The pulmonary disease in patients with cystic fibrosis (CF) reflects genetic mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene that produce defective epithelial ion transport. The CF airway epithelial ion transport abnormalities lead to a well described pathophysiological cascade that adversely affects the innate defense mechanism of the lungs. Impaired anion secretion by the CFTR channels and absence of CFTR inhibition on the epithelial sodium channel (ENaC) are major contributors toward the pathogenesis of CF (Boucher et al., 1988; Quinton, 1989; Welsh, 1990; Rowe et al., 2005). The CF ion transport abnormalities lead to a decrease in hydration of the ASL (Matsui et al., 1998a). The loss of ASL volume/height degrades cilial function, and it causes mucus adhesion, which ultimately produces the chronic inflammation and infection that lead to a significant decline in pulmonary function (Tarran et al., 2005; Boucher, 2007; Livraghi and Randell, 2007).

Aerosol pharmacotherapies specifically targeting the ion transport defect in CF patients have been tested. These therapies include hypertonic saline (HS) (Donaldson et al., 2006; Elkins et al., 2006), amiloride (Pons et al., 2000), INS37217 (Deterding et al., 2005), and MOLI1901 (Grasemann et al., 2007). Clinical studies demonstrated that aerosolized HS (6–7% NaCl solution) in CF adults was safe, it enhanced MC, it decreased the number of pulmonary exacerbations, and it improved pulmonary function when administered two to four times a day (Donaldson et al., 2006; Elkins et al., 2006). In a recent study, aerosolized HS (7% NaCl) was found to be safe and well tolerated in CF infants, supporting a preventative treatment strategy (Subbarao et al., 2007). The mechanism of action of aerosolized HS therapy reflects the direct addition of sodium and chloride ions to the airway surface, generating an osmotic gradient that expands the ASL and increases MC (Hirsh, 2002; Donaldson et al., 2006; Boucher, 2007; Tarran et al., 2007). The multidose regimen (two to four times a day) in these studies was necessary to offset the rapid rate of salt and water absorption from the ASL in CF patients and produce the sustained increase in MC required for an overall improvement in pulmonary function.

Amiloride was designed as an orally active potassium-sparing diuretic, but when administered as an aerosol to CF patients, the transient block of ENaC resulted in an increase in ASL volume and a short-acting enhancement of MC in CF patients (Köhler et al., 1986; App et al., 1990; Hirsh et al., 2004). To potentially improve the durability of HS in CF patients, a combination therapy of amiloride and HS was tested. It is interesting to note that the expected synergistic effect (combination > either therapy alone) was not found (Donaldson et al., 2006). The proposed mechanism for the amiloride block of the HS-induced increase in MC reflected the action (inhibition) of amiloride on the aquaporin channels in the respiratory epithelium (Donaldson et al., 2006). This amiloride-sensitive aquaporin channel hypothesis was challenged by Levin et al. (2006) who reported amiloride had no effect on AQ-3, -4, and -5 in AQP-transfected Fisher rat thyroid cells, and in CF and normal bronchial epithelia cells. Regardless of the amiloride aquaporin controversy, previous reports are consistent with lack of amiloride selectivity for ENaC. For example, it has been reported that amiloride inhibits the sodium-proton exchanger (NHE) at higher concentrations (Kleyman and Cragoe, 1988) and to a lesser extent the sodium-calcium exchanger (Kleyman and Cragoe, 1988; Antolini et al., 1993).

To improve ENaC blocker CF therapy, a novel chemical entity should more efficiently block ENaC, be more selective, maintain an increase in MC better than amiloride, and when combined with HS, outperform itself or HS alone. To design and synthesize an ENaC blocker with the aforementioned attributes, we modified the pyrazinoylguanidine structure of amiloride (Hirsh et al., 2006). A focused library of novel ENaC blocker compounds were tested for potency (IC₅₀) and for drug recovery (reversibility of short-circuit current), identifying Parion compound 552-02 as a potential lead compound for CF lung disease (Hirsh et al., 2006).

In this study, we tested the hypothesis that a novel ENaC blocker, 552-02, is more efficacious and more selective on ENaC, promoting and maintaining an expansion in the ASL and increasing and maintaining elevated MC as an aerosolized therapy compared with amiloride. Further studies were performed to investigate whether compound 552-02 when used in combination with HS acted synergistically, expanding ASL volume to values greater than measured with 552-02 or HS alone.

Materials and Methods

Cell Culture

Canine bronchial epithelial tissue used for primary culture was provided by Marshall Bio Resources (North Rose, NY) from animals undergoing scheduled sacrifice approved by the veterinary staff to ensure the humane care and treatment of experimental animals. Human bronchial epithelial cells were provided by the Tissue Culture Core of the Cystic Fibrosis Center at University of North Carolina at Chapel Hill under the auspices of protocols approved by the Institutional Committee on the Protection of the Rights of Human Subjects. Human bronchial tissue was harvested from excess donor lung tissue at the time of lung transplantation from a portion of the main stem or lobar bronchus. Protocols for primary human or canine bronchial epithelial culture are similar to previously described methods (Hirsh et al., 2004). Canine or human bronchi were incubated in minimal essential medium containing 0.1% protease (type XIV; Sigma-Aldrich, St. Louis, MO) and 50 μg/ml DNase at 4°C for a minimum of 24 h. Fetal bovine serum (10%) was added to the medium, and the epithelial layer was scraped and rinsed to improve cell yield. Cells were then centrifuged for 5 min at 500g. Resuspended cells were seeded at a density of 0.25 to 0.4 × 10⁶/cm² on 0.4-μm porous collagen-coated (human placenta type VI; Sigma-Aldrich) Snapwell or Transwell (Corning Life Sciences, Acton, MA) membranes (1.13 and 4.7 cm²), and they were maintained at an air-liquid interface in hormonally defined medium supplemented with penicillin and streptomycin (Matsui et al., 1998b).

