The airway surface liquid (ASL) is the thin fluid layer lining the airways whose depth may be reduced in cystic fibrosis. Prior measurements of ASL depth have been made in airway epithelial cell cultures. Here, we established methodology to measure ASL depth to ∼1-μm accuracy in ex vivo fragments of freshly obtained human and pig tracheas. Airway fragments were mounted in chambers designed for perfusion of the basal surface and observation of the apical, fluorescently stained ASL by scanning confocal microscopy using a high numerical aperture lens immersed in perfluorocarbon. Measurement accuracy was verified using standards of specified fluid thickness. ASL depth in well-differentiated primary cultures of human nasal respiratory epithelium was 8.0 ± 0.5 μm (SE 10 cultures) under basal conditions, 8.4 ± 0.4 μm following ENaC inhibition by amiloride, and 14.5 ± 1.2 μm following CFTR stimulation by cAMP agonists. ASL depth in human trachea was 7.0 ± 0.7 μm under basal conditions, 11.0 ± 1.7 μm following amiloride, 17.0 ± 3.4 μm following cAMP agonists, and 7.1 ± 0.5 μm after CFTR inhibition. Similar results were found in pig trachea. This study provides the first direct measurements of ASL depth in intact human airways and indicates the involvement of ENaC sodium channels and CFTR chloride channels in determining ASL depth. We suggest that CF lung disease may be caused by the inability of CFTR-deficient airways to increase their ASL depth transiently following secretory stimuli that in non-CF airways produce transient increases in ASL depth.
- cystic fibrosis
- fluorescence microscopy
- cystic fibrosis transmembrane conductance regulator
the airway surface liquid (ASL) is the thin layer of aqueous fluid bathing the airways at the interface between airway epithelial cells and the air space. The ASL provides a suitable chemical and physical environment for mucociliary clearance, bacterial killing, and airway epithelial cell homeostasis (2, 3, 22, 33). ASL composition and volume (depth) depend on secretion/absorption by airway surface epithelial cells, fluid secretion by submucosal glands, and fluid convection up the airways.
Altered ASL composition and volume in cystic fibrosis (CF) have been proposed as primary mechanisms responsible for chronic lung infection in CF, where excessive, viscous mucus accumulates in the airways, impairing mucociliary clearance and bacterial killing. The defect in CF is mutation of the cystic fibrosis transmembrane conductance regulator (CFTR) Cl− channel, which reduces Cl− permeability in various cell types including the airway epithelium (3, 21). An early hypothesis relating defective CFTR function to CF lung disease was increased ASL sodium chloride concentration in CF resulting from impaired Cl− absorption (“high-salt hypothesis”, Ref. 25); however, the near-isotonicity of CF and non-CF ASL measured by several laboratories provided direct evidence against this hypothesis (12, 15, 18, 27). Matsui and colleagues (18) have proposed “ASL dehydration” in CF, where ENaC Na+ channel hyperactivity, perhaps due to abrogation of an inhibitory CFTR-ENaC interaction, produces a thinned ASL that impairs mucociliary clearance (4, 31). The ASL dehydration hypothesis has become widely accepted and has provided the scientific rationale for ongoing development of CF therapies aimed at increasing ASL depth, including inhibitors of ENaC (9) and activators of alternative (non-CFTR) Cl− channels (6, 7, 11, 14, 35).
Evidence for reduced ASL volume (depth) in CF has come primarily from measurements of ASL depth in cultures of human airway surface epithelial cells grown on porous filters (18, 30, 33). However, it is recognized that cell cultures may not be a good surrogate for intact airways with regard to ASL properties because of differences in morphology, cellular composition and mucociliary clearance dynamics, as well as the absence of submucosal glands, surface goblet cells, and humoral factors (5, 15, 24, 33, 36, 37). Furthermore, the phenotype of airway cell cultures is quite variable depending on the precise culture conditions and passage number. Transepithelial resistances in a range of 125–1,000 Ω/cm2 have been reported by different labs using different culture media and conditions (16–18, 38, 39), with a median resistance of ∼300 Ω/cm2 for cultures used for ASL depth measurements (18, 33). Studies using airway cell cultures suggest complex regulatory mechanisms for ASL depth, involving activities of ion channels, ATP and adenosine concentrations, surface shear forces, ciliary surface tension, protease concentration, and hormone concentrations (5, 19, 30, 32, 33). It has not been tested whether conclusions about ASL regulatory mechanisms and ASL dehydration in CF derived from cell culture studies can be extrapolated to intact airways.
