Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Although chest disease is primarily caused by environmental triggers such as foreign antigens, viruses, bacteria, and pollutants, there are clearly person-to-person differences in the sensitivity to these agents (reviewed in Reference 1). It has, for example, been known for a long time that there is familial clustering of asthma. Asthma is characterized clinically by chronic, intermittent airway obstruction with wheezing, coughing, and breathlessness, and the production of excess mucus. Many, but not all, smokers develop chronic bronchitis where there is also excessive production of mucus. The normal role of this mucus is to lubricate and protect the epithelia of the respiratory tract and to clear particulate matter together with the action of the epithelial cilia (2). The major macromolecular constituents of mucus are mucins that are the highly glycosylated glycoproteins produced, in the case of the airways, by the surface epithelial cells and mucus glands of the bronchial tree.

To date, 12 genes that encode epithelial mucins have been described (5). The family of four genes on chromosome 11p15.5 (MUC2, MUC5AC, MUC5B, and MUC6) is thought to encode the major gel-forming mucins, which give mucus its viscoelastic properties (13). The MUC1, MUC3, and MUC4 glycoproteins and that encoded by the newly described gene MUC12 are apparently, at least in part, membrane associated (12, 18) and probably have a dual function of protection as well as signaling cell-to-cell contact and possibly differentiation (19). MUC7 encodes a low molecular weight-secreted mucin that is present in saliva (22, 23). MUC8 has been reported to be expressed in the bronchial tree (10), but these observations have not been substantiated by any other group. MUC9 encodes a low molecular weight glycoprotein that is expressed in the oviduct (11).

Most of the MUC genes tested have been found to show a high level of genetically determined polymorphism (5, 7, 24, 25). The polymorphism is mainly due to variation in the number of copies of tandemly repeated sequence (VNTR) within a large domain, which composes a major part of their coding region. The tandem repeats along the length of the array are very similar to each other but not identical and vary in length from 24 nucleotides in MUC5AC to 507 nucleotides in MUC6 (26). They are generally a multiple of three so that they encode repeats of an amino-acid sequence. In some cases, the variation in repeat numbers is so great that it can be responsible for a twofold difference in the length of the coding sequence and size of the protein (14, 19, 25).

There is considerable evidence that mucins are overexpressed as a secondary consequence of inflammatory disease (27, 28). However, we are interested in determining whether genetic differences in allele length or sequence affect susceptibility to inflammatory disease.

It has been shown by analysis of messenger RNA (mRNA) or protein, by several groups, and by a variety of means that MUC1, MUC4, MUC5AC, and MUC5B are all expressed in the bronchial tree (16, 17, 29). The expression of MUC2 is more controversial. MUC2 mRNA was not detected by Northern blot analysis of normal bronchus (34) but was cloned from a complementary DNA (cDNA) library and detected by Northern blots in RNA from a chronic bronchitic patient (35). Here, we confirm the expression of MUC1, MUC4, MUC5AC, and MUC5B as well as showing MUC2 in both inflamed and noninflamed tissue at both the mRNA and protein levels.

MUC1, MUC2, MUC4, and MUC5AC, which all show length variation, were thus all clear candidates for the study of chest disease susceptibility. Although our previous studies showed no evidence of allelic variation in the number of the tandem repeats of coding sequence of MUC5B in healthy individuals (25), it seemed possible that patients with severe asthma might have rare variant length alleles of MUC5B that might account for the large accumulation of MUC5B described in a patient who died in status asthmaticus (15). We have also examined a polymorphic locus in intron 36 of MUC5B (36). This region, which also comprises tandem repeats of sequence, has been shown to contain a protein binding site for a 42-kD protein (NF1-MUC5B) (37) and may be involved in the regulation of MUC5B. Thus, it seemed possible that variation in the number of tandem repeats of this binding site would affect the level of MUC5B expression.

