INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effective mucociliary clearance of the viscoelastic mucus atop the cilia in the tracheobronchial airways depends on the maintenance of an aqueous periciliary fluid layer. There is evidence that the airway epithelium maintains the homeostasis of the airway surface liquid by responding individually to perturbations of its osmotic and ionic composition. When Man and colleagues subjected the luminal surfaces of native canine tracheal epithelia to hyperosmotic, mannitol-induced stresses or to isosmotic decreases in [Na⁺] and [Cl], they observed changes in transepithelial potential difference and short-circuit current (1) indicative of cellular responses designed to normalize ion concentrations in the airways. In response to hyperosmotic or hyposmotic saline stresses in native ferret trachea, Price and colleagues observed regulatory changes in the concentrations of ions in the airway surface liquid, as indicated by changes in osmolality (2). However, net water transport in response to a hyperosmotic saline challenge was not observed, and the predicted net water transport in response to a hyposmotic saline challenge was observed only after some normalization of the ion concentrations. In the absence of ionic and osmotic stresses, Phillips and colleagues demonstrated that water fluxes across airway epithelium were bidirectional, with 30% of the luminal-to-basolateral water flux being associated with acetylstrophanthidin-inhibited luminal-to-basolateral Na⁺ transport, and 15% of basolateral-to-luminal water flux being associated with furosemide- and diphenylamine-2-carboxylate (DPC)-inhibited basolateral-to-luminal Cl transport (3, 4). Thus, the lack of or delay in net transepithelial water transport observed by Price and colleagues (2) could be explained by a model in which the water fluxes associated with actively driven Na⁺ and Cl ion fluxes are distinct from the water flux associated with an imposed osmotic gradient. In this model, it is possible that under the hyperosmotic conditions imposed by Price and colleagues (2), the sum of these water fluxes was zero, and that as a result, no net fluid transport was observed. Similarly, when Price and colleagues imposed a hyposmotic stress (2), the delay in net water transport could be explained by an initial water flux associated with ion transport opposing the water flux induced by the osmotic gradient. We therefore hypothesized that net water transport into the airway lumen is the vectorial sum of the water fluxes associated with actively driven intracellular Na⁺ and Cl ion transport and the water flux due to osmosis.

To evaluate this model, we used an anesthetized, mechanically ventilated canine model in which the humidity of the inspired air was regulated and the airway surface liquid was collected quantitatively at the posterior commissure to give respiratory tract fluid output (RTFO). We measured [Na⁺] and [Cl] in the RTFO, as well as the osmolality in the supernatant of the RTFO. To determine the airway secretory output (ASO), we subtracted the volume contributions of the administered mannitol solutions to the collected RTFO, indicated by fluorescein-labeled dextran tracers, from the RTFO. To demonstrate that a reduction of [Na⁺] and [Cl] in the airway surface liquid in the absence of an osmotic gradient causes concurrent increases of Na⁺ and Cl fluxes as well as in water flux toward the airway lumen, we administered a near-isosmotic (250 mOsm), ion-free mannitol solution (250 µl) as an aerosol to the airway lumen. To demonstrate that net water transport into the airway lumen is the sum of that associated with ion transport and that due to osmosis, we administered a hyperosmotic (950 mOsm), ion-free mannitol solution (250 µl) as an aerosol to the airway lumen.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

System Design

Humidity-controlled ventilation system. The ventilation system shown schematically in Figure 1 was used to enable positive-pressure ventilation of the dog along with a precisely controlled inspired humidity. Compressed air (~ 3.0 mg/L in absolute humidity) was humidified with a humidifier (Bird, Palm Springs, CA) to 24.5 ± 0.6 (mean ± SEM) mg/L (i.e., ~ 70% of relative humidity at 33° C). The inspiratory conduit was warmed with heat tapes and maintained at ~ 30° C, such that the temperature of the inspired air was slightly less than 34° C when it reached the trachea. The ventilation system was adjusted to give a measured tidal volume (VT) of 200 ml at a respiratory rate (RR) of 20 breaths/min in each dog. A mixing chamber in the humidified inspiratory conduit was equipped with a humidity probe (HMI36; Vaisala, Helsinki, Finland) to monitor the humidity of the humidified inspired air. A second humidity probe (HMI36; Vaisala) was used in the expired mixing chamber, which was maintained at ~ 45° C, to measure the humidity of the expired air. At 45° C, the relative humidity of the expired air ranged from 60% to 80%, the most sensitive and reliable range of the humidity probe. The humidity data from the probes were processed at 0.2 Hz by a humidity processor (HMI36A; Vaisala). The data for pulmonary parameters such as VT, RR, peak flow, and pressure, as well as for absolute humidity, were stored in a personal computer.