CBE and HBE Short-Circuit Current and Reversibility Measurements

Bronchial epithelial monolayers from 6- to 14-day-old cultures grown on permeable membrane supports were mounted in modified Ussing chambers (Physiologic Instruments, San Diego, CA). All experiments were performed in Krebs-Ringer bicarbonate (KRB), pH 7.4, containing 140 mM Na⁺, 120 mM Cl^–, 5.2 mM K⁺, 1.2 mM Ca²⁺, 1.2 mM Mg²⁺, 2.4 mM HPO ²⁺₄, 0.4 mM H₂PO₄^–, 25 mM HCO₃^–, and 5 mM glucose. The epithelium was bathed on both sides with warmed (37°C) KRB circulated by gas lift with 95% O₂, 5% CO₂, maintaining the pH at 7.4. The transepithelial voltage was clamped to 0 mV, except for 0.2-s pulses (+5 mV) every 20 s to calculate transepithelial resistance (R_t). The short-circuit current (I_sc) and R_t were digitized and recorded on a computer. Data were acquired and analyzed using Acquire and Analysis version 1.2 software (Physiological Instruments, San Diego, CA). The 50% inhibition of I_sc concentration (IC₅₀) was calculated from apical drug additions ranging from 10^–11 to 10^–4 M (approximately half-log increments), and it was analyzed using nonlinear regression (Prism version 3; GraphPad Software Inc., San Diego, CA) using the following equation: where X is the logarithm of concentration and Y is the response. Y starts at the bottom of the curve and increases with a sigmoid shape (identical to the “four-parameter logistic equation”). Stocks of ENaC blocker were dissolved in dimethyl sulfoxide (DMSO) at a concentration of ∼10 mM, and they were stored at –10°C or below until use.

After a full concentration-effect study, the percentage of recovery of I_sc from apical sodium channel blocker exposure was measured 3 min after completion of three mucosal bath replacements with KRB. The percentage of recovery was calculated as recovered current after the third wash divided by predrug I_sc × 100.

Drug Off-Rate Kinetic Measurements in Xenopus Oocytes

Oocytes were harvested from adult frogs with approved animal use protocols (Institutional Animal Care and Use Committee, University of North Carolina at Chapel Hill). The oocytes follicular cell layer was removed enzymatically with collagenase and hyaluronidase (4 mg/ml type 2; 1 h; 23°C) in modified Barth's solution (MBS) containing 88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO₃, 0.82 mM MgSO₄, 0.33 mM Ca(NO₃)₂, 0.41 mM CaCl₂, and 20 mM HEPES, pH 7.5, with NaOH. Then, oocytes were washed in 5 volume-equivalents of enzyme-free MBS solution containing 10 mM EGTA and 1% (w/v) bovine serum albumin to quench residual enzyme activity. In subsequent experiments, oocytes were shrunken in hypertonic MBS solution (supplemented with 230 mM sucrose) for 0.5 to 1 h, and mature oocytes (stage V–VI) with clear separation of their plasma and vitelline cell membranes were visually selected and stored overnight in isotonic MBS solution. The selected oocytes were injected the next day with cRNAs of rat αβγ-ENaC subunits (0.3 ng of each subunit in 50 nl/oocyte; Nanojet; Drummond Scientific, Broomall, PA) prepared from cDNA constructs (a gift from Dr. Bernard Rossier, University of Lausanne, Lausanne, Switzerland). Oocytes were used for patch-clamp analysis of drug off-rate kinetics 16 to 48 h after injection of ENaC cRNAs. Patch-clamp measurements of ENaC-mediated currents were made after the oocyte vitelline membrane was removed (Roboz #5 forceps; Roboz, Rockville, MD) in hypertonic MBS solution (15 min). The patch-clamp bath solution contained 110 mM Li-aspartate, 1 mM MgCl₂, 1.8 mM CaCl₂, and 10 mM HEPES, titrated to pH 7.35 with LiOH. The patch pipette was filled with solution containing 75 mM Tris-aspartate, 17 mM NaCl, 3 mM MgATP, 0.2 mM Na₂GTP, 0.1 mM CaCl₂, 1 mM EGTA, and 10 mM HEPES titrated to pH 7.35 with NaOH. Patch-clamp measurements of ENaC-mediated currents were recorded from outside-out macropatches using an EPC-7 patch-clamp amplifier (List Electronics, Darmstadt, Germany). Current was digitized at 1 kHz (16-bit, ITC; Instrutech, Long Island, NY) after low-pass filtering (0.1 kHz; –3-dB; Four-Pole, Bessel filter; Ithaca Filter Corporation, Ithaca, NY), and it was acquired with PC running HEKA-PULSE acquisition software (Bruxton Corp., Seattle, WA). Patch-pipettes (borosilicate; Warner Instruments, Hamden, CT) were fabricated from thin-walled glass using a three-stage pull routine (DMZ universal puller; Zeitz, Augsburg, Germany). Pipette resistance with the indicated patch solutions was 7.6 ± 0.2 mΩ (mean ± S.E.M.; n = 24). Currents were recorded at 0-mV membrane potential, with flowing bath conditions. Drug washout was achieved with a Fast-step solution exchanger (Warner Instruments). Exchange of bath solution at the patch was ∼70 ms (Caldwell et al., 2005). The membrane potential was not adjusted for an ∼2-mV diffusion potential between pipette and bath solutions as reported previously (Caldwell et al., 2004). A silver/silver chloride electrode connected to the bath via a 3% agar bridge containing 1 M KCl served as the ground electrode. Experiments were performed at ∼23°C.