Here we report our polymorphism data for MUC1, MUC2, MUC4, MUC5AC, and MUC5B in asthma. The majority of patients with asthma are also atopic, and it is known that genes such as the immunoglobulin (IgE) receptor (1) that are involved in atopy are likely to influence asthma susceptibility. However, not all atopic individuals develop asthma and other genetic factors are clearly involved. We have compared atopic individuals with and without asthma to specifically test the hypothesis that mucins affect the susceptibility of atopic individuals to asthma.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Patients' Blood Samples

Patients in this study were recruited from St. Mary's Hospital Chest and Allergy Clinic, a smoking addiction and general practice clinic, and fell into four categories: patients with asthma, patients with allergic conditions but not asthma, smokers with chronic obstructive lung disease, and smokers without chronic obstructive lung disease. We also tested healthy volunteers (laboratory personnel and spouses of patients). All gave fully informed consent and all the hospital patients and most of the control subjects completed questionnaires on smoking history, atopy, and breathlessness. Atopy was defined by positive skin prick tests (two or more positive reactions greater than the histamine control or one very large reaction) and generally confirmed by raised IgE levels (more than 120 IU/liter). Full spirometry (peak expiratory flow rate [PEFR], forced expiratory volume [FEV₁], and vital capacity [VC]) was done on all the atopic individuals. Patients were classified on the basis of their medical history (obtained from their notes) and questionnaire response as well as spirometry and atopy tests. Asthma was diagnosed clinically as follows: on the basis of a clear and characteristic history of episodic breathlessness, chest tightness, and wheeze either occurring spontaneously or as a result of exposure to characteristic inciting agents (allergens, irritants, exercise, or infection), and spontaneous resolution or a resolution after bronchodilator therapy; and/or demonstration of a 20% or greater fluctuation in serial peak flow measurements; and/or 20% or greater improvement in FEV₁ or peak flow after administration of bronchodilator in the clinic. The nonasthmatic groups comprised 34 patients referred for allergic rhinitis, 10 for urticaria, three for eczema, and three for food allergies, and had no recent or confirmed medical history of asthma. A total of 202 atopic volunteers were recruited from the St. Mary's clinic. From this cohort, we selected 50 pairs of age- and sex-matched atopic, asthmatic and atopic, nonasthmatic individuals of Northern European descent. The smaller nonasthmatic set was matched in turn for age and sex with individuals from the larger asthmatic set without consultation of other clinical data or smoking history. Thereafter, the complete set of 86 Northern European patients with asthma was also classified in terms of severity on clinical grounds, based on the nature of drug treatment received. Severe asthmatics were defined as those treated with regular oral steroids in addition to standard asthma medication, moderate asthmatics as those with occasional steroid use in addition to -agonists and/or inhaled steroids, and mild as those usually treated with -agonists alone. PEFR, FEV₁, and VC obtained for the matched series and nonmatched atopic asthmatics are shown in Table 1.

Tissues

Surgical specimens of lung and bronchus were taken from normal areas of lobectomies from patients with lung cancer. Bronchial biopsy sections were from routine diagnostic bronchoscopies. The histologic assessment made on parallel sections was taken from the histopathology report. All samples were obtained from St. Mary's Hospital.

Ethics

Ethical committee permission was obtained for this study from Parkside Health Authority (Ethical Committee no. 2893).

Polymorphism Analysis

Genomic DNA samples were prepared from whole blood using a Puregene kit (Flowgen, Sittingbourne, UK). A total of 5 to 7 µg of DNA was treated with HinfI or PvuII in a final volume of 25 µl (as recommended by the manufacturer [GIBCO-BRL, Life Technologies Ltd., Paisley, UK]). The HinfI fragments were separated by agarose electrophoresis (0.8% agarose in 0.89 M Tris, 0.1 M borate, 0.002 M ethylenediaminetetraacetic acid [EDTA] buffer, pH 8.3 [TBE]) on 20 × 25 cm gels with 30 samples per gel (Horizon 20:25 apparatus; GIBCO-BRL) for 24 h at 2 V/cm. The PvuII fragments were separated using 0.5% agarose, 1× TBE, and run at 2 V/cm for 24 h, followed by a complete change of the tank buffer and continued electrophoresis at 1.2 V/cm for a further 19 h. Three kinds of markers were applied to each gel: Raoul markers (Appligene, Durham, UK), 1-kb ladder (GIBCO-BRL), and lamda HindIII (GIBCO-BRL), and were detected with ethidium bromide. The Raoul markers were also visible after transfer because they were usually revealed with the MUC probes. Three DNA samples with alleles of known size were also used as further size standards, with two of these samples being mixed and loaded in the same well.

