Cystic fibrosis and the relationship between mucin and chloride secretion by cultures of human airway gland mucous cells

Walter E. Finkbeiner, Lorna T. Zlock, Masatoshi Morikawa, Anna Y. Lao, Vijay Dasari, Jonathan H. Widdicombe


—We investigated how cystic fibrosis (CF) alters the relationship between Cl and mucin secretion in cultures of non-CF and CF human tracheobronchial gland mucous (HTGM and CFTGM, respectively) cells. Biochemical studies showed that HTMG cells secreted typical airway mucins, and immunohistochemical studies showed that these cells expressed MUC1, MUC4, MUC5B, MUC8, MUC13, MUC16, and MUC20. Effects of cumulative doses of methacholine (MCh), phenylephrine (Phe), isoproterenol (Iso), and ATP on mucin and Cl secretion were studied on HTGM and CFTGM cultures. Baseline mucin secretion was not significantly altered in CFTGM cells, and the increases in mucin secretion induced by mediators were unaltered (Iso, Phe) or slightly decreased (MCh, ATP). Across mediators, there was no correlation between the maximal increases in Cl secretion and mucin secretion. In HTGM cells, the Cl channel blocker, diphenylamine-2-carboxylic acid, greatly inhibited Cl secretion but did not alter mucin release. In HTGM cells, mediators (10−5 M) increased mucin secretion in the rank order ATP > Phe = Iso > MCh. They increased Cl secretion in the sequence ATP > MCh ≈ Iso > Phe. The responses in Cl secretion to MCh, ATP, and Phe were unaltered by CF, but the response to Iso was greatly reduced. We conclude that mucin secretion by cultures of human tracheobronchial gland cells is independent of Cl secretion, at baseline, and is unaltered in CF; that the ratio of Cl secretion to mucus secretion varies markedly depending on mediator; and that secretions induced by stimulation of β-adrenergic receptors will be abnormally concentrated in CF.

  • trachea
  • bronchi
  • mucus
  • ion transport

in the conducting airways, the tracheobronchial glands are a major source of mucus, the heterogeneous secretion rich in mucins and containing various antimicrobial and immunological products that is vital to maintaining healthy lungs (15). These branched tubuloacinar glands contain two primary secretory cell types: mucous and serous cells. Serous cells are generally found as an acinar demilune, whereas mucous cells line the secretory tubules leading from the acini to the airway surface (45). Like the surface goblet cells, the primary secretory products of mucous cells are mucins (50). By contrast, the predominant secretions of serous cells are believed to be various antimicrobial compounds (16).

In addition to macromolecules, mucous glands secrete significant volumes of water, secondary to active secretion of Cl and HCO3 (3, 64, 68). On the basis of their location it has generally been assumed that serous cells are responsible for the bulk of gland fluid production, a view supported by early immunocytochemical data showing the predominant location of CFTR in the lung is gland serous cells (14, 31, 49). Although this result has not always been confirmed by others (38), recent studies have convincingly demonstrated CFTR in the apical membrane of serous cells by both immunocytochemistry and functional analysis (39–41). Recent studies suggest that mucous cells also secrete fluid. Thus we have shown that gland cultures of mucous phenotype secrete as much Cl as those of serous phenotype (20). Also suggestive of secretion of liquid by mucous cells secrete is the finding in intact glands that fluid flow along the gland lumen continues to increase at 50 μm from the termination of the tubule, a point presumably some distance beyond the serous demilune (68).

Here, we further characterize our human airway gland mucous cell model. Metabolic labeling and biochemical analysis show that the cells synthesize and release mucin glycoprotein, and studies of mucin gene and glycoprotein expression demonstrate that they resemble their native counterparts. In CF, gland secretions are abnormally viscous and there is evidence from cell cultures that Cl and liquid secretion are less than normal (33, 34, 69). Given the evidence discussed above that mucous cells may contribute to gland liquid secretions, we have also investigated the effects of CF on Cl secretion by mucous gland cultures. Finally, if mucous cells do secrete water, then this may be coordinated in some way with mucus secretion, and we asked whether the quantitative relationship between mucus secretion and Cl secretion is altered in CF.


Isolation and culture of HTGM and CFTGM cells.

The Committee on Human Research at the University of California San Francisco approved the use of human tissues for these studies. Nondiseased human bronchial tissues used for immunohistochemistry were obtained from excess donor tissue following lung transplantation surgery. For cell culture, we obtained tracheas and mainstem bronchi from non-CF and CF patients from autopsies performed within 24 h after death. All non-CF postmortem tissues were from individuals without significant pulmonary disease. Strips of epithelium were removed from the trachea and bronchi, and the gland-rich submucosal tissues were then dissected from between the cartilaginous rings. From these pieces of submucosal tissue, we isolated small segments of gland tubules and acinar structures by enzymatic digestion as described previously (60). These gland fragments were plated into T25 flasks in 1:1 mixture of DMEM and Ham's F-12 medium (DMEM/F12) supplemented with 20% FBS, penicillin (105 U/l), streptomycin (100 mg/l), gentamicin (100 mg/l) and amphotericin B (2.5 mg/l). During the first hours of culture, the gland fragments attached. After 8–24 h, cultures were rinsed with PBS and plating medium was replaced with bronchial epithelial growth medium (BEGM; Lonza, Basel, Switzerland). Subsequently, medium was changed every 24 h for 3 days and every 2 days thereafter. Cultures derived from CF patients received additional PBS rinses prior to media changes to prevent microbial contamination. In BEGM, there was robust growth of cells from the attached gland fragments. After the cultures became ∼80% confluent, cells were removed by trypsinization (0.05% trypsin, 0.02% EDTA) and plated (3 × 105 cells) onto 12-mm cell culture inserts (Transwell polycarbonate membranes, 0.4-μm pore diameter; Corning, Corning, NY) coated with human placental collagen (11). Plating medium used for the isolated cells was the same as for acini but lacked penicillin, streptomycin, and amphotericin B and contained less gentamicin (50 mg/l). After 12–18 h, cultures were rinsed with PBS and “HTGM medium” was added to the outside of the insert (basolateral side of the cells). HTGM medium was composed of DMEM/F12 supplemented with insulin (10 μg/ml), transferrin (5 μg/ml), retinoic acid (10−6 M), hydrocortisone (0.5 μg/ml) triidothyronine (20 ng/ml), epidermal growth factor (25 ng/ml), bovine serum albumin (2 mg/ml), 0.1% Ultroser G serum substitute (Pall, Port Washington, NY). All cultures were maintained in a humidified incubator (5% CO2, 37°C). Confluency of cells on the inserts was revealed when medium ceased to leak through to the inside of the insert, and the mucosal surface of the cells appeared dry. Only confluent, “dry” cultures were used, and previous studies have shown us that such cultures invariably have transepithelial resistances (Rte) >100 Ω·cm2. Typically, we studied the first passage HTGM and CF-HTGM cells between days 10 and 14, by which time the cell sheets had not only a well-developed Rte but also scattered cells whose apical zones contained the prominent electron lucent secretory granules typical of native mucous gland cells (20).