Model Fitting. Following drug washout from the bath, the time-dependent membrane current (I_mem) in outside-out macropatches was fitted with a single exponential kinetic model for analysis of drug off-rates: where I_wash is the time-dependent drug-sensitive Na⁺ current, and I_ss is the steady-state current achieved after drug washout. The time course of drug washout is described with time constant τ_wash.I_mem was fitted beginning ∼100 ms after switching to drug-free patch superfusion solution. Fit parameters were first adjusted manually and then automatically for convergence by minimization of least squares using the Levenburg-Marquardt algorithm (Table Curve 2D).

The drug off-rate constant k_off inverse seconds (s^–1) was calculated as the inverse of the fitted time constant:

The fractional current recovery (frct rcvry) after drug washout was calculated from measured steady-state current before (I_sspredrug) and 10 to 120 s after drug exposure (I_sspostdrug) as shown in the following equation:

Sodium Channel Absorption and Biotransformation by Human Tracheobronchial Epithelial Cells

The rate of serosal appearance of amiloride and 552-02 (both 100 μM) by HTBE (MatTek, Ashland, MA) was measured after administering ∼200 μl/cm² drug to the apical surface. The rate of serosal appearance of drug was measured using cultures grown at an ALI for no less than 24 days on PTFE Millicells (0.4 μm; 30 mm; Millipore Corporation, Billerica, MA) using hormonally defined media (MatTek). The average R_t before the start of the experiment was 422 Ωcm². Samples (250 μl) were collected from the serosal compartment over 240 min, and they were applied to an Atlantis column (4.6 × 150 mm i.d.; dC18; 5 μm; Waters, Milford, MA) maintained at 40°C. Elution of sodium channel blockers was achieved with a 12-min mobile phase consisting of a 0.02% trifluoroacetic acid, pH 3.0:0.02% trifluoroacetic acid acetonitrile, linear gradient (95:5–56:64), at a flow rate of 1.5 ml/min. The column was monitored by an online fluorescence detector 474 (λ = 362 nm, excitation; 412 nm, emission) and a mass spectrometer Micromass ZQ (Waters). Concentrations are determined by measuring intrinsic fluorescence of each compound using high-performance liquid chromatography. Quantitative analysis uses a seven-point standard curve generated from authentic reference standard materials of known concentration and purity.

Sodium-Proton Exchange Assay

A Chinese hamster lung fibroblast cell line (PS120), lacking all endogenous Na⁺/H⁺ exchangers, was stably transfected using Lipofectin (Invitrogen, Carlsbad, CA) with human NHE1 or NHE2 or NHE3. Cells were grown in medium (Dulbecco's modified Eagle's medium) supplemented with 25 mM NaHCO₃, 10 mM HEPES, pH 7.4, 50 IU/ml penicillin, 50 mg/ml streptomycin, 10% fetal bovine serum, and 800 mg/ml G-418 (Invitrogen) in a 5% CO₂, 95% O₂ incubator at 37°C, and they were grown on glass coverslips until they reached 50 to 70% confluence. The cells were loaded with the acetoxymethyl ester 29,79-bis(2-carboxyethyl)-5(6)-carboxyfluorescein (5 mM) in Na1 medium (130 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 1 mM MgSO₄, 1 mM NaH₂PO₄, 25 mM glucose, and 20 mM HEPES, pH 7.4) for 20 min at 22°C, and then they were washed with TMA1 medium (130 mM tetramethylammonium chloride, 5 mM KCl, 2 mM CaCl₂, 1 mM MgSO₄, 1 mM NaH₂PO₄, 25 mM glucose, and 20 mM HEPES, pH 7.4) to remove the extracellular dye. The coverslip was mounted at an angle of 60° in a 100-μl fluorometer cuvette designed for perfusion and maintained at 37°C. The cells were pulsed with 40 mM NH₄Cl in TMA1 medium for 3 min, with or without 100 μM amiloride or 552-02 followed by TMA1 medium, which resulted in the acidification of the cells. Na1 medium was then added, which induced alkalinization of the cells. The Na⁺/H⁺ exchange rates (H⁺ efflux) were calculated, as the product of Na⁺-dependent change in pHi and the buffering capacity at each pHi, and they were analyzed with the use of a nonlinear regression data analysis program (Origin version 6.0; OriginLab Corp., Northampton, MA) and fitted using a general allosteric model described by the Hill equation: where [H⁺] is proton concentration calculated from pH =–log[H⁺], V is velocity, k is affinity constant, n is apparent Hill coefficient, and V_max is the maximum velocity.

Animal Preparation for MC Studies

All procedures were approved by the Mount Sinai Animal Research Committee to ensure the humane care and treatment of experimental animals. The methods were similar to those described previously (Hirsh et al., 2004). In brief, adult ewes (∼40 kg) were restrained in an upright position. The animals' heads were immobilized and the nasal passages anesthetized with 2% lidocaine. The animals were nasally intubated with a 7.5-mm i.d. ETT. The cuff of the ETT was placed just below the vocal cords, and its position was verified by bronchoscopic visualization. After intubation, the animals were allowed to equilibrate for approximately 20 min before initiating MC measurements.