The purpose of this study was to develop methodology for accurate measurement of ASL depth in ex vivo airways in order to measure ASL depth in human and pig trachea and investigate effects of Na+ and Cl− channel modulators. The goal was to measure ASL depth to 1-μm accuracy in fragments of intact airways having a stable ASL. Our studies required the construction of custom perfusion chambers and the development of procedures for tissue processing, fluorescent dye labeling, microscopy, and data analysis.
Primary cultures of human nasal epithelial cells were grown as described (17, 39). Fragments of nasal epithelium, obtained by brushing the nasal mucosa of healthy non-CF volunteers and CF subjects, were collected in 15-ml centrifuge tubes containing PBS supplemented with penicillin (105 U/l), streptomycin (100 mg/l), gentamicin (50 mg/l), and amphotericin B (2.5 mg/l), hereafter referred to as “PBS with antibiotics.” Samples were rinsed three times in PBS with antibiotics containing 5 mM dithiothreitol to reduce any contamination with mucus. The tissue was then incubated at 4°C overnight in PBS with antibiotics containing 0.025% protease (from Streptomyces griseus, type XIV, Sigma). The following day, the enzyme solution was removed and replaced by a 1:1 mixture of DMEM and Ham's F-12 containing 5% FCS and antibiotics. Cell number and viability (>90%) were determined by Trypan blue staining, and the cells were plated at a density of 5 × 105 per cm2 onto 12-mm-diameter, 0.4-μm pore polycarbonate cell culture inserts (Snapwell; Corning, Lowell, MA) precoated with human placental collagen (15 mg/cm2; Sigma). Cultures were grown at an air-liquid interface in ALI medium (17) at 37°C, 5% CO2/95% air to develop functionally differentiated cell sheets. Medium was changed every 2–3 days. Transepithelial resistance was measured using an epithelia volt-ohmmeter (World Precision Instruments, Sarasota, FL). Cultures were used 21–23 days after plating, at which time resistance was 400–1,000 Ω/cm2 and a thin ASL film was seen. All procedures involving human subjects were approved by the Institutional Review Board, Office of Research, University of California, San Francisco, California.
Human and pig airways.
Human airways were obtained as transplant rejects from the California Lung Transplantation Donation Network. Fragments of trachea were used for ASL measurements, and the remaining lung tissue was used for other studies. Human tracheas used in this study were kept on ice in a plastic bag and transported to the lab within 12 h after lung harvest. Pig tracheas were obtained from Pork Farm (Stockton, CA) and transported on ice in plastic bags to the lab within 12 h after harvest.
Human nasal respiratory epithelial cells grown on Snapwell permeable supports (1.12-cm2 surface area) were mounted in Ussing chambers (Physiologic Instruments, San Diego, CA). Cells were bathed for a 30-min stabilization period in HCO3−-buffered solution containing (in mM): 120 NaCl, 5 KCl, 1 MgCl2, 1 CaCl2, 10 d-glucose, 5 HEPES, and 25 NaHCO3 (pH 7.4) and aerated with 95% O2/5% CO2 at 37°C. For measurements in human and pig trachea, the posterior mucosa, which is membranous and lacks cartilage, was isolated and cut into small fragments (∼1.5 × 1.5 cm), rinsed in PBS, and mounted in Ussing chambers. Human and pig tracheas were bathed for a 120-min stabilization period in HCO3−-buffered solution and aerated with 95% O2/5% CO2 at 37°C. Baseline potential difference (PD) was −5.8 ± 0.4 mV (apical side negative, n = 4), and resistance was 259 ± 19 Ω/cm2 in pig trachea; PD and resistance were −3.9 mV and 240 Ω/cm2, respectively, in human trachea. Short-circuit current (Isc Isc was recorded continuously with agonists/inhibitors added at specified times.
Perfusion chambers for ASL measurements.