View larger version (13K): [in this window] [in a new window]   Figure 1.   Schematic diagram of humidity-controlled ventilation system. Compressed air is conditioned as shown prior to delivery to the dog. The thick solid line indicates the pathway of compressed air, and the thin dashed line indicates the transmission of signals and data. (1) Humidifier, (2) inspired-air mixing chamber (2 L), (3) expired-air mixing chamber (0.6 L), (4) bypass solenoid valve (normally open when not activated) (8267C23; Automatic Switch Co., Florham Park, NJ) to vent humid air when it was not being delivered to the dog, (5) inspired-air solenoid valve (normally closed when not activated) (8267C19; Automatic Switch Co.), (6) expired-air solenoid valve (normally-closed when not activated) (8267C23; Automatic Switch Co.), (7) pneumotachograph (Model 1; Fleisch, Lausanne, Switzerland) and pressure transducer (Model MP45-14; Validyne, Northridge, CA), (8) safety valve to ensure that pressure remains under 25 cm H2O to avoid any potential barotrauma, (9) Vaisala humidity processor to collect outputs from two humidity probes, (10) logic valve controller to receive signals from the pneumotachograph and to trigger the solenoid valve's operating sequence. The inspired mixing chamber as well as the intervening tubing were maintained at 30 ± 0.5° C with flexible heating tapes (Omegalux, CT) in conjunction with two temperature controllers (CN8500; Omega). To prevent condensation, the expired-air conduit was maintained at ~ 45° C with heating tapes. The expired-air mixing chamber was also maintained at ~ 45° C, with a water bath.

Modified endotracheal tube with a suction catheter. A double- lumen endobronchial tube (size 32 French; Mallinckrodt Medical, St. Louis, MO) was modified, as described by Chen and Yeates (5), to facilitate quantitative collection of the RTFO with an airtight seal to enable the use of positive-pressure ventilation. As shown in Figure 2, a proximal cuff in the oropharynx was used to form an airtight seal with the larynx, and a distal cuff placed just caudal to the posterior commissure (or interarytenoid groove), was used to secure the position of the endotracheal tube. Intermittent suction was applied to the sampling vial connected to the collection catheter to transport airway surface liquid in the posterior commissure to the vial via the catheter.

View larger version (64K): [in this window] [in a new window]   Figure 2.   A sagittal section of the oropharynx, larynx, and trachea, showing the positioning of the proximal and distal cuffs of the endotracheal tube and the positioning of the collection port in the catheter (polytetraflouroethylene; 1.29 mm I.D. and 1.90 mm O.D.). The catheter is sealed at the tip. Two ports are located symmetrically on lateral sides of the catheter in juxtaposition with the posterior commissure. This design minimizes any invagination of the mucosa into the ports. The vocal folds are located between the two cuffs. The distal cuff is confined on the dorsal side of the tube, so that mucus transport to the posterior commissure was not interrupted.

Animal Preparation

The study was approved by the Research and Development Committee of the Veterans Affairs (VA) Chicago Health Care System, West Side, Chicago, IL. The dogs used in the study were used and housed in the research facility of the West Side VA Hospital, Chicago, which is approved by the American Association of Accreditation for Laboratory Animal Care. The U.S. Government Animal Welfare Regulations and the Guide for the Care and Use of Laboratory Animals were followed throughout the study.

Eight male beagle dogs (Covance, Kalamazoo, MI), aged 1 to 2 yr and weighing 12 to 16 kg, were used. Each dog was fasted overnight but allowed water ad libitum. The dog was anesthetized with intravenous propofol (Diprivan, 7 mg/kg; Zeneca Pharmaceuticals, Wilmington, DE) and secured in the supine position. Anesthesia was maintained by continuous infusion of propofol at a rate of 800 to 1,000 µg/kg/min until the dog's jaw was relaxed. The modified double-lumen, size 32 French endotracheal tube with a suction catheter described earlier was inserted into the trachea under direct laryngoscopic visualization. The collection ports of the catheter were carefully placed in juxtaposition to the posterior commissure to collect the airway surface liquid. To avoid retardation of mucus transport into the posterior commissure, the dorsal (upper) lumen of he endotracheal tube was used for inspiration and the ventral (lower) lumen of the tube was used for expiration.

After intubation, intravenous etomidate (Amidate; Abbott Laboratories, North Chicago, IL) was administered at a rate of 5 to 10 µg/kg/ min, and the rate of propofol administration was reduced to 400 to 500 µg/kg/min. Etomidate sensitizes the carotid body, resulting in improved blood chemistry, and propofol, a respiratory depressant, was used to suppress the myotonic and clonic effects of etomidate (6). The combination of these two short-acting hypnotic agents allowed the maintenance of pH at above 7.35 and of Pa_CO2 at < 35 mm Hg. An arterial catheter (20-gauge, 2-in. long; Becton Dickinson, Sandy, UT) was placed subcutaneously into a femoral artery to facilitate the monitoring of blood pressure and the withdrawal of arterial blood samples. A Micro Sprayer catheter (Penn-Century, Philadelphia, PA), 40-cm long and 0.64 mm in diameter, was inserted via an Opti-Port (Mallinkcrodt Medical) through the ventral (inhalation) lumen of the endotracheal tube, so that the atomizing nozzle at the tip of the catheter protruded ~ 5 mm past the distal end of the endotracheal tube (inside the trachea). The Opti-Port was sealed with the catheter in place by using a custom- designed sleeve. The modified endotracheal tube was connected to a ventilation system. A 15-W heat lamp was used to prevent any condensation on the exposed parts of the endotracheal tube and fittings. Water-heated underpads and a blanket were used to maintain the dog's rectal temperature at 38 ± 0.5° C. Other physiologic monitoring included electrocardiography pulse oximetry for hemoglobin oxygen saturation (Sa_O2) and measurement of rectal temperature (SpaceLabs Medical, Redmond, WA).

