Transforming Growth Factor-β1 Decreases Expression of the Epithelial Sodium Channel αENaC and Alveolar Epithelial Vectorial Sodium and Fluid Transport via an ERK1/2-dependent Mechanism*

Acute lung injury (ALI) is characterized by the flooding of the alveolar airspaces with protein-rich edema fluid and diffuse alveolar damage. We have previously reported that transforming growth factor-β1 (TGF-β1) is a critical mediator of ALI after intratracheal administration of bleomycin or Escherichia coli endotoxin, at least in part due to effects on lung endothelial and alveolar epithelial permeability. In the present study, we hypothesized that TGF-β1 would also decrease vectorial ion and water transport across the distal lung epithelium. Therefore, we studied the effect of active TGF-β1 on 22Na+ uptake across monolayers of primary rat and human alveolar type II (ATII) cells. TGF-β1 significantly reduced the amiloride-sensitive fraction of 22Na+ uptake and fluid transport across monolayers of both rat and human ATII cells. TGF-β1 also significantly decreased αENaC mRNA and protein expression and inhibited expression of a luciferase reporter downstream of the αENaC promoter in lung epithelial cells. The inhibitory effect of TGF-β1 on sodium uptake and αENaC expression in ATII cells was mediated by activation of the MAPK, ERK1/2. Consistent with the in vitro results, TGF-β1 inhibited the amiloride-sensitive fraction of the distal airway epithelial fluid transport in an in vivo rat model at a dose that was not associated with any change in epithelial protein permeability. These data indicate that increased TGF-β1 activity in the distal airspaces during ALI promotes alveolar edema by reducing distal airway epithelial sodium and fluid clearance. This reduction in sodium and fluid transport is attributable in large part to a reduction in apical membrane αENaC expression mediated through an ERK1/2-dependent inhibition of the αENaC promoter activity.

Acute lung injury (ALI) 1 is a devastating syndrome characterized by flooding of alveolar spaces with a protein-rich exudate that impairs pulmonary gas exchange, leading to arterial hypoxemia and respiratory failure (1). Epithelial injury can contribute to alveolar flooding, because the epithelial barrier is much less permeable under normal conditions than the endothelial barrier. Injury to alveolar epithelial cells can also disrupt normal epithelial fluid transport, impairing the removal of edema fluid from the alveolar space. Clinical studies have demonstrated that impaired alveolar fluid clearance is a characteristic feature of clinical lung injury (2,3), but the mechanisms for this decrease in epithelial fluid transport have not been well worked out. The removal of edema fluid from the airspaces occurs via an active transport-dependent sodium concentration gradient across the distal lung epithelium. The ratelimiting step in the transport of fluid across the lung epithelium is the movement of sodium and chloride across the apical plasma membrane, specifically the movement of sodium through amiloride-sensitive and -insensitive channels (4). Among the sodium channels at the apical membrane of lung epithelial cells, amiloride-sensitive channels represent 50 -60% of the sodium transport, particularly in rat and human lungs (4). Molecular identification of the proteins involved in amiloride-sensitive sodium flux achieved in the last few years revealed that three homologous subunits, ␣-, ␤-, and ␥ENaC, correspond to the pore-forming subunits that form a high affinity amiloride-sensitive sodium-selective pore (5). The critical role of ␣ENaC in the absorption of salt and fluid by lung epithelial cells has been confirmed by the generation of ␣EnaC knockout mice. Knockout neonates develop respiratory distress syndrome and die within 48 h of birth from failure to clear their lungs of fluid (6). In contrast, ␤and ␥ENaC knockout mice were able to clear fluid from the lungs at birth, although at a slower rate than in the wild-type control (7,8). The absence of intact distal lung sodium and fluid transport has also been associated with worse respiratory failure and higher mortality in patients with acute lung injury (3); however, the mechanisms for the regulation of the distal airspace lung sodium and fluid transport in acute lung injury are not fully understood.
The cytokine transforming growth factor ␤1 (TGF-␤1) plays a critical role in the resolution of tissue injury in multiple organs, including the lung (9). Following acute lung injury, TGF-␤1 has been most thoroughly evaluated during the late phases of tissue repair, where it plays a critical role in the development of pulmonary fibrosis (10,11). However, in a recent study evaluating global patterns of gene expression following bleomycin-induced lung injury, we found that the expression levels of several TGF-␤1-inducible genes were dramatically increased as early as 2 days after the induction of injury (12), a time point that precedes the maximal increase in alveolar flooding in this experimental model. Furthermore, the TGF-␤1-inducible gene, procollagen III, is one of the earliest predictors of the severity of acute lung injury in humans (13,14). In vitro, TGF-␤1 directly increases the permeability of endothelial monolayers (15). Using mouse models of bleomycininduced and E. coli-induced acute lung injury, we previously found that active TGF-␤1 is a critical mediator of alveolar edema (19). In vitro, TGF-␤1 increased permeability across epithelial and endothelial monolayers, suggesting mechanisms by which this cytokine could contribute to alveolar flooding.
In addition to its known effect on the protein permeability across the lung endothelial and epithelial barriers, TGF-␤1 was shown to affect the ion channels at the apical membrane of the alveolar epithelium. For example, using a patch clamp technique, a recent study indicated that TGF-␤1 reduced the activity of highly selective cation (HSC) channels at the apical membrane of ATII cells (16). However, the mechanisms underlying these effects and their in vivo relevance were not examined. HSC channels are formed by a combination of the three ENaC subunits when coexpressed in Xenopus oocytes (17). We therefore hypothesized that activation of a TGF-␤1-dependent signaling pathway would decrease basal vectorial fluid transport across the lung epithelium by affecting expression or function of the amiloride-sensitive sodium channel ENaC in lung epithelial cells.
We now report that TGF-␤1 does indeed inhibit fluid alveolar epithelial sodium and fluid transport, both in vitro and in vivo, and that this effect is due to an ERK1/2-dependent repression of ␣ENaC mRNA and protein expression, resulting in a marked reduction in the steady state level of ␣ENaC at the apical surface of polarized alveolar epithelial cells.

