We have reported that TNF, a proinflammatory cytokine present in several lung pathologies, decreases the expression and activity of the epithelial Na+ channel (ENaC) by ∼70% in alveolar epithelial cells. Because dexamethasone has been shown to upregulate ENaC mRNA expression and is well known to downregulate proinflammatory genes, we tested if it could alleviate the effect of TNF on ENaC expression and activity. In cotreatment with TNF, we found that dexamethasone reversed the inhibitory effect of TNF and upregulated α, β, and γENaC mRNA expression. When the cells were pretreated for 24 h with TNF before cotreatment, dexamethasone was still able to increase αENaC mRNA expression to 1.8-fold above control values. However, in these conditions, β and γENaC mRNA expression was reduced to 47% and 14%, respectively. The potential role of TNF and dexamethasone on αENaC promoter activity was tested in A549 alveolar epithelial cells. TNF decreased luciferase (Luc) expression by ∼25% in these cells, indicating that the strong diminution of αENaC mRNA must be related to posttranscriptional events. Dexamethasone raised Luc expression by fivefold in the cells and augmented promoter activity by 2.77-fold in cotreatment with TNF. In addition to its effect on αENaC gene expression, dexamethasone was able to maintain amiloride-sensitive current as well as the liquid clearance abilities of TNF-treated cells within the normal range. All these results suggest that dexamethasone alleviates the downregulation of ENaC expression and activity in TNF-treated alveolar epithelial cells.
- Na+ channel
- transepithelial current
active sodium ion absorption by polarized lung alveolar epithelial cells is the main driving force of lung liquid clearance at birth and lung edema clearance in adulthood. Vectorial Na+ gating at the apical membrane and its extrusion by Na+-K+-ATPase at the basolateral side create osmotic pressure that drives water from the alveoli to the interstitium (5, 56). Although several types of Na+ channels and cotransporters are expressed in alveolar epithelial cells (54), experimental evidence indicates that the epithelial sodium channel (ENaC) plays a prominent role in the lung liquid clearance process. This channel is composed of three related subunits (α, β, and γ) (11, 12) that are expressed in the lung from alveoli to the nasal epithelium (10, 27, 55). Inactivation of αENaC expression in knockout (KO) mice produces a lethal phenotype; these mice die shortly after birth, unable to clear liquid from their lungs (36). Rescuing αENaC(−/−) KO by αENaC expression driven by the cytomegalovirus promoter leads to viable mice but with reduced ENaC expression in the kidney, colon, and lung (37, 61). These mice present increased lung edema and diminished lung liquid clearance in lung injury produced by thiourea or hyperoxia (22). In humans, respiratory distress syndrome has been reported in premature babies with type 1 pseudohypoaldosteronism, a rare genetic disease where ENaC is defective (1, 53). Downregulation of ENaC expression and activity is, therefore, associated with susceptibility to lung edema.
Inflammatory cytokines and growth factors, such as interleukin-1β (IL-1β) (72), IL-4 (34), interferon-γ (IFN-γ) (33), and transforming growth factor-β1 (TGF-β1) (31), present during lung inflammation and linked with acute respiratory distress syndrome, have been found to modulate ENaC expression and activity in lung epithelial cells. We reported recently that tumor necrosis factor (TNF) decreases ENaC expression and activity in primary cultures of alveolar epithelial cells (18). TNF is a proinflammatory cytokine that plays an important role in the activation of host defense by promoting the production of a wide spectrum of other cytokines (IL-1, IL-6, IL-8, and GM-CSF) in the inflammatory process (30, 77). It is not known so far how the cytokine modulates ENaC activity in lung epithelial cells and what is the physiological significance of this modulation.
Dexamethasone (Dex) is a synthetic steroid that has been shown to upregulate ENaC expression via a very conserved glucocorticoid regulatory element (GRE) in its promoter (15, 17, 63, 74). Dex is also a well-known anti-inflammatory agent that has been used or proposed as a therapeutic approach to inflammatory disease, including lung injury (29). It exerts its anti-inflammatory effect by repressing NF-κB and activator protein-1 (AP-1) proinflammatory gene stimulation (48). For all these reasons, we tested here if Dex could alleviate the impact of TNF on ENaC expression in alveolar epithelial cells. First, the modulation of α, β, and γENaC mRNA expression was studied by Northern blotting in alveolar epithelial cells treated with TNF and Dex. The consequence of such cotreatment on αENaC promoter activity was then evaluated after transient transfection of a plasmid where the αENaC promoter was driving luciferase (Luc) gene expression. To complete our study, we investigated the impact of these treatments on transepithelial and apical membrane amiloride-sensitive current generated by the cells as well as the liquid clearance ability of cell monolayers. Together, our results revealed that Dex in cotreatment with TNF normalizes ENaC expression and activity in alveolar epithelial cells.
