Stereology Toolbox (Morphometrix, Davis, CA), and the Computer-Assisted Stereology Toolbox software system (C.A.S.T-Grid; Olympus, Albertslund, Denmark). Random sampling of high-power fields of airway epithelium was achieved by labeling areas of intact epithelium with a line on the underlying basement membrane. Two to six images per section (minimum of two, maximum of six, depending on the length of the intact epithelium) at a magnification of ×1,200 were sampled randomly at equal distances apart on the basement membrane line. Adequate biopsies were those with at least two high-power fields of intact epithelium. The median number of adequate biopsies analyzed in both groups was five (range, three to seven).

The volume density of mucin in goblet cells in the airway epithelium and the surface area of the epithelial basal lamina were determined by point and intersection counting (cycloid grid, Stereology Toolbox) (Figure 2). The number of goblet cells per volume of epithelium was calculated as follows: Nv_gob,epi = Vv_gob,epi/V_gob, where Vv_gob,epi is the volume density of goblet cells in epithelium and V_gob is the number-weighted volume of the goblet cell (22). The number-weighted volume of 20 goblet cells per biopsy (randomly sampled) was measured with a rotator, a local volume estimator in the C.A.S.T-Grid system (23).

View larger version (12K): [in this window] [in a new window]   Figure 2.   The volume of stored mucin in goblet cells in the airway epithelium in 13 subjects with asthma and 12 healthy subjects. The volume density of mucin stored in goblet cells in the airway epithelium was determined by point counting, using a cycloid grid consisting of points and lines (Stereology Toolbox). Points within the grid falling on areas staining positive by Alcian blue and PAS were counted as points on mucin (Ptmuc). Points within the grid falling on epithelial cells of all cell types were counted as points on epithelium (Ptepi). Lines within the grid intersecting the basal lamina were counted as lines intersecting the basal lamina (Lbl). The volume of mucin per volume of epithelium (Vvmuc,epi) was calculated as The surface of basal lamina per unit volume of epithelium (Svbl,epi) was calculated as where l/Pt is the length per test point. The volume of mucin per unit surface area of basement membrane (in mm3/mm2) was calculated as

Immunohistochemistry Studies

The MUC5B-like antibody (11C1; diluted 1:100) was generated from a hybridoma in which the immunogen was a secretory product of primary human tracheobronchial epithelial cells, and the specificity of the antibody was demonstrated by ELISA and Western blot. 11C1 is an IgG1 monoclonal antibody specifically reactive to submucosal gland cells rather than surface epithelial cells. Western blot demonstrated the recognition of the same mucin molecule recognized by the 17Q2 antibody. However, ELISA suggested that the 11C1 activity is blocked by a synthetic peptide representing the naked region of human MUC5B. In addition, the epitope was sensitive to protease treatment, suggesting that the antibody recognizes the unglycosylated MUC5B peptide. The ELISA was carried out in two ways. One involved the use of a sputum mucin-coated plate for the 11C1 reaction; before the reaction, the antibody was preincubated with 5 µg of MUC5B peptide. The competition was ~ 50%. The second ELISA approach used MUC5B peptide to coat the plate, which was then reacted with 11C1 antibody. 11C1 reacts with MUC5B in this solid phase whereas the 17Q2 antibody does not react. Subsequently, Western blots (in 1% agarose gel) were used and both 17Q2 and 11C1 recognized the same mucin species (as represented by smears in the very high molecular weight region). However, in additional Western blot experiments 17Q2 recognized protease-treated sputum mucin (purified from a CsCl gradient and two cycles of gel filtration of the void volume region), but 11C1 did not, indicating that 11C1 recognizes the naked peptide region of sputum mucin.

The MUC1 monoclonal antibody (1:50 dilution) was purchased from Neomarkers (clone GP1.4; Neomarkers, Fremont, CA); data from the manufacturer indicate that this IgG1 antibody was raised against human milk fat globule membranes and is produced by a fusion between BALB/c splenocytes and mouse myeloma P3-X63-Ag8.653 cells. In Western blotting it recognizes two glycoproteins in the molecular weight range of 265K-400K, identified as the epithelial membrane antigen, or mucin 1 (MUC1). MUC1 antibody reacts with the DTRP epitope in the tandem repeats.