Six identical stainless steel perfusion chambers were constructed to perfuse the serosal surface of cell cultures and tracheal fragments in which the fluorescently stained ASL at the apical (mucosal) surface is observed through an objective lens (Fig. 1, A and B). Stainless steel was used because of its excellent thermal conductivity, inertness, and mechanical strength. The chamber consists of a bottom portion for perfusion upon which the cell layer/tissue is mounted and a top portion (lid) that is secured by four recessed screws. Four pins are used to secure the tissue on the bottom portion of the chamber with mucosal side facing upward. A rubber O-ring between the lid and bottom portion of the chamber prevents fluid leakage around the undersurface of the cell layer/tissue. The diameter of the circular cut-out in the lid is 12 mm to allow for ASL imaging using an immersion objective lens. The surface area of the exposed mucosa is ∼1 cm2, so that addition of 20 μl of fluid on top of the dried mucosa produced a 16- to 20-μm increase in ASL depth in the central area of the filter. In all studies, the objective lens was heated to 37°C using a band-heater (Bioptech). For ASL measurements in cell cultures, the flat porous membrane was excised from the insert using a scalpel and mounted in the perfusion chamber with cells facing up and the porous membrane facing the perfusate. For ASL measurements on tracheal mucosal fragments, the posterior airway mucosa was isolated from underlying serosal fat tissue by gentle dissection. The mucosa was cut into small fragments (∼1.5 × 1.5 cm) and mounted in perfusion chambers with mucosa facing up and serosal surface contacting the perfusate. The perfusion chamber was placed in a 37°C microincubator (Harvard Apparatus) on the microscope stage.
Scanning confocal microscopy.
The ASL was imaged using a Nikon EZ-C1 confocal microscope equipped with a water-immersion objective lens (Nikon Plan APO, ×60, numerical aperture 1.2, working distance 270 μm). Generally, a stack of 100 images (1 μm apart) was acquired over 40–60 s. The x-y scanned area was 236 μm × 236 μm at a resolution of 256 × 256 pixels. The ASL was covered with a high-boiling point, water-immiscible, inert perfluorocarbon (FC-70, refractive index 1.303, density 1.9 g/ml; 3M) to prevent evaporation and for immersion of the objective lens. Although the density of FC-70 is greater than that of water, the ASL is not perturbed because of surface tension effects. As explained below, the ASL was stained red with rhodamine B-dextran (10 kDa, Invitrogen), an aqueous-phase dye, and the epithelial cell cytoplasm was stained green with the cell-trappable dye Cell Track Green (Invitrogen).
ASL depth measurements.
For measurements in airway cell cultures, the ASL was stained with rhodamine B-dextran by depositing a very small amount of solid (powder form) dye (∼2 μg) on the cell apical surface 2 h before initial measurements; cell cytoplasm was stained green by incubation during the 2-h time with Cell Track Green in culture medium (20 μM) in the perfusate bathing the undersurface of the porous filter. After the 2-h incubation, various ion channel activators and/or inhibitors (100 μM amiloride, 20 μM forskolin + 100 μM IBMX, 20 μM CFTRinh-172) were added to the culture medium and incubated at 37°C for an additional 30–120 min. FC-70 (20 μl) was added to cover the ASL before the snapwells were removed from the incubator.
For measurements in pig and human trachea, fragments of posterior mucosa as described above were mounted in perfusion chambers. Twenty microliters of PBS (containing 20 μM Cell Track Green) was added onto the mucosal surface. After incubation at 37°C for 45 min, the surface fluid was removed carefully using a pipette, and the mucosal surface was rinsed twice with PBS and dried briefly with an air stream. Twenty microliters of PBS containing rhodamine B-dextran (1 mg/ml) and activators/inhibitors (listed above) was added to the mucosal surface, which was then covered with 50 μl of FC-70. Because of surface tension effects, much of the added fluid accumulates at the edge of the filter at the interface between the filter and curved chamber wall. For CFTR inhibition, CFTRinh-172 was used for human cell cultures and trachea. The glycine hydrazide MalH-2, synthesized as described (26), was used for pig trachea because of its greater potency for inhibition of pig CFTR than CFTRinh-172. The serosal bathing solution contained the same concentrations of activators/inhibitors and was changed hourly. ASL depth was measured just after FC-70 addition and at specified times thereafter. Between measurements, the perfusion chamber was maintained at 37°C in an incubator. In some studies, ASL depth in pig trachea was measured at 0, 1, 2, 4, 6, 8, and 12 h after adding fluid on top of the mucosa. In some studies, to test amiloride and CFTR activator effects on steady-state ASL depth, pig tracheas were incubated at 37°C for 6 h in the absence of agonists/inhibitors, followed by addition of amiloride or forskolin + IBMX to the ASL (1 mg of dry powders dispersed by sonication in FC-70) and to the serosal perfusate (100 μM amiloride or 20 μM forskolin + 100 μM IBMX).