In Vitro Studies
Cell Culture Primary cultures of rat and human alveolar epithelial cells were used for the in vitro studies. Rat ATII cells were isolated by elastase digestion and selective adhesion plating on rat immunoglobulin G, as previously reported (18). Rat ATII cells were plated on 6.5-mm polycarbonate transwells (Corning Costar Co., Cambridge, MA) with a 0.4-m pore size. Cells were seeded at a concentration of 1.5 ϫ 10 6 cells/cm 2 in DMEM-H21 medium containing 10% low endotoxin fetal bovine serum, 1% penicillin, and streptomycin and kept at 37°C in a humidified 95% air, 5% CO 2 environment. 24 h later, nonadherent epithelial cells were removed by washing with PBS, and fresh medium was added to the lower compartments of the transwells, thus maintaining the ATII cell monolayers with an air-liquid interface on their apical side. After 72-96 h, cells form confluent monolayers that reach a transepithelial electrical resistance greater than 1500 ohms⅐cm 2 .
Following approval of the University of California, San Francisco, Committee on Human Research, human alveolar epithelial type II cells were isolated using a modification of methods previously described (19). Briefly, alveolar type II cells were isolated from human lungs that were not used by the Northern California Transplant Donor Network. Our studies indicate that these lungs were in good condition, both physiologically and pathologically (20). Cells were isolated after the lungs had been preserved for 4 -8 h at 4°C. A lobe of the human lung was selected that had no evidence of injury on the preharvest chest radiograph, could be normally inflated, and had no area of consolidation or hemorrhage.
The pulmonary artery for this segment was perfused with 37°C PBS solution, and the distal airspaces of a segmental bronchus were lavaged 10 times with 37°C Ca 2ϩ -, Mg 2ϩ -free PBS solution containing 0.5 mM EGTA and EDTA. 60 -90 ml of pancreatic porcine elastase (8 units/ml) diluted in a Ca 2ϩ -, Mg 2ϩ -free Hanks' balanced salt solution was instilled into the airspaces of 50 g of the chosen segment of lung tissue. The lung was incubated in a water bath for 30 min at 37°C and minced finely in the presence of fetal bovine serum and DNase I (500 g/ml). The cell-rich fraction was filtered sequentially through one-layer gauze, two-layer gauze, and 150-and 30-m nylon meshes. The cell suspension was then layered onto a discontinuous Percoll density gradient 1.04 -1.09 g/ml solution and centrifuged at 400 ϫ g for 20 min to remove red blood cells. The cells that accumulated at the interface of the solution and the Percoll were a mixture of type II pneumocytes and alveolar macrophages. These cells were recovered by centrifugation at 200 ϫ g for 10 min at 4°C. The pellet was resuspended in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. The cells were incubated in Dulbecco's modified Eagle's medium containing magnetic beads coated with an anti-CD-14 antibody (Dynabeads M/450 CD14; Dynal, Oslo, Norway) at 4°C for 40 min under constant mixing to eliminate macrophages. The cell viability was assessed by trypan blue exclusion. The purity of isolated human alveolar type II cells was checked by Papanicolaou staining or by staining with anti-human type II cell antibody (obtained from Leland Dobbs, University of California, San Francisco), and their purity has been more than 90% (data not shown). Human alveolar type II cells were seeded on collagen I-coated transwells at a density of 1 ϫ 10 6 cells/cm 2 . Five days after the cells were seeded, the monolayer developed a transepithelial electrical resistance greater than 1500 ohms⅐cm 2 , as reported for rat ATII cell monolayers.

Reagents
All cell culture media were prepared by the University of California, San Francisco, Cell Culture Facility using deionized water and analytical grade reagents. Amiloride, ouabain, and protease inhibitors were obtained from Sigma. Human TGF-␤1 was obtained from R&D Systems (Minneapolis, MN). 22 Na was obtained from PerkinElmer Life Sciences at a specific activity of 582 mCi/mg and at a concentration of 1 mCi/ml. Sulfo-NHS-SS-Biotin and streptavidin-agarose beads were obtained from Pierce. The MAPK inhibitors PD98059 (an inhibitor of the kinase upstream of ERK1/2), SB203580 (an inhibitor of the ␣and ␤-isoforms of p38 MAPK), and SP600125 (a reversible inhibitor of the c-Jun Nterminal kinase), were obtained from Calbiochem. All of these MAPK inhibitors have been shown to block their respective MAPK in lung epithelial cells (21)(22)(23). Mifepristone (RU486), a known glucocorticoid antagonist, has been purchased from BIOMOL (Plymouth Meeting, PA).
Monoclonal antibody against the three isoforms of TGF-␤ was obtained from R&D Systems (Minneapolis, MN). Mouse monoclonal actin antibody was obtained from Chemicon international (Temecula, CA), and antibodies recognizing ERK and phospho-ERK were from Cell Signaling (Beverly, MA). Polyclonal and affinity-purified ␣ENaC antibodies were provided by C. Canessa (Yale University School of Medicine, New Haven, CT). Antibodies and phosphoantibodies for MAPKs were purchased from Cell Signaling Technology (Beverly, MA). Goat anti-rabbit and anti-mouse secondary antibodies conjugated with horseradish peroxidase were purchased from ICN (Costa Mesa, CA). Cationic liposomes (Fugene 6) were obtained from Roche Applied Science, and the chemiluminescence kit was obtained from Amersham Biosciences. The following three plasmids were a kind gift from A. Dagenais (Universite de Montreal, Montreal, Canada): (a) a plasmid (pGL-3) containing the luciferase gene under the control of the fulllength murine ␣ENaC promoter (pGL3-basic/␣mENaC promoter), (b) a plasmid (pGL-3) containing the luciferase gene under the control of a murine ␣ENaC promoter with a deletion in the glucocorticoid regulatory element (GRE) (pGL3-basic/␣mENaC del AVR-XhO I promoter), and (c) a plasmid (pGL-3) containing the luciferase gene under the control of the full-length murine ␣ENaC promoter and containing a SV40 enhancer. Protein concentration of cell lysates was determined using the Bio-Rad protein assay kit with BSA as the standard.

Northern Blot Analysis
Total RNA from rat ATII cells was extracted by a modification of the guanidinium-phenol technique (24). 10 g of total RNA was electrophoresed on a 1% agarose-formaldehyde gel and transferred to Gene-Screen nylon membrane (PerkinElmer Life Sciences) by overnight blotting with a 10ϫ saline/sodium citrate solution. Hybridization was performed in a solution containing 0.5 M sodium phosphate, pH 7.2, 7% SDS (w/v), and 1 mM EDTA, pH 8. The nylon membranes were hybridized successively with ␣ENaC and 18 S rRNA cDNA probes labeled with [ 32 P]dCTP. ␣ENaC mRNA was detected with a 764-bp mouse ␣-ENaC cDNA (His 445 to the stop codon), which has a high homology with rat ␣ENaC cDNA (24). For quantitative studies, ␣-ENaC mRNA was normalized to 18 S rRNA expression to ensure that the same amount of RNA was present in each lane. The 18 S rRNA probe consisted of a 640-bp cDNA fragment that has been amplified by reverse transcription-PCR between nucleotides 852 and 1492 of the rat 18 S rRNA sequence (24). After hybridization, the membranes were washed successively for 30 min with 100 mM sodium phosphate (pH 7.2), 0.1% (w/v) SDS; 40 mM sodium phosphate (pH 7.2), 0.1% (w/v) SDS; and 40 mM sodium phosphate (pH 7.2), 1% (w/v) SDS. The blots were exposed either to Kodak X-AR film, using an intensifying screen, or to a Phos-phorImager (Amersham Biosciences) for densitometric analysis. For reproducibility and statistical reasons, Northern blotting was repeated several times with RNA extracted from cells isolated from different animals.

Western Blot Analysis
Western blot analysis was performed as described previously (25). After equal amounts of protein were loaded in each lane and separated in 10% SDS-PAGE gels, proteins were transferred to Immobilon-P membranes (Millipore Corp., Bedford, MA), and ␣ENaC was detected using an affinity-purified antibody at a 1:1,000 dilution and horseradish peroxidase-GAR at a 1:10,000 dilution. Actin, MAPKs, and phospho-MAPKs were detected with commercially available antibodies (mouse monoclonal actin antibody at a dilution of 1:1,000, MAPK and P-MAPK antibodies at a dilution of 1:500, and horseradish peroxidase-GAM at a 1:2,000 dilution). The protein bands were visualized using chemiluminescence. Quantification was done using a digital image analysis system (ChemiImager; Alpha Innotech).