The MUC2 monoclonal antibody (1:50 dilution) was also purchased from Neomarkers (clone CCP58); data from the manufacturer indicate that this IgG1 antbody was raised against a synthetic peptide of 29 amino acids (KYPTTTPISTTTMVTPTPTPTGTQTPTTT) from MUC2 protein, coupled to keyhole limpet hemocyanin (KLH). It is produced by a fusion between BALB/c splenocytes and mouse p3-NS-1/Ag4-1 (NS1) cells. It recognizes a single protein of 520 kDa, identified as MUC2 glycoprotein. For MUC2 an additional rabbit polyclonal antibody was used (anti-MRP, 1:50 dilution, courtesy of J. Gum, UCSF).

The MUC5AC monoclonal antibody (1:50 dilution) was also purchased from Neomarkers (clone 1-13M1); data from the manufacturer indicate that this IgG1 antbody was raised against mucin M1from the fluid of an ovarian mucinous cyst from an O Le(a-b-) patient. It is produced by a fusion between BALB/c splenocytes and mouse myeloma Sp2/0 cells. It recognizes the peptide core of gastric mucin M1.

For the immunohistochemistry experiments, preimmune horse serum and a mouse purified plasmacytoma IgG1 (1:50 dilution, MOPC-31c; Sigma) were used as controls. Antigen retrieval was achieved by trial incubations with untreated, trypsin-treated, or sodium citrate-treated sections. Sections for MUC5AC and MUC5B staining were incubated with trypsin (Zymed Laboratories, South San Francisco, CA) at a 1:1 dilution for 45 min at 37° C, whereas sections for MUC2 staining were boiled in 10 mM sodium citrate buffer, pH 6, for 15 min and a deglycosylation step with 1% periodic acid was also included in some experiments (24). No antigen retrieval step was used for MUC1 staining. Nonspecific binding was blocked by incubation in 1.5% horse serum. Primary antibody incubation was for 16-24 h, biotinylated secondary antibody (Vector Laboratories, Burlingame, CA) incubation was for 2 h, avidin-peroxidase antibody complex (Vector Laboratories) incubation was for 2 h, diaminobenzidine (DAB) chromogen (Zymed Laboratories) incubation was for 10 min, and counterstaining with Gill's hematoxylin (Sigma) was for 15 s. The intensity of immunostaining was scored visually (0 to 5+ scale).

Measurement of Mucin Gene Expression

Biopsies were immersed in 20% sucrose at 4° C for 2-4 h and placed in O.C.T. compound (Sakura Finetek U.S.A., Torrance, CA) before being snap frozen in liquid nitrogen. A 5-µm section from each biopsy was stained with Gill's hematoxylin and biopsies with intact epithelium but lacking submucosal gland tissue were selected; 23 of the 25 subjects had biopsies with these criteria. Total cellular RNA was obtained from these biopsies by trimming off the O.C.T. compound, sonicating the biopsy in 0.5 ml of RNAzol (Tel-Test, Friendswood, TX), and isolating RNA according to the method of Chomczynski and Sacchi (25). RNA samples were rigorously tested using -RT (reverse transcriptase) controls in order to evaluate RNA contamination with DNA. In most cases the difference between -RT and +RT data points was within 10-20 threshold cycles (C_t, the threshold cycle in real-time polymerase chain reaction [PCR] that reflects mRNA copy numbers in a given RNA sample), indicating less than 0.1% contamination with genomic DNA. Four of 23 patient samples were judged to have excessive genomic DNA contamination. In samples from the remaining 8 healthy subjects and 11 subjects with asthma transferrin receptor delta C_t values (+RT versus -RT controls) were at least 10 and did not interfere with RNA analysis. Duplicate biopsies were available from 3 of the 8 healthy subjects and from 4 of the 11 subjects with asthma; these duplicates were also analyzed and averaged.