After electrophoresis, the markers were visualized by poststaining with ethidium bromide. After measuring the migration of the marker bands, the gels were depurinated (0.25 M HCl for 30 min), denatured (1.5 M NaCl, 0.5 M NaOH for 30 min), neutralized (0.5 M Tris HCl, 0.001 M EDTA, pH 7.2, for 30 min), and the digested DNA was then transferred onto Hybond N+ membranes (Amersham Pharmacia Biotech, Buckinghamshire, UK) by capillary blotting overnight or vacuum blotting (Vacugene XL; Amersham Pharmacia Biotech) for 2 h, all as recommended by the manufacturers. The DNA was then fixed onto the filters by baking at 80°C for at least 2 h.

Probing and washing of the filters were conducted by standard procedures. The MUC genes were detected using tandem repeat cDNA probes: PUM24P for MUC1 (38), SMUC41 for MUC2 (39), JER64 for MUC4 (40), JER58 or JER 47 for MUC5AC, and JER57 for MUC5B (41). A total of 25 ng was labeled by random primed labeling using the Multiprime DNA labeling kit (Amersham Pharmacia Biotech). Filters were prehybridized in 200 ml of 6× saline sodium citrate (SSC) (stock, 20× SSC = 3 M NaCl, 0.3 M trisodium citrate), 5× Denhardts (100× Denhardts = 2% wt/vol Ficoll, 2% wt/vol polyvinylpyrrolidone, 2% wt/vol bovine serum albumin [BSA]), 0.5% wt/vol sodium dodecyl sulfate (SDS) in a shaking water bath at 65°C for approximately 4 h, and 500 µg of sonicated Herring sperm DNA and labeled probe boiled together for 5 min were added to the prehybridization solution for overnight hybridization. The filters were washed down in several changes of SSC, with a final stringent wash of 0.1× SSC, 0.1% SDS at 65°C for 10 min. Autoradiography was carried out with Fuji HRE-30 Medical X-ray film and Fuji screens (FG8) for 1 to 7 d at 70°C. The HinfI filters were sequentially probed for MUC1, MUC2, and MUC5AC, whereas the PvuII filters were probed for MUC4 and MUC5B. Thus, the invariant MUC5B provides an additional mobility control for MUC4, whereas MUC1, MUC2, and MUC5AC provide mobility controls for each other.

The position of the wells was marked onto the autoradiographs by using luminescent marks (Glo-bug X-ray marking solution; Radleys, Saffron Walden, UK) to line up the filter. The autoradiographs were recorded using a UVP image analysis system (Cambridge, UK) and the alleles were sized in comparison with markers using the Gelworks Image Analysis software package (UVP). The standard curve was plotted using the MUC alleles of known size as well as the Raoul markers, the other markers being used as an extra check for gel quality. MUC5AC, which shows simpler allelic variation, was analyzed as two size class alleles as described previously (25) that were named MUC5AC*a and MUC5AC*b (6.6 and 7.4 kb, respectively).

The intronic MUC5B polymorphism was examined by polymerase chain reaction (PCR) using primers that amplify across the entire VNTR region. Each PCR reaction for MUC5B contained 2 µl 10× PCR buffer (500 mM KCl, 100 mM Tris-HCl, pH 9.0, 15 mM MgCl₂, 0.1% Triton X-100), 2 µl 2 mM deoxynucleotide triphosphates, 1 µM sense and antisense primers, 0.5 µl (0.25 µg) DNA, 0.1 µl (0.5 U) Taq polymerase (Promega, Madison, WI), and 3 µl 50% (vol/vol) glycerol, with added water to a final volume of 20 µl. After a 3-min denaturation at 94°C, 30 cycles were performed consisting of a 30-s denaturation at 95°C, 30-s annealing at 64°C, 2-min extension at 70°C, and a final extension of 6 min at 72°C (sense primer, 5'-AGTGTGCAGTGACTGGCGAG-3' nt 3967-3986 in Genbank Y09788; antisense primer, 5'-CTAGAGTTGCAGGTGGCAGG-3' nt 4655-4674 in Genbank Y09788). The PCR products were electrophoresed on 2% agarose gels and visualized with ethidium bromide staining under ultraviolet light.

In Situ Hybridization

The 48-mer oligonucleotide probes that were used were as described by Audie and colleagues (29) and are designed from the tandem repeat region of each of the mucin genes. A 48-mer oligo dT probe (which hybridizes to polyA mRNA tails) was used as a positive control to test for the quality of the tissue. A total of 100 pmol of each probe was labeled with digoxygenin (DIG) using a standard oligonucleotide 3'-tailing reaction and the efficiency of the labeling reaction was estimated in accordance with the manufacturer of the DIG oligonucleotide tailing kit (Boehringer Mannheim, Indianapolis, IN).