For amino acid analysis, secretions from cell and organ cultures were compared. To prepare organ cultures, tracheobronchial submucosal tissues were minced and then incubated for 12 h in a metabolic shaker in 5% CO2-95% air in medium consisting of DMEM/F12 and penicillin, streptomycin, gentamicin, and amphotericin B at the concentrations described for the primary cell cultures.

Metabolic radiolabeling and gel filtration chromatography.

Dissected tracheobronchial submucosal tissue fragments and 12–16 cell culture inserts from HTGM cell cultures prepared from each of three different individuals were metabolically labeled by adding Na2[35S]O4 (60 μCi/cm2), or [3H]-glucosamine (5 μCi/cm2) to the bathing media (basal side only for cell cultures). After 24 h, the apical surfaces of the cell culture inserts were rinsed three times with PBS. The PBS rinses were collected, dialyzed overnight against distilled water to remove free isotope (molecular mass cutoff 12,000–14,000 Da), and lyophilized. Airway submucosal tissue organ culture supernatant was collected after 12 h, centrifuged at 1,000 g to remove tissue fragments, and then dialyzed as described above. Gel filtration chromatography with Sepharose Cl-4B was performed by use of a 1.6 × 84 cm column equilibrated in PBS, 0.1% SDS, and 0.5% βmercaptoethanol. Flow was 30 ml/h, and 5-ml fractions were collected. An aliquot of each fraction was counted for radioactivity via a beta scintillation counter (LS7500 Beckman Instruments, Irvine, CA). The void volume (Vo) peaks from each of the three individual experiments were separately pooled and dialyzed as described by Rose (51) to remove SDS and lyophilized before further characterization. The separate samples were tested for enzymatic sensitivity, buoyant density, and amino acid content, as described in CsCl density gradient centrifugation and Amino acid analysis.

Enzymatic digestions.

Aliquots of the pooled high-molecular-weight Vo secretions from one HTGM cell culture were separately exposed to the following enzymatic digestions: chondroitinase ABC (Sigma-Aldrich; St. Louis, MO), 0.4 U/ml at 37°C for 18 h in 0.1 M Tris acetate, pH 7.3 (37); heparinase from Flavobacterium heparinum (MP Biomedicals; Solon, OH), 0.05 U/ml at 35°C for 16 h in 0.1 M sodium acetate, 1 mM CaCl2, pH 7.0 (30); heparitinase from F. heparinum (MP Biomedicals), 0.03 U/ml at 43°C for 16 h in 0.1 M sodium acetate, 1 mM CaCl2, pH 7.0 (30); hyaluronidase from Streptomyces hyalurolyticus (EMD Chemicals; Gibbstown, NJ), 30 U/ml at 37°C for 16 h in 0.1 M Na acetate, 0.15 M NaCl, 2 mM PMSF, 1 mM CaCl2 adjusted to pH 6.0 with 0.1 M acetic acid (46); peptide-N4-[N-acetyl-β-glucosaminyl] asparagine amidase 10 U/ml from Flavobacterium meningosepticum (N-glycanase; ProZyme; Hayward, CA) for which the sample is first denatured by boiling in 0.2 M NaPO4, 0.5% SDS, and 0.05 M β-mercaptoethanol, pH 7.5. Nonidet P-40 was added to a sevenfold excess over SDS and N-glycanase treatment proceeds at 37°C for 18 h (48); and pronase (Sigma-Aldrich), from Streptomyces griseus, 1% by weight at 37°C for 72 h, with further additions of enzyme (0.5% by weight) at 24 and 48 h, in 0.1 M Tris·HCl buffer, pH 8.0, containing 1 mM CaCl2 (21). Pronase was preincubated at 60°C to inactivate contaminating enzymes. Following all digestions, chromatography was performed as described above. Fractions (5 ml) were collected, and an aliquot of each fraction was counted for radioactivity.

CsCl density gradient centrifugation.

The lyophilized cell culture Vo fractions obtained from different HTGM cultures were treated with hyaluronidase as described above, dialyzed, and lyophilized, and the fractions were subjected to density gradient centrifugation in CsCl adjusted to a density of 1.52 g/ml (51). Centrifugation (32,000 g, 20°C, 72 h) was performed by using a Beckman SW 41 Ti rotor (Beckman Coulter, Brea, CA). Sequential fractions were collected from the bottom of the tube, weighed, and counted for radioactivity.

Amino acid analysis.

The amino acid compositions of hyaluronidase-treated Vo fractions prepared from HTGM cell cultures from a third specimen and from an organ culture of tracheobronchial submucosal tissue fragments were determined by using a Beckman 6300 amino acid analyzer (Beckman Instruments, Palo Alto, CA) by the method of Bidlingmeyer et al. (6).

