Search

This article traces the beginnings of the various areas of physiological research on airway epithelium. First mentioned in 1600, it was not until 1834 that it was found to be ciliated. Goblet and basal cells were described in 1852, to be followed by ~10 other epithelial cell types (the most recent in 2018). It also contains nerve endings and resident leukocytes. Mucociliary clearance was documented in 1835, but the first studies on the ciliary beat cycle did not appear until 1890, and a definitive description was not published until 1981. It was established in 1932 that goblet cells in the cat trachea were unresponsive to cholinergic agents; but only since 1980 or so has any significant progress been made on what does cause them to degranulate. Active transfer of salts across epithelia creates local osmotic gradients that drive transepithelial water flows. Vectorial salt transport was first described for airway epithelium in 1968, and the associated volume flows were measured in 1981. Evidence that airway epithelium releases signaling molecules first appeared in 1981. Since then, scores of molecules have been identified. The pace of research in most areas increased dramatically after the development of confluent, polarized cultures of airway epithelium in the early 1980s.

—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 (10−5M) 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.

Salt and water absorption and secretion across the airway epithelium are important for maintaining the thin film of liquid lining the surface of the airway epithelium. Movement of Cl across the apical membrane involves the CFTR Cl channel; however, conductive pathways for Cl movement across the basolateral membrane have been little studied. Here, we determined the regulation and single-channel properties of the Cl conductance ( GCl) in airway surface epithelia using epithelial cultures from human or bovine trachea and freshly isolated ciliated cells from the human nasal epithelium. In Ussing chamber studies, a swelling-activated basolateral GCl was found, which was further stimulated by forskolin and blocked by N-phenylanthranilic acid (DPC) = sucrose > flufenamic acid = niflumic acid = glibenclamide > CdCl2 = 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB) = DIDS = ZnCl2 > tamoxifen > 4,4′-dinitro-2,2′-stilbene-disulfonate disodium salt (DNDS). In whole cell patch-clamp experiments, three types of GCl were identified: 1) a voltage-activated, DIDS- (but not Cd-) blockable and osmosensitive GCl; 2) an inwardly rectifying, hyperpolarization-activated and Cd-sensitive GCl; and 3) a forskolin-activated, linear GCl, which was insensitive to Cd and DIDS. In cell-attached patch-clamp recordings, the basolateral pole of isolated ciliated cells expressed three types of Cl channels: 1) an outwardly rectifying, swelling-activated Cl channel; 2) a strongly inwardly rectifying Cl channel; and 3) a forskolin-activated, low-conductance channel. We propose that, depending on the driving force for Cl across the apical membrane, basolateral Cl channels confine Cl− secretion or support transcellular Cl− absorption.

Lactoferrin and lysozyme are important antimicrobial compounds of airway surface liquid, derived predominantly from serous cells of submucosal glands but also from surface epithelium. Here we compared release of these compounds from the following human cell cultures: primary cultures of tracheal epithelium (HTE), Calu-3 cells (a lung adenocarcinoma cell line frequently used as a model of serous gland cells), 16HBE14o- cells (an SV40 transformed line from airway surface epithelium), T84 cells (a colon carcinoma cell line), and human foreskin fibroblasts (HFF). For lysozyme, baseline secretory rates were in the order Calu-3 > 16HBE14o- > HTE ≈ T84 > HFF = 0; for lactoferrin, the only cell type showing measurable release was HTE; for mucus, HTE > Calu-3 > 16HBE14o- ≈ T84 > HFF = 0. A wide variety of neurohumoral agents and inflammatory stimuli was without effect on lactoferrin and lysozyme release from HTE or Calu-3 cells, although forskolin did stimulate secretion of water and lysozyme from Calu-3 cells. However, the concentration of lysozyme in the forskolin-induced secretions was much less than in airway gland secretions. Thus our data cast doubt on the utility of Calu-3 cells as a model of airway serous gland cells but do suggest that HTE could prove highly suitable for studies of mucin synthesis and release.

Elevated levels of Na and Cl in airway surface liquid may play a major role in the airway pathology of cystic fibrosis (CF) (J. J. Smith, S. M. Travis, E. P. Greenberg, and M. J. Welsh. Cell85: 229–236, 1996) and could be caused by block of transcellular Cl absorption due to lack of a functional CF transmembrane conductance regulator (CFTR). To test for transcellular absorption of Cl across non-CF epithelium, we studied how fluid absorption was affected by the opening and closing of Cl channels. Forskolin (an activator of CFTR) tripled fluid absorption across primary cultures of bovine tracheal epithelium but had no effect on human cells. However, in both species, fluid absorption was markedly inhibited by 5-nitro-2-(3-phenylpropylamino)benzoate, a blocker of CFTR. Microelectrode studies suggested that the magnitude of the absorptive response to forskolin in bovine cells depended on the size of an inwardly directed electrochemical driving force for Cl movement across the apical membrane. Patch-clamp measurements of bovine cells revealed CFTR in the apical membrane and a cAMP-activated, inwardly rectifying Cl channel in the basolateral membrane. We conclude that a significant fraction of absorbed Cl passes transcellularly in bovine tracheal epithelial cultures, with CFTR as the path of entry in the apical membrane and a novel cAMP-activated Cl channel as the exit route in the basolateral membrane. Our data further indicate that a similar pathway may exist in non-CF human tracheal epithelium.