MC Measurement

Test articles, either 4 ml of H₂O (vehicle), 4 ml of 3 mM amiloride, or 4 ml of 3 mM 552-02, were aerosolized using the Pari LC JetPlus nebulizer to free-breathing sheep. The nebulizer had a flow rate of 8 l/min, and the time to deliver the nebulized test article solution was 10 to 15 min. Depending on the interval of pharmacologic action being tested (immediate or long-acting), the radiolabel ^99mTechnicium-sulfur colloid (^99mTc-SC; 3.1 mg/ml; containing approximately 20 mCi) was administered immediately after test article (T = 0 format) or 4 h after test article (T = 4 format). The radiolabel was aerosolized using a Raindrop Nebulizer that produced a droplet with a mass median aerodynamic diameter of 3.6 μm. The nebulizer was connected to a dosimetry system consisting of a solenoid valve and a source of compressed air (20 pounds per square inch). The output of the nebulizer was directed into a plastic T connector, one end of which was connected to the ETT, and the other end was connected to a respirator. The system was activated for 1 s at the onset of the respirator's inspiratory cycle. The respirator was set at a tidal volume of 300 ml, with an inspiratory-to-expiratory ratio of 1:1, and a rate of 20 breaths/min. Radiolabeled aerosol was administered for ∼5 min. The sheep were then extubated, and data capture by gamma camera was initiated. This first time point established the baseline deposition image, and it was assigned the value of 0% clearance. Total radioactive counts in the lung were collected every 5 min during the 1-h observation periods and every 15 min thereafter. A washout period of at least 7 days (half-life of ^99mTechnicium = 6 h) separated studies for each animal.

Canine Airway Surface Liquid Drug Durability Assay

Using a gravimetric procedure, mucosal surface liquid mass was measured on canine bronchial cells that had been grown at ALI culture on 12-mm PTFE Transwell membranes (Corning Incorporated Life Sciences, Lowell, MA) for a minimum of 6 days. At the start of the experiment, each Transwell insert was removed from the Transwell plate (lower compartment), blotted dry, and weighed. Then, 50 μl of vehicle (0.1% DMSO), 10 μM ENaC blocker, 1.5% HS, or 1.5% HS with 10 μM 552-02 (before and after HS), was applied to the mucosal surface, and the change in mass was recorded. After weighing, the cells were immediately returned to a Transwell plate (lower chamber; 500 μl; KRB, pH 7.4) and placed in a 37°C, 5% CO₂ incubator. Experiments were performed in KRB at 37°C. To reduce artifact due to an apical carbohydrate osmotic gradient formed during volume loss, glucose was not included in the apical KRB. The mass of mucosal surface liquid was monitored serially over 8 h. The mass of surface liquid was converted to volume in microliters. Data are reported as percentage of initial volume (100% = 50 μl).

Effect of 552-02 on Human CF Bronchial Epithelial ASL Height Measurements under Static and Sheer Stress Conditions

Primary human CF bronchiolar epithelia were plated on 12-mm T-clear inserts coated with human placental collagen, and they were grown at an ALI. For all experiments, cells were allowed to differentiate for a minimum of 3 weeks, and they were highly ciliated. Cultures that had an R_t < 200 Ωcm² were not included in this study. To study the effect of sheer stress, a mechanical device imposing in vivo-like forces directly to the respiratory epithelium (phasic motion; 0.5 dynes/cm²) was applied 30 min before measuring ASL height and for 4 h after the initial measurement. To apply sheer stress to the cultures, the culture plates housing the cells were phasically rotating in the incubator to deliver a phasic shear stress of 0.5 dynes/cm² (Tarran et al., 2005). The ASL was labeled by adding 10 μl of 2 mg/ml fluorescein isothiocyanate-dextran in Hanks' balanced salt solution for 1 min, after which the culture was placed in an incubator at 37°C to re-equilibrate for 1 h. To prevent ASL dehydration, 100 μl of perfluorocarbon (Fluorinert FC-77; 3M) was added to the apical surface.

For fluorescent microscopy, cultures were removed from the incubator, and they were placed on a glass coverslip with a basolateral volume of 200 μl of Hanks' balanced salt solution containing 10 mM glucose and 10 mM HEPES, pH 7.4. The ASL height was measured by an X-Z scanning confocal microscope (Leica Microsystems, Inc., Deerfield, IL). Based on previous observations, basal ASL volumes were estimated to be between 1 and 10 μl, for all experiments, we assumed the resting ASL volume to be 5 μl. Both compound 552-02 and/or 0.2 mg of NaCl were added to the apical membrane as dry powder in perfluorocarbon to yield approximate concentrations of 10 μM for 552-02 and 4% hypertonic saline.

To minimize variability, the ASL height measurements were made in five separate locations. The measurements were averaged, and the standard deviation was calculated. The average ASL height for each image was determined using ImageJ software version 1.36b (W. S. Rasband, National Institutes of Health, Bethesda, MD).

Materials

Racemic compound 552-02 methanesulfonate was synthesized by Albany Molecular Research Inc. (Albany, NY) (Hirsh et al., 2006). Cell culture media, bovine serum albumin, fetal bovine serum, bovine pituitary extract, epidermal growth factor, penicillin, retinoic acid, DNase, human placenta collagen VI, streptomycin, amiloride (Midamor), and lidocaine were purchased from Sigma-Aldrich. The ^99mTc-SC was purchased from Mallinckrodt (St. Louis, MO). Salts and solvents were of analytical or high-performance liquid chromatography grade, and they were purchased from VWR (West Chester, PA).