Membrane Protein Biotinylation
Biotinylation and recovery of apical membrane proteins were performed as previously described (26) using Sulfo-NHS-SS-Biotin. The protein concentration of the lysates was determined using the BCA kit. Streptavidin-agarose beads were added to equal amounts of total protein to recover the biotinylated proteins. The amount of added beads was adjusted to ensure a complete recovery of biotinylated proteins from lysates. Biotinylated proteins were eluted from the beads by heating to 90°C in SDS-PAGE sample buffer and analyzed by Western blotting.

Measurement of Monolayer Bioelectric Properties
Transepithelial resistance (R t ; kiloohms⅐cm 2 ) and spontaneous potential difference (SPD; mV, apical side as reference) were measured using the Millicell-ERS (Millipore). Transepithelial current (I sc ; A/ cm 2 ) was calculated from the relationship I sc ϭ SPD/R t (Ohm's law). The effect of TGF-␤1 (0.1-10 ng/ml for 24 h) or its vehicle on bioelectric properties of ATII cell monolayer was evaluated on day 4 in culture. Because it has previously been shown that TGF-␤1 can induce a polarized regulation of its dependent genes in lung epithelial cells (27), TGF-␤1 was also added only to either the apical or basolateral side of the cell monolayer. 22 Na Uptake-The activity of sodium transport pathway on the apical membrane of rat and human ATII cells was determined by unidirectional tracer uptake measurement using the technique described by Mairbaurl et al. (28). Briefly, after exposure to TGF-␤1 or to its vehicle on both sides of the transwell, the cells were washed two times with washing medium (150 mM NaCl and 2 mM HEPES, pH 7.4) at 37°C and were equilibrated to the flux medium for 10 min. The flux medium was composed of 140 mM NaCl, 5 mM KCl, 1 mM Na 2 HPO 4 , 1 mM MgCl 2 , 0.2 mM CaCl 2 , 10 mM glucose, and 20 mM HEPES, pH 7.4, at 37°C. After equilibration, the medium was replaced with fresh flux medium containing 22 Na at a final activity of 5 Ci/ml and ouabain (3 ϫ 10 Ϫ3 M). After 6 min, cell monolayers were washed four times by overflow with cold washing medium to eliminate 22 Na not taken up by the cells and to stop further uptake. We verified that the last wash solution did not contain any radioactivity. The cells were then lysed with 0.1% NaOH, and the radioactivity in the lysate was measured in a ␤-counter. Protein determination was used for normalization of the results. Control experiments have shown that the tracer uptake was linear over a period range of 10 min. Thus, only one 6-min time point was taken. All fluxes were measured in triplicate.

Sodium Uptake and Fluid Transport Measurements
Fluid Transport-Fluid transport across rat and human ATII cell monolayers was measured as described previously (29). Briefly, after rat ATII cells were plated on transwells and grown in an air-liquid interface for 24 h, 150 l of medium containing 0.5 Ci/ml 131 I-albumin was added to the apical side of the monolayers, and the transwells were placed in a humidified tent. 5 min later, 20 l of medium containing 131 I-albumin were sampled from the apical side (initial sample). Transwells were then incubated for 24 h at 37°C with 100% humidity in a CO 2 incubator. Medium in upper and lower compartments of the transwell were kept at the same level to avoid any effect of the hydrostatic pressure. 24 h later, a second 20-l sample was removed from the apical side of the monolayers (final sample). Collected samples were weighed and counted in a ␥-counter (Beckman). Fluid absorption was calculated as in our prior in vivo studies (30). Fluid absorption ϭ (1 Ϫ radioactivity in the initial sample/radioactivity in the final sample) ϫ 100%. In pilot experiments, we were able to detect a 2-fold increase in the fluid transport across these ATII cell monolayers after c-AMP stimulation (10 Ϫ5 M forskolin plus 4 ϫ 10 Ϫ4 M isobutylmethylxanthine) and a 50% inhibition of the fluid transport after exposure to amiloride (10 Ϫ4 M) (data not shown).

␣ENaC Promoter Reporter Cells and Luciferase Assay
Rat ATII cells were transiently transfected with one of the following plasmids: (a) a plasmid (pGL-3) containing the luciferase gene under the control of the full-length murine ␣ENaC promoter and containing a SV40 enhancer that induces a stronger expression the promoter (pGL3enhancer/␣mENaC promoter); (b) a plasmid (pGL-3) containing the luciferase gene under the control of the murine ␣ENaC promoter (pGL3-basic/␣mENaC promoter), or (c) a plasmid (pGL-3) containing the luciferase gene under the control of a murine ␣ENaC promoter with a deletion in the GRE (pGL3-basic/␣mENaC del AVR-XhO I promoter). After rat ATII cells were plated on transwells (5 ϫ 10 5 cells/transwell) and grown in an air-liquid interface for 24 h, cells were transfected in triplicate by incubation with cationic liposomes (6 l of Fugene 6, 1 g of DNA per transwell). 48 h later, the cells were exposed to TGF-␤1 or its vehicle for 8 or 24 h. In some experiments, an inhibitor of ERK 1/2, JNK or p38 MAPK, or dexamethasone (100 nM) with or without mifepristone (RU486) (5 M), a known glucocorticoid antagonist, was added to the cell medium 1 h prior to exposure to TGF-␤1 or its vehicle. Then cell extracts were prepared and analyzed for luciferase activity according to the manufacturer's instructions (Promega, Madison, WI) using a luminometer. Luciferase activity was corrected for total cellular protein and reported as a percentage of the control cells (cells that were transfected but not treated with TGF-␤1). Cells were also co-transfected with a ␤-galactosidase plasmid, and ␤-galactosidase activity was determined in order to verify that the efficiency of transfection was comparable within and between experiments.

Cell Viability
Cell viability after exposure to different experimental conditions was measured by an Alamar Blue assay. Cell media were replaced by medium containing 10% Alamar Blue and placed at 37°C in the cell incubator for 2-3 h. The medium was then collected and read on a plate reader at 530 nm.

In Vivo Studies Animal Preparation and Experimental Protocol
This protocol was approved by the University of California, San Francisco, Committee on Animal Research. Sprague-Dawley rats weighing 300 -350 g were anesthetized with an intraperitoneal injection of ketamine/xylazine (90 and 10 mg/kg, respectively). Rats were placed supine in a declined position, and a curved needle was placed into the trachea under direct vision. TGF-␤1 or its vehicle was instilled into the trachea through a feeding catheter. After the procedure, supplemental oxygen was given until the rats were fully awake. The first series of experiments was designed to measure the effect of active recombinant TGF-␤1 or its vehicle (n ϭ 4 in each group) on the lung endothelial and alveolar epithelial permeability to protein.
In the second series of experiments, distal airspace fluid clearance was measured 24 h after the intratracheal administration of TGF-␤1 or its vehicle (n ϭ 11 in the TGF-␤1 group, and n ϭ 8 in the control group), as we have done before (31). Pilot experiments indicated that active TGF-␤1 instilled into the airspaces could affect lung epithelial fluid clearance at doses that did not have any effect on the protein permeability across this barrier. In contrast, higher doses caused both a decrease in lung epithelial fluid transport and increase in lung endothelial and epithelial permeability to protein. Thus, the dose of active TGF-␤ 1 chosen for these studies was low enough (250 ng) so that we could study the effect of active TGF-␤ 1 on the in vivo vectorial lung epithelial fluid transport without the confounding effect of increase in epithelial permeability to protein. To assess the validity of our in vivo measurement of the lung endothelial and alveolar epithelial permeability to protein, we used the intratracheal instillation of IL-1␤ (50 ng) as a positive control.

