cystic fibrosis (CF) is caused by defects in an anion channel called CF transmembrane conductance regulator (CFTR) that is found primarily in the apical membranes of epithelial cells. Most CF symptoms, such as high salt in sweat and obstruction of the pancreatic ducts, intestine, or vas deferens occur because of defective electrolyte transport. In most organs, the link between pathophysiology and loss of CFTR-mediated anion conductance is relatively easy to explain as obstruction following the loss of CFTR-mediated fluid secretion. This is not the situation in the airways, where the major symptom is chronic airway infection. These infections provoke massive neutrophilic inflammation that relentlessly destroys the lungs. Because chronic lung infection is by far the major cause of death in CF patients, it is essential to understand precisely how the loss of CFTR compromises the innate defenses of the airways.
A growing consensus holds that the old idea of defective airway mucus (8) is correct, but crucially, because of work by many researchers in the early 1980s, the hypothesis has shifted away from searching for defective mucin molecules and is focused instead on the effects of altered fluid and electrolyte transport by airway surface epithelia and glands (26). Electrolyte transport is seen as essential to the innate defense system of the airways, which has at least two components. Defects in the physical removal of pathogens from the airways by mucociliary and cough clearance play a large role (15). In addition, the early onset and severity of CF airway disease, in contrast with the milder courses of other diseases that reduce airway clearance, suggest that the loss of CFTR function in the airways also compromises the effectiveness of the vast array of innate defense molecules found in the airways, from mucins to a host of antimicrobials, anti-proteases, and anti-inflammatory compounds (9, 23, 26).
Most innate defense molecules are secreted by submucosal glands. Glands are important for airway health, and CFTR is important for proper submucosal gland function (1–3, 5, 7, 10, 13, 14, 24–27). For example, comparison of tracheal xenografts with and without glands revealed much higher levels of lysozyme and greater resistance to infection in the gland-containing grafts (5, 25), whereas measures of secretion from individual human airway glands showed that, in marked contrast to normal glands, CF glands do not secrete to VIP or forskolin (14). In contrast with this absolute defect, gland secretion to acetylcholine persists in CF glands, but appeared to be diminished in quantity and to have increased viscosity, but no other changes in ions or pH (22).
These findings, and other work indicating equivalent pH of gland mucus stimulated either by forskolin or carbachol, are at odds with a model of gland function based on the Calu-3 cell line, which shares many features of gland serous cells. In the model by Devor et al. (6), forskolin stimulates Calu-3 cells to secrete but does not activate Na+-K+-2Cl, the main transport pathway by which Cl− is moved into the cell across the basolateral membrane, leaving HCO3− as the major anion mediating fluid secretion; this model was supported by direct measures of Calu-3 fluid secreted in response to forskolin or VIP, which had average pH >8 and calculated HCO3− concentrations of 70–90 mM (11, 17). In addition, there is good evidence that CF surface epithelia, at least in culture, produce a fluid that is more acidic than normal (4).
To reinvestigate the question of whether the pH of gland mucus is indeed unchanged in CF, Song et al. (Ref. 23a; see p. C741 in this issue) studied nasal biopsies from six subjects (10–22 yr old) with minimal CF lung disease. The biopsies were embedded in agar containing 50 μM pilocarpine, and their surfaces were wiped clean and covered with an oil layer. Bubbles of mucus were secreted by the glands into the oil layer, where their pH was measured using ratiometric imaging of injected BCECF-dextran. In contrast to a prior report (12), the pH of freshly secreted mucus was ∼6.6 in the CF subjects, significantly lower than ∼7.2 measured in biopsies from eight control subjects.
Although the oil layer should isolate the mucus bubbles from the surface epithelium, Song et al. used a clever method as an additional precaution against surface cell modulation of the gland secretions. They inserted BCECF-coated micropipettes directly into the gland duct openings and drew the secreted mucus into the pipettes. These experiments were not done with CF subjects. Instead, they compared mucus generated in the presence or absence of inhibitors of CFTR, including two specific inhibitors that had been developed in the Verkman laboratory (18, 21), and found that the inhibitors reduced the pH of mucus stimulated with pilocarpine from ∼6.9 to ∼6.6. One interesting feature of these experiments is that the authors were able to stimulate gland secretion with forskolin even in the presence of CFTR inhibitors, which is unexpected given the absence of forskolin-stimulated secretion from CF glands (14).