MATERIALS AND METHODS
Alveolar epithelial cell isolation and experimental conditions.
Alveolar epithelial cells were isolated from male Sprague-Dawley rats as described previously and according to a procedure approved by our institutional animal care committee (17). Perfused lungs were digested with elastase, and the cells were purified by a differential adherence technique on bacteriological plastic plates coated with rat IgG (17). The cells were maintained in minimum essential medium (MEM) (Invitrogen, Burlington, Ontario, Canada) containing 10% FBS (Invitrogen), 0.08 mg/l gentamicin, 0.2% NaHCO3, 0.01 M HEPES, and 2 mM l-glutamine. They were plated at 4 × 105 cells/cm2 density in 25-cm2 plastic flasks or at 1 × 106 cells/cm2 on polycarbonate filters (Costar Transwell, Toronto, Ontario, Canada) and cultured at 37°C with 5% CO2 in a humidified incubator. The medium was supplemented with Septra (3 μg/ml trimethoprim and 17 μg/ml sulfamethoxazole) for the first 3 days. After this time, the medium was replaced, and the cells were cultured without Septra.
To study the influence of TNF and Dex on the αENaC steady state mRNA level, alveolar epithelial cells cultured for 3 days were treated with 100 ng/ml recombinant mouse TNF (Calbiochem, La Jolla, CA) for 6, 8, or 24 h in the presence or absence of 100 nM Dex. Total RNA from these cells was extracted by a modification of the guanidinium-phenol technique (19). For Northern blotting, 10 μg of total RNA were electrophoresed on 1% agarose-formaldehyde gel and transferred to GeneScreen nylon membranes (NEN, Boston, MA) after overnight blotting with 10× SSC. Hybridization was performed, as reported previously, in Church buffer [0.5 M sodium phosphate, pH 7.2, 7% SDS (wt/vol), 1 mM EDTA, pH 8] (19). The nylon membranes were hybridized successively with a 764-bp mouse αENaC cDNA probe (His445 to stop codon) (19), a 1,403-bp rat βENaC probe (ATG to Trp486), a 1,794-bp rat γENaC probe (Gly141 to 3′-UTR), and, finally, with an 18S rRNA probe consisting of a 640-bp fragment cloned from nt 852–1492 of rat 18S rRNA after reverse transcription-polymerase chain reaction (RT-PCR) amplification (17). The Northern blots were scanned and analyzed in a Typhoon PhosphorImager (Molecular Dynamics, Sunnyvale, CA). mRNA expression was always compared with matching untreated cells for each time period of the study. For reproducibility and statistical reasons, Northern blotting was repeated several times with RNA extracted from cells isolated from different rats. At least three animals (n = 3) were used for each time point.
Transient transfection and Luc assay.
The modulation of αENaC promoter activity by 100 ng/ml TNF, 100 nM dexamethasone (Dex), and TNF+Dex cotreatment was tested in A549 cells, a human alveolar epithelial cell line. The cells were transiently transfected with a 2.9-kb BamHI-MscI fragment of the mouse αENaC promoter (17) cloned upstream of the Photinus pyralis Luc reporter gene (αENaC-Luc). A 1.17-kb GRE deletion mutant (αENaC Δ GRE) not responding to dexamethasone was also tested. The generation of these constructs was described previously (31). In some experiments, the pHH-Luc plasmid with a mouse mammary tumor virus (MMTV) promoter highly responding to dexamethasone treatment (57) was also transfected to test for steroid resistance in A549 cells. pRL-SV40 (Promega, Madison, WI), a plasmid expressing Renilla reniformis luciferase, was cotransfected in every condition to allow normalization of the Luc response. Twenty-four hours before transfection, 1.25 × 105 cells were plated in 1.9-cm2 wells of 24-well plates and cultured without antibiotic in DMEM, 10% (vol/vol) FBS, and 2 mM l-glutamine. For each well, 1.4 μg of αENaC-Luc plasmid was mixed with 250 pg pRL-SV40 and 2.5 μl Lipofectamine 2000 (Invitrogen) in 100 μl OptiMEM. For transfection with the αENaC Δ GRE mutant, 550 ng of the plasmid was mixed with 80 pg pRL-SV40 and 1.38 μl Lipofectamine 2000. Similar conditions were used for pHH-Luc transfection. After 20 min of incubation at room temperature, the mix was added drop by drop to the cells. After 6 h of transfection, the cells were rinsed, and fresh medium was added either with or without 100 ng/ml TNF, 100 nM dexamethasone, or both combined for 24 h. Firefly and Renilla luciferase assays were undertaken with the Dual-Luciferase Reporter Assay System as specified by the manufacturer (Promega). Luminometry measurements were performed in a Turner TB/2020 luminometer. For response curves, the cells were incubated with 0, 1, 2.5, 10, or 25 nM dexamethasone in the presence or absence of 100 ng/ml TNF. Ten micromolar RU-486, an antagonist of the glucocorticoid receptor (GR), was tested in some transfections.