Total RNA isolation and RT-PCR.

Ten-day-old HTGM cell cultures from three separate individuals were used for the analysis of mucin gene transcripts by RT-PCR for expression of MUC1, 2, 3, 4, 5B, 5AC, 6, 7, 13, 15, 16, 17, 19, and 20 with β-actin as a quality control (Table 1). Total RNA was isolated and purified according to standard procedure using TRIzol reagent (Invitrogen, Carlsbad, CA). The quality of isolated RNA was determined by agarose gel electrophoresis followed by spectrophotometry. RNA samples were then incubated with oligo(dT) and random primers in a 12-μl reaction volume. The samples were heated to 70°C for 10 min and immediately placed on ice. A mixture containing 1× reaction buffer (GIBCO-BRL) and deoxynucleoside triphosphates was added to each sample. Superscript II TM reverse transcriptase was added to the reaction mixture to a final volume of 20 μl. The reaction was then incubated for 1 h at 42°C and terminated by heating at 70°C for 10 min. Next, the RT samples were used for the mucin gene PCR reactions. The PCR components included 2 μl of template DNA, 1× Taq polymerase reaction buffer, 2 μM of each primer, 200 μM of dNTPs, 1% DMSO, and Taq polymerase. The amplification conditions were 94°C for 5 min, 94°C for 30 s, 62°C for 30 s, and 72°C for 30 s. After 20 cycles, the reaction mixtures were further incubated for 5 min at 72°C. The amplified products were then analyzed by agarose gel electrophoresis.

View this table:
Table 1.

PCR primer sets used for detection of mucin gene expression by HTGM cells

Immunohistochemistry for mucins.

Human bronchi and HTGM cultures from four different individuals were stained with antibodies against MUC1 (Invitrogen), MUC2 (Invitrogen), MUC3 (Santa Cruz Biotechnology, Santa Cruz, CA), MUC4 (Invitrogen), MUC5B (Santa Cruz Biotechnology), MUC5AC (Invitrogen), MUC6 (Thermo Fisher Scientific, Fremont, CA), MUC7 (Abcam, Cambridge, UK), MUC8 (Santa Cruz Biotechnology). MUC13 (Sigma), MUC15 (Sigma), MUC16 (Santa Cruz Biotechnology), MUC17 (Sigma), and MUC20 (Sigma). Tissues and cultures were fixed in 10% neutral buffered formalin and processed for paraffin embedding by standard techniques. Tissue sections (5 μm) were dewaxed in two changes of Clear-Rite (Thermo Scientific, Waltham, MA) then rehydrated through a series of graded alcohols. Slides were then submerged in 3% hydrogen peroxide to quench endogenous peroxidase activity. When necessary, heat-induced antigen retrieval was performed using a specialized pressure heater (Decloaking Chamber, Biocare Medical, Concord, CA) to unmask epitopes by boiling slides for 10 min at 125°C in either low (Reveal Decloaker) or high (Borg Decloaker) pH antigen retrieval buffers (Biocare Medical). Primary antibodies were incubated 30–60 min. Antibody detection was done using the SuperPicture Kit (Invitrogen, Camarillo, CA) or the LSAB+ Universal Kits (DAKO, Carpinteria, CA). Stains were developed using 3,3′-diaminobenzidine (DAKO), which was applied for 5 min. Slides were then counterstained with hematoxylin, blued, dehydrated through a series of graded alcohols, cleared in two changes of xylene, and mounted with coverslips. Prior to the peroxidase quenching step for the MUC5B antibody, tissues were deglycosylated by using a 1:1 solution of sodium acetate, pH 4.5 and sodium thiosulfate, pH 7.6 for 30 min at 4°C (9). For each antibody, we used positive control tissues recommended by the manufacturer. Negative controls consisting of omission of primary antibody and either isotype-matched irrelevant antibodies (for monoclonal antibodies) or normal species-specific serum (for polyclonal antibodies) were performed on each tissue type. Table 2 provides additional information regarding the antibodies and specific staining reactions.

View this table:
Table 2.

Mucin antibodies and specific immunohistochemical reaction protocols

Measurement of mucin secretion.

Mucin secretion by HTGM and CFTGM cells was studied between days 10 and 14 after plating onto the cell culture inserts. Cultures were washed three times with PBS to remove mucin accumulated on the cell surfaces. Next, both the apical (luminal) and basal (serosal) sides of the cells were incubated with 1 ml of DMEM/F12 for 30 min. At 30-min intervals, the apical medium was collected and stored. We added experimental drugs prior to the fourth time period and compared the amount of mucin secreted during this period with the previous 30-min period. The ratio of mucin release after drug exposure during the fourth period to release during the third period was determined for each cell sheet and compared with control cell sheets not receiving drugs. Thus secretion is expressed as (P4/P3)test/(P4/P3)control, where P4 and P3 are the amounts of mucus released during the third and fourth collection periods, respectively. For example, suppose in control tissues the amount of mucin released during period 4 was 77% of that released in period 3, whereas in test tissues treated with secretagogue the amount of mucin released in period 4 was 150% of that released in period 3. Then the stimulated secretory rate was 194% of control (i.e., 150/0.77). Secretagogues tested included methacholine (MCh), phenylephrine (Phe), isoproterenol (Iso), and ATP at concentrations ranging from 10−7 to 10−4 M. These agonists were selected for the following reasons: MCh because acetylcholine is the most potent stimulatory of fluid and mucin secretion by native glands; Phe because it has negligible effects on fluid secretion by human glands; Iso because in native glands it stimulates mucin release but has relatively minor effects on fluid secretion; and ATP because it is an efficacious secretagogue of goblet and gland cells (36, 43, 57, 59). In some studies of non-CF mucous cells, diphenylamine-2-carboxylic acid (DPAC) at 10−3 M was added during the final 30-min period.