Measurement of Pulmonary Edema
Measurement of pulmonary edema (excess lung water). Extravascular gravimetric lung water was measured on the left lung using standard methods (31).

Measurement of Endothelial and Epithelial Permeability to Albumin
20 h after intratracheal instillation of TGF-␤1 or its vehicle, 1 Ci of 125 I-labeled albumin in 1 ml of saline was administered intraperitoneally. After 2 h and at the end of the experiment (4 h after intraperitoneal administration of 125 I-labeled albumin), plasma samples were obtained. After exsanguinations, the left lung was processed for measurement of lung endothelial permeability to albumin, expressed as extravascular plasma equivalents (EVPE) in ml, calculated using the equation (31), where C H represents the 125 I counts/min/g in the homogenized lung, C Pend represents the counts/min/g in plasma at the end of the experiment, and C Pave represents the average counts/min/g in the plasma samples collected at 2 h and at the end of the experiment. Q B is the blood volume in the lungs determined by the gravimetric method.
To measure the effect of TGF-␤1 or its vehicle on the lung epithelial permeability to protein, the right lung was lavaged three times with 4 ml of saline, and the total radioactivity of the lavage was counted. Lung epithelial permeability to protein was expressed as distal airspace plasma equivalents (DAPE) in l calculated using the equation, where C A represents the total 125 I counts/min in the bronchoalveolar lavage, and C Pave represents the average counts/min/g in the plasma samples collected at 2 h and at the end of the experiment.

Measurement of Distal Airspace Fluid Clearance
As in our previous studies (31), we determined distal airspace fluid clearance (DAFC) by measuring the increase in protein tracer concentration ( 131 I-labeled albumin, 1 Ci) over a 30-min period following instillation of an iso-osmolar, 5% albumin solution in normal saline solution into the trachea using an in situ model previously described (31). The increase in protein concentration provides a good estimate of the liquid volume removed from the distal airspaces of the lung (31). Briefly, rats were anesthetized with pentobarbital (100 mg/kg intraperitoneally) and then exsanguinated by transection of the abdominal aorta. 6 ml/kg isotonic 5% albumin solution in NS warmed to 37°C was then instilled into the trachea. An airway pressure of 8 cm of H 2 O was maintained for 30 min with 100% oxygen. 5 min after instillation, a base-line sample (0.1 ml) was aspirated from the distal airways using PE50 tubing (T ϭ 0 sample). 30 min later, a second sample was collected from the distal airways. Body temperature was carefully maintained at 37°C throughout the experiment with an external heating lamp and warming blanket. Core body temperature was monitored with a digital thermometer (Fisher) placed in the midesophagus. DAFC, expressed as the percentage of alveolar fluid volume cleared during the first 1 h, was determined using the equation, where P I is the initial 131 I-albumin activity/g at T ϭ 0, and P F is the 131 I-albumin activity/g 1 h later.

Statistics
All of the data are summarized as mean Ϯ S.D. One-way analysis of variance and Fisher's exact t test were used to compare experimental with control groups. A p value of Ͻ0.05 was considered statistically significant.

TGF-␤1 Decreases Sodium and Fluid Transport across Rat
Lung Epithelial Cell Monolayers-We previously reported that active TGF-␤1, at concentrations of 5-10 ng/ml and higher, caused concentration-dependent increases in the permeability of rat ATII monolayers (18). Thus, we sought to determine whether lower concentrations of active TGF-␤1 would selectively affect sodium and fluid transport. Exposure to active TGF-␤1 on both sides of the cell monolayer at concentrations as low as 0.5 ng/ml for 24 h significantly decreased the transepithelial current across primary cultures of rat ATII cell monolayers calculated from the measurement of transepithelial resistance and spontaneous potential difference with an epithelial ohmvoltmeter with Ag-AgCl electrodes (Fig. 1, A-C). This TGF-␤1-mediated inhibition of the transepithelial current across ATII cell monolayers occurred only when TGF-␤1 was placed on the basolateral side of the cell monolayer (Fig. 1,  D-F).
TGF-␤1 at a concentration of 1 ng/ml had no effect on the total transepithelial resistance, suggesting that cellular ion transport shows no effect on transepithelial resistance. In fact, measurements of potential difference (PD) and transepithelial resistance in ATII monolayers before and after the addition of amiloride showed a decrease in PD with no change in resistance (data not shown). Thus, the resistance of ATII monolayers is determined mostly by the tight junctions. We hypothesized that the TGF-␤1-mediated decrease in the calculated transepithelial current was explained by a decrease in transepithelial ion transport. Therefore, we measured the effect of TGF-␤1 on the sodium uptake across the apical membrane of rat ATII cells. This technique measures only the transcellular, and not the paracellular, transport of sodium across rat ATII cell monolayers. Active TGF-␤1 caused a 40% decrease in sodium uptake by rat ATII cells. Time-and dose-response studies indicated that the decrease in sodium uptake by lung epithelial cells was comparable for concentrations of active TGF-␤1 from 1 to 30 ng/ml ( Fig. 2A) and was observed from 6 to 24 h after exposure to active TGF-␤1 (Fig. 2B). Moreover, the TGF-␤1-mediated decrease in sodium uptake by rat ATII cells was quantitatively similar to and unaffected by the addition of a saturating concentration of amiloride (10 Ϫ4 M), suggesting that TGF-␤1 largely abolished amiloride-sensitive sodium uptake in these cells (Fig. 2C). This effect was specific for TGF-␤1, since it was blocked by a specific antibody against active TGF-␤1, but it was not blocked using an isotypic irrelevant antibody (Fig. 2D).
To corroborate the results of 22 Na uptake experiments, we used a recently described method to measure net fluid transport across monolayers of cultured rat ATII cells (29). Active TGF-␤1 (1 ng/ml) decreased the fluid absorption across rat lung epithelial cell monolayers by 40% over 24 h (Fig. 2E). The net decrease in transepithelial fluid transport after exposure to active TGF-␤1 is comparable with the results obtained with radiolabeled sodium (Fig. 2A). This system makes it possible to measure directly net fluid transport in vitro, an important advance, since prior studies have only measured short circuit current and ion uptake across cultured lung epithelial cells.
TGF-␤1 Inhibits in Vivo Vectorial Alveolar Epithelial Fluid Transport in Rats-To determine whether active TGF-␤1 would affect in vivo basal fluid clearance across the distal airspace epithelium in rats, rats were instilled with 250 ng of active TGF-␤1 or with its vehicle via a feeding catheter placed in the trachea. This dose of TGF-␤1 was chosen because it caused a decrease in lung epithelial fluid clearance but did not have any effect on the protein permeability across this barrier. Active TGF-␤1 caused a 40% decrease in the vectorial fluid transport across the distal airspaces in rats (Fig. 3A). Active TGF-␤1 affected mostly the amiloride-sensitive fraction of the fluid clearance, since no further decrease in fluid clearance was observed when amiloride (10 Ϫ3 M) was added to the protein solution instilled into the distal airspaces (Fig. 3A). However, additional experiments in which a protein tracer labeled with 131 I-albumin was injected intravenously 20 h after administration of TGF-␤1 and 4 h before exsanguination of the animal showed that active TGF-␤1 caused only a modest increase in the extravascular accumulation of labeled albumin (Fig. 3C) and did not increase the extravascular lung water content measured by the gravimetric method (Fig. 3B). Moreover, active TGF-␤1 did not affect the protein permeability across the distal lung epithelium (Fig. 3D). Higher doses of TGF-␤1 (Ͼ1 g) increased both epithelial and endothelial permeability (data not shown). Taken together, these results indicate that active TGF-␤1 instilled in normal noninjured lungs inhibits the amiloride-sensitive fraction of the basal fluid transport across the distal airspace epithelium at a dose that did not affect the lung epithelial permeability to protein.