Six-µm sections from paraformaldehyde or formalin-fixed, paraffin wax-embedded tissue were mounted onto chrome gelatin- treated slides and left to dry completely. The slides were then rehydrated and incubated in 0.1 M glycine, 0.2 M Tris-HCl, pH 7.4, for 10 min followed by 1 µg/µl proteinase K (Boehringer Mannheim) in 0.1 M Tris-HCl, pH 8.0, 50 mM EDTA at 37°C for 30 min and then washed in distilled water for 5 min. The sections were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS), pH 7.4, for 15 min at room temperature, washed in PBS for 5 min, and then treated with 0.25% acetic anhydride, 0.1 M triethanolamine, pH 8, in 4× saline sodium phosphate EDTA (SSPE) for 10 min. The sections were washed again in distilled water for 5 min, prehybridized in 4× SSPE (Promega), 1× Denhardts solution (Sigma, St. Louis, MO) for 1 h at 42°C, dehydrated by passing through a series of increasing concentrations of alcohol, and then air-dried. The appropriate DIG-labeled probe was added to the hybridization solution (16.2 mM dithiothreitol [DTT], 3.3× SSPE, 42% formamide, 0.08 M NaH₂PO₄, 0.83× Denhardts, 0.83% sarkosyl, 6.25 µg/ml transfer RNA) to give a final concentration of 5 µl probe/ml hybridization mix, and 100 µl was applied to the sections. Coverslips were applied to each slide and placed in a humid chamber overnight at 42°C. Posthybridization washes were performed as follows: 4× SSPE, 0.2 mM DTT for 30 min, 4× SSPE for 30 min, 1× SSPE three times for 15 min, 1× SSPE for 30 min at 42°C, and finally two washes of 0.1× SSPE at 42°C for 30 min each. The slides were then washed in 30, 70, and 100% ethanol for 5 min each and air-dried. A blocking solution of 1% BSA (Sigma)/0.25% horse serum in PBS was applied to the sections and incubated for 3 h at room temperature. They were then washed in PBS for 5 min before -DIG-alkaline phosphatase antibody (diluted 1:1,000 in block; Boehringer Mannheim) was applied under a coverslip and left overnight at 4°C. This was followed by four changes of PBS for 10 min each before treatment with substrate buffer (0.1 M Tris-HCl, pH 9.5, 50 mM MgCl, 0.1 M NaCl) for 5 min. Detection buffer (0.33 mg/ml nitroblue tetrazolium (NBT) [Promega], 0.17 mg/ml 5-bromo-4-chloro-3-indolyl phosphate (BCIP) [Promega] in substrate buffer) was then added to each section and the slides left in the dark at room temperature overnight. The slides were then washed for 20 min in several changes of PBS, rinsed in distilled water, and mounted in an aqueous mountant.

Immunohistology

Antibodies. The monoclonal antibody LICRLonM8, obtained from the Lister Institute for Cancer Research (London, UK), was used to detect MUC1. This antibody recognizes MUC1 tandem repeat sequence both in the glycosylated and unglycosylated state (42, 43). The monoclonal antibody 996/1, used for MUC2, recognizes the repeat unit of MUC2 in the newly synthesized glycoprotein (44, 45) and was kindly supplied by L. Durrant and M. R. Price (Nottingham University, UK). The polyclonal antiserum for MUC4 was raised against synthetic peptides corresponding to the tandem repeat sequence and was purified by affinity chromatography (A. Lopez-Ferrer, V. Curull, C. Barranco, M. Garrido, J. Lloreta, F. X. Real, C. de Bolos, unpublished observation) and recognizes MUC4 on immunohistology of airway tissue without any deglycosylation treatment (46). The polyclonal sera for MUC2 (LUM2-3) and MUC5AC (LUM5-1), which recognize nontandem repeat domains of these molecules and thus fully glycosylated as well as unglycosylated mucin, were kindly supplied by I. Carlstedt (Lund University, Sweden) (14, 16).