Protocol

After the preparation procedure (which lasted about 40 to 50 min), each dog was stabilized for 20 min on the basis of the ventilatory parameters used throughout the protocol (i.e., mechanical ventilation with humid air [~ 70% at 33° C] at a rate of 200 ml/breath and 20 breaths/min). The RTFO during this period were collected over 1-min periods at the beginning of each of three consecutive 6-min periods, and the collected outputs were pooled (Samples A). During the 1-min collection period, intermittent suction was applied to the collecting catheter to minimize any evaporation of fluid from or condensation of fluid into the collected samples. To minimize or avoid any artifact induced by contribution to the RTFO of stimulated secretions induced by intubation and initial instrumentation, observed by Winters and Yeates (7), we did not quantitate the RTFO collected during this stabilization period, but used it to estimate baseline values of [Na⁺], [Cl], and osmolality. Following the stabilization period, a 250-mOsm mannitol solution (250 µl), containing fluorescein-labeled, 10,000-D dextran (1 mg/ml), was delivered to the trachea. Delivery was accomplished by depression of a 300-µl stainless steel syringe (Penn-Century) attached to the Micro Sprayer catheter during two consecutive inspiratory maneuvers. The mannitol solution was atomized into the trachea axially. This provided a reproducible deposition pattern of an aerosol with a volume median aerodynamic diameter of 20 µm (Penn-Century) that approached 100% efficiency. RTFO were collected 5 min after the mannitol aerosol challenge, and therafter at 6-min intervals, through use of the collection protocol described earlier (Samples B1 to B5). Immediately after the fifth RTFO collection, a 950-mOsm mannitol solution (250 µl), containing rhodamine B-labeled, 10,000-D dextran (1 mg/ml), was administered to the trachea in the same manner as the 250-mOsm mannitol solution. An arterial blood sample was taken at this time for analysis of blood gases and pH. RTFO was collected five times after the 950-mOsm mannitol challenge, through use of the same collection protocol previously described (Samples C1 to C5).

Immediately after the experiment, the RTFO collected after the 250-mOsm and 950-mOsm mannitol challenges (Samples B1 to B5 and C1 to C5, respectively) were weighed and centrifuged (Model J2-MI; Rotor JA-17; radius_Max = 123 mm; Beckman Instruments, Fullerton, CA) at 16,000 rpm (35,300 × g) at 4° C for 30 min. To prevent any mucus gel from getting into the pipette tip after separation of the centrifugate layers, only ~ 90% of the sol phase on the top of the RTFO was extracted for analysis. The unextracted RTFO (mucus gel with some liquid) was not used for any analysis. Five-microliter aliquots of the centrifugate supernatant were used to determine the percentage of mannitol solution in the RTFO samples; the remaining supernatants were sealed and stored at 70° C. Each of the RTFO samples collected after intubation (Samples A) was divided into two parts. One part was sealed and stored at 70° C (Samples AI), and the other part was centrifuged according to the same procedure as previously described, and the supernatant was separated from the remainder of the centrifugate (Samples AII).

Determination of percentage of mannitol solution in supernatant of RTFO. As compared with the absorption rate of mannitol (t_1/2 = 65 min) (8) and 10,000-D dextran (t_1/2 = 1,150 min) (9) in the lungs, and given the time scale of these experiments (30 min after each challenge), the absorption of mannitol and fluorescein-labeled dextran in the trachea were probably minimal. Dextran therefore represents a good tracer of mannitol and macromolecular secretions in the airways. The supernatants of Samples B1 to B5 and C1 to C5 (5 µl) were assayed for either fluorescein-labeled dextran or rhodamine B-labeled dextran, respectively. To determine the maximum fluorescence from each of the labeled dextran tracers, 5 µl of each of the ion-free mannitol solutions containing either fluorescein-labeled dextran (1 mg/ml) or rhodamine B-labeled dextran (1 mg/ml) was mixed with 5 µl of the supernatants of Samples AII. The background fluorescence was determined with 5 µl of the supernatants of Samples A2. Each sample was added to an individual well in a 96-well (8 × 12-well) plate (Nalge Nunc International, Rochester, NY) and was diluted to 60 µl with purified water (Milli-Qplus; Millipore, Bedford, MA). The 96-well plate was then shaken for 20 min. The fluorescence of each sample was measured with a fluorometer (Cytofluor II; Applied Biosystems, Foster City, CA). The fluorescence produced by fluorescein-labeled dextran was measured with excitation and emission filters at wavelengths of 488 ± 10 nm and 530 ± 13 nm, respectively. The fluorescence produced by rhodamine B-labeled dextran was measured with excitation and emission filters at wavelengths of 530 ± 13 nm and 620 ± 20 nm, respectively. The percentages of the mannitol solutions in each of Samples B1 to B5 and C1 to C5 were calculated from the fluorescence intensity of the supernatant of each of Samples B1 to B5 and C1 to C5 divided by the maximum fluorescence intensity of the respective mannitol solutions, with subtraction of background fluorescence.