Culture inserts and tracheas were fixed in 10% buffered formalin and embedded in paraffin. Sections were cut at 5 μm using a rotary microtome and transferred to glass slides. Sections were stained with hematoxylin-eosin and photographed on a light microscope equipped with a digital camera. To show overall tracheal morphology, whole trachea including membranous and cartilaginous portions, was trans-sectioned and stained with hematoxylin-eosin.
ASL data were expressed as number vs. ASL depth histograms using fluorescence intensity data from all pixels in image stacks (generally 100 images at different z-positions, each 256 × 256 pixels). The depth histograms provide a summary of all ASL depths computed for each of 2562 = 65,536 z-fluorescence profile corresponding to each x,y-position. To construct histograms, z-fluorescence profiles were extracted from data sets for each of the 65,536 distinct x,y-positions in image stacks. The z-fluorescence profiles were smoothed using a five-point smoothing routine to reduce noise effects. ASL depth for each z-fluorescence profile was computed using a 70% threshold to identify ASL “edges” (see results); this threshold was determined empirically to give correct thicknesses of solution standards. All 65,536 ASL depth values were summarized in number histogram format, in which the number of ASL values in 0.01-μm bins were plotted on the y-axis. Averaged ASL depths were computed from histograms by averaging the central 70% of depth values, excluding the highest and lowest 15% of values. In general, selecting the central 70% of depth values gave nearly the same average as using all depth values.
Statistical differences among different groups in cell culture and airway studies were determined using ANOVA. P less than 0.05 was considered to be statistically significant.
Validation of methodology for ASL depth measurement.
Our measurement approach, developed following much trial and error, involves scanning confocal microscopy of the ASL using a high numerical aperture lens immersed in a layer of perfluorocarbon covering the ASL. The challenge in accurately measuring ASL depth in intact airways was the need to image the ASL, without perturbing its properties, with a high-numerical aperture, short-working distance, immersion lens. ASL depth measurement with ∼1-μm resolution was accomplished by z-scanning laser confocal microscopy in which the fluorescently stained ASL was visualized through perfluorocarbon using a ×60 water-immersion lens with a numerical aperture of 1.2 and a working distance of 270 μm. This objective was chosen after screening a series of objectives with different magnifications, numerical apertures, physical sizes, and working distances. The perfluorocarbon used for immersion is chemically inert, water-immiscible, high-boiling point and oxygen-permeable, and has a refractive index (1.303) similar to that of water (1.33) to allow acquisition of nondistorted images using a lens engineered for water immersion. The perfusion chamber was designed to immobilize small airway fragments with temperature control, basal surface perfusion, and apical surface imaging with the immersion lens. Six identical perfusion chambers were constructed to carry out six parallel experiments on different samples over an extended time period in which samples were moved between a 37°C cell culture incubator and a microincubator on the stage of the confocal microscope. A schematic of the perfusion chamber is shown in Fig. 1A and a photograph in Fig. 1B. The mucosal surface with fluorescently stained ASL and overlying perfluorocarbon fluid layer faces upward, and the serosal surface contacts the perfusate from below.