Six-µm sections were mounted onto TESPA (3-aminopropyltriethoxysilane; Sigma) treated slides and left to dry. The sections were then rehydrated. After a 5-minute wash in PBS, the sections were treated in 1% H₂O₂/methanol for 30 min at room temperature. The sections were then rinsed in distilled water and PBS for 5 min each and incubated in goat serum (1:5 in PBS) for 15 min. Primary antibody was added at an appropriate dilution, and the sections were covered with a coverslip and left at room temperature for 1 h. After rinsing in PBS for 5 min, biotinylated goat  mouse secondary antibody (DAKO, Ely, Cambridgeshire, UK) was added at a dilution of 1:300 in PBS and the slides were left to incubate for 1 h. The sections were then treated with peroxidase-conjugated streptavidin (1:600 in PBS; DAKO) for 30 min after washing in PBS for 5 min. A final wash in PBS for 5 min was followed by development of the color reaction with 3-3'-diaminobenzidine (Sigma). The sections were then washed in distilled water for 5 min, counterstained in hematoxylin, washed in running water for 5 min, dehydrated, cleared in histoclear, and mounted in DPX (Merck Ltd., Poole, UK) mountant.

Families from the Centre d'Etude du Polymorphisme Humain (CEPH)

DNA samples from large two- and three-generation families were obtained from CEPH (47).

Statistics

Allele length distribution was compared by the Mann Whitney U test (48) using the SPSS computer package (SPSS Inc., Chicago, IL). In the case of the MUC5B intronic polymorphism where discrete alleles can be defined, a Monte Carlo method, CLUMP, was used (49).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of MUC Genes in the Bronchial Tree

Expression of MUC1, MUC2, MUC3, MUC4, MUC5AC, and MUC5B in the bronchial tree was studied in surgical specimens of bronchus and lung tissue and in bronchoscopy samples by immunohistochemistry and/or in situ hybridization (Table 2). Of these, only MUC3 was not detected. The pattern of expression of MUC1, MUC4, MUC5AC, and MUC5B was in broad agreement with other reports: MUC1 was mainly expressed in the alveoli in the pneumocytes; MUC5AC in mucus-secreting cells of the epithelial surface and some submucosal glands; MUC5B in the glands but not in normal surface epithelium; and MUC4 throughout the bronchial epithelial surface and in some submucosal glands. MUC2 mRNA was not expressed in alveoli or bronchiolar epithelium of the lung tissue but was present in some cells (both mucosecretory and ciliated) of the bronchial epithelium of all the biopsy samples and the one surgical sample tested. The pattern and extent of expression varied from weak staining of very occasional cells to a strong signal in a substantial proportion of the cells (Figure 1). Biopsies that showed inflammation seemed to show more staining. The same result was obtained with the antibody 996/1, which detects the MUC2 tandem repeat sequence within the MUC2 apoprotein and thus detects the newly synthesized, relatively unglycosylated MUC2 (Figure 1, Table 2) and was confirmed on a few samples with LUM2-3, which detects fully glycosylated MUC2 (data not shown).

View larger version (160K): [in this window] [in a new window]   Figure 1.   Expression of the MUC genes in an area of normal epithelium from a bronchial biopsy specimen. (A and B) Two different magnifications show staining with 996 antibody, which detects MUC2 apoprotein in serial sections from the same bronchial biopsy specimen. Note that staining is observed in the supranuclear region of the cells and represents newly synthesized (unglycosylated) MUC2. (C-F) In situ hybridization results for MUC2, MUC4, MUC5AC, and MUC5B are shown. Areas of granulomatous inflammation (not shown) were observed in this specimen.

VNTR Polymorphism of MUC1, MUC2, MUC4, and MUC5AC

We examined the distribution of allele length in the matched set of 50 atopic patients with asthma and 50 atopic controls. To control for any gel-to-gel variation, matched and unmatched control and patient groups were distributed across each of the gels. The results were plotted in histogram form grouping the fragment sizes in 500-bp steps as described previously (25). In the case of MUC5AC, most of the samples showed the common MUC5AC alleles (MUC5AC*a and MUC5AC*b) as described previously. One rare larger allele MUC5AC*c was detected in this cohort.

The data are shown in Figure 2 and some representative Southern blot results are shown in Figure 3. MUC1 showed a bimodal distribution of allele length as previously reported (50) but no difference between patients and control subjects. MUC4 showed a rather similar multimodal distribution of alleles in the two groups, though there were five rare large alleles in the atopic, nonasthmatic group (four of which are found in two apparent homozygotes) compared with two in the asthma group. The overall difference in length distribution was not statistically significant (P = 0.500, Mann Whitney test), but the occurrence of these rare large alleles is noteworthy. No difference in allele frequency was detected for MUC5AC, the smaller allele (MUC5AC*a) being the most frequent in both groups.