Measurement of [Cl], [Na⁺], and osmolality. [Na⁺] and [Cl] in RFTO were measured with ion-sensitive microelectrodes at less than 3 mo after completion of the remaining studies. A Na⁺-selective glass electrode (MI-420; Microelectrodes, Bedford, NH) and a solid-state electrode for Cl (MI-200, Microelectrodes) were coupled with a double-junction reference electrode (MI-403; Microelectrodes) for this purpose. The reference electrode was composed of an internal glass reference barrel containing a wire coated with AgCl equilibrated with a KCl solution (3 M), and of an outer reference chamber filled with 0.9% saline. This arrangement minimized the diffusion of the KCl solution into the microliter sample, producing a stable reading. The ion-selective microelectrodes used in this part of the study measure unbound [Na⁺] and [Cl]. The electrodes were calibrated at room temperature (RT), using NaCl solutions of 10, 100, 200, and 400 mM, both before the measurements and immediately after the measurements. The results were fitted to a semilogarithmic plot to yield the slope and intercept of voltage versus concentration for each ion: 56.1 ± 0.2 mV/ln(mM) and 149.3 ± 0.4 mV, respectively, for the Na⁺ electrode, and 49.8 ± 1.2 mV/ln(mM) and 175.3 ± 2.2 mV, respectively, for the Cl electrode. To determine whether addition of 250 µl of ion-free solution would affect [Na⁺] and [Cl] in the airway surface liquid as a result of dissociation of Na⁺ or Cl from mucus gel (mucins), we carefully divided each of the untreated secretions (Samples A1) into five samples with a microliter syringe. These samples were then diluted to 100%, 80%, 60%, 40%, 20%, and 10% of their original volumes with purified water in a total volume of 100 µl. Measurements of [Na⁺] and [Cl] were made in the supernatants of Samples B1 to B5 and C1 to C5. This enabled homogeneous samples to be assayed, and facilitated the quantitative retrieval of the samples after measurements of [Na⁺] and [Cl]. It also minimized any artifact caused by adhesion of mucus gel to the microelectrodes. Because mucins are anionic, some Na ions in the RTFO could bind to the mucins and thus be inaccessible for measurement with the ion-selective electrode. As a result, the [Na⁺] of the RTFO in the study would be lower than if the [Na⁺] were measured with flame photometry.

The osmolality of Samples A2, B1, and C1 was measured with a vapor pressure osmometer (Model 5520; Wescor, Logan, UT) after the [Na⁺] and [Cl] measurements were made.

Na⁺ and Cl contents. By assuming the density of the RTFO as 1 g/ ml, we calculated the unbound Na⁺ and Cl contents in the RTFO by multiplying the respective unbound ion concentrations by the volume of the RTFO.

Enzyme-linked immunosorbent assay. A primary antibody, 10G5, developed to bind to a mucinlike antigen, was kindly provided by Dr. Carol Basbaum (University of California, San Francisco). The enzyme-linked immunosorbent assay (ELISA) of mucinlike antigen in RTFO was similar to that described by Steiger and associates (10). The supernatants of Samples AII, B1, and C1 were used. Each sample was dialyzed at RT on a shaker for 1 h, using 0.05 M sodium bicarbonate solution. Three groups of samples were prepared, at the ratios of 1:500, 1:1,000, and 1:2,000. A 96-well flat-bottom plate (Nunc) was scanned for background absorbance at a wavelength of 405 nm, using a microplate reader (Model 2001; BioWhittaker, Walkersville, MI) prior to the assay. Each dialyzed sample in each group was divided into two wells, with the sample in one well serving as a positive control and that in the other well serving as a negative control. Each well contained 50 µl of dialysate. The plate was placed in a vacuum container and dried overnight at RT. The dried, dialyzed samples were reconstituted at RT on a shaker for 1 h, using 100 µl of PNT solution (phosphate-buffered saline containing 1% N goat serum [blocking serum from Vectastain ABC kit (Vector Laboratories, Burlingame, CA)] and 0.3% Triton X-100). The PNT solution was used to block nonspecific mucinlike antigens. The solutions in each well were removed with suction tube, and the plate was rinsed with PNT solution. After the blocking procedure, the primary antibody (10G5; 50 µl) was aded to the positive control well of the sample, and PNT solution (50 µl) was added to the negative control well of the sample. The plate was incubated at RT on a shaker for 1 h. Following the incubation, each well was washed with PNT solution. The incubation and washing were done after each addition of secondary antibody (biotinylated antimouse IgG; Vectastain) and the avidin-biotinylated enzyme complex reagent (AK5002; Vectastain). After the labeling procedure, substrate (p-nitrophenylphosphate; Sigma, St. Louis, MO), buffered in the solution of 0.1 M NaHCO₃ and 20 mM MgCl at pH 9.5, was added (100 µl) to each well and incubated for 12 min. The plate was scanned at a wavelength of 405 nm with a microplate reader. The background absorbance in each well was subtracted from the measured absorbance of the sample in each respective well. The absorbance measured in each negative control sample was subtracted from the absorbance measured in the respective positive control sample to give the absorbance due to mucinlike antigen in each sample.

Statistics

Results are presented as mean ± SEM. The statistical significances (p) for RTFO, ion concentrations, and ion contents were calculated using two-way analysis of variance (ANOVA). For the data on osmolality and the ELISA, the statistical significances were calculated using one-way ANOVA. Values of p < 0.05 were considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The air inspired by the animals used in the study was conditioned to 24.5 ± 0.6 (mean ± SEM) mg/L at 33° C, which is equivalent to 69 ± 2% relative humidity. Under these experimental conditions, the humidity of the expired air was remarkably stable throughout the 65-min experiment, only decreasing from 40.6 ± 0.2 mg/L at the beginning of the experiment to 40.1 ± 0.1 mg/L at the end of the experiment. An example of the expired air humidity in one dog is shown in Figure 3. Small transient decreases in the humidity of the expired air were observed during the collection of RTFO. The mean arterial oxygenation (Pa_O2), Pa_CO2, pH, and rectal temperature measured after the 950-mOsm mannitol challenge were 125 ± 4 mm Hg, 34.2 ± 2.0 mm Hg, 7.360 ± 0.022, and 37.8 ± 0.1° C, respectively.