We used a sandwich-type ELISA to quantify mucin secretion (61). We coated 96-well microtiter plates (Immulux HB; Dynex Technologies, Chantilly, VA) with 100 μl of a 5 μg/ml solution of purified monoclonal antibody A10G5 (17). The plates were incubated overnight (4°C) and then washed five times with PBS. Next, the wells were incubated (2 h, room temperature) with 0.1% gelatin in PBS to block nonspecific binding of antigen. After five washes with PBS containing 0.1% Tween 20 (PT), undiluted experimental samples (100 μl) were added in triplicate wells and incubated for 1 h at room temperature. After five washes with PT, wells were incubated (1 h, room temperature) with 100 μl of 10 μg/ml biotinylated A10G5 (28) diluted in ELISA buffer (3% normal goat serum, 0.05% Tween 20 in PBS). The plates were washed five times in PT and then incubated (1 h, room temperature) with streptavidin-β-galactosidase conjugate (Rockland Immunochemicals, Gilbertsville, PA) (1:1,000 in ELISA buffer). Following five washings with PT, 100 μl of 2-nitrophenyl-β-d-galactopyranoside, 1 mg/ml in 0.05 M NaPO4, pH 7.2, 1.5 mM MgCl2 was applied to each well. After 30 min, absorbance values were read at 405 nm, via a microplate reader (Model 3550, Bio-Rad, Richmond, CA). The mean value of the triplicate samples was used to estimate the amount of secreted mucin from a standard curve as described previously (61). The limit of detection was 0.5 ng.

Measurement of Isc and Rte.

Studies were performed on confluent sheets of HTGM and CFTGM cells between days 10 and 14 after plating on inserts. Entire filters and their overlying cell sheets were mounted in modified Ussing chambers and bathed in bicarbonate-buffered Krebs-Henseleit solution (pH 7.4), bubbled with 95% O2-5% CO2, at 37°C. Transepithelial potential difference was clamped to zero and the resulting short-circuit current (Isc) was continuously displayed on a pen recorder. Rte was determined from the size of the current deflections caused by 0.2-s voltage pulses of constant amplitude (0.2–1 mV) imposed on short-circuited cell sheets every 20 s. Drugs were added to both sides of the tissue in random order as 100-fold concentrated stock solutions made on the day of the experiment. The contribution of Na+ absorption to the Isc was eliminated with amiloride before applying the secretogogues and in some studies DPAC (10−3 M) was added prior to the addition of mediators.


Cell culture media, FBS, and antibiotics were obtained from the Cell Culture Facility, University of California, San Francisco, CA. Growth factors were obtained from BD Biosciences (Franklin Lakes, NJ) or Sigma-Aldrich. All other reagents were purchased from Sigma-Aldrich unless indicated otherwise.

Statistical analysis.

Data are presented as means ± SE. Test of differences between means were performed with Student's t-test with P < 0.05 being taken as statistically significant.


Enzymatic digestions.

Airway glands maintained in vitro incorporate glucose, sulfate, and other precursor molecules into mucin glycoproteins (58). Analysis of the metabolically radiolabeled HTGM cell secretions by gel filtration on Sepharose Cl-4B under the reducing and dissociating conditions used prevented mucins from associating with other macromolecules and revealed that the majority of 3H- and 35S-labeled material eluted in the high-molecular-weight Vo. Identical chromatographic results were obtained from cultures derived from three different tracheobronchial specimens. A representative result is shown in Fig. 1A. The material present in the Vo (fractions 9-12) of each specimen was separately pooled, dialyzed as described above, and subjected to additional analysis. From one sample, we tested for the presence of glycopeptides by treating the sample with pronase, which causes nonspecific cleavage of glycoprotein peptide chains (21). Digestion of the samples confirmed the presence of high-molecular-weight glycoproteins (Fig. 1B). Next, we tested for the presence of proteoglycans by incubating the high-molecular-weight material with various proteoglycan-digesting enzymes (Fig. 1, CG). The high-molecular-weight material isolated by Sepharose Cl-4B chromatography was resistant to digestion with chondroitinase ABC, which digests chondroitin, chondroitin sulfates, dermatan sulfate, and hyaluronic acid (Fig. 1C). The action of chondroitinase on hyaluronic acid is slow; however, we confirmed the absence of hyaluronic acid with Streptomyces hyaluronidase (Fig. 1D). The absence of heparan and heparan sulfate was shown by the lack of any digestion with heparinase (Fig. 1E). The high-molecular-weight material was also resistant to heparitinase, which digests heparatan sulfate (Fig. 1F). The Vo was resistant to digestion with N-glycanase, confirming the absence of significant amounts of N-linked proteins in the high-molecular-weight secretions of airway epithelial cultures (Fig. 1G). The resistance of the excluded volume material to treatment with a panel of enzymes selective for proteoglycans and to N-glycanase demonstrated that this material does not contain proteoglycans or N-linked glycoproteins (47).

Fig. 1.

Gel filtration chromatography of secretions from non-cystic fibrosis (non-CF) human tracheobronchial gland mucous cell (HTGM) cultures. Cell culture inserts were incubated on their basal sides for 24 h with medium containing Na2[35S]O4 or [3H]glucosamine. Apical secretions were obtained by rinsing cultures three times with PBS. Samples were pooled and dialyzed against distilled water, lyophilized, and applied to a Sepharose Cl-4B column (1.6 × 84 cm) as described in methods. Significant amounts of material remained in the void volume. The void volume fractions (9–12) were pooled and used for enzymatic digestions. In all panels void volume (Vo) and total volume (Vt) refer to elution positions of blue dextran and Na2[35S]O4, respectively. A: representative undigested sample. BG: void volume material from pooled cultures from a single individual treated with pronase (B), chondroitinase ABC (C), hyaluronidase (D), heparinase (E), heparitinase (F), and N-glycanase (G).