TGF-␤1 Decreases the Amiloride-sensitive Fraction of Sodium Uptake across the Apical Membrane of Primary Cultures of Polarized Human ATII Cell Monolayers-
The studies on the effect of TGF-␤1 on sodium uptake by rodent ATII cells was extended to primary cultures of human ATII cells due to the concerns of possible differences in the regulation of sodium uptake between rodents and humans. Exposure to TGF-␤1 (10 ng/ml) for 24 h caused a significant decrease (35%) in sodium uptake across the apical membrane of human lung epithelial cells (Fig. 4). Active TGF-␤1 affected mostly the amiloridesensitive fraction of the apical sodium uptake, since no further decrease in sodium uptake was observed when amiloride (10 Ϫ4 M) was added with TGF-␤1 (Fig. 4). Taken together, these data indicate that active TGF-␤1 produced a comparable inhibition of the vectorial movement of sodium across the apical membrane of rat and human lung epithelial cell monolayers (Figs.  2C and 4).

TGF-␤1 Decreases ␣ENaC Gene and Protein Expression in Rat Lung Epithelial Cell Monolayers-
The previous series of experiments demonstrated that TGF-␤ 1 significantly affected the function of amiloride-sensitive sodium channels, so the next series of experiments was designed to determine whether active TGF-␤1 would alter the ␣ENaC gene and total protein expression. ␣ENaC gene expression was measured by Northern blot after exposure to active TGF-␤1 (10 ng/ml) for 24 h. The results showed that TGF-␤1 caused more than a 2-fold decrease in the ␣ENaC gene expression compared with cells treated with its vehicle (Fig. 5A). Also, active TGF-␤1 reduced the total ␣ENaC protein expression in lung epithelial cells by 40% (Fig.  5B). Preliminary studies from other investigators have suggested that active TGF-␤1 decreased functional HSC channels at the surface of rat lung epithelial cells (16). Previous studies have demonstrated that these HSC channels correspond to the pore-forming channels composed of the multiple subunits of ENaC (32); thus, the next series of experiments was designed to determine whether active TGF-␤ 1 would decrease the expression of ␣ENaC protein at the apical cell membrane. As

FIG. 1. TGF-␤1 decreases transepithelial potential difference and current across rat lung epithelial cell monolayers.
A-C, active TGF-␤1 induces a dose-dependent decrease in transepithelial PD and transepithelial current (TEC) across primary cultures of polarized rat ATII cell monolayers. Rat ATII cell monolayers cultured at an air-liquid interface for 4 days were exposed to TGF-␤1 (0.1-10 ng/ml) for 24 h; PD, TEC, and resistance were measured with an epithelial ohmvoltmeter with Ag-AgCl electrodes; results are the mean Ϯ S.D. of at least four experiments done in triplicate; *, p Ͻ 0.05 from monolayers exposed to TGF-␤1 vehicle. D-F, active TGF-␤1 induces a polarized regulation of transepithelial PD and TEC across primary cultures of polarized rat ATII cell monolayers. Rat ATII cell monolayers cultured at an air-liquid interface for 4 days were exposed for 24 h to TGF-␤1 (10 ng/ml) that was placed either on both sides or on the apical or basolateral side only of the cell monolayer; PD, TEC, and resistance were measured with an epithelial ohmvoltmeter with Ag-AgCl electrodes. Results are the mean Ϯ S.D. of at least four experiments done in triplicate; *, p Ͻ 0.05 from monolayers exposed to TGF-␤1 vehicle.
shown in Fig. 5C, apical membrane biotinylated ␣ENaC protein was detected by Western blot. Exposure to active TGF-␤1 (10 ng/ml) for 24 h decreased the amount of biotinylated ␣-ENaC protein by ϳ80% (Fig. 5C). Control experiments showed that actin, an intracellular protein, was not detected in the pool of apical membrane biotinylated proteins but was detected in the whole cell lysate (Fig. 5D). Taken together, these data demonstrate that active TGF-␤ 1 markedly decreases both ␣ENaC gene expression and ␣ENaC protein expression at the apical cell membrane of rat ATII cells, thus providing one explanation for the inhibition by TGF-␤1 of the amiloridesensitive fraction of the vectorial sodium transport across the lung epithelium.