Cytokine detection in the medium.
The profiles of cytokines secreted in the conditioned medium of TNF-treated cells were investigated with a “protein array” approach using a rat cytokine array (rat cytokine array I; RayBiotech, Norcross, GA), which enables the simultaneous detection of 19 cytokines. The apical and basolateral media of control, TNF-, and TNF+Dex-treated cells were collected after 24-h stimulation. After 10 min of 2,800 g centrifugation, the supernatants were incubated overnight at 4°C with the membranes. The membranes were then incubated sequentially with biotinylated-antibodies and labeled streptavidin before being revealed by chemiluminescence according to the manufacturer's protocol. The membranes were exposed for different time periods to Kodak X-omat AR, which was developed with a film processor.
The impact of TNF on the electrophysiology of alveolar epithelial cells was studied, as reported previously (17), by successively recording potential differences (Pd; mV) and transepithelial resistance (Rte; Ω/cm2) across cell monolayers with an epithelial voltohmeter (EVOM; World Precision Instruments, Sarasota, FL). Alveolar cells plated at 1 × 106 cells/cm2 density on polycarbonate membranes (cat. no. 3401, Costar Transwell; 1.0 cm2) were cultured for 3 days, until they formed a tight epithelium and were treated or not for 24 h with 100 ng/ml TNF, 100 nM Dex, or TNF+Dex added at both the apical and basolateral sides. Transepithelial current (Ite) across the monolayers was calculated according to the following formula: Ite = Pd/Rte, where Pd is the transepithelial potential difference across cell monolayers, and Rte is the transepithelial resistance. To also quantify the amount of amiloride-sensitive current generated by the monolayers, Pd and Rte were measured successively before and after 5 min of incubation at 37°C with 1 μM amiloride. At this concentration, amiloride is a specific inhibitor of ENaC. At least 50 filters (n = 50), all from different animals, were investigated for every condition tested.
Short-circuit current measurement after amphotericin B permeabilization of the basolateral membrane.
Cells grown on 24-mm Costar filters (cat. no. 3412, Costar Transwell; 4.5 cm2) were cultured for 3 days and treated or not for 24 h with 100 ng/ml TNF, 100 nM Dex, or TNF+Dex. For short-circuit current (Isc) measurements, the filters were mounted in an Ussing chamber, and Isc52), with a few modifications. The apical side was bathed with the following physiological buffer: 141 mM NaCl, 5.4 mM KCl, 0.78 mM NaH2PO4, 1.8 mM CaCl2, 0.8 mM MgCl2, 5 mM glucose, and 15 mM HEPES, pH 7.4, at 37°C, whereas for the basolateral side 116 mM NaCl was replaced by an equivalent amount of N-methyl-d-glucamine (NMDG)-Cl. After stabilization of Isc for 5–10 min, the basolateral medium was replaced by a solution containing 7.5 μM amphotericin B. Amphotericin B stock solution was freshly prepared at a 10 mM concentration in DMSO and diluted to 7.5 μM in the basolateral solution. Amiloride-sensitive current was measured after the apical addition of 10 μM amiloride. Five filters (n = 5) seeded with cells from different animals were analyzed for every condition tested.
Semiquantitative RT-PCR estimation of αENaC mRNA expression for dose response curves to Dex.