The expression of mucin genes was measured by using real-time quantitative PCR (26). Existing procedures were modified to improve the sensitivity of the assays and to allow simultaneous measurements of nine mucin genes. Specifically, a two-step procedure was designed that involved multiplex gene-specific PCR (nine mucin RT-PCR primer sets in addition to a housekeeping gene and other genes related to asthma and atopy) followed by real-time PCR on the generated cDNA product. The procedure was based on the PE Biosystems (Foster City, CA) ABI Prism 7700 sequence detection system, using a 96-well plate format. This approach was validated with 34 sets of genes, and no skewed representation of mRNAs was observed even after 25 cycles of RT-PCR amplification of cDNAs (data not shown). All the reagents including TaqMan universal PCR master mix and optical plates were purchased from PE Biosystems. The primers and probes for RT and TaqMan were designed with Primer Express software (PE Biosystmes), based on sequencing data from National Center for Biotechnology Information (NCBI, NIH) databases (Tables 2 and 3) and purchased from Biosearch Technologies (Novato, CA). Both sets of the primers were nested, and RT-PCR products were within a range of 250 bp. All PCR conditions were as recommended by the manufacturer, and the results are expressed as the number of mucin gene copies normalized to the number of the transferrin receptor gene copies (the housekeeping gene with constant expression in our system; in preliminary experiments the expression levels of transferrin receptor gene correlated well with other housekeeping genes, including glyceraldehyde-3-phosphate dehydrogenase [GAPDH]). The copy number of each gene was derived using calibration curves, where C_t values (threshold cycle in real-time PCR that reflects mRNA copy numbers in a given RNA sample) were plotted against amplicon concentration in copy numbers (K. Livak, Sequence Detector User Bulletin 2, PE Biosystems) (26). Dilution curves confirmed the linear dependence of the C_t signal on the concentration of template RNAs, and revealed that the linear response of the TaqMan real-time PCR was most reliable when the C_t values were less than 38 (> 1 copy per reaction).

Statistics

All measurements were made on coded samples that hid the identity of the subject. Data were entered into a computer spreadsheet program (Microsoft Excel; Microsoft, Redmond, WA) and exported to a statistics program (StatView; Abacus Concepts, Berkeley, CA) for descriptive and comparative statistics. The Mann-Whitney U test was used to compare data on goblet cell mucin stores in the airway epithelium and mucin concentrations in induced sputum. The unpaired t test was used to compare the copy numbers of genes for the nine mucin genes in the two populations. The Spearman rank order test was used to determine correlations between data. A probability value of < 0.05, using two-tailed tests, was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Stored mucin in goblet cells, expressed as the volume density of mucin in the airway epithelium, was higher in the subjects with asthma than in the healthy subjects. Although there was overlap between the two groups for this outcome, the mean value for stored mucin was three times higher than normal in the subjects with asthma (Table 4, Figure 1, and Figure 2). The results were similar whether the volume of epithelial cells or the surface area of the basal lamina was used as the reference compartment (Table 4). The goblet cell size was similar in both groups whereas the goblet cell number was significantly increased in the subjects with asthma (Table 4 and Figure 3), indicating that increase in stored mucin in goblet cells was secondary to goblet cell hyperplasia. The volume of the epithelium per surface area of basement membrane was similar in healthy subjects and subjects with asthma (Table 4), indicating that the epithelium as a whole was not hypertrophied.

View larger version (113K): [in this window] [in a new window]   Figure 1.   Photomicrographs of representative histologic sections from endobronchial biopsies from a healthy subject (left) and a subject with asthma (right). The biopsies were taken from large central airways, fixed in 4% paraformaldehyde, embedded in glycomethacrylate, cut into 2-µm sections, and stained sequentially with Alcian blue, periodic acid-Schiff reagent, and hematoxylin. Mucin stores in goblet cells appear as a mixture of blue and purple staining. The sections demonstrate increased Alcian blue and PAS staining in the section from the subject with asthma. Original magnification: ×200.

View larger version (11K): [in this window] [in a new window]   Figure 3.   Left: Comparison of the volume of mucin per volume of goblet cells in 12 healthy subjects and 13 subjects with asthma demonstrates that goblet cell size is similar in healthy subjects and subjects with asthma. Right: Comparison of goblet cell number in 12 healthy subjects and 13 subjects with asthma demonstrates an increase in goblet cell number in the subjects with asthma. Data represent means ± SD; an asterisk (*) denotes significantly greater than in healthy subjects, p < 0.005.