The luminal surface of airways is lined by a thin film of airway surface liquid (ASL). Physiological regulation of the depth of ASL has not been reported previously. In this paper, we have used low-temperature scanning electron microscopy of rapidly frozen specimens of bovine tracheal epithelium to demonstrate alterations in the depth of ASL in response to the cholinergic agonist methacholine. We first established that methacholine selectively stimulated airway glands, with maximal secretion at approximately 2 min and a return to baseline within approximately 5 min. A 2-min exposure to methacholine increased the depth of ASL from 23 to 78 microns. Thereafter, depth decreased linearly with time, reaching 32 microns at 30 min. The initial increase in depth was blocked by bumetanide, an inhibitor of active chloride secretion, whereas the slow decline back to baseline was inhibited by amiloride, a blocker of active sodium absorption. We conclude that the methacholine-induced changes in ASL depth reflect transient gland secretion followed by liquid absorption across the surface epithelium.

We have tested two hypotheses: 1) the cystic fibrosis transmembrane conductance regulator (CFTR) represents the predominant Cl conductance in the apical membrane of human tracheal epithelium, and 2) CFTR in this tissue is close to maximally activated under baseline conditions. In support of the first hypothesis, we found 1) when the level of differentiation of cultures was varied by varying the culture conditions, there was a significant positive correlation between the levels of CFTR and the magnitude of mediator-induced Cl secretion. 2) Amiloride-insensitive baseline short-circuit current (Isc) and mediator-induced increases in Isc were inhibited by diphenylamine-2-carboxylic acid (DPAC) but not by 4,4'-diisothiocyanostilbene-2,2'-disulfonic acid (DIDS), a pharmacology consistent with passage of apical membrane Cl current through CFTR; Ca-activated Cl channels are inhibited by DIDS but not by DPAC. 3) Raising temperature from 22 degrees to 37 degrees C increased 125I efflux, and this increase was inhibited by DPAC and blockers of protein kinase A, but not by DIDS or 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid-acetoxymethyl ester. In support of the second hypothesis, we have earlier shown [M. Yamaya, W.E. Finkbeiner, S.Y. Chun, and J.H. Widdicombe. Am. J. Physiol. 262 (Lung Cell. Mol. Physiol. 6): L713-L724, 1992] that adenosine 3',5'-cyclic monophosphate (cAMP)-elevating agents are essentially without effect on Isc across primary cultures of human tracheal epithelium. Here, we further show that these agents are also usually without effect on 125I efflux; the mean increase in efflux in response to elevating cAMP was approximately 20% that of raising temperature from 22 degrees to 37 degrees C.

Cells from the acini of human tracheal glands were grown in culture to produce confluent cell sheets of mucous or mixed seromucous phenotype. Levels of mediator-induced Cl secretion in mucous cells were 2-18% those of seromucous cells. Levels of the cystic fibrosis transmembrane conductance regulator (an apical membrane Cl channel) were also much less in mucous than in seromucous cells. These results suggest that serous cells are more important than mucous cells in providing the fluid component of gland secretions.

Of 12 cell lines derived from human lung cancers, only Calu-3 cells showed high transepithelial resistance (Rte) and increases in short-circuit current (Isc) in response to mediators. Calu-3 cells formed polarized monolayers with tight junctions and Rte of approximately 100 omega.cm2. Baseline Isc was approximately 35 microA/cm2 and was increased by approximately 75 microA/cm2 on elevation of intracellular adenosine 3',5'-cyclic monophosphate (cAMP) by isoproterenol. Flux studies showed that the increase in Isc was due to Cl- secretion. Forskolin and permeant analogues of cAMP also increased Isc. Consistent with the presence of cAMP-dependent Cl- secretion, immunoprecipitation demonstrated the presence of the cystic fibrosis transmembrane conductance regulator (CFTR). Bradykinin, methacholine, trypsin, and histamine all transiently (15–30 s) elevated Isc, probably by increasing intracellular Ca concentration. Experiments in which the basolateral membrane was permeabilized with nystatin indicated that CFTR was substantially activated under baseline conditions and that Ca-activated Cl- channels were absent from the apical membrane. We anticipate that Calu-3 cells will prove useful in the study of Cl- secretion and other functions of human airway epithelial cells.