Alveolar epithelial cells were cultured in 24-wells plates at a concentration of 1 × 106 cells/cm2. After 3 days of culture, the cells were incubated with 0, 1, 2.5, 10, or 25 nM Dex in the presence or absence of 100 ng/ml TNF. For each condition tested, total RNA was isolated from two wells (3.8 cm2 total surface area) by directly lysing the cells with TRIzol reagent according to the manufacturer's protocol (Invitrogen). αENaC mRNA expression was examined by semiquantitative RT-PCR. Five micrograms of total RNA were reverse-transcribed with MMLV reverse transcriptase (Invitrogen), as reported previously (18). An amount of 1.5 out of 20 μl of the cDNA reaction was subjected to PCR amplification with Taq polymerase (18). For αENaC amplification, 50 pmol (1 μM) of the sense primer 5′-GAG CCT GCC TTT ATG GAT GA-3′ located in exon 5 and 50 pmol (1 μM) of the antisense primer 5′-GAG CTT TGC AAC TCC GTT TC-3′ present in exon 11 were used for PCR. For β-actin cDNA amplification, 12.5 pmol (250 nM) of the sense primer 5′-CTA AGG CCA ACC GTG AAA AG located in exon 3 and 12.5 pmole (250 nM) of the antisense primer 5′-GCC ATC TCT TGC TCG AAG TC-3′ present in exon 4 were employed. The amplification conditions were 1 min at 94°C for denaturation of DNA, 2 min at 58°C for annealing of the primers, and 2 min of incubation at 72°C for elongation in a DNA engine thermal cycler (MJ Research, Reno, NV). PCR amplifications for αENaC and β-actin were performed side by side. The reactions were stopped after 19 and 14 cycles, respectively, when amplification was still in the exponential phase. Ten microliters of the reactions were run on 1% agarose gel. After electrophoresis, the gels were stained by ethidium bromide, and the amplified products were quantified by scanning with a Typhoon fluorescence scanner (Molecular Dynamics). The results were reported as percent of control expression after normalization with the β-actin signal. The experiments were repeated four times (n = 4) for every condition tested on alveolar epithelial cells purified from different animals.
Fluid transport across alveolar epithelial cell monolayers.
Fluid transport across alveolar monolayers was measured according to a method adapted from Fang et al. (26). Alveolar epithelial cells grown on Costar filters were cultured at the air-liquid interface (ALI) as reported previously (9). On day 4 of culture, 1.5 ml of medium containing 125I-albumin (0.5 μCi/ml) was added to the apical side. The cells were grown as usual in a humidified incubator at 37°C with 5% CO2. Two-hundred fifty microliters of the apical medium was taken after 5 min (t5) and 24 h (t24) of incubation. The samples were weighed and counted in a gamma counter (Cobra Auto-gamma; Packard, Downers Grove, IL). Fluid transport was calculated according to the following formula: fluid absorption (in %) = [1 − (CPM of t5 sample/weight of t5 sample)/(CPM of t24 sample/weight of t24 sample)] × 100, where CPM is counts per minute. The absorbed volume was calculated in microliters per square centimeter per hour and expressed as a percent of untreated controls for each series. Permeability to protein (Pp) was also estimated in each monolayer by collecting and counting the apical and basolateral media as well as the H2O wash of the monolayers. Pp was calculated with the following formula: Pp = (CPM of the basolateral compartment)/(CPM of the apical compartment) where CPM of the apical compartment was CPM of the t5 sample + CPM of the t24 sample + CPM of the remaining apical volume + CPM of the apical washing solution. Monolayers with Pp > 3% were discarded. Free 125I, measured in the presence of 20% trichloroacetic acid, was lower than 2.5%. Evaporative loss of fluid measured in a parafilm-sealed Transwell membrane was nil. In parallel with each experiment, the cells were treated with 100 ng/ml TNF, 100 nM Dex, 100 ng/ml TNF + 100 nM Dex, and 10 μM amiloride to also test the importance of amiloride-insensitive transport in the process. Filters of at least six animals (n ≥ 6) were tested for each condition.
The data are presented as means ± SE. Comparisons between different conditions were made by paired t-test, analysis of variance (ANOVA), and post hoc analysis (Fisher protected least significant difference) with Statview software (SAS Institute, Cary, NC). A probability of P < 0.05 was considered to be significant. The Dex response curve on αENaC promoter activity in A549 cells and the response of αENaC mRNA in alveolar epithelial cells were determined by nonlinear regression with Prism3 software, according to the following equation: Y = (B + Bmax) × X/(Kd + X), where Bmax is the Y maximum value, B is Y at X = 0, and Kd is X when Y = Bmax/2. The Dex concentration needed to overcome the decrease evoked by TNF was estimated by intrapolation and mathematical solving of the equation.
Dex inhibits the effect of TNF on ENaC expression.