Figure 1C shows x-z confocal reconstructions of fluorescence intensities for 4-, 5-, 6-, 10-, and 20-μm-thick layers of aqueous fluid (between cover glasses) containing rhodamine B-dextran. In addition, data for “0 μm” (dried layer of TMR-dextran in place of solution) is included as a measure of the z-point-spread-function of the optical system and hence the smallest measurable ASL depth. Measurements were done under conditions identical to those used for subsequent ASL measurements in which a water-immersion lens is immersed into perfluorocarbon overlying the fluorescently stained fluid. The bright fluorescent fluid layer is well demarcated in each case. Figure 1C shows overlaid z-fluorescence intensity profiles measured for various fluid thicknesses. Intensity profiles were nearly identical for different x,y-positions in the image plane. Fluorescence is near zero well above or below the fluid layer and maximum at the center of the fluid layer. Intensity profiles are curved rather than ideal (rectangular-shaped) because of the non-delta function point-spread-function of a real optical system. To deduce fluid thickness from intensity profiles, an empirically determined threshold of 70% was used to identify the boundaries of the fluorescent layer. Using this threshold, the deduced and specified fluid thicknesses were in agreement for multiple samples of different thicknesses.
Figure 1D shows number vs. ASL z-depth histograms for all pixels in image stacks. We use a number histogram, which summarizes the number of ASL z-depths in predefined bins, to display each of the 65,536 ASL z-depths corresponding to each x,y-location. These number histograms are constructed from the 65,536 ASL z-depths by standard binning procedures. Narrow distributions were found in each case, with differences in thickness of 1 μm readily resolved. We reasoned that a histogram analysis of data from all pixels in full image stacks would provide a rigorous, unbiased approach to analyze cell/airway ASL data and allow assessment of possible heterogeneity in ASL depth in different areas of a sample.
ASL measurements in well-differentiated human nasal respiratory epithelial cell cultures.
Our measurement method was applied first to study the determinants of ASL depth in well-differentiated primary cultures of human nasal respiratory epithelial cells. Histology of non-CF and CF cultures showed similar well-differentiated, pseudostratified ciliated epithelial sheets (Fig. 2A). Isc analysis showed amiloride-sensitive ENaC activity greater in CF than in non-CF cultures (Fig. 2B), in agreement with many prior studies. The increase in CFTR current in response to forskolin in the non-CF cultures was fully inhibited by CFTRinh-172. Forskolin did not increase current in the CF cultures. UTP produced a transient increase in Isc in both the non-CF and CF cultures, resulting from stimulation of calcium-activated chloride channels.
For confocal fluorescence imaging, the ASL was stained red with the aqueous-phase dye rhodamine B-dextran, and cell cytoplasm was stained green by 2-h incubation with Cell Track Green, followed by incubations with various CFTR/ENaC activators/inhibitors added in the culture medium for an additional 2 h. The flat porous membrane containing the epithelial cell layer was mounted in a perfusion chamber, and perfluorocarbon was added on top of the cells. Figure 2C shows representative x-z confocal reconstructions and z-fluorescence profiles. The overlaid z-fluorescence profiles from randomly chosen x,y-coordinates were remarkably similar to one another. Figure 2C shows deduced number vs. z-depth histograms, which were generally unimodal and nearly symmetrical, although broader than histogram distributions in Fig. 1D measured for solution standards. The broader histograms represent a combination of effects of cell surface heterogeneity, imperfect surface flatness, and reduced fluorescence signal compared with the solution standards. Average ASL depth in non-CF nasal cells under control conditions (no activators or inhibitors) was 8.0 ± 0.5 μm (SE, 10 cultures), similar to values of ∼7 μm reported previously for human tracheal epithelial cell cultures (33). We found that ASL depth remained constant for at least 4 h during measurements.
Figure 2D summarizes averaged ASL depths from cell culture studies. ASL depth in CF cell cultures under control conditions was 6.0 ± 0.3 μm, slightly although significantly less than that of 8.0 ± 0.5 μm in non-CF nasal cell cultures. The reduced ASL depth in CF cultures could be a consequence of Na+ hyperabsorption and/or Cl− hyposecretion, as well as of differences in the activities of other ion channels. CFTR activation by forskolin/IBMX increased ASL depth in non-CF cultures but produced little effect in CF cultures. Figure 2D shows the time course of increasing ASL depth in the non-CF cell cultures following addition of forskolin/IBMX to the perfusate. CFTR inhibition by CFTRinh-172 largely abolished the increase in ASL depth caused by forskolin/IBMX. ENaC inhibition by amiloride incubation for 2 h did not significantly increase ASL depth in non-CF cultures, but did in the CF cultures. The greater amiloride effect in CF cells may be related to higher ENaC activity.