Here we describe the conditions which allow cultured human tracheal epithelial cells to retain the ion transport properties and ultrastructure of the original tissue. The order of potency of growth supports and media additives in elevating baseline short-circuit current (Isc) and responses to mediators were vitrogen gel (VIT) greater than extracellular matrix from bovine corneal endothelial cells (ECM) greater than human placental collagen (HPC), and 2% Ultroser G serum substitute (USG) greater than 5% fetal calf serum (FCS) greater than defined growth factors (GF). For all combinations of medium and growth supports, an air interface (AIR) gave better electrical properties than immersion feeding (IMM). As opposed to our earlier conditions (HPC/FCS/IMM), the best new combination (VIT/USG/AIR) produced higher baseline Isc (58.0 +/- 10.6 vs. 5.1 +/- 1.0 microA/cm2) and increased Isc responses to isoproterenol (6.1 +/- 1.5 vs. 0.8 +/- 0.3 microA/cm2) and bradykinin (9.6 +/- 2.0 vs. 1.0 +/- 0.2 microA/cm2), while retaining high transepithelial resistance (227 +/- 5 omega.cm2). VIT/USG/AIR led to the appearance of cilia, an increase in the depth of the cell sheets (50 vs. 10 microns), longer and more frequent apical microvilli, and increased interdigitations of the basolateral membrane. Protein and DNA content were also significantly increased. Secretory granules were present which stained with antibody to goblet cells, but not to serous or mucous gland cells. CF cells grown in VIT/USG/AIR showed high baseline Isc (69 +/- 18 microA/cm2) and a proportionately larger inhibition of Isc by amiloride (70 +/- 10 vs. 34 +/- 3%). Isc did not respond to isoproterenol, and the response to bradykinin was 22% normal.

The role of adenosine 3',5'-cyclic monophosphate (cAMP) and protein phosphorylation during beta-adrenergic receptor stimulation of bovine tracheal gland serous cells was investigated in vitro. Isoproterenol, a beta-adrenergic agonist, increased the secretion of 35S-labeled molecules. Intracellular cAMP levels were increased within 1 min after stimulation of bovine tracheal gland serous cells with isoproterenol. The dose-response relationship for isoproterenol-stimulated generation of cAMP correlated with the dose-response relationship for isoproterenol-stimulated secretion of 35S-labeled molecules. The phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine potentiated both isoproterenol-evoked secretion of 35S-labeled molecules and the production of intracellular cAMP, and the beta-adrenergic receptor antagonist propranolol completely blocked both effects. The secretory response of the cells to isoproterenol could be mimicked by the cAMP analogues 8-bromoadenosine 3',5'-cyclic monophosphate and dibutyryl adenosine 3',5'-cyclic monophosphate. Activity of cAMP-dependent kinase was measured in soluble and particulate cell extracts. cAMP effected the state of phosphorylation of proteins associated with the soluble but not the particulate fraction. These studies are consistent with the hypothesis that beta-adrenergic stimulation of secretion from bovine tracheal gland serous cells occurs via a cAMP-mediated pathway and that one of the molecular events in this pathway is cAMP-dependent protein phosphorylation.

In primary cultures of dog tracheal epithelium, isoproterenol produced a transient increase in short-circuit current (Isc) (duration 30 s; maximal increase, 32 +/- 5 microA/cm2). This was followed by a more slowly developing sustained increase (9 +/- 3 microA/cm2), which mimicked the response to N6, 2'-O-dibutyryladenosine 3',5'-cyclic monophosphate (DBcAMP). The transient and sustained responses had dissociation constants for isoproterenol of 2 x 10(-8) and 2 x 10(-9) M, respectively. Bradykinin (in the presence of indomethacin), substance P, histamine, and thrombin produced only transient increases in Isc. The time courses of these transients closely paralleled changes in concentration of intracellular Ca ([Ca2+]i) as measured with fura 2. For different mediators, there was a significant correlation between the maximal transient increase in Isc and the maximal increase in [Ca2+]i. The transients in Isc were not associated with elevation of adenosine 3',5'-cyclic monophosphate (cAMP) and were unaffected by pretreatment with DBcAMP, which abolishes the steady-state increase in response to isoproterenol. Both the transient increases in Isc and [Ca2+]i were inhibited by pretreatment with the Ca chelator 1,2-bis(2-aminophenoxy)ethane N,N,N',N'-tetraacetic acid. The phorbol ester 12-O-tetradecanoylphorbol 13-acetate abolished the transient increases in [Ca2+]i and Isc in response to isoproterenol but not to bradykinin. These results provide evidence that 1) isoproterenol and bradykinin elevate [Ca2+]i by different mechanisms, and 2) Ca elevation is associated with a transient increase in Isc, whereas increased cAMP is associated with a smaller sustained increase.