Song et al. (23a) also used a clever method to study acid/base transport by the surface epithelium. After the surface was cleaned with a cotton swab and coated with oil, 300-nl fluid droplets were placed under the oil in such a way that they contacted the surface yet retained their spherical shapes. These fluid droplets contained BCECF-dextran and were buffered either with HCO3− or HEPES to match the bath solutions, and were then gassed appropriately. Both solutions and bath also contained various inhibitors or activators and were initially set to either pH 7.6 or 6.4. Some of the results were surprising. In the absence of HCO3−, the surface epithelium was still able to alkalinize acidic droplets, and this alkalinization was reduced by CFTR inhibitors, being almost abolished by GlyH-101. This unexpected result may be of considerable importance: what CFTR-dependent mechanism is at work here in the absence of bath HCO3− or CO2?
Song et al. (23a) also showed that if the droplets initially had a high pH, the surface airway acidified them, and this process was unaffected by CFTR inhibitors or by amiloride. Acid secretion by the surface epithelium was further stimulated by ATP + histamine, and was not inhibited by any of the proton transport inhibitors they tried when used singly. Combinations of inhibitors were not used, nor were they able to inhibit proton channels because of interference with the BCECF measurements. Hence, the mechanisms of both unstimulated and stimulated proton secretion in these experiments remain undefined.
This work is of interest because of the new methods introduced, which should be widely adopted by investigators interested in understanding how surface epithelia and glands alter the pH of the airway surface liquid. Yet the work also illustrates the difficulties inherent in working with such complex, relatively inaccessible systems. The use of biopsies from relatively healthy patients is an important addition to the use of tissues from other sources, but the small amounts of tissue available and problems inherent in its acquisition curtail what can be done: indeed, the authors were unable to apply either the gland duct cannulation method or the surface drop method because of the few biopsy samples available. In the absence of appropriate large animal models of CF airway disease, specific inhibitors of CFTR are presently the best hope we have for reproducing CF physiology and pathophysiology in intact tissues. Thus the development of potent and specific CFTR inhibitors by high-throughput screening in the Verkman laboratory (18, 21) and their informative use herein is of considerable importance. It remains to be determined why these inhibitors did not seem to reproduce the complete block of forskolin-stimulated secretion observed with CF glands; it is essential that the development of even more potent and specific inhibitors of CFTR continue.
What about the issue of the relative importance of glands and surface airway epithelia in controlling the pH of the air surface liquid? Although these results appear to show clear evidence for alterations of gland fluid pH in CF and following inhibitors, they do not resolve the issue of whether CFTR in the gland serous cells (see Fig. 1) is important in this regard, or whether the gland fluid is being modified as it passes through the ciliated duct, which appears to be a continuation of the surface epithelium, and has recently been shown to express abundant CFTR (16). This could in principle be solved by cannulation experiments, but only if the cannulating pipette is placed deep enough to collect mucus directly from the nonciliated regions of the gland duct. This is difficult to do because of the variable pathway taken by most ducts as they snake through the musculature to the surface epithelium, and it is not clear that it was done routinely in these experiments.
Despite these issues, these results provide additional evidence for an intrinsic defect in the ability of CF tissues to secrete base (4), but they do not conclusively demonstrate that CF airway surface liquid is intrinsically hyperacidic because, as the authors point out, the steady-state pH of the airway surface liquid under more physiological conditions might be capable of compensating for the lack of CFTR. If such a pH difference can be established beyond doubt, it will then be necessary to determine its contribution to airway health. In this regard, it may be informative that even healthy airways maintain a surface liquid pH that is acidic relative to serum.
This study was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-51817, the Cystic Fibrosis Foundation, and Cystic Fibrosis Research, Inc.
The author thanks Mauri Krouse for valuable discussions and insights.
- Copyright © 2006 the American Physiological Society