Statistical Analysis

All values are depicted as the mean ± S.D. unless otherwise specified. Data from in vitro assays (potency, reversibility comparisons of k_off, and frct rcvry) were tested for significance (*, P < 0.05) using a paired t test, GraphPad Prism version 4.0, and SigmaStat version 2.03 (Systat Software, Inc., Point Richmond, CA). Amiloride off-rate analysis was performed before or after exposure to 552-02. A one-way analysis of variance followed by a Dunnett's post-hoc analysis was used to test for significance for Na⁺/H⁺ exchanger activity. A two-way analysis of variance with repeated measures followed by a Bonferroni post-hoc analysis was used to test for significance for surface liquid volume ASL height with more than two treatments groups. Ovine MC versus time data (0–1 and 4–6 h) was fitted using a linear regression model, and the slopes of each line were compared using an analysis of covariance, for the immediate MC activity. For the durable MC format (T = 4), due to unequal time collection from 5 to 6 h in some of the animals, significant differences between amiloride and 552-02 were determined at specific time points using a paired t tests. Nonlinear, linear regression, and statistical analysis were performed using the program GraphPad Prism version 4.0.

Results

ENaC Blocker Potency, Maximal Efficacy, and Reversibility. Potency, maximal efficacy, and reversibility of the novel sodium channel blocker 552-02 on ENaC in CBE and HBE were measured and compared with the prototypic ENaC blocker amiloride (Fig. 1; Table 1). The differences in the aforementioned properties of the two ENaC blockers were determined by recording the change in I_sc in response to increasing concentrations of drugs (10^–11–10^–4 M) in the apical bath and after apical wash. Active sodium transport was the dominant component of I_sc in CBE and HBE (Fig. 2, A and B, respectively), as evidenced by the observation that the maximal effective concentrations of selected blockers decreased the I_sc (>96%) from the baseline value. The shape of the concentration-effect curves for amiloride and 552-02 were sigmoidal, and they were similar to a drug-receptor binding complex curve exhibiting saturable binding. The calculated IC₅₀ and percentage of I_sc recovery (reversibility of drug to binding site complex) from maximal block after a full concentration-effect response of 552-02 and amiloride are shown in Table 1.

Compound 552-02 was two orders of magnitude more potent (100-fold) than amiloride in blocking epithelial sodium channel-generated I_sc in CBE and approximately 60-fold more potent in normal HBE. The significant difference in IC₅₀ values between the two ENaC blockers suggests that 552-02 is accessing additional auxiliary binding sites in the channel pore, which would provide for a more stable drug-channel interaction (Hirsh et al., 2006).

To calculate the reversibility of I_sc after full-block, the apical compartment was washed three times, and the I_sc was recorded for 15 min thereafter. Compound 552-02 was less reversible compared with amiloride for CBE (19 versus 90% recovery, respectively), and 38 versus 90% of basal I_sc recovered in HBE (Table 1). If a more vigorous washout were applied (more volume exchanges or longer duration), the I_sc from 552-02 would fully recover to values that are similar to what was measured after washout from amiloride-induced block (Table 1). Using this protocol, no significant decay in basal current was noted over the duration of exposure or recovery intervals for control (untreated) tissues.

Off-Rate Kinetics and I_Na Recovery. To assess whether the slow recovery of I_sc after 552-02 washout could be explained by slow dissociation of the drug from ENaC, patch-clamp recordings from Xenopus oocytes injected with cRNAs of rat αβγ-ENaC subunits were used to measure kinetics of drug dissociation from the channel. Figure 3A depicts a current record showing baseline current and drug-induced inhibition of I_Na. Bath superfusion of 1 μM 552-02 reversibly inhibited I_Na. Rapid switching to drug-free patch superfusion solution resulted in a slow (τ_wash = 26.8 s), near-complete recovery of the 552-02-inhibited current as shown in Fig. 3A. In subsequent experiments, a saturating amiloride concentration (10 μM) was applied, and it inhibited I_Na to the same extent as 552-02 (Fig. 3A). Amiloride washout (τ_wash = 236 ms; Fig. 3A, inset) resulted in a rapid and complete recovery of I_Na. In Fig. 3B, summary data for multiple experiments revealed an approximately 170-fold slower off rate for 552-02 than amiloride. The k_off value for 552-02 was 0.039 ± 0.007 s^–1, whereas the k_off value for amiloride was 6.57 ± 0.98 s^–1, reflecting a more stable 552-02 channel complex. Summary data for the fractional I_Na recovery after drug washout is shown in Fig. 3C. Like amiloride, 552-02 is a reversible blocker.

Drug Absorption by HTBE. The retention of ENaC blockers in the ASL is a property that contributes to the durability of a pharmacological effect. The rate of serosal appearance of amiloride or 552-02 using HTBE cultures was measured after administering 100 μM of the test article to the apical surface. Compound 552-02 was absorbed by the HTBE at a significantly (P < 0.05) slower rate compared with amiloride (Fig. 4). This finding was also confirmed using a starting concentration of 10 μM of the test article on CBE (data not shown). Furthermore, no significant (<5%) biotransformation of either amiloride or 552-02 by HTBE cultures was detected on the mucosal or serosal side with liquid chromatography-mass spectrometry. The slower rate of 552-02 absorption was consistent over the duration of the study.