View larger version (26K): [in this window] [in a new window]   Figure 2.   Histograms showing the allele distribution for MUC1, MUC2, MUC4, and MUC5AC for the series of 50 atopic, asthmatic and 50 matched atopic, nonasthmatic individuals. For MUC1, MUC2, and MUC4, band sizes are calculated by reference to standards and binned in 500-bp groups as described. For each individual, two data points are obtained that correspond to the tandemly repeated length alleles on each of the chromosome homologues. In the case of tandemly repeated length homozygotes, the single band size is counted twice. The total number of chromosomes counted therefore corresponds to twice the number of individuals. For MUC5AC, which shows fewer alleles, the data are shown as genotypes (aa, ab, bb, ac). Solid bars = atopic, asthmatic individuals; open bars = atopic, nonasthmatic individuals.

View larger version (66K): [in this window] [in a new window]   Figure 3.   A representative experiment showing VNTR variation of MUC1 and MUC2. The two autoradiographs are from the same Southern blot filter probed with PUM24P for MUC1 and SMUC41 for MUC2. Atopic, asthmatic samples are labeled A and atopic, nonasthmatic samples, C. The arrows denote other laboratory samples included as internal size control alleles. The presence of multiple bands in one of these lanes is due to the deliberate loading of a mixture of two samples in the same well. The marker shown, M, is the Raoul marker.

However, a dramatic difference between the matched control and asthmatic groups is seen in the case of MUC2. The distribution of allele length shows that the mean and median lengths of MUC2 is smaller in the asthma group than in the atopic control group (P = 0.000, Mann Whitney test). When these distributions are compared with those of our Northern European healthy controls who are not defined for atopy or smoking status, and with the other chest disease groups (nonmatched asthma, bronchitic smokers, and smoker controls, all of Northern European extraction), it is clear that it is not the asthmatic group that is significantly different. Rather it is the atopic nonasthmatics that are significantly different from all other groups (Figure 4). This distribution was also compared with our previously published (51) and unpublished work on inflammatory bowel disease and gastritis. This showed that these data sets are also significantly different from the atopic nonasthmatics but do not differ significantly from the asthmatic group (data not shown). It can thus be seen clearly that it is the distribution in the atopic nonasthmatics that is unusual. We also examined the distributions of alleles in the complete group of asthma patients classified according to severity, but there was no observed difference in distribution (data not shown).

View larger version (36K): [in this window] [in a new window]   Figure 4.   Histograms showing the allele distributions for MUC2 in the 50 atopic, nonasthmatic (A) and 50 matched atopic, asthmatic individuals (B) in comparison with other groups. (C) Additional nonmatched atopic asthmatics are shown; (D) healthy volunteers; (E) bronchitic smokers; and (F ) nonbronchitic smokers. The P values shown in B to F are those obtained from comparison with the atopic nonasthmatics (data from A) using the Mann Whitney test. The binning and numbers of chromosomes are determined as described in Figure 2.

Search for Rare Allele Length Variants of MUC5B

A total of 229 patients (including 103 asthmatics) was tested for evidence of length variation in MUC5B by Southern blot analysis. This included 10 cases in which asthma was classified as severe. Although variants were found, digestion with other restriction enzymes (BglII, SacII/ HindIII) failed to reveal the same variant patterns, showing that they were not due to differences in length of the tandem repeat region but rather restriction site polymorphisms. Furthermore, these variants were observed in both patients and control subjects.

VNTR Polymorphism in Intron 36 of MUC5B

Using slightly different primers from those described by Desseyn and colleagues (36), we found the same alleles as those described previously in a French population (3, 5, 6, 7, and 8 repeat units). To determine whether this polymorphism can be considered as an independent genetic marker from MUC5AC, we examined possible association of the MUC5B intron 36 variation with the length variation of MUC5AC. Haplotypes were determined for all the unrelated chromosomes in the CEPH families. The other MUC genes had previously been tested on these families (25). MUC5B intron 36 VNTR was tested on sufficient individuals (grandparents and/or children) from the CEPH families in order to assign unambiguous haplotypes. The alleles occurred at rather similar frequencies to those observed in the French population (36). The distribution of MUC5B intron 36 alleles on chromosomes carrying the large and small alleles of MUC5AC is shown in Figure 5. No very obvious association is observed. The very slight difference in frequency was shown not to be statistically significant (P = 0.141 [10,000 iterations CLUMP test]).