Density gradient centrifugation.

As shown in Fig. 2, the hyaluronidase-resistant Vo fractions obtained from HTGM secretions had buoyant densities of ∼1.54 g/ml, consistent with their identification as mucin (51).

Fig. 2.

CsCl density-gradient centrifugation of hyaluronidase-treated Vo fractions from HTGM cultures from a single individual. Centrifugation (32,000 g, 20°C, 72 h) was performed by using a Beckman SW40.1 rotor. Sequential fractions were collected from the bottom of the tubes, weighed, and counted for radioactivity. HTGM cell secretion samples have buoyant densities of ∼1.54 g/ml, consistent with their identification as mucin.

Amino acid analysis.

The amino acid composition of hyaluronidase-resistant Vo fractions obtained from secretions of HTGM cells contained a predominance of serine, threonine, and proline and was similar to that of the material collected from short-term organ cultures of tracheal submucosal tissues (Table 3). This provides additional evidence that the majority of the high-molecular-weight secretions consist of mucins.

View this table:
Table 3.

Amino acid composition of hyaluronidase-treated, high-molecular-weight glycoconjugates released from HTGM cultures from a single individual compared to secretions from submucosal tissue organ culture

Semiquantitative RT-PCR for mucin genes.

Amplicons of the expected size indicated expression of MUC1, MUC4, MUC5B, MUC13, MUC16, and MUC20 genes by HTGM cell cultures obtained from three different individuals, although MUC13 expression was weak (Fig. 3). There was no detectable expression of MUC2, MUC3, MUC5AC, MUC6, MUC7, MUC15, or MUC19.

Fig. 3.

PCR for expression of various mucin genes by HTGM cells from 3 different individuals. Lane A, β-actin; lane B, MUC1; lane C, MUC2; lane D, MUC3; lane E, MUC4; lane F, MUC5B; lane G, MUC5AC; lane H, MUC6; lane I, MUC7; lane J, MUC13; lane K, MUC15; lane L, MUC16; lane M, MUC17; lane N, MUC19; lane O, MUC20; lane P, ladder. Cultured mucous gland cells express MUC1, MUC4, MUC5B, MUC13 (weakly), MUC16, MUC17, and MUC20 RNA.

Immunohistochemistry for mucins.

Control stains were positive for all antibodies and omission of the anti-mucin antibodies or substitution with irrelevant antibodies yielded absence of staining in all tissues and cultures (data not shown). Table 4 provides a summary of the results of immunostaining in native airway tissues and HTGM cells for all mucins tested and photomicrographs of representative staining for native airway and HTGM cells are shown in Figs. 4 and 5, respectively.

Fig. 4.

Representative photomicrographs of immunohistochemical staining for various mucins in human bronchi. A: MUC1. Ciliated cells, goblet cells, and gland duct cells show staining; gland cells are negative. B: MUC2. Only a subpopulation of goblet cells show staining. C: MUC3. Staining is absent. D: MUC4. Staining is present in the apical membranes of ciliated cells and in endothelial cells. E: MUC5B. Mucous gland cells and some goblet cells show staining. F: MUC5AC. Only the goblet cells are positive. G: MUC6. Staining is absent. H: MUC7. Only serous gland cells show staining. I: MUC8, surface epithelium. The ciliated cells show prominent staining in their subapical regions. J: MUC8, glands. There is diffuse but weak staining of both serous and mucous gland cells. K: MUC13. All surface epithelial cells, serous cells, and mucous cells are positive. L: MUC15. Ciliated cells show staining. Although the serous cells in this microscopic field are positive, staining for serous cells was variable. M: MUC16. Both goblet and mucous cells show staining. N: MUC17. Staining is absent. O: MUC20. Staining is diffusely positive. Scale bars, 40 μm.

View this table:
Table 4.

Comparison of mucin glycoprotein expression by native bronchial cells and HTGM cells. HTGM cultures from four different individuals were studied. The numbers in parenthesis indicate the number of cultures showing positive or negative staining

Staining for seven mucins (MUC2, MUC3, MUC5AC, MUC6, MUC7, MUC15, MUC17) was absent in native tracheobronchial mucous cells, and identical results were obtained in all four cultures of HTGM cells tested. Immunohistochemical staining for four mucins (MUC8, MUC13, MUC16 and MUC20) was positive in the native mucous cells as well as in each of the HTGM cultures. As expected MUC5B staining was positive in native mucous cells. However, it was detected in only two of the four HTGM cultures. MUC4 was not detected in native mucous cells although it was present in HTGM cell cultures. Finally, MUC1 staining, positive in native ciliated cells, goblet cells, and gland duct cells and negative in native gland serous and mucous cells, was weakly positive in each of the HTGM cell cultures.

The pattern of staining in the cultured mucous cells varied (Fig. 5). Staining for three mucins (MUC1, MUC5B, MUC16) was found primarily in “luminal” cells at the air-liquid interface of the HTGM cell sheets corresponding to the site where electron-lucent secretory granules are located (20). Staining for MUC4 was restricted primarily to the apical membrane of some of the luminal cells. Finally, staining for three mucins (MUC8, MUC13, MUC20) was located diffusely throughout the HTGM cell sheets.

Fig. 5.

Representative photomicrographs of immunohistochemical staining for various mucins in HTGM cells: MUC1 (A), MUC2 (B), MUC3 (C), MUC4 (D), MUC5B (E), MUC5AC (F), MUC6 (G), MUC7 (H), MUC8 (I), MUC13 (J), MUC15 (K), MUC16 (L), MUC17 (M), MUC20 (N). There is staining of cells at the air-liquid interface for MUC5B (E), MUC16 (L), and weakly for MUC1 (A). Staining for MUC4 (D) is seen at the apical membrane of the cells at the air-liquid interface. There is diffuse staining of the cell culture sheets for MUC8 (I), MUC13 (J), and MUC20 (N). Staining is negative for MUC2, MUC3, MUC5AC, MUC6, MUC7, MUC15, and MUC17. Scale bar = 20 μm.