Na⁺/H⁺Exchange Activity. To test the effect of amiloride or 552-02 on the NHE, human NHE 1, NHE2, and NHE3 activity was monitored over a wide pH range. Amiloride (100 μM) significantly blocked NHE1 and NHE2 activity as represented by an ∼90% decrease in the V_max value compared with control, whereas 552-02 (100 μM) only decreased NHE1 V_max value by 27%, and it had no affect on NHE2 activity (Fig. 5, A and B). 552-02 (100 μM) produced a decrease (27%) in NHE3 V_max compared with control, whereas amiloride had no effect.

Ovine MC. MC was expressed as the percentage of inhaled radiolabel ⁹⁹Tc-SC cleared from a central region of the right lung over time. To test the immediate action of ENaC blockers on MC, the radiolabel was administered immediately after vehicle or drug aerosolization, and the clearance of radiolabel was monitored for 60 min thereafter (T = 0 format; Fig. 6A). The clearance versus time data for test articles (vehicle, amiloride, and 552-02) were fitted using a linear regression analysis, and the slope of the fitted line for each test article was calculated. Aerosol dosing of vehicle (H₂O_sterile; 4 ml), or equimolar concentrations (3 mM; 4 ml) of amiloride, or 552-02 all were associated with positive slope values: 12.8, 26.0, and 43.3, respectively (Fig. 6A). The difference between the three slopes was analyzed using analysis of covariance, and it was found to be significantly different (P < 0.0001). Amiloride approximately doubled the slope (26.0) compared with vehicle control (P < 0.05), similarly to previously reported data (converted to retention) (Hirsh et al., 2004). Compound 552-02 showed the greatest MC efficacy (a slope of 43.3, reflecting a >40% particle clearance), and it was significantly (P < 0.05) better at enhancing MC than amiloride (Fig. 6A).

To test the durability of drug activity on enhancing MC, test articles were first administered by aerosol and 4 h later the radiolabel ⁹⁹Tc-SC was administered (T = 4 format). MC was monitored for 2 h thereafter, yielding a total duration of action test of 6 h. MC measured 4 h after vehicle administration was very slow (Fig. 6B), with a slope equal to 1.8, that was significantly different from t = 0 vehicle slope. This observation suggested that the vehicle was active on MC immediately after dosing. Furthermore, the MC after amiloride delivery was not significantly different from vehicle (Fig. 6B). However, compound 552-02 produced a persistent acceleration of MC when measured 4 h after delivery (slope = 12.3), which was significantly different compared with either vehicle or amiloride (P < 0.05).

Durability of 552-02 Activity with or without Hypertonic Saline on Surface Liquid Retention in CBE. To test the in vitro durability of ENaC blockers on ASL volume, CBE cells were used, and the change in surface liquid volume over 8 h was measured (gravimetrically). A 50-μl volume was administered to the surface that consisted of KRB (control with 0.1% DMSO) with or without 10 μM 552-02 or 10 μM amiloride. The CBE absorbed >80% of the delivered amount of buffer over 8 h (Fig. 7A). Amiloride (10 μM) did not significantly slow the rate of absorption compared with vehicle (Fig. 7A). In contrast, 10 μM 552-02 significantly inhibited the rate of absorption, with approximately 75% of the initial volume retained on the CBE surface at 8 h (Fig. 7A).

Using the same protocol, administration of 1.5% HS produced an immediate “secretion” (20% increase in surface liquid volume), which persisted for 2 h, followed by rapid volume loss (rate of fluid loss –4.2 ± 0.7 μl/cm²/h) (Fig. 7B). In contrast, administration of equivolume of buffer (vehicle) resulted in a monotonic absorption of fluid (–3.1 μl/cm²/h), and 10 μM 552-02 significantly decreased absorption of isotonic buffer (–1.3 μl/cm²/h). It is noteworthy that combining 10 μM 552-02 and 1.5% HS, produced an immediate and sustained increase in surface liquid volume. The net fluid movement for the combination over the entire experimental interval was secretory (+0.7 ± 1.0 μl/cm²/h), which was significantly different compared with 1.5% HS (–4.2 μl/cm²/h) and 552-02 alone (–1.3 μl/cm²/h) (Fig. 7B).

To test whether 552-02 nonselectively blocks aquaporin channels as reported for amiloride (Donaldson et al., 2006), 10 μM 552-02 was added before or after administration of 1.5% HS. Both 10 μM 552-02 before 1.5% HS or after 1.5% HS produced an ∼35% increase in surface liquid volume at 8 h compared with the initial starting volume 50 μl (Fig. 7C). This increase was significantly different from control (<15%) (P < 0.0001), it was 70% greater than 10 μM 552-02 alone (Fig. 7B), and it was 80% greater than HS alone (Fig. 7B). These data suggest that, unlike amiloride, 552-02 does not inhibit water flow.

Effect of Sheer Stress with and without 552-02 on ASL Height in CF HBE. To test whether 552-02 only blocks liquid absorption, or whether it can also induce liquid secretion, we measured the effects of compound 552-02 on ASL height (volume) under thin-film conditions (Tarran and Boucher, 2002). Cultures used for these experiments were maintained under standard static conditions or under phasic-motion conditions that possibly more closely mimic the in vivo condition (Tarran et al., 2005). CF bronchial epithelial cell ASL height was measured using confocal microscopy in the X-Z plane. Images of the ASL height (green) for cultures maintained under static or under phasic-motion (sheer stress) conditions, with or without 10 μM 552-02, are shown in Fig. 8A. Under static conditions, ∼10 μM 552-02 maintained ASL height at 5 μm, but it did not increase ASL height compared with control after 4 h (Fig. 8B). In CF cultures maintained under phasic motion conditions without drug, ASL height was significantly greater (P < 0.05) compared with cells under static conditions. It is noteworthy that the ASL height in CF cultures maintained under phasic-motion treated with 552-02 was significantly increased compared with control cultures under phasic motion, consistent with ASL secretion induced by 552-02.