TGF-␤1 Decreases Sodium and Fluid Transport across Rat Lung Epithelial Cell Monolayers via an ERK1/2-dependent
Mechanism-Previous studies have shown that activation of the MAPKs, ERK1/2, can inhibit ␣ENaC gene expression (33). Because TGF-␤1 signaling has been shown to activate multiple MAPK pathways in other cell types, we sought to determine whether TGF-␤1-induced inhibition of ␣ENaC expression and 22 Na uptake in rat ATII cells was mediated by ERK1/2. Preincubation with PD98059 (10 M), an inhibitor of the kinase upstream of ERK1/2, blocked the TGF-␤1-dependent decrease in sodium uptake across rat ATII cells (Fig. 6A). In contrast, pretreatment with SB203580, an inhibitor of the ␣and ␤-isoforms of p38 MAPK, or with SP600125, a reversible inhibitor of the c-Jun N-terminal kinase, did not affect the TGF-␤1-mediated inhibition of sodium uptake (Fig. 6A). Moreover, active TGF-␤1 induced the phosphorylation of ERK1/2, p38, and JNK in rat lung epithelial cells, an effect blocked by pretreatment with PD98059, SB203580, or SP600125, respectively, all known inhibitors of one of the three MAPKs (Fig. 6B).
It has previously been shown that the activation of the ERK1/2 signaling pathway caused a down-regulation of the transcription of the ␣ENaC by inhibiting its promoter activity (33). Thus, rat ATII cells were transiently transfected with a plasmid (pGL-3) containing the luciferase gene under the control of the murine ␣ENaC promoter and a SV40 enhancer. 48 h FIG. 2. TGF-␤1 decreases sodium and fluid transport across rat lung epithelial cell monolayers. A, dose response of TGF-␤1 on sodium uptake across the apical membrane of primary cultures of polarized rat ATII cell monolayers. Rat ATII cell monolayers cultured at an air-liquid interface for 4 days were exposed to TGF-␤1 (0.1-30 ng/ml) for 24 h; 22 Na uptake was measured as described under "Experimental Procedures." Results are the mean Ϯ S.D. of at least four experiments done in triplicate; *, p Ͻ 0.05 from monolayers exposed to TGF-␤1 vehicle. B, time response of TGF-␤1 on sodium uptake across the apical membrane of primary cultures of polarized rat ATII cell monolayers. Rat ATII cell monolayers cultured at an air-liquid interface for 4 days were exposed to TGF-␤1 (10 ng/ml) for 6 -48 h; 22 Na uptake was measured as described under "Experimental Procedures." Results are the mean Ϯ S.D. of at least four experiments done in triplicate; *, p Ͻ 0.05 from monolayers exposed to TGF-␤1 vehicle. C and D, active TGF-␤1 decreases the amiloride-sensitive fraction of sodium uptake across the apical membrane of primary cultures of polarized rat ATII cell monolayers. Rat ATII cell monolayers cultured at an air-liquid interface for 4 days were exposed to TGF-␤1 (10 ng/ml) for 24 h; 22 Na uptake was measured as described under "Experimental Procedures." In some experiments, amiloride (10 Ϫ4 M) was added to the flux medium; in other experiments, antibodies either against TGF-␤1 or an isotypic irrelevant antibody were added to the culture medium at the same time as TGF-␤1. Results are the mean Ϯ S.D. of at least four experiments done in triplicate; *, p Ͻ 0.05 from monolayers exposed to TGF-␤1 vehicle. E, active TGF-␤1 decreases fluid absorption across polarized rat ATII cell monolayers. Rat ATII cell monolayers cultured at an air-liquid interface for 4 days were exposed to TGF-␤1 (1 ng/ml) for 24 h; fluid absorption was calculated by direct volume weight and independently by the concentration of 131 I-albumin; the surface area of these cultured transwell monolayers is 0.33 cm 2 , and the data are shown as fluid absorption expressed as l/cm 2 /h. Results are the mean Ϯ S.D. of at least four experiments done in triplicate; *, p Ͻ 0.05 from monolayers exposed to TGF-␤1 vehicle. posttransfection, the cells were exposed to TGF-␤1 or its vehicle for 24 h. In some experiments, an inhibitor of ERK1/2, JNK, or p38 MAPK was added to the cell medium 1 h prior to exposure to TGF-␤1 or its vehicle. Then cells were harvested, and luciferase activity was measured, corrected for total cellular protein, and reported as a percentage of the control cells (cells that were transfected but not treated with TGF-␤1). Cells were also co-transfected with a ␤-galactosidase plasmid, and ␤-galactosidase activity was determined in order to verify that the efficiency of transfection was comparable within and between experiments (data not shown). The results showed that active TGF-␤1 decreased the activity of the ␣ENaC promoter by 40%, an effect that was blocked by preincubation with the inhibitor of the ERK1/2 pathway (PD98059) (10 M) (Fig. 6C). In contrast, pretreatment with the inhibitor of p38 MAPK or of the c-Jun N-terminal kinase, did not reverse the TGF-␤1-mediated inhibition of ␣ENaC promoter activity in lung epithelial cells (Fig. 6C).
Glucocorticoids increase ␣ENaC gene expression in alveolar epithelial cells via binding to the glucocorticoid responsive element on the ␣ENaC promoter (24) and stimulate in vivo FIG. 4. TGF-␤1 decreases the amiloride-sensitive fraction of sodium uptake across the apical membrane of primary cultures of polarized human ATII cell monolayers. Human ATII cell monolayers cultured at an air-liquid interface for 4 days were exposed to TGF-␤1 (10 ng/ml) or its vehicle for 24 h; 22 Na uptake was measured as described under "Experimental Procedures." In some experiments, amiloride (10 Ϫ4 M) was added to the flux medium. Results are the mean Ϯ S.D. of at least four experiments done in triplicate; *, p Ͻ 0.05 from control monolayers exposed to TGF-␤1 vehicle.