Effects of secretagogues on Cl secretion.

After mounting in Ussing chambers, it took Rte and Isc ∼5 min to stabilize. At this point, Rte of HTGM and CF-HTGM were not statistically different (HTGM cells, 202 ± 6 Ω·cm2, n = 207; CFTGM cells, 207 ± 9 Ω.cm, n = 136). Baseline Isc values were lower in the CF cells (HTGM cells, 21.7 ± 0.7 μA/cm2, n = 129; CFTGM cells, 12.9 ± 0.7 μA/cm2, n = 56). Once Isc had stabilized, amiloride was added to remove active absorption of Na+, resulting in a reduction in Isc in both HTGM (−2.4 ± 0.2 μA/cm2, n = 129) and CFTMG (−4.9 ± 0.5 μA/cm2, n = 56). In the presence of this agent, Isc is an index of active Cl secretion. Added after amiloride, Phe had no statistically significant effects on Cl secretion by HTGM or CF-HTGM (Fig. 6A). By contrast, MCh, ATP, and Iso all stimulated Isc of HTGM with Kd values of ∼3 × 10−7 M and maximal responses by 10−5 M. The effectiveness of secretagogues on increasing Isc across HTGM was ATP > MCh ≫ Iso > Phe (Fig. 6A). Responses of CFTGM to MCh, ATP, and Phe were not significantly different from those of HTGM (Fig. 6A). By contrast, the response of CFTGM to Iso was significantly reduced to ∼20% that of HTGM (Fig. 6A).

Fig. 6.

Chloride and mucin secretion by HTGM and cystic fibrosis (CF) human tracheobronchial gland mucous (CFTGM) cells in response to adrenergic, cholinergic, and purinergic agonists (10−5 M). A: chloride secretion measured in Ussing chambers under short-circuit conditions (HTGM cells, n = 26–32 inserts from 7–8 individuals; CFTGM cells, n = 31 inserts from 7 individuals). Untreated cells showed no significant change in short-circuit current (Isc) over the time interval involved. B: mucin secretion measured by ELISA (release of A10G5 antigen) expressed as percent increase (HTGM cells, n = 21–23 inserts from 12 individuals; CFTGM cells, n = 7 inserts from 3–4 individuals). * Significantly different from HTGM cells.

Effects of secretagogues on mucin secretion.

The basal rate of mucin secretion, measured in the period immediately before addition of secretagogue (i.e., the third 30-min collection period), was not statistically different between HTGM (68 ± 11 ng·cm−2·h−1, n = 56) and CFTGM (100 ± 16 ng·cm−2·h−1, n = 14). However, for both HTGM and CFTGM, control (untreated) tissues showed a significant decline in secretion over time, such that during the fourth collection period (the period during which drugs were added to test tissues), mucin secretion from HTGM was 77 ± 3% of that in the third period, whereas in CFTGM cells it was 79 ± 7%. Therefore, as described in methods, the secretory responses of test tissues were corrected for this time-dependent decline. Preliminary experiments showed that the responses of CFTGM to mediators had Kd values much the same as for normal cells. Therefore, responses were tested in detail for 10−5 M only. Figure 6B shows that responses to MCh and ATP were reduced in CFTGM cells, but that the responses to Iso and PE were unaltered.

Effects of Cl channel blockade.

In HTGM cells, the nonspecific blocker of Cl channels, DPAC (1 mM), significantly reduced the response in Isc to all stimulatory secretagogues. In eight pairs of tissues, Iso-induced chloride secretion was reduced from 9.3 ± 2.0 to 0. 8 ± 0.2 μA/cm2 (P < 0.05), ATP-induced chloride secretion was reduced from 15.0 ± 1.8 to 1. 5 ± 0.5 μA/cm2 (P < 0.05), and MCh-induced secretion was reduced from 10.5 ± 1. 4 to 4.3 ± 1.3 μA/cm2 (P < 0.05). The response to Phe was not significantly different from zero, either with or without the blocker. Despite its actions on the Isc responses to mediators, DPAC had little or no effect on mediator-induced mucin secretion. Thus, in eight sheets from three tracheas, the secretory index for Iso was 155 ± 12 in the absence of blocker and 144 ± 15 in the presence. Corresponding values for Phe were 123 ± 14 and 133 ± 12; for ATP, 226 ± 15 and 177 ± 40; and for MCh, 133 ± 22 and 83 ± 8, respectively.


Airway gland mucous secretions are abnormally viscous in CF (33). Anion channel blockers mimic the effects of CF in pig airway glands (4), and it has been reported that gland serous cells contain the highest levels of CFTR in the airways (14). Therefore, it has generally been assumed that the increased viscosity of gland secretions in CF is caused by reduced fluid secretion by serous cells coupled with normal levels of mucin secretion by mucous cells. However, cell culture models of gland mucous cells secrete Cl to approximately the same degree as serous cell cultures (20). There is evidence for fluid secretion by mucous cells of native glands (67). Finally, mucous cultures contain both CFTR and calcium-activated chloride channels as revealed by patch clamping and Ussing chamber studies (22). Therefore, it is possible that alterations in secretion of Cl (or mucins) by mucous cells may contribute to the abnormally concentrated gland secretions in CF. Here, we have tested this hypothesis using primary cultures of human airway gland cells of mucous phenotype.

To obtain cultures of HTGM cells, we first isolate small fragments of gland tissue through a combination of tissue dissection and enzymatic digestion. Microexplants of glandular cells attach to collagen-coated tissue culture flasks, and this is followed by outgrowths of seromucous glandular cells. The growth medium (BEGM) encourages robust cell growth while inhibiting growth of potentially contaminating fibroblasts and endothelial cells. It is not known whether the propagating gland cells arise from mucous cells, serous cells, myoepithelial cells, or perhaps cells with stem-cell like properties residing in gland ductular tissue as described by Borthwick et al. (7).