Effect of Hypertonic Saline with or without 552-02 (before and after Addition) on ASL Height in CF HBE. To test whether hypertonic saline and compound 552-02 act synergistically on CF bronchial cell ASL height (volume), and to determine whether compound 552-02 selectively blocks ENaC and not aquaporin channels in CF airway epithelia, ∼10 μM 552-02 was added to the apical surface before and after apical administration of ∼ 0.2 mg of NaCl as a mimic of HS (Fig. 9). The change in ASL height was measured in the X-Z plane using confocal microscopy for 240 min. Representative confocal images of ASL height for each treatment are shown in Fig. 9A. Figure 9B displays summary data for these experiments. ASL height remained constant without salt addition over 4 h. Adding 0.2 mg of NaCl to the surface of the epithelia caused an immediate 8-fold increase in ASL height compared with control. However, the significant ASL expansion decreased steadily to a value not significantly different from control by 4 h. To determine whether 552-02 enhanced the osmotic effect of HS, we tested the effects of 0.2 mg of NaCl in combination with ∼10 μM 552-02 administered before or after NaCl. In both cases, an immediate increase in ASL height was produced using the combination therapy, with a slower decline in the loss of ASL height over the time course (Fig. 9B). The greatest effect was observed with 552-02 delivered first, followed by NaCl (11-fold increase in ASL height after 10 min compared with control.) The overall time course experiments indicated that the predose 552-02/NaCl combination had the most significant effect, not only in causing the largest increase but also in maintaining the expansion of ASL. The observation that 552-02 pretreatment was most effective in adding to the osmotic action of HS again demonstrates that 552-02 did not block osmotically driven water flow like amiloride.

Discussion

Emerging evidence suggests that CF lung pathophysiology is linked to dehydration of airway surfaces caused by epithelial ion transport defects (Baconnais et al., 2005; Boucher, 2007; Donaldson et al., 2007). One pharmacologic approach to rehydrate airway surfaces is aerosol ENaC blocker therapy. The ENaC blocker standard for CF pharmacotherapy has been amiloride, an orally active potassium-sparing diuretic. Although exerting short-term improvements in MC or lung function (Köhler et al., 1986; App et al., 1990; Knowles et al., 1990), it has not consistently improved overall pulmonary function when used with current complex medical regimens (Graham et al., 1993; Pons et al., 2000; Burrows et al., 2006). A review of ENaC blocker CF pharmacotherapy suggested a lack of respiratory benefits with amiloride (Burrows et al., 2006). It is noteworthy that amiloride preadministered before HS aerosol therapy was not effective (Donaldson et al., 2006).

Compound 552-02 was specifically designed for aerosol delivery to the pulmonary system as a more selective, potent, long-acting ENaC blocker to promote an expansion in ASL volume, with or without hypertonic saline. Compound 552-02 (Fig. 1B) was selected from a series of novel ENaC blockers to produce a selective block on airway epithelial sodium channels, with a potency up to 2 orders of magnitude greater than amiloride (Hirsh et al., 2006). To overcome the in vivo shortcomings in CF using amiloride (Hofmann et al., 1997; Hirsh et al., 2004), a selection criterion for ENaC blockers was formed: greater intrinsic activity at ENaC (lower IC₅₀), less reversibility (much slower k_off), a decrease in drug absorption, maintained expansion of ASL, and compatibility with HS therapy.

In this study, we demonstrated that 552-02 blocked the majority (>95%) of I_sc, with a calculated IC₅₀ value of ∼7 nM (Table 1). The increase in potency of 552-02 compared with amiloride provides greater efficacy, a critical consideration for drugs delivered via aerosol. The higher IC₅₀ value reported in this study for amiloride is within the calculated S.D. for amiloride for human airway epithelium compared with past studies (Hirsh et al., 2004). The greater potency of 552-02 is probably a function of more efficient binding of the molecule associated in the pre-M2 region (Schild et al., 1997; Kellenberger et al., 2003; Kashlan et al., 2005) to possibly three auxiliary binding sites in the vicinity of the proposed site for amiloride (Hirsh et al., 2006). The lower percentage of recovery after washing (Table 1) and more specifically, the 170-fold slower k_off for 552-02 measured in patch-clamp studies (Fig. 3), are congruent with a tighter 552-02/channel complex compared with amiloride.

An important ENaC blocker criterion is slower transepithelial drug absorption. In this study, using human airway epithelia, penetration of 552-02 into the serosal compartment was significantly slower than amiloride (Fig. 4), consistent with a slower transepithelial absorption.

To test whether 552-02 is more selective for ENaC in contrast to amiloride, Na⁺/H⁺ exchanger activity was measured in the presence and absence of 552-02 or amiloride. Although 552-02 contains the amiloride structure, 552-02 diminished the effect on NHE1 V_max by ∼27%, and it was inactive on NHE2 activity compared with 90% by amiloride. In addition, 552-02 decreased the V_max value by 27% on NHE3, whereas amiloride had no effect. The reduced pharmacological activity of 552-02 on the NHEs and the predicted maximum human plasma concentration of 552-02 from aerosol delivery of ∼3.5 nM (Doran et al., 2006) predict a minimal block of 552-02 on NHE activity.