View larger version (22K): [in this window] [in a new window]   Figure 3.   An example of expired-air absolute humidity in an experiment in which a dog was ventilated with humid air. The x marks indicate the RTFO collections following each challenge, and the arrows indicate 250-mOsm and 950-mOsm ion-free mannitol challenges. A blood sample was withdrawn after the 950-mOsm mannitol challenge.

The impact of the 250-mOsm and 950-mOsm ion-free mannitol aerosols on osmolality of RTFO at 5 min after the respective challenges, compared with the osmolalities of blood and the airway secretions taken after intubation, can be seen in Figure 4. The osmolality of the RTFO collected immediately after intubation (Samples AII) was 282 ± 7 mOsm. The osmolality of the blood samples was 299 ± 3 mOsm. At 5 min after the 250-mOsm mannitol challenge, the osmolality of the RTFO samples was 313 ± 7 mOsm (Sample B1). The osmolality of the RTFO samples at 5 min after the 950-mOsm mannitol challenge was 350 ± 15 mOsm (Samples C1). This was significantly greater than that after the 250-mOsm mannitol challenge.

View larger version (13K): [in this window] [in a new window]   Figure 4.   Comparison of osmolality in the supernatant of blood samples, RTFO collected after intubation, and RTFO collected at 5 min after 250-mOsm and 950-mOsm ion-free mannitol challenges (250 µl). Data are mean ± SE for eight beagle dogs, with the exception that RTFO following intubation was collected from six beagle dogs. *Statistically significant difference at p < 0.05.

As shown in Figure 5, the weights of the sequential RTFO after the 950-mOsm mannitol challenge were significantly larger than those of the corresponding RTFO after the 250-mOsm mannitol challenge. After subtraction of the portion of mannitol solution in the RTFO, the airway secretion outputs (ASO) induced by the 950-mOsm mannitol challenge were still significantly larger than those after the 250-mOsm mannitol challenge. In 1-h control experiments described in a previous report by our group (5), the weight of the RTFO without any aerosol challenge was < 10 mg in each collection during each experiment, and averaged 5 to 6 mg per collection (0.9 mg/min). It is notable that the increases in RTFO and ASO caused by the mannitol aerosol were maximal in the first 5 min after the challenges, and rapidly decreased toward the baseline level over the subsequent 30 min.

View larger version (19K): [in this window] [in a new window]   Figure 5.   Temporal responses of RTFO and ASO following mannitol challenges. The time below each sample represents the end of each collection period. The 250-mOsm mannitol aerosol (250 µl) was administered at 5 min and the 950-mOsm mannitol aerosol (250 µl) was administered at 35 min after a 20-min stabilization period on mechanical ventilation. Data are mean ± SE for eight beagle dogs.

As shown in Figure 6, [Na⁺] and [Cl] in the RTFO at 5 min after the 250 mOsm mannitol challenge (Sample B1) were 82 ± 9 mM and 125 ± 10 mM, respectively. [Na⁺] and [Cl] in the RTFO after the 250-mOsm mannitol challenge remained relatively constant for the subsequent 30 min (Samples B1 to B5). At 5 min after the 950-mOsm mannitol challenge, larger decreases of [Na⁺] and [Cl] in the RTFO were observed (Sample C1 versus B1). Over the next 30 min, [Na⁺] and [Cl] in the RTFO collected at 5 min after the 950-mOsm challenge increased slowly from 32 ± 2 mM and 42 ± 5 mM, respectively, in Sample C1 to 53 ± 8 mM and 73 ± 10 mM, respectively, in Sample C5.

View larger version (28K): [in this window] [in a new window]   Figure 6.   Temporal responses of unbound [Na+] and unbound [Cl] in each RTFO after 250-mOsm (at 5 min) and 950-mOsm (at 35 min) mannitol challenges. The time below each sample represents the end of each collection period. Data are mean ± SE for eight beagle dogs.

[Na⁺] and [Cl] in the RTFO collected immediately after intubation (without separating the sol and gel phases; i.e., Samples AI) were 101 ± 5 mM and 161 ± 9 mM, respectively (Figure 7). [Na⁺] and [Cl] in these RTFO decreased in proportion to the volume of added purified water. Thus, [Na⁺] and [Cl] in the RTFO were not substantially influenced by any dissociation of Na⁺ and Cl from mucus gel, but instead resulted from changes in transepithelial ion transport and mucociliary transport.

View larger version (19K): [in this window] [in a new window]   Figure 7.   Illustration of changes in (a) unbound [Na+] and (b) unbound [Cl] in RTFO collected after the addition of purified water. The solid line predicts linear decreases of [Na+] and [Cl] in the samples in proportion to the quantity of purified water added. The samples were collected after intubation from six beagle dogs. Data are means ± SE.