View larger version (16K): [in this window] [in a new window]   Figure 5.   Histogram showing the distribution of alleles for the MUC5B intron 36 polymorphism (3, 5, 6, 7, or 8 tandem repeats) on chromosomes carrying the large alleles for MUC5AC (MUC5AC*b) and the small alleles for MUC5AC (MUC5AC*a). The number of chromosomes represents twice the number of individuals tested.

Analysis of the matched atopic series also showed no difference in the distribution of the MUC5B intron 36 alleles between the asthmatic and nonasthmatic groups (Figure 6). Tests on the extended data set of asthmatics also showed no difference between the groups classified in terms of severity (not shown).

View larger version (14K): [in this window] [in a new window]   Figure 6.   Histogram showing the distribution of alleles for the MUC5B intron 36 polymorphism in the matched atopic, asthmatic and nonasthmatic populations. The data are shown as genotypes (53, 73, 75, 76, 77, 78, 83), the numbers referring to the number of copies of tandem repeat detected as in Figure 5. Solid bars = atopic, asthmatic individuals; open bars = atopic, nonasthmatic individuals.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this study, we examined allelic variation of putative functional significance in the MUC genes as candidates for susceptibility to asthma. Our main hypothesis was that variation in length of the glycosylated tandem repeat domain is likely to affect mucin function by altering the physicochemical propertiesthrough the change in size and extent of glycosylation and thus the accessibility of the cell surface to bacteria and other environmental agents, as well as the consistency of the mucus. We selected genes shown by our own (Figure 1) and previous studies to be expressed in the bronchial tree.

In addition to testing VNTR variation within the coding region of these MUC genes, we examined an intronic polymorphism of MUC5B with a putative role in MUC5B expression. Our previous studies had shown that despite the location of MUC5AC, MUC2, and MUC6 in a gene cluster on 11p15.5, this is a recombination rich region, and there was little indication of allelic association between them (25). In this study, we tested for evidence of association between the MUC5B intronic polymorphism and MUC5AC to determine whether these polymorphisms could also be considered independent genetic markers for disease association studies. Although statistically significant association might be detectable on a larger data set, the data presented here indicate that there is not a high level of linkage disequilibrium.

No significant differences in allele distribution were found in the asthmatic and nonasthmatic groups for MUC1, MUC4, MUC5AC, and MUC5B. However, we report here for the first time a highly significant difference in the distribution of MUC2 alleles. The atopic, nonasthmatic group showed significantly larger mean tandem repeat sequence lengths than all the other groups. This represents a genetically determined constitutional difference between the two groups of individuals. Although all the samples were taken from people of Northern European ancestry, it is hard to exclude hidden population stratification. Because of the observation of more rare large alleles for MUC4 in the atopic, nonasthmatic group, we checked that the individuals with these alleles did not also have unusually large MUC2 alleles that might have distorted the data set. This was not the case. It will, however, be important to reproduce the MUC2 findings on separate cohorts of patients. If substantiated, the data suggest that "longer" MUC2 alleles protect against development of asthma in people otherwise predisposed because of their atopy.

The possible biological explanation for this finding is not clear, particularly because there have been doubts about the quantity of MUC2 mucin secreted in the bronchial tree. Both MUC5AC and MUC5B have been found in respiratory secretions but very little MUC2 (16, 17). However, the expression studies reported here on a rather limited number of samples would tend to suggest that MUC2 is upregulated in inflammatory disease. This idea is supported by in vitro evidence, which shows the upregulation of MUC2 in response to pathogens and inflammatory cytokines (52, 53). It seems possible that there are allelic differences in the amount of MUC2 secreted in response to inflammation. This could be a direct effect of the length of the MUC2 RNA transcript or MUC2 protein or be due to association with a linked polymorphism, possibly within a regulatory element of the MUC2 gene. Experiments are underway to attempt to identify MUC2 in sputum from individuals of known MUC2 genotype and to determine whether the quantity of MUC2 is variable. It is tempting to speculate that this might operate by interaction of the small amounts of MUC2 glycoprotein with the more abundant MUC5B or MUC5AC glycoprotein to produce a mucin gel with altered physicochemical properties. In such a case, shorter, perhaps more abundant, MUC2 alleles might for example generate a denser network of mucus that is more difficult to clear from the airways. However, experiments to date have failed to show crosslinking of the mucin gene products (54).