To investigate whether more potent, less reversible, and less permeable ENaC blockers are more effective in accelerating MC than amiloride, a whole animal sheep model was used to compare the effects of equimolar concentrations (3 mM) of amiloride or 552-02. An increase in MC in this assay reflects an increase in ASL volume (Mentz et al., 1986), and it was used to test the assumption that ENaC blockers increase ASL volume under physiologic, phasic respiratory motion conditions. Compound 552-02 significantly outperformed amiloride when assayed for acute effects (Fig. 6A), with clearance rates closely resembling values obtained in human patients with genetically complete absence of ENaC function (pseudohypoaldosteronism) (Kerem et al., 1999).

To test whether 552-02 was more durable than amiloride in vivo, MC was measured 4 to 6 h after drug delivery. No differences were measured in MC over this interval between control and amiloride; however, an acceleration of MC was maintained in animals exposed to 552-02, with clearance rates significantly greater than amiloride over the 4 to 6-h interval (Fig. 6B). These in vivo data strongly support the hypothesis that by blocking epithelial sodium channels with potent, less reversible, less permeant compounds, a sustained increase in MC can be obtained. Studies in CF patients will be required to assess what duration of effect on MC is required for clinical benefit.

To test whether beneficial effects are achievable when combining HS and 552-02, ASL volume retention and ASL height measurements were performed in CBE (Fig. 7), and CF bronchial airway epithelia (Fig. 9). After adding Krebs' buffer to CBE, the buffer was rapidly absorbed by the epithelium (Fig. 7A). Comparing effects of equimolar concentrations of ENaC blockers on the retention of surface liquid by CBE, 552-02 was far superior at maintaining surface volume compared with amiloride. Compared with control, equivolume HS (1.5%) additions to CBE surfaces produced an initial ASL volume expansion (osmotically driven secretion) during the first 2 h (Fig. 7B). However, once the surface liquid became isotonic, fluid absorption dominated (–4.2 μl/cm²/h) (Fig. 7B). In addition, when HS was combined with 552-02, before or after HS treatment, virtually all the buffer and osmotically driven secreted liquid was retained on CBE surfaces for the duration of the assay (8 h). These data support the hypothesis that a greater more durable MC effect might be measured using a combination therapy consisting of 552-02 and HS.

To investigate whether ENaC blockers alone, and particularly 552-02, could not only conserve ASL volume on airway surfaces but also induce ASL volume secretion measurement of ASL volume under thin-film conditions were performed on CF bronchial epithelia. Previous studies using confocal imaging indicated that normal ASL volume homeostasis in cell cultures under static conditions is maintained in part by the accumulation of adenosine in the ASL (Lazarowski et al., 2004). CF airway epithelia cannot maintain ASL height consistent with normal epithelial cultures under static conditions, due to the absence of CFTR. In contrast, phasic motion-induced shear stress produced sufficient ATP onto the surface of the epithelia to raise the ASL ATP concentration that activates the purinergic receptor, activating a calcium-activated chloride channel and partially inhibiting ENaC, to produce volume secretion. Consistent with these reports, CF cultures under static conditions could not maintain adequate ASL height (5 μm) for MC. Furthermore, the addition of 552-02 to the ASL of CF cultures under static conditions did not increase ASL height (Fig. 8). However, ASL height in CF cultures under phasic motion conditions was significantly raised, consistent with the actions of elevated ATP concentrations in the phasic motion cultures. As predicted, the addition of 552-02 to the CF culture under phasic motion conditions produced a larger anion induced ASL volume secretion than observed with no 552-02 treatment (Fig. 8). These data suggest that ENaC block under physiologic conditions not only conserves ASL volume but also induces ASL volume secretion. Similar interactions are predicted for normal epithelia under static conditions, where ENaC block should increase both adenosine- and ATP-stimulated chloride and volume secretion.

To improve the clinical efficacy of amiloride, a pilot study testing the combination therapy (aerosolized amiloride followed by HS) in CF patients measuring pulmonary function was performed. Unexpectedly, amiloride with HS was less effective than HS alone (Donaldson et al., 2006). The lack of synergism between HS and amiloride was explained by an apparent nonselective action of amiloride on epithelial aquaporin water channels (Donaldson et al., 2006). Levin et al. (2006) challenged the data, reporting that amiloride had no affect on AQ-3, -4, and -5 in AQP-transfected cells, and in CF and normal bronchial epithelial cells. Regardless of the amiloride aquaporin controversy, we investigated whether ENaC blocker 552-02 exhibited any activity on ASL volume expansion when delivered before or after HS administration. Both in CBE and in CF bronchial epithelia, irrespective of the sequence, 552-02 administration with HS was significantly superior with respect to ASL volume expansion than HS treatment alone (Figs. 7, B and C, and 9). Thus, the combined data from CBE and CF bronchial epithelial argue strongly that 552-02 does not promote a block of transepithelial osmotically driven water flux, and indeed the sequence of 552-02 followed by HS may be the preferred arm for clinical trials.

In summary, compound 552-02 is a novel potent, more selective, less epithelial-permeant, more durable epithelial sodium channel blocker than amiloride. When administered as an aerosol, 552-02 produced a sustained increase in MC in vivo for periods greater than 5 h. Furthermore, when used as a combination therapy with HS in vitro, compound 552-02 produced a sustained increase in ASL volume that was greater than 552-02 or HS alone. The use of 552-02 aerosol therapy alone, or in combination with HS, could be clinically beneficial by hydrating airway surfaces, and hence restoring the efficacy of the primary innate defense mechanism to the CF lung.