We have used cultures of human airway gland cells grown as described here in previous studies and termed them “mucous” on the basis of several lines of evidence. First, in transmission electron micrographs they show electron-lucent granules typical of native mucous cells but lack the smaller electron-dense granules typical of native gland serous cells (20). Second, they stain positively with A1F8, an antibody specific for mucous cells of native glands, and do not stain with either B1D8, an antibody that stains serous cells, or antibody against lactoferrin (19, 20). Here, we show that the macromolecular secretions from these mucous cells consist predominantly of mucins on the basis of their size, their sensitivity (or lack of) to enzymes, their buoyant density and their amino acid compositions. Furthermore, we characterize the expression of specific mucins by the HTGM cell cultures.

Mucins consists of the cell surface or membrane-tethered mucins with tandem repeats (MUC1, MUC3, MUC4, MUC11, MUC12, MUC13, MUC16, MUC17, MUC20), the polymeric secreted, gel-forming mucins with cysteine-rich tandem repeats (MUC2, MUC5B, MUC5AC, MUC6, MUC19), the nonpolymeric, secreted, mucins with cysteine-poor tandem repeats (MUC7, MUC8, MUC9), and the mucins without tandem repeats (MUC14, MUC15, MUC18) (52). Analysis of mucin gene and glycoprotein expression in HTGM cells demonstrates that they express MUC5B but not MUC5AC or MUC7, consistent with the expression of these airway mucins in the native mucous cells and distinguishing them from goblet and serous cells (29, 55). As expected, HTGM cells also expressed MUC8 and MUC16 (13, 54). Here, we also show that the cell surface mucins MUC13 and MUC20 are present in the native airway mucous cells as well as in HTGM cell cultures. Although another cell surface mucin, MUC17, was detected in HTGM cells by RT-PCR, there was no evidence of its glycoprotein expression in either the HTGM cells or in the native airway tissue. However, cultures did show RNA expression and glycoprotein staining for other cell-tethered mucins, MUC1 and MUC4, which was not evident in the native mucous cells. Perhaps a reflection of a somewhat weak RT-PCR signal, expression of MUC5B glycoprotein was detected in only two of four separate HTGM cell cultures. Refinements in mucous gland cell culture conditions may identify components that upregulate MUC5B gene and glycoprotein expression in vitro similarly to the effects of that mediators of inflammation on goblet cell MUC5AC expression in cultures of human bronchial epithelial cells (62).

The pattern of staining in the cultured mucous cells varied (Fig. 5). Thus three mucins (MUC1, MUC5B, MUC16) were localized to cells at the air-liquid interface (luminal cells) of the HTGM cell sheets corresponding to the site of where electron lucent secretory granules appear (20). In the three cell sheets positive for MUC4 product, the staining was focal and restricted the apical membrane of luminal cells. Finally, three mucins (MUC8, MUC13, MUC20) were located diffusely throughout the HTGM cell sheets. Little is known about MUC13 and MUC20 or their role in the airways (26). In our studies, MUC8, a nonpolymeric, secreted mucin, was detected, albeit weakly, in both mucous and serous cells of the native airway glands. It was not seen in goblet cells although it was strongly detected in the apical and subapical regions of ciliated cells. The role of MUC8 in the airways is also poorly understood, and, to date, there are no data suggesting that it is altered in lung disease. Expression of MUC8 glycoprotein in upper airways (nasal mucosa and nasal polyps) has been shown to be quite variable and not altered in CF, asthma, or allergic rhinitis (44). The cell surface mucins MUC1, MUC4, and MUC16 are present in both the periciliary and mucus layers of the ASL and are detected in human sputum, accounting for 10% of its mucin content, with the remaining consisting of the gel-forming mucins MUC5AC and MUC5B (56). Functions attributed to the cell tethered mucins include hydration, lubrication, protection from proteases, pathogen defense, and cell signaling (26). In the native airways, we find rather diffuse staining of MUC1 in ciliated and goblet cells whereas MUC4 shows staining in the apical membranes of ciliated cells and the membranes of the goblet cell secretory granules. Staining for MUC16 appears located on the membranes of the secretory granules of both goblet and mucous cells. The HTGM cultures should aid in further studies of glandular mucins and their roles in airway homeostasis.

The ELISA secretion assay used in these studies relies on a monoclonal antibody (A10G5) that recognizes a mucin epitope present within secretory granules of mucous, serous, and goblet cells along with material located along the surface of airway mucosa (17, 61). The cellular distribution of the 10G5 epitope suggests binding to several different mucins, perhaps including both gel-forming and tethered types. In studies of the secretion from cultures of tracheal or bronchial surface epithelial cells stimulated with ATP (10−9 −10−3 M), the A10G5 ELISA measures secretory rates nearly identical to that calculated by release of 35SO4-labeled glycoconjugates (W. E. Finkbeiner and L. T. Zlock, unpublished data).

Somewhat surprisingly, amiloride, a blocker of active Na+ absorption, inhibited Isc. However, mRNA for both the α and β subunits of the epithelial sodium channel (ENaC) is found in both mucous and serous cells of human airway glands (8). Whether this ENaC functions is unknown, and it is possible that the inhibitory effect of amiloride is one of the ways in which our cultures differ from native cells. However, the inhibitory responses to amiloride were small (2–5 μA/cm2) compared with the stimulatory responses to MCh (10–15 μA/cm2); acetylcholine is the most efficacious secretagogue of native glands (66). Thus Cl secretion in our cells seems to be quantitatively much more important than Na+ absorption, as is presumably the case in native mucous cells.