Six of the subjects with asthma had mild asthma (FEV₁  80%pred) and seven had moderate asthma (FEV₁ < 80%pred). Surprisingly, the amount of stored mucin in goblet cells was higher in the subjects with mild asthma, although this difference was not significant (Figure 4). In contrast, the levels of secreted mucin in induced sputum were significantly higher in the subjects with moderate asthma than in the subjects with mild asthma (Figure 4). Furthermore, the levels of secreted mucin were inversely correlated with the levels of stored mucin in goblet cells for the asthma group as a whole (r_s = 0.78, p = 0.007). Secreted mucin levels varied widely in the healthy subjects; the median mucin concentration in induced sputum from the 12 healthy subjects was approximately half that of the 13 subjects with asthma (19.6 [interquartile range; 9.9-60.4 µg/ml] versus 34.0 [29.9-56.0 µg/ml], p = 0.17).

View larger version (13K): [in this window] [in a new window]   Figure 4.   The volume of stored mucin in goblet cells in the airway epithelium in six subjects with mild asthma and seven subjects with moderate asthma (left), and the level of mucin glycoprotein in induced sputum in the same subjects with mild and moderate asthma (right). Data represent means ± SD; an asterisk (*) denotes significantly greater than subjects with mild asthma, p < 0.05.

Analysis of relative mucin gene-specific mRNA copy numbers in homogenates of endobronchial biopsies from both groups showed, first, that MUC5AC was the most frequently expressed mucin gene in both groups (Figure 5). Taking the sum of mucin gene-specific mRNA copy numbers in biopsy homogenates from both groups, the majority were for MUC5AC (median, 76.5%; interquartile range, 64.2-78.4%). The corresponding values for MUC1, MUC2, MUC3, MUC4, MUC5B, MUC6, MUC7, and MUC8 were 11.2 (7.6-14.1%), 1.0 (0.1-2.7%), 0.0 (0.0-0.0%), 5.7 (3.7-8.6%), 5.3 (0.9-8.7%), 0.1 (0.0-0.2%), 0.01 (0.0-0.04%), and 0.04 (0.01-0.08%), respectively. MUC5AC gene expression was increased in the subjects with asthma, but not significantly so; the expression levels of the MUC2 and MUC4 genes were significantly increased in the subjects with asthma (p < 0.02), whereas the expression of MUC5B was significantly decreased (p = 0.004) (Figure 5).

View larger version (18K): [in this window] [in a new window]   Figure 5.   Expression of nine different mucin genes in homogenates of endobronchial biopsies from 8 healthy subjects and 11 subjects with asthma. Gene expression was measured by real-time RT-PCR, and the expression of mucin genes is represented as the ratio of mucin gene-specific mRNA copy numbers to those of the transferrin receptor (housekeeping gene). Data represent means ± SD; an asterisk (*) denotes mRNA expression in subjects with asthma is significantly greater than in healthy subjects, p < 0.005.

MUC1, MUC5AC, and MUC5B proteins were detectable by immunohistochemistry in both healthy subjects and subjects with asthma, and staining was localized to goblet cells (Figure 6); the levels of MUC1 and MUC5AC were increased as assessed by semiquantitative methods (Table 5). MUC2 glycoprotein was not detectable immunohistochemically in either group by our methods (Table 5 and Figure 6). The lack of available specific antibodies prevented examination of the expression of MUC3, MUC4, MUC6, MUC7, or MUC8 at the protein level.

View larger version (92K): [in this window] [in a new window]   Figure 6.   Photomicrographs of immunostaining of representative sections from endobronchial biopsies from healthy subjects (left) and subjects with asthma (right). Murine monoclonal antibodies directed against specific mucin gene products were used in experiments using the ABC- immunoperoxidase method (DAB substrate). The counterstain was hematoxylin. The protein products of four mucin genes were analyzed: MUC1 (A and B), MUC2 (C and D) MUC5AC (E and F ), and MUC5B (G and H ). The MOPC-21 antibody was used for the isotype control experiments (I and J ). Bar: 10 µm. Original magnification: ×200.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our hypothesis that epithelial mucin stores are increased in mild and moderate asthma proved correct, and we determined that goblet cell hyperplasia, not hypertrophy, accounted for the increase. Surprisingly, however, our hypothesis that secreted mucin in the airway lumen would be increased in proportion to these abnormalities proved incorrect. In fact, there was an inverse rather than a direct relationship between stored mucin in the epithelium and secreted mucin in induced sputum. Stored mucin levels were similarly increased in the mild and moderate asthma subgroups, but secreted mucin was increased only in the moderate subgroup.