The Na⁺ and Cl contents in the RTFO after the mannitol challenges were maximal in the first 5 min. Despite the significant difference in weight of the RTFO after the 250-mOsm and 950-mOsm mannitol challenges (Figure 5), the Na⁺ and Cl contents of the respective RTFO were similar (Figure 8).

View larger version (22K): [in this window] [in a new window]   Figure 8.   Temporal responses of unbound Na+ and unbound Cl contents in RTFO collected after 250-mOsm (at 5 min) and 950-mOsm (at 35 min) mannitol challenges. The time of each sample represents the end of each collection period. Data are mean ± SE for eight beagle dogs.

The rate of RTFO of the control group (RTFO is the same as ASO when no challenges are administered) was estimated to be 0.9 ± 0.1 mg/min from a previous study (5). The total Na⁺ and Cl contents in the control RTFO, calculated from RTFO in the control group multiplied by [Na⁺] and [Cl] in the RTFO collected after intubation (Samples AI), were calculated to be 3.2 µmol and 4.1 µmol, respectively. In addition, after the aerosolized mannitol challenges, the ASO and the total Na⁺ and Cl contents of the RTFO were significantly increased compared with the control (Figures 9a and 9c). The 950-mOsm mannitol challenge caused a 164% greater increase in ASO than did the 250-mOsm mannitol challenge (Figure 9a). Longitudinal mucociliary transport contributed only partly to this increase, since the total recovery of mannitol solution in the five RTFO after the 950-mOsm mannitol challenge was only 54% higher than its total recovery after the 250-mOsm mannitol challenge (Figure 9b). As shown in Figure 9c, there were no significant differences in the total Na⁺ and Cl contents of RTFO after the 250-mOsm and 950-mOsm mannitol challenges. Thus, the further increase in ASO by 164% with the 950-mOsm mannitol challenge was probably due to an increase in transepithelial water flux.

View larger version (28K): [in this window] [in a new window]   Figure 9.   Summary of (a) ASO in mg/min, (b) the percentage of mannitol recovery in the five RTFO samples collected after each challenge, and (c) the total unbound Na+ and unbound Cl contents in the five RTFO samples collected after each challenge. Data are mean ± SE for eight beagle dogs. *Statistically significant difference at p < 0.05. Control values were estimated from a previous report (5).

In the supernatant samples diluted to 1:1,000, the ELISA used for mucinlike antigen showed an absorbance of 1.88 ± 0.19 in RTFO at 5 min after the 250-mOsm mannitol challenge (Sample B1) and 1.93 ± 0.13 in RTFO at 5 min after the 950-mOsm mannitol challenge (Samples C1), as compared with 1.92 ± 0.35 in RTFO after intubation (Samples AII). There were no significant differences in mucinlike antigen concentration in the supernatants of the RTFO at 5 min after the 250-mOsm and 950-mOsm mannitol challenges.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have provided evidence in vivo that an imposed reduction of [Na⁺] and [Cl] in airway surface liquid, in the absence of an osmotic gradient, causes increases in ASO, as well as in its Na⁺ and Cl contents. A greater increase in ASO was induced by an imposed reduction of [Na⁺] and [Cl] in the airway lumen in the presence of an osmotic gradient toward the airway lumen, and was not associated with a further increase in Na⁺and Cl content. We interpret these data to indicate that the airways respond individually to ionic and osmotic stresses in the airway lumen, and that net transepithelial water transport in the airways is the sum of the water fluxes associated with active Na⁺ and Cl ion fluxes and the water flux caused by osmosis.

The airway surface liquid in the trachea immediately upon being challenged with 250 µl of 250-mOsm, ion-free mannitol in our study should be nearly isosmotic, with a reduction by a factor of perhaps two in [Na⁺] and [Cl]. Since the [Na⁺] and [Cl] in the RTFO collected at 5 min after this challenge were 82 ± 8 mM (Figure 6) and 125 ± 10-mM (Figure 6), respectively, as compared with 101 ± 5 mM (Figure 7a) and 161 ± 9 mM (Figure 7b), respectively, in the samples collected after intubation, an initial rapid return of [Na⁺] and the [Cl] toward homeostatic conditions would seem to have occurred, as indicated by Price and colleagues (2). It is likely that the rapid increase in potential difference (more negative) observed by Man and coworkers (1) after replacement of 50 mM NaCl in a buffer by 100 mOsM mannitol was related to an increase in basolateral-to-luminal Cl flux. Since both [Na⁺] and [Cl] in the airway surface liquid were reduced, the increase in net fluid transport after the 250-mOsm mannitol challenge in our study was probably due to this increased basolateral-to-luminal, Cl-associated water flux, as well as to a decrease in luminal-to-basolateral, Na⁺-associated water flux.

The increases in ASO after the ion-free mannitol challenges were due to the induced changes in transepithelial fluid secretion and mucociliary transport. After administration of 250 µl of an aerosol of 0.9% saline, we found that the rate of RTFO (ASO + additional aerosolized mannitol solution) increased from a baseline value of 0.9 mg/min to 2.3 mg/min (5), whereas administration of the same volume of 250-mOsm mannitol caused the rate of RTFO to increase to 4.8 ± 1.1 mg/ min (5). This latter value is comparable to the 4.8 ± 0.5 mg/ min that we found for the same challenge. Although this observed larger increase in RTFO could have been partly due to a further increase in mucociliary clearance, we propose that the major portion of the additional fluid in the ASO after the 250-mOsm mannitol challenge originated from an induced basolateral-to-luminal fluid flux.