ASL measurements in ex vivo trachea.
Studies were done on fresh pig tracheas that were delivered to the lab on ice within 12 h after harvest. Pig tracheas obtained in this way remained viable for at least 24 h as evidenced by active ciliary beating, submucosal gland fluid secretion, and stability of transepithelial potential differences and Isc.
Figure 3A shows a whole mount of pig trachea. Because the posterior membrane lacks cartilaginous rings, it could be mounted quite flat in the perfusion chamber. Figure 3A also shows histology of posterior membranous trachea representative of that used for ASL measurements, showing typical mucociliary epithelium with submucosal tissue containing primarily serous glands and underlying trachealis muscle. Isc data in Fig. 3B shows functional activity of ENaC, CFTR, and calcium-activated chloride channels. ENaC was inhibited with amiloride, CFTR-dependent Cl− current was inhibited by MalH-2, an inhibitor of pig CFTR, calcium-activated chloride current was increased by carbachol, and the remaining current was inhibited slowly by bumetanide.
For ASL depth measurements, the tracheal mucosa was rinsed with PBS and cut into small fragments, and each fragment was mounted in a separate perfusion chamber. The ASL was stained red and epithelial cell cytoplasm green with the same dyes used for cell culture studies. The ASL was covered with perfluorocarbon, and the perfusion chambers were kept in a 37°C incubator for up to 6 h (Fig. 3C), being transferred at specified times to a microincubator on the stage of the confocal microscope for measurements of ASL depth. We chose 6 h as the investigation time based on initial studies showing a relatively stable ASL depth by 6 h following addition of 20 μl of PBS on the mucosal surface (Fig. 3D), and maintenance of full tissue viability as assessed by vital dye exclusion, gland fluid secretion, mucociliary activity, and tissue PD/resistance. Figure 3E shows a representative x-z confocal reconstruction and z-fluorescence profile. Figure 3E shows number vs. z-depth histograms for measurements made on two pig tracheas. Histograms were generally wider in the ex vivo tracheal fragments than in cell cultures. The average ASL depth in pig trachea was 7.9 ± 0.6 μm (SE, n = 3) under control conditions, similar to that in cell cultures.
Challenges in ASL depth measurements in intact airways include the non-flat airway surface and heterogeneity in different locations of the airway mucosa. The airway mucosa in intact airways is smooth, although curved, due to stretching effects from the C-shaped cartilage, but wrinkles after isolation from the cartilage and underlying musculature. A nearly smooth, flat mucosa was produced by mounting the mucosa in the perfusion chamber, with the mucosa stretched mildly over pins and immobilized between the flat plates. To reduce heterogeneity related to different locations in the airway mucosa, only the posterior, cartilage-free portion of the tracheal mucosa was used for ASL depth measurements. Also, five locations in each tissue sample were scanned to report averaged ASL depth values. Examination of histograms allowed direct visual evaluation of heterogeneity. Data were excluded in ∼15% of scans, where heterogeneity was found to be excessive based on very wide (>5 μm) and/or markedly asymmetric depth distributions.
Similar measurements were made on human tracheas obtained through Northern California Transplant Donor Network. ASL depth measurements were made within 24 h after trachea harvest, at which time we previously showed that human tracheas remain viable (34). Histology of a representative specimen in Fig. 4A shows normal tracheal structure with intact respiratory epithelium and mixed seromucous airway glands. Isc analysis in Fig. 4B shows amiloride-sensitive ENaC activity and forskolin-stimulated CFTR current that was inhibited by CFTRinh-172. Figure 4C shows representative x-z confocal reconstructions, z-fluorescence profiles, and number histograms. The average ASL depth was 7.0 ± 0.7 μm under control conditions.
Studies were done to determine the influence of ENaC and CFTR modulators on ASL depth in pig and human trachea. For these studies, as described in methods, channel modulators were present in the small volume of fluid initially added onto the mucosal surface and in the perfusate throughout the 6-h study.