FIG. 3. TGF-␤1 inhibits vectorial alveolar epithelial fluid transport in rats.
A, active TGF-␤1 inhibits the amiloride-sensitive fraction of the basal vectorial fluid transport across the distal lung epithelium. Rats were instilled intratracheally with 250 ng of active TGF-␤1 or its vehicle; 24 h later, the animals were exsanguinated, and distal airspace fluid clearance was measured over 30 min by tracheal instillation of a 5% protein solution labeled with 131 I-albumin while the lungs were maintained inflated with a positive pressure of 8 cm of H 2 O. In some experiments, amiloride (10 Ϫ3 M) was added to the protein solution instilled in the trachea. Results are the mean Ϯ S.D. of 11 experiments for the TGF-␤1 group and of 8 experiments for the control group; *, p Ͻ 0.05 from controls instilled with TGF-␤1 vehicle. B, active TGF-␤1 does not increase the extravascular lung water content of the lung. Rats were instilled intratracheally with 250 ng of active TGF-␤1 or its vehicle; extravascular lung water content was measured by the gravimetric method. Results are the mean Ϯ S.D. of five experiments; *, p Ͻ 0.05 from controls instilled with TGF-␤1 vehicle. C and D, active TGF-␤1 causes a modest increase in the lung endothelial permeability to protein but does not affect the barrier function of the distal airspace epithelium. Rats were instilled intratracheally with 250 ng of active TGF-␤1 or its vehicle. A protein tracer labeled with 131 I-albumin was injected intravenously 20 h after administration of TGF-␤1 and 4 h before exsanguination of the animal. Labeled albumin was measured in the extravascular and distal airspaces of the lung as plasma equivalents. Intratracheal instillation of IL-1␤ (50 ng), used as a positive control, caused a significant increase in both the lung endothelial and alveolar epithelial permeability to protein. Results are the mean Ϯ S.D. of four experiments; *, p Ͻ 0.05 from controls instilled with TGF-␤1 vehicle.
FIG. 5. TGF-␤1 decreases ␣ENaC gene and protein expression in rat lung epithelial cell monolayers. A, active TGF-␤1 causes a 2-fold decrease in ␣ENaC gene expression in polarized rat ATII cell monolayers. Rat ATII cell monolayers cultured in an air-liquid interface for 4 days were exposed to TGF-␤1 (10 ng/ml) for 24 h. ␣ENaC gene expression was measured by Northern blot. One representative experiment is shown. Densitometry analysis of four comparable Northern blots for detection of ␣ENaC gene expression showed a significant decrease in the expression of ␣ENaC gene expression in rat ATII cell monolayers that were pretreated with TGF-␤1 for 24 h compared with control monolayers exposed to TGF-␤1 vehicle; *, p Ͻ 0.05 from control monolayers. B, active TGF-␤1 causes a 2-fold decrease in ␣ENaC total protein expression in polarized rat ATII cell monolayers. Rat ATII cell monolayers cultured in an air-liquid interface for 4 days were exposed to TGF-␤1 (10 ng/ml) for 24 h. ␣ENaC total protein expression was detected by Western blot. One representative experiment is shown. Densitometry analysis of four comparable Western blots for detection of ␣ENaC total protein expression showed a significant decrease in the expression of ␣ENaC total protein expression in rat ATII cell monolayers that were pretreated with TGF-␤1 for 24 h compared with control monolayers exposed to TGF-␤1 vehicle; *, p Ͻ 0.05 from control monolayers. C and D, active TGF-␤1 decreases ␣ENaC expression at the apical cell membrane of polarized rat ATII cell monolayers. Rat ATII cell monolayers cultured in air-liquid interface for 4 days were exposed to TGF-␤1 (10 ng/ml) vectorial fluid transport across the lung epithelial barrier (34). Moreover, oxidative stress disrupts GRE-dependent transcription of ␣ENaC in lung epithelial cells through an ERK1/2-dependent mechanism (33). Thus, the next series of experiments was designed to determine whether active TGF-␤1 would affect the dexamethasone-dependent stimulation of ␣ENaC transcription via an ERK1/2-dependent inhibition of the GRE elements present on the ␣ENaC promoter. Rat ATII cells were transiently transfected with a plasmid (pGL-3) containing the luciferase gene under the control of the murine ␣ENaC promoter (pGL3-basic/␣mENaC promoter) or with a plasmid (pGL-3) containing the luciferase gene under the control of a murine ␣ENaC promoter with a deletion in the GRE (pGL3basic/␣mENaC del AVR-XhO I promoter). Cells were also cotransfected with a ␤-galactosidase plasmid, and ␤-galactosidase activity was determined in order to verify that the for 24 h. ␣ENaC protein expression at the apical cell membrane was measured by cell surface biotinylation. Actin, an intracellular protein, was not detected in the pool of biotinylated proteins but was detected in the whole cell lysate (positive control). One representative experiment is shown; densitometry analysis of four comparable blots for detection of ␣ENaC protein expression at the apical cell membrane showed a significant decrease in the expression of ␣ENaC protein expression at the apical cell membrane in rat ATII cell monolayers that were pretreated with TGF-␤1 for 24 h compared with control monolayers exposed to TGF-␤1 vehicle; *, p Ͻ 0.05 from control monolayers.
FIG. 6. TGF-␤1 decreases sodium and fluid transport across rat lung epithelial cell monolayers via an ERK1/2-dependent mechanism. A, active TGF-␤1 decreases sodium uptake across the apical membrane of primary cultures of polarized rat ATII cell monolayers via an ERK1/2-dependent mechanism. Rat ATII cell monolayers cultured at an air-liquid interface for 4 days were exposed to TGF-␤1 (10 ng/ml) or its vehicle for 24 h; 22 Na uptake was measured as described under "Experimental Procedures"; in some experiments, PD98059 (10 M) (an inhibitor of the kinase upstream of ERK1/2), SB203580 (10 M) (an inhibitor of the ␣and ␤-isoforms of p38 MAPK), or SP600125 (10 M) (a reversible inhibitor of the c-Jun N-terminal kinase) was added to the cell medium 1 h before exposure to TGF-␤1 or to its vehicle. Results are the mean Ϯ S.D. of at least four experiments done in triplicate; *, p Ͻ 0.05 from monolayers exposed to TGF-␤1 vehicle. B, active TGF-␤1 causes a time-dependent increase in the phosphorylation of the three MAPKs in polarized rat ATII cell monolayers. Rat ATII cell monolayers cultured in an air-liquid interface for 4 days were exposed to TGF-␤1 (10 ng/ml) or its vehicle for 24 h. Total protein expression was determined by Western blot. In some experiments, PD98059 (10 or 20 M) (an inhibitor of the kinase upstream of ERK1/2), SB203580 (10 or 20 M) (an inhibitor of the ␣and ␤-isoforms of p38 MAPK), or SP600125 (10 or 20 M) (a reversible inhibitor of the c-Jun N-terminal kinase) was added to the cell medium 1 h before exposure to TGF-␤1 or to its vehicle. One representative experiment is shown; four additional experiments gave comparable results. C, active TGF-␤1 inhibits the activity of ␣ENaC promoter in polarized rat ATII cell monolayers via an ERK1/2-dependent mechanism. After rat ATII cells were plated on transwells (5 ϫ 10 5 cells/transwell) and grown in an air-liquid interface for 24 h, cells were transiently transfected with a plasmid (pGL-3) containing the luciferase gene under the control of the murine ␣ENaC promoter and containing a SV40 enhancer that induces a stronger expression of the ␣ENaC promoter. 48 h later, the cells were exposed to TGF-␤1 or its vehicle for 24 h; in some experiments, PD98059 (10 M) (an inhibitor of the kinase upstream of ERK1/2), SB203580 (10 M) (an inhibitor of the ␣and ␤-isoforms of p38 MAPK), or SP600125 (10 M) (a reversible inhibitor of the c-Jun N-terminal kinase) was added to the cell medium 1 h prior to exposure to TGF-␤1 or its vehicle. Results are the mean Ϯ S.D. of four experiments done in triplicate; *, p Ͻ 0.05 from monolayers exposed to TGF-␤1 vehicle. efficiency of transfection was comparable within and between experiments (data not shown). The results indicated that active TGF-␤1 decreased the dexamethasone-dependent ␣ENaC promoter activity by 80%, an effect that was blocked by prior exposure to PD98059 (10 M) (Fig. 7). This effect was not observed in lung epithelial cells transiently transfected with a plasmid that did not respond to dexamethasone (pGL3-basic/ ␣mENaC del AVR-XhO I promoter). A comparable lack of response to dexamethasone was also observed by pretreating the cell monolayers with mifepristone (RU486), a known glucocorticoid antagonist, before exposing the cells to dexamethasone. Interestingly, TGF-␤1 did not further inhibit the dexamethasone-dependent ␣ENaC promoter activity in cells that have been pretreated with mifepristone (Fig. 7). In addition, pretreatment with SB203580, an inhibitor of the ␣and ␤-isoforms of p38 MAPK or with SP600125, a reversible inhibitor of the c-Jun N-terminal kinase, did not reverse the TGF-␤1-mediated inhibition of the dexamethasone-dependent ␣ENaC promoter activity in lung epithelial cells (Fig. 7). Finally, the base-line luciferase activity values obtained after transfection with the empty vector were within the range of the background values measured in cells that were not transfected. DISCUSSION The cytokine TGF-␤1 plays a critical role in the resolution of tissue injury in multiple organs, including the lung (9). Following ALI, TGF-␤1 has been thoroughly evaluated during the late phases of tissue repair, where it plays a critical role in the development of pulmonary fibrosis (10,11). We have recently reported that the activation of TGF-␤1 in the distal airspace of the lung is a critical step in the development of pulmonary edema after intratracheal administration of bleomycin or E. coli endotoxin in mice (18). TGF-␤1 directly increased alveolar epithelial permeability after its activation by the epithelial integrin ␣ v ␤ 6 , since mice lacking this integrin were completely protected from pulmonary edema after airspace instillation of bleomycin (18). These results indicate that active TGF-␤ 1 may contribute to the development of alveolar edema in ALI. Since impaired alveolar epithelial sodium and fluid transport contributes to pulmonary edema, we hypothesized that the activation of the TGF-␤1-dependent signaling pathway could be associated with a decrease in the vectorial ion and water transport across the distal lung epithelium.