Whereas the total percentage of mannitol solution recovered in the five RTFO samples collected after the 950-mOsm mannitol challenge was 54% greater than the total recovered after the 250-mOsm mannitol challenge (Figure 9b), the increase in ASO after the 950-mOsm mannitol challenge was 164% greater than the increase in ASO after the 250-mOsm mannitol challenge (Figure 9a). Thus, it is evident that osmotically driven water transport across the epithelium contributed to the further increase in ASO after the 950-mOsm mannitol challenge. Since the total Na⁺ and Cl contents in the RTFO collected after each of the 250-mOsm and 950-mOsm challenges were not significantly different for each challenge (Figures 8 and 9), we interpreted this as indicating that the transepithelial water transport processes involved were not associated with additional transepithelial Na⁺ or Cl transport. This suggests that this additional water flux was probably associated with increased transcellular water fluxes through aquaporins that preferentially transport water, rather than resulting from an increase in water and ion fluxes through tight junctions (11).

We now propose an internally consistent model of transepithelial water and ion fluxes that explains some otherwise inexplicable observations regarding the responses of the airway epithelium to ionic and osmotic stresses. This model is shown schematically in Figure 10. Disregarding any contribution to the airway surface liquid by diffusion, mucociliary transport, or hydrostatic pressure differences across the epithelium, we propose that

View larger version (101K): [in this window] [in a new window]   Figure 10.   Proposed model for regulation of ion-associated water fluxes and osmotically induced water fluxes caused by perturbations of [Na+], [Cl], and osmolarity of airway surface liquid. Only the dominant pathways involved in ion and water flux are shown. A premise of this model is that airway epithelial cells can sense changes in [Na+], [Cl], or osmolarity in the airway surface liquid. This model is illustrated as if all these events occurred within a single cell type. While this is possible, especially under different physiologic conditions, it is also likely that different cell types execute a preponderance of one or more of these functions according to whether their function is primarily absorptive or secretory.

where J_W^Net is the net water flux across the epithelium, J_W^Osm is the water flux due to osmosis, J_W^Ion is the water flux associated with ion fluxes, f is a function operator, is the osmotic gradient, J_W is the unidirectional water flux, J_Na+ is the intracellular Na⁺ flux, J_Cl is the intracellular Cl flux, B refers to the basolateral side of the epithelium, and L refers to the luminal side of the epithelium. In this model, the vectorial intracellular Na⁺ and Cl fluxes are associated with their respective vectorial transepithelial water fluxes, and any paracellular water fluxes associated with any paracellular transport of the counter ions are assumed to be negligible and are therefore omitted. If the lack of fluid flow observed across tight junctions in Madin-Darby canine kidney (MDCK) cells (12) is found to be true for airway epithelia, this assumption will be validated.

The application of this model results in an internally consistent explanation of the findings reported herein and those of other investigators, as well as those in our own previous studies. Yankaskas and coworkers showed that when [Na⁺] and [Cl] were increased on the surface of the canine tracheal lumen under short-circuit conditions, there were increases in Na⁺ and Cl fluxes from the lumen to the submucosa (13). The model predicts that this results in a luminal-to-basolateral, Na⁺-associated water flux that opposes the water flux caused by the imposed osmotic gradient. The presence of such an opposing fluid flux could explain why Price and colleagues did not observe an increased water flux toward the airway lumen when a 600-mOsm NaCl solution was added to the ferret tracheal lumen (2). In the case of the present study, in which the airway lumen was exposed to a hyperosmotic stress with reductions in [Na⁺] and [Cl], the decrease in [Cl] in the airway lumen would induce an increase in transcellular Cl flux into the airway lumen, together with a basolateral-to-luminal Cl-associated water flux. The water flux due to the osmotic gradient is in the same direction as the Cl-associated water flux. In addition, the reduction of [Na⁺] in the airway surface liquid would tend to reduce the luminal-to-basolateral Na⁺ flux together with its associated water flux. The predicted substantial increase in airway hydration is consistent with the observed marked increase in ASO. Extending this logic, a reduction of [Na⁺] and [Cl] in the airway lumen, together with a hyposmotic challenge, would result in an increase in Cl-associated water flux into the tracheal lumen, as well as a decrease in the luminal-to-basolateral water flux associated with Na⁺ flux. However, the water flux due to the osmotic gradient in this case is in the opposite direction. This could explain why Price and colleagues did not observe any transport of water from the airway lumen to the submucosa until the Cl flux into the tracheal lumen decreased (after approximately 5 to 10 min) (2). In the experiments reported herin with 250-mOsm, ion-free mannitol solution, there was an obligatory decrease in ion concentrations in the airway surface liquid, with little change in the osmotic gradient. As proposed, these conditions would increase water flux into the airway lumen resulting from the basolateral-to-luminal Cl flux, and decrease water flux associated with a reduction in luminal-to-basolateral Na⁺ flux, without any osmotically induced water flux. Such a challenge would therefore be expected to result in a smaller increase in ASO than would a similar depletion of ions in the airways in the presence of a hyperosmotic (mannitol) challenge, which is an accord with the results reported herein.