Figure 5, A and B, shows representative x,z-confocal reconstructions and corresponding number histograms for effects of amiloride, CFTR inhibitors (MalH-2 or CFTRinh-172), and CFTR activators (IBMX+forskolin). Histograms were generally unimodal, although in some cases, tails were seen, and the histograms were relatively wide. Figure 5, C and D, summarizes average ASL depths for each of the four conditions tested. In both pig and human trachea, ASL depth was increased by amiloride and CFTR activators, but was not significantly changed by CFTR inhibitors. In pig trachea, CFTR inhibition prevented the increase in ASL depth produced by forskolin/IBMX.
We postulated that some or most of the increased ASL depth in amiloride-treated tracheas was due to slowed fluid absorption and hence residual (non-steady-state) ASL at 6 h. To test whether amiloride would alter “steady-state” ASL depth in pig trachea, a 6-h incubation in the absence of transport modulators was first done so that ASL depth reached its near steady-state value. Amiloride was then added onto the ASL in powder form (sonicated in FC-70) and in the perfusate. ASL depth did not change significantly at 2 h after amiloride addition (Fig. 6A). In contrast, treatment with forskolin + IBMX significantly increased ASL depth at 2 h (Fig. 6A). In a control study to verify the efficacy of amiloride under the conditions of the experiment, amiloride was confirmed to inhibit airway fluid absorption (Fig. 6B). These results suggest that CFTR activation is more effective than ENaC inhibition in increasing steady-state ASL depth in pig trachea.
An important accomplishment of this study was the development of methodology to measure ASL depth in tracheal fragments ex vivo. Our approach utilized laser scanning confocal microscopy with a high-numerical aperture, high-magnification lens that was immersed into perfluorocarbon overlying the fluorescently stained ASL. Custom chambers were constructed for perfusion of the tissue/cell layer basal surface, exposure of the apical surface, and temperature control. Test standards of specified fluid thickness indicated the ability to detect differences in fluid thickness of more than 1 μm.
We note technical differences in the methodology used here compared with that used previously for ASL measurements in cell cultures. Prior studies used an inverted confocal microscope with z-scanning to image the fluorescently stained ASL from the bottom, through the translucent porous filter and cell layer (30, 32, 33). Visualization of the fluorescently stained ASL from below is not possible for the relatively thick, optically opaque airways. To visualize the ASL from above, the fluorescently stained ASL was covered with a high-boiling point perfluorocarbon, which prevented evaporation and allowed ASL imaging using a water-immersion lens. Unlike prior ASL measurements, we used a histogram method to analyze ASL depth data from confocal image stacks, which avoided investigator bias in selection of x,z-planes by utilizing information from all pixels in all images. Accurate fluid thickness measurement was verified using solution standards of known thickness, which produced narrow, unimodal histogram distributions of thickness. Use of histogram analysis rather than user-selected planar reconstructions, as done previously, provides a quantitative, unbiased determination of ASL depth and allowed direct inspection of data for heterogeneity and technical artifact.
Measurements in well-differentiated primary cultures of human nasal respiratory epithelial cells showed a small, although significant, reduction in ASL depth in CF cultures compared with non-CF cultures, in general agreement with prior data. IBMX and forskolin significantly increased ASL depth in non-CF cultures, but not in CF cultures, as expected. Interestingly, amiloride significantly increased ASL depth only in the CF cultures, which may be related to greater ENaC activity as a consequence of differences in electrochemical driving forces and/or intrinsic ENaC hyperactivity.
The motivations for making measurements in pig trachea included the recent generation of a CF pig model (23), the considerable existing body of information on pig airway and submucosal gland physiology (1, 13, 34), and the availability of high-quality fresh pig tracheas having consistent phenotype. ASL depths in both pig and human tracheal fragments were comparable to that in the cell cultures. As in the cell culture studies, ASL depth in trachea was increased by CFTR activators. Although CFTR inhibition largely prevented the increase in ASL depth produced by CFTR activators, CFTR inhibitors alone did not affect basal (unstimulated) ASL depth. Amiloride, when present throughout the study, produced a greater ASL depth than in basal conditions, which is likely due to slowing of absorption of excess fluid added onto the mucosal surface at the start of the 6-h incubation. Addition of amiloride to a preformed ASL in pig trachea did not affect ASL depth, whereas CFTR activation by forskolin + IBMX increased ASL depth. We consider these initial measurements as primarily descriptive in nature, with many further studies needed to establish the driving forces and transport mechanisms responsible for our observations.