The first series of experiments demonstrated that active TGF-␤1, at concentrations much lower than those required to affect epithelial paracellular permeability, significantly decreased 22 Na uptake across the apical membrane of rat and human distal lung epithelial cells, suggesting that active TGF-␤1 could modulate the function of apical ion channels involved in sodium uptake by these cells. These results were confirmed by directly measuring fluid transport across both rat and human ATII cell monolayers after exposure to active TGF-␤1. TGF-␤1-mediated inhibition of the transepithelial current across ATII cell monolayers was only observed when TGF-␤1 was placed on the basolateral side of the polarized cell monolayer. This finding is in accordance with previous results from our laboratory that have shown that TGF-␤1 can induce a polarized regulation of the secretion of fibronectin in lung epithelial cells (27). Additional studies performed in the presence of amiloride indicated that active TGF-␤1 inhibited most or all of the amiloride-sensitive fraction of sodium uptake by these cells.
Recent work has demonstrated that the most important amiloride-sensitive channel in lung epithelial cells is ENaC. Among the three ENaC subunits, ␣ENaC appears to play a critical role in maintaining fluid homeostasis in the distal airspaces of the lung. Indeed, the results of gene deletion experiments in mice indicate that ␣ENaC knockout mice develop respiratory distress and die within 48 h after birth because of their inability to clear fluid from the airspaces (6). In contrast, ␤and ␥ENaC knockout mice were able to clear fluid from the lungs at birth, although at a slower rate than in the wild-type control (7,8). We therefore examined whether the inhibition of FIG. 7. TGF-␤1 inhibits dexamethasone-dependent ␣ENaC promoter activity in polarized rat ATII cell monolayers via an ERK1/ 2-dependent mechanism. Rat ATII cells were transiently transfected with a plasmid (pGL-3) containing the luciferase gene under the control of the murine ␣ENaC promoter (pGL3-basic/␣ENaC promoter) or with a plasmid (pGL-3) containing the luciferase gene under the control of a murine ␣ENaC promoter with a deletion in the GRE (pGL3-basic/␣ENaC ⌬GRE). In contrast to the pGL3-basic/␣ENaC promoter, pGL3-basic/ ␣ENaC ⌬GRE is not responsive to dexamethasone; 48 h later, the cells were exposed to TGF-␤1 or its vehicle in the presence of dexamethasone (100 nM) for 8 h with or without mifepristone (RU486) (5 M). In some experiments, PD98059 (10 M) (an inhibitor of the kinase upstream of ERK1/2), SB203580 (10 M) (an inhibitor of the ␣and ␤-isoforms of p38 MAPK), or SP600125 (10 M) (a reversible inhibitor of the c-Jun N-terminal kinase) was added to the cell medium 1 h prior to exposure to TGF-␤1 or its vehicle. Results are the mean Ϯ S.D. of four experiments done in triplicate; *, p Ͻ 0.05 from monolayers exposed to TGF-␤1 vehicle. the vectorial epithelial ion and fluid transport by active TGF-␤1 was secondary to a decrease in ␣ENaC gene and protein expression in lung epithelial cells. Exposure of rat ATII cells to TGF-␤1 for 24 h significantly decreased total ␣ENaC gene and protein expression in these cells. Also, TGF-␤1 inhibited most of the apical membrane expression of ␣ENaC, as shown by the results of cell surface biotinylation experiments. This finding is in accordance with the fact that TGF-␤1 completely inhibited the amiloride-sensitive 22 Na uptake by lung epithelial cells and suggests that the expression of ENaC at the ATII cell apical membrane could account for most of the amiloride-sensitive sodium uptake by these cells. Indeed, a recent study demonstrated a direct relationship between cAMP-mediated increase in amiloride-sensitive sodium transport at the apical membrane of Madin-Darby canine kidney cells transfected with the three ENaC subunits and an increase in ENaC expression at the apical membrane of these cells (35).
What are the mechanisms by which TGF-␤1 regulates amiloride-sensitive sodium transport and ␣ENaC expression in lung epithelial cells? Canonical TGF-␤ signaling involves the activation of a family of transcriptional activators called SMADs that translocate to the nucleus in response to activation of TGF-␤ receptors and induce downstream effects through modulating expression of target genes. In many cell types, TGF-␤ isoforms can also activate MAPKs, including ERK1/2, p38, and JNK (36). In the current study, TGF-␤1 was found to activate all three MAPKs in rat ATII cell monolayers, but only the inhibition of the ERK1/2 pathway prevented the TGF-␤1-mediated inhibition of amiloride-sensitive sodium uptake. Our results are consistent with those of previous studies that have reported that the activation of ERK1/2 inhibits ␣ENaC gene and protein expression in several epithelial cell lines including lung epithelial cells (33). In contrast, the activation of the p38 MAPK pathway has been associated with a stimulation of ␣ENaC gene expression in renal collecting duct cells (37).
What are the mechanisms by which the activation of ERK1/2 inhibits ␣ENaC expression in lung epithelial cells? Previous studies have shown that the activation of ERK1/2 inhibits the activity of the ␣ENaC promoter. For example, the induction of an oxidative stress with exogenous H 2 O 2 inhibited the activity the ␣ENaC promoter via ERK1/2-dependent and thioredoxinsensitive pathways in A549 cells, an alveolar epithelial cell line (33). In this cell line, ERK1/2 inhibition of the ␣ENaC promoter occurred via the repression of ␣ENaC promoter GRE activity (33), which has been shown to be one of the strongest positive regulators of the ␣ENaC promoter (24). Our finding that TGF-␤1 inhibited the dexamethasone-dependent increase in ␣ENaC promoter activity suggests that the effect of TGF-␤1 on ␣ENaC expression in primary lung epithelial cells works by a similar mechanism.
These results did not exclude the existence of other levels of regulation of ␣ENaC by TGF-␤1, such as trafficking, channel activity, or degradation. Our data indicate that TGF-␤1 decreased the cell membrane expression of ␣ENaC more than the total ␣ENaC protein population, suggesting that TGF-␤1 could also regulate the trafficking as well as the turnover of ␣ENaC at the cell membrane. Recent studies have reported that ENaC protein expression at the apical cell membrane of lung epithelial cells may be affected by experimental conditions that are relevant for acute lung injury, such as hypoxia or c-AMP stimulation (38,39). Hypoxia is associated with a decrease in the cell membrane expression of the three ENaC subunits without affecting the total ␣ENaC protein population (39), whereas c-AMP increases sodium uptake across these cells by increasing the trafficking of ENaC to the cell membrane (35). Therefore, additional studies will be needed to determine how TGF-␤1 may affect trafficking and turnover of ENaC at the cell membrane of lung epithelial cells.
In Vivo Relevance-Our initial in vivo observation indicated that active TGF-␤1 is a critical mediator of alveolar edema and epithelial damage in two in vivo experimental models of acute lung injury (18). We therefore sought to determine whether TGF-␤1 could also affect basal in vivo alveolar fluid clearance in rats. Consistent with our in vitro results, we found that active TGF-␤1 instilled in normal noninjured lungs inhibited the amiloride-sensitive fraction of the basal fluid transport across the distal airspace epithelium at a dose that did not affect the lung epithelial permeability to protein. Taken together, the in vivo data from both studies indicate that active TGF-␤1 promotes the development of alveolar edema by increasing lung epithelial permeability to protein and by inhibiting the active fluid removal from the airspaces by the lung epithelium. What is the role of active TGF-␤1 in human ALI? Earlier clinical studies had shown that the TGF-␤1-inducible gene, procollagen III, is one of the earliest and best predictors of the severity of acute lung injury in humans (13,14). Clinical studies of the role of TGF-␤1 in the development of ALI in humans are still lacking. However, a recent study including a small number of patients showed the presence of functionally active TGF-␤1 in the bronchoalveolar lavage fluid from patients within 24 h of the diagnosis of ALI, whereas none was detected in control subjects (40). These data suggest that activation of TGF-␤1 could also be an important mechanism in the early phase of acute lung injury in humans.
In summary, the results of the current study demonstrate that increased TGF-␤1 activity in the distal airspaces during ALI promotes alveolar edema by reducing distal airspace epithelial sodium and fluid clearance. This reduction in sodium and fluid transport is attributable in large part to a decrease in apical membrane ␣ENaC expression in lung epithelial cells mediated through an ERK1/2-dependent inhibition of the ␣ENaC promoter activity. Clinical studies have demonstrated that impaired alveolar fluid clearance is a characteristic feature of clinical lung injury (2,3). Since the TGF-␤1-mediated inhibition of alveolar fluid clearance occurs at a concentration 5-10-fold lower than the one necessary to increase lung epithelial permeability to protein, and since functionally active TGF-␤1 is present within the airspaces of patients during the early phase of ALI (40), activation of TGF-␤1 in the distal airspaces is likely to be a critical mechanism for impaired fluid clearance in patients with ALI.