To our knowledge, our finding of significant goblet cell hyperplasia in subjects with mild and moderate asthma has not been described previously. Most previous reports that have described goblet cell abnormalities in asthma have been postmortem histologic studies of patients who died of acute severe asthma. We are aware of only one published study that has analyzed airway goblet cells in endobronchial biopsies from subjects with mild asthma and healthy control subjects. In that study, Lozewicz and coworkers (29) counted the number of goblet cells per 20 epithelial cells in two areas of epithelium stained with hematoxylin and found that the number of goblet cells was no different in healthy subjects and subjects with asthma. In contrast, our study finds clear evidence of a more than 2-fold increase in the number of goblet cells in subjects with mild and moderate asthma. Methodological differences probably explain the differences in results in our studies. We applied a design-based, stereological approach to the biopsy analysis, and we measured multiple outcomes related to goblet cells in multiple biopsies from each subject. We discovered that mucin stores in goblet cells in the airway epithelium were 3-fold higher than normal in the subjects with asthma and that this abnormality was not explained by goblet cell hypertrophy. It was an increase in goblet cell number that explained the increase in epithelial mucin stores in the subjects with asthma.

In the subjects with asthma the severity of disease, as assessed by FEV₁ %pred, was not correlated with the goblet cell mucin stores in the airway epithelium, and some subjects with asthma had values for mucin stores within the range of the healthy control subjects. The mean value for stored mucin in the subjects with mild asthma was higher than that of the subjects with moderate asthma. This finding may have at least two pathophysiologic implications. First, increased goblet cell mucin stores in mild asthma raises the possibility that acute degranulation of goblet cells leading to mucin hypersecretion might be an important mechanism for asthma exacerbations in these subjects. This mechanism might even explain the finding that 15-30% of patients who die from acute severe asthma are found to have histories suggesting mild disease prior to their fatal attack (30). Second, the finding that goblet cell mucin stores in moderate asthma are higher than normal but not higher than in mild asthma suggests that chronic goblet cell degranulation may be occurring in moderate asthma. This possibility is supported by our finding that mucin levels in induced sputum were higher in the subgroup of subjects with moderate asthma than in the subgroup with mild asthma. Chronic mucin hypersecretion secondary to goblet cell degranulation may therefore contribute to the pathophysiology of chronic airway narrowing in subjects with moderate asthma.

The most frequently expressed mucin gene in both healthy subjects and subjects with asthma was MUC5AC. The relative mucin-specific mRNA copy number for MUC5AC was approximately 60% higher than normal in the subjects with asthma. Although this increase was not significant in this relatively small sample size, the increase is consistent with the increase in MUC5AC protein expression found by immunohistochemistry. Taken together, these findings implicate MUC5AC as the principal airway mucin in both health and asthma and suggest that upregulation of MUC5AC may account for increased mucin stores in asthma. This upregulation of MUC5AC may be mediated by helper T cell type 2 cytokines, especially IL-4, IL-13, and IL-9 (31).

Our finding that the mucin genes MUC2 and MUC4 were overexpressed at the mRNA level is difficult to interpret, because the expression levels of these genes were much lower than that of MUC-5AC. In addition, MUC2 protein was not detectable, and lack of an available antibody prevented immunohistochemical studies of the MUC4 protein. Although our inability to detect MUC2 protein expression may be a methodological problem, the finding is consistent with that of Hovenberg and coworkers (35) and Sheehan and colleagues (36), who found no evidence of MUC2 protein in airway tissue or secretions from healthy subjects or subjects with asthma.

In summary, we found that even mild asthma is associated with goblet cell hyperplasia and increased stored mucin in the airway epithelium. Moderate asthma has increased levels of stored mucin similar to mild asthma but higher levels of secreted mucin. We speculate that acute degranulation of hyperplastic goblet cells may represent a mechanism for asthma exacerbations in mild and moderate asthma and that chronic degranulation of goblet cells may contribute to chronic airway narrowing in moderate asthma. Therefore, specific targeting of goblet cell abnormalities in asthma may represent a novel strategy for decreasing symptoms and exacerbation rates and for reversing airway narrowing.