We found that the neurohumoral regulation of mucous cell secretion was little altered in CF. Cl secretion in response to Ca2+-elevating agents was unaltered in CF, but Cl secretion in response to Iso, an agent that acts predominantly through cAMP, was significantly reduced. Also, block of mediator-induced Cl secretion with DPAC had no effect on mucin secretion by non-CF mucous cells. Across mediators there was no correlation between stimulation of mucous secretion and stimulation of Cl secretion. We conclude that secretions of mucins and of Cl (and water) by mucous cells are not linked to each other by intracellular signaling mechanisms. However, as demonstrated by the immunostaining for mucins, there is heterogeneity in the HTGM cell cultures. Thus ion transport and mucin secretion could be the products of different cell phenotypes within the cell sheets.

Stimulation of mucin release by Iso or Phe was unaltered in CF. The responses to MCh and ATP, however, were significantly reduced, most markedly for MCh. CF cells have a higher baseline secretory rate than non-CF, and this could reduce their response to mediators. However, the secretory response of non-CF cells to MCh showed no significant dependence on baseline secretory rates (best least-squared linear regression), and the range of baseline rates of the CF cells fell within the range for non-CF cells. In other words, the regression of response on baseline secretion for CF cells was roughly parallel to, but below, the regression for non-CF cells. Thus, for any given rate of baseline secretion, CF cells showed a smaller response to MCh than did non-CF cells. Thus the small response of CF cells to MCh is not due to their secretory rates already being near maximal under baseline conditions. Similar conclusions apply to the effects of CF on the response to ATP. These results suggest that there may be minor abnormalities in Ca2+ signaling in CF, as proposed by others (1).

The finding that MCh was the least potent mediator at stimulating mucin secretion is somewhat surprising given that cholinergic agents are the most potent stimulators of human airway glands in terms of volume of secretion (66). However, in the cat, the viscosity and protein of gland secretions vary considerably depending on secretagogue (42), and in humans the gland secretions evoked by cholinergic agents may have lower concentrations of mucins than those evoked by other agents.

In vitro, the potency sequence for stimulation of Cl secretion in non-CF cells by mediators was MCh ≈ Iso ≈ ATP > Phe. In both CF and non-CF cells, responses to Phe were essentially zero. Responses to MCh and ATP were unaltered in CF (Fig. 6). In surface epithelial cells the response to Ca2+-elevating agents is reduced in CF (65, 69, 70), a result that is explainable if, in non-CF cells, elevation of intracellular Ca2+ concentration ([Ca2+]i) opens basolateral K channel (63), hyperpolarizes the apical membrane, and drives Cl through constitutively open CFTR. Our gland cell cultures contain CFTR as evidenced by the reduced Cl secretory response to Iso in CF. Therefore one might also expect the responses to MCh and ATP to be reduced in CF by the same mechanism as proposed for surface cultures. However, such effect would tend be counterbalanced by the upregulation of calcium-activated Cl channels reported for airway epithelium in CF (25, 35), resulting in no net change in the effects of these mediators. By contrast, we found that the response to Iso was significantly reduced in CF (Fig. 6). This reflects the fact that this agent acts predominantly through cAMP and has smaller effects on [Ca2+]i than do MCh or ATP (65, 71).

Figure 6 compares the maximal changes in mucin secretion with the maximal changes in Cl secretion in response to the various mediators for both CF and non-CF cells. Two related facts are particularly apparent. First, the ratio of maximal Cl secretion to maximal mucous secretion varies considerably between mediators. MCh produces the secretions with the lowest ratio of mucus secretion to Cl secretion, and Phe with the highest. The mucous gland cell secretions induced by Phe are therefore presumably more concentrated than those produced by MCh. The plot suggests that mucous gland cell secretions induced by ATP should be of intermediate concentration. The same is true for Iso in non-CF tissues, but the reduction in Iso-induced Cl secretion in CF increases the ratio of mucous secretion to Cl secretion, making it as high as that seen with Phe. However, it is important to point out that in the native airway mucous cell secretions would be mixed with that of serous cells and perhaps further modified by the cells of the gland ducts.

The second notable feature of Fig. 6 is that across mediators there is no clear-cut relationship between the level of mucin secretion induced and the level of Cl secretion. In fact, across mediators least-squared linear regressions showed no significant correlation between the mean change in mucin secretion and the mean change in Cl secretion in either CF or non-CF cells. This would suggest that the secretion of mucus is not coordinated with the secretion of chloride. This is consistent with the finding reported here that inhibition of Cl secretion with DPAC does not alter mucus secretion by non-CF gland cells. By blocking Cl secretion in a variety of ways, Jarry et al. (32) also concluded that cultured human colonic goblet cells also show uncoupling of Cl and mucin secretion. Finally, Lethem et al. (43) have shown that discharge of mucous granules from airway goblet cells is unaltered in CF. Although we don't know how to account for the CF-related declines in the mucin secretory responses to ATP and MCh, the important conclusion is that the loss of the Cl secretory response to Iso does not affect mucin secretion. More generally, across mediators there is no correlation between the mediator-induced change in Cl secretion and the mediator-induced change in mucin secretion in either HTGM or CFTMG cells.

In summary, these studies suggest that mucin secretion is independent of Cl movement in cultured tracheobronchial gland mucous cells. They also support previous studies indicating that specific agonists exert differential regulation of mucus secretion and the secretion of Cl and water. Our results also suggest that mucous cell secretions induced by stimulation of purinergic, cholinergic, or α-adrenergic receptors will be little altered in CF. However, mucous cell secretions induced by stimulation of β-adrenergic receptors should be abnormally concentrated in CF.


This research was supported by National Institutes of Health Grants DK72517 and HL73856, a Research Development Program and Drug Discovery grant from the Cystic Fibrosis Foundation, and a gift from Pam Fair and Glen Sullivan.


None of the authors have commercial associations that pose a conflict of interest.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 68.
  68. 69.
  69. 70.
  70. 71.
View Abstract