Acini of tracheal glands from 12 humans without cystic fibrosis (CF) and from two CF patients were isolated by enzymatic digestion. They were plated on flasks coated with human placental collagen (HPC) in media containing Ultroser G serum substitute and a variety of growth factors. Confluent cell sheets formed after 20 days. Cells were then isolated by trypsinization and replated at 10(6) cells/cm2 onto porous-bottomed inserts coated with HPC. Confluent sheets formed on day 1 after replating and were studied on day 10 in Ussing chambers. Transepithelial resistance (Rte) and baseline short-circuit current (Isc) of CF cultures (171 +/- 67 omega.cm2, 3.8 +/- 0.8 microA/cm2; n = 5) were significantly less than non-CF (541 +/- 116 omega.cm2, 9.9 +/- 2.3 microA/cm2; n = 14). Responses in Isc to mediators were also significantly reduced in CF: isoproterenol (10(-5)M) = 0.04 +/- 0.04 vs. 1.9 +/- 0.6; methacholine (10(-5)M) = 0.2 +/- 0.1 vs. 7.1 +/- 1.7; bradykinin (10(-6)M) = 0.4 +/- 0.1 vs. 5.0 +/- 1.0 microA/cm2; n = 5 for CF, n = 14 for non-CF. When CF and non-CF cells were matched for baseline Isc and Rte, the responses of CF cells to mediators still remained statistically lower than normal. The reduced responses of CF cells to bradykinin and methacholine, in addition to isoproterenol, suggest that both Ca-dependent and adenosine 3',5'-cyclic monophosphate-dependent regulation of Cl secretion are defective in CF tracheobronchial glands. The resulting reduction in fluid secretion by glands may contribute to the accumulation of airway mucus in CF.

Cultures of dog tracheal epithelium have proved very useful in studies of ion transport. Their short-circuit current (Isc), however, is usually much less than the original tissue. We have tested a variety of conditions in an attempt to produce large numbers of cells with electrical properties comparable with the original tissue. Of several growth supports, human placental collagen (HPC) gave the best results. When plated at 2.5 x 10(5) cells/cm2 onto HPC, cells grown in serum-free, growth factor-supplemented medium (GF medium) showed increases in cells per unit area, thickness of cell sheet, numbers of domes, numbers of apical microvilli, and degree of basolateral membrane interdigitation compared with cells grown in medium containing 5% fetal calf serum (FCS medium). Transepithelial resistance (Rte) and the increases in Isc and intracellular Ca in response to isoproterenol were also increased. However, baseline Isc and adenosine 3',5'-cyclic monophosphate levels were not changed. The improved electrical properties were maintained for up to 4 mo. GF medium combined with an air interface produced further increases in Rte, Isc, and changes in Isc in response to amiloride and isoproterenol. Ultrastructural features such as the presence of cilia, greater thickness of the cell sheet, and increased amplification of apical and basolateral membranes also indicated improved differentiation. Our results show that GF medium and an air interface can be combined with a simple growth support and a relatively low-plating density to allow the easy production of greater than 500 cm2 of cultured cells from a single trachea, with a level of differentiation similar to that of the original tissue.

Apical membranes from cow tracheal epithelium were prepared in a two-step process. First, most of the unwanted membranes in the crude homogenate were aggregated with Mg and removed by a low-speed spin. The membranes remaining in the supernatant were pelleted by a high-speed spin, resuspended, and exposed to ouabain-affinity chromatography. This step removed approximately 50% of the protein, all the Na-K-adenosinetriphosphatase, but had no effect on total levels of alkaline phosphatase (a marker for apical membranes). The specific activity of the apical membrane marker, alkaline phosphatase, was 21 +/- 7-fold (mean +/- SD) greater in the apical membranes than in the homogenate. Markers for nuclei, mitochondria, and basolateral membranes were excluded compared with the homogenate. Similar results were obtained with primary cultures of cow tracheal epithelium. The vesicular nature of the membranes was demonstrated in isotope uptake studies that revealed an osmotically active space.

In 1984, primary cultures of dog tracheal epithelium were shown to retain the ion transport processes of the original tissue. Since then, ion transport in both nasal and tracheal cell cultures has been studied with Ussing chambers, conventional and ion-sensitive microelectrodes, radiotracers, Cl-sensing fluorescent dyes, and patch clamping. Much of the resulting information could not have been obtained using the native tissues and has been especially useful in determining the ion transport defect(s) in cystic fibrosis. Recent improvements in growth conditions should make highly differentiated cell cultures routinely available, further increasing the usefulness of cultured cells for the study of ion transport by airway epithelium.