Winters and Yeates showed that mucociliary clearance after the administration of aerosolized hyperosmotic Na⁺-gluconate solution or hyperosmotic choline-Cl solution to the airway lumen was significantly faster than that following the administration of hyperosmotic NaCl solution (7). This is also consistent with our model. When hyperosmotic Na⁺-gluconate aerosol is administered, this model predicts that the net water flux will consist of an increased luminal-to-basolateral Na⁺-associated water flux, an increased basolateral-to-luminal Cl-associated water flux, and an increased water flux toward the airway lumen resulting from the imposed osmotic gradient. The net result will be a marked increase in mucociliary clearance. When hyperosmotic choline-Cl aerosol is administered, the model predicts a decrease in luminal-to-basolateral Na⁺-associated water flux, a decrease in basolateral-to-luminal Cl- associated water flux, and an increase in water flux toward the airway lumen due to the imposed osmotic gradient. The net result will again be a marked increase in mucociliary clearance.

Some underlying cellular mechanisms are consistent with the foregoing model. We propose that the coupled transcellular ion and water transport are regulated by active cellular processes. In support of this claim, Phillips and colleagues demonstrated that the activation energy for luminal-to-basolateral water transport was considerably greater than that for diffusion, and that this water transport was associated with luminal-to-basolateral Na⁺ transport by the Na⁺-K⁺ pump (3). Also, Price and colleagues showed that the change in osmolality in the airway lumen caused by the perturbation of [Na⁺] and [Cl] in the lumen was markedly reduced by cooling the tracheal membrane to 4° C (2). These data suggest that luminal-to-basolateral Na⁺ transport in the presence of a hyper- or hypoosmotic gradient is active and is associated with water transport. In vivo, inhibition of the Na⁺-K⁺ pump by administration of acetylstrophanthidin was reflected by marked increase in RTFO (5) and increased mucociliary clearance (14).

The homeostatic regulation of ion concentrations in the airway surface liquid requires that transcellular ion fluxes respond either to an electrochemical gradient or to changes in ion concentrations in the airway surface liquid, or both. There is evidence that the Cl ion has the ability to self-regulate Cl channels in MDCK cells (15), as well as those expressed in Xenopus oocytes (16). Also, a reduction in [Cl] in the airway surface liquid increases the electrochemical gradient for a Cl efflux across the apical membrane of sweat ducts (17). Assuming that [Cl] in the airway surface liquid is in dynamic equilibrium with intracellular [Cl], then a decrease in [Cl] in the airway surface liquid will decrease intracellular [Cl]. Such an induced decrease in intracellular [Cl] will cause an increase in the activity of the basolateral Na⁺-K⁺-2Cl cotransporter. It is possible, but not demonstrated, that this Cl flux has a stoichiometric relationship with an associated water flux, as claimed for the brush border Na⁺/glucose cotransporter (18). An increase in Cl flux into the airway is predicted to be associated with an increase in airway hydration and mucociliary transport.

The presence of a polarized epithelium is essential to the utility of this model. The data presented earlier, from in vivo experiments and in vitro experiments on native epithelia, are consistent with the proposed model. This model, however, is neither applicable to experiments with cultured cells in which osmotic and ionic challenges were applied directly to both the luminal and basolateral sides of the cells (19), nor to studies of cell volume regulation in which the culture media surrounded whole cells (20, 21). Since increases in cyclic adenosine monophosphate have been shown to change the Na⁺ permeability of tight junctions in the intestine (22), it is possible that paracellular ion and water fluxes assume increased importance in "activated" airways.

These vectorial ion and water fluxes may or may not be associated with the same cell type. The preponderance of cystic fibrosis transmembrane regular Cl channels in serous cells as well as in the linings of ducts of mucous glands (23) suggests that these channels are especially suited to secrete Cl and water to hydrate mucous granules as these granules are exocytosed from the corresponding cells (24). The goblet cell of the airway epithelium is also a secretory cell (25, 26). The presence of microvilli on the apical membrane of ciliated cells is typical of absorptive epithelia. It is also quite likely that individual airway epithelial cells have the inherent ability to either absorb or secrete ions and water, depending on the physiologic conditions and the receptors activated. It has been clearly demonstrated that in native ovine airway epithelia, the integrated responses of the cell types present can both promote absorption of water and can be induced to secrete water (4).

The consequences of this model are the obligatory stoichiometric relationships between the active-transport-driven luminal-to-basolateral Na⁺ flux and the concurrent water flux, as well as between the transcellular basolateral-to-luminal Cl flux and its associated water flux. Such a stoichiometry could be associated with the Na⁺-K⁺ pump and the Na⁺-K⁺-2Cl cotransporter, respectively. These could provide the mechanisms that render ion-associated water transport processes independent of osmotically driven transepithelial water fluxes in airway epithelium. The osmotically driven water fluxes are probably facilitated by the insertion of aquaporins into the plasma membrane (27, 28). Because ion channels partly desolvate the ions passing through these channels (29), aquaporins preferentially transport water molecules (11), and because the Na⁺-K⁺ pump and the Na⁺-K⁺-2Cl cotransporter potentially transport both ions and water, these processes would provide the underlying regulatory mechanisms for a tight regulation of water and ion transport across the tracheobronchial airways. The physiologic consequence of this would be the robust control of airway hydration, in accord with the observed maintenance of airway hydration under a wide variety of adverse physiologic conditions.