The relative contribution of the airway surface epithelium vs. submucosal glands in regulating ASL depth was not determined in this study, whose focus was on the development of methodology and the measurement of ASL depth in fragments of intact trachea, which most closely mimics native trachea. Addressing the longstanding question of the role of submucosal glands on ASL properties presents a considerable challenge, which perhaps can be addressed by selectively inactivating submucosal glands in ex vivo airways.
The data reported here represent the first direct measurements of ASL depth in viable fragments of intact trachea and so advance prior studies of the ASL done in cell culture models. Although measurements in tracheal fragments recapitulate the in vivo trachea better than cell cultures, tracheal fragments are not subject to neurohumoral influences as in intact trachea, nor are ex vivo tracheal fragments exposed to convective effects or shear forces. Several technical caveats of our method deserve mention as well. The use of a high-numerical aperture immersion objective required bathing the ASL in an inert, although relatively dense, perfluorocarbon, which we assume, based on observations here and prior cell culture studies, does not affect ASL properties. Another technical limitation is that geometric constraints imposed by the perfusion chamber and lens permit measurements only near the center of the sample, which is the most desirable region where edge effects are minimized. Finally, we note that drug testing in tracheal fragments requires relative prolonged incubations because of limited penetration of some agents from the basal surface, and the need to add compounds to the ASL in a manner that does not perturb its depth. Notwithstanding these limitations, accurate ASL depth measurements in tracheal fragments are quite practical, as is testing of putative modulators of ASL depth.
From the findings here, we speculate on a potential mechanism for the pathogenesis of CF lung disease that emphasizes the Cl− transporting function of CFTR rather than CFTR-dependent ENaC hyperfunction. Based on the much more substantial effect of CFTR activators than of amiloride on increasing steady-state ASL depth in tracheal fragments, we propose that an important determinant of CF lung disease is the inability of CFTR-deficient airways to increase their ASL depth following various secretory stimuli that in non-CF airways produce transient increases in ASL depth. CFTR activation greatly increased ASL depth in the 2-h time point studied here, with further studies needed to define the early time kinetics of ASL depth. Transient increases in ASL depth in non-CF airways might greatly facilitate the clearance of bacteria and particulate matter by reducing ASL viscosity, increasing convective transport, and increasing mucociliary action (28). The beneficial action of hypertonic saline is likely to involve such transient increases in ASL depth (8, 10, 29). The challenges in testing this hypothesis will include demonstrating transient changes in ASL depth in non-CF airways in vivo and a lack thereof in CF airways, and providing evidence to link impairment in ASL depth elevation to CF lung disease. Although CFTR dysfunction and consequent impaired airway chloride secretion is not a new mechanism to explain CF lung disease, the idea of transient increases in ASL depth is new and different from the prevailing view that tonic ENaC hyperfunctioning is the principal determinant of CF lung disease.
In conclusion, we have established methodology for measurement of ASL depth in tracheal fragments ex vivo. Our results using Na+ and Cl− channel modulators indicate the involvement of Na+ and Cl− channels as key determinants of ASL depth. The methodology developed here is applicable to study mechanisms of regulation of ASL depth in airways, for examination of the role of submucosal glands, and, using CF airways, for testing the ASL dehydration hypothesis.
This work was supported by National Institutes of Health Grants HL-73856, EB-00415, DK-72517, DK-86125, DK-35124, and EY-13574, and Research Development Program and Drug Discovery grants from the Cystic Fibrosis Foundation (to A. S. Verkman). Y. Song was supported in part by a UCSF Academic Senate Individual Investigator Grant; W. Namkung was supported by the Postdoctoral Fellowship Program of Korea Research Foundation Grant MOEHRD, KRF-2007-357-C00091, funded by the Korean Government.
No conflicts of interest are declared by the author(s).
We thank Dr. Songwan Jin (Dept. of Mechanical Engineering, Korea Polytechnic Univ.) for help with custom software for analysis of confocal images. We also thank Dr. Andrew Bullen at the Biological Imaging Development Center of UCSF for technical support on confocal imaging.
- Copyright © 2009 the American Physiological Society