αENaC mRNA expression was studied in primary alveolar epithelial cells treated for 6, 8, or 24 h with 100 ng/ml TNF in the presence and absence of 100 nM Dex. As shown previously (18), TNF gradually decreased αENaC mRNA expression to ∼30% of untreated cells at 24 h (Fig. 1). Dex raised, at every time point studied, the level of αENaC transcript to ∼4.5-fold of the untreated controls (Fig. 1). In cotreatment experiments, Dex in the presence of TNF was still able to increase αENaC transcript expression by 3.2-fold compared with the controls (Fig. 1). The β and γENaC mRNA expressions were also investigated at the 24-h time point. TNF downregulated the expression of both subunits to ∼30% and 9%, respectively (Fig. 1). Although the increase induced by Dex was smaller than for αENaC, in cotreatment experiments the expression of β and γENaC mRNA was maintained to 1.7-fold and 1.1-fold, respectively, compared with the controls (Fig. 1). αENaC mRNA expression was also tested in alveolar epithelial cells pretreated for 24 h with TNF before cotreatment with Dex for 6, 8, or 24 h. In such conditions, Dex was still able to heighten αENaC mRNA expression to 1.8-fold of untreated cells at 24 h (TNF:TNF+Dex) (Fig. 2). If TNF was removed after pretreatment and the cells were treated for 24 h with Dex alone, αENaC mRNA expression increased 2.34-fold compared with untreated controls (TNF:Dex) (Fig. 2). After 24 h of TNF pretreatment, Dex in the presence of TNF could not maintain β and γENaC mRNA expression since the two subunits were decreased to 47% and 14% of untreated controls, respectively (P < 0.05; TNF:TNF+Dex) (Fig. 2). Dex was also less effective in upregulating β and γENaC mRNA expression after removal of TNF (TNF:Dex) (Fig. 2).
Effect of TNF and Dex on αENaC promoter activity.
To study the effect of TNF on αENaC promoter activity in the presence and absence of Dex, A549 alveolar epithelial cells were transfected with αENaC-Luc, a construct where a 2.9-kb αENaC promoter fragment was cloned upstream of the firefly Luc reporter gene. TNF decreased Luc expression by ∼25% upon 24-h treatment (P < 0.05) (Fig. 3A). As reported previously (17), Dex induced a fivefold increase in Luc activity (P < 0.05) (Fig. 3A) but was reduced to 2.77-fold when the cells were exposed to TNF+Dex cotreatment (Fig. 3A). Despite such a decrease, Dex was still able to raise ENaC promoter activity above control cell values (P < 0.05) (Fig. 3A). Transfection of αENaC Δ GRE-Luc, a 1.17-kb deletion mutant not responding to Dex because of the absence of GRE (Fig. 3B), or the use of RU-486 (mifepristone), a steroid receptor antagonist (Fig. 3C), evoked a similar Luc expression pattern. In these transfections, Dex could not raise Luc activity (Fig. 3, B–C). Luc activity was still significantly decreased (P < 0.05) in TNF-treated cells (Fig. 3, B–C), suggesting that the ∼25% decline induced by TNF is steroid independent. The decrease in Dex stimulation by TNF did not involve inhibition of the steroid transduction pathway in A549 cells since the cytokine had no impact on the strong Dex response following transfection of pHH-Luc, a plasmid where Luc expression was driven from a MMTV promoter (Fig. 3D). Conversely, GR overexpression in A549 cells could not prevent TNF from decreasing the extent of Dex stimulation on Luc activity (data not presented).
Dex concentration required to reverse the effect of TNF on αENaC expression.
Dose response curves of ENaC promoter activity at different Dex concentrations are presented in Fig. 4A. Although Dex stimulation was markedly inhibited in the presence of TNF (P < 0.05), a Dex concentration above 1.35 nM was able to normalize ENaC promoter activity to the level of untreated controls. A similar dose response curve was seen when αENaC mRNA expression was tested in alveolar epithelial cells at different Dex concentrations in the presence or absence of TNF (Fig. 4B). Stimulation by Dex was significantly lower in the presence of TNF (P < 0.05). However, a 1.65-nM Dex concentration was sufficient to normalize αENaC mRNA expression in the presence of TNF (Fig. 4B).
Effect of Dex on the cytokine secretion of alveolar epithelial cells treated with TNF.
Dex is a known anti-inflammatory agent that downregulates the production of cytokines by its inhibition of NF-κB signaling (48). To test the nature of the proinflammatory cytokines secreted in TNF-treated cells and to determine if Dex could downregulate such secretion, the apical and basolateral media of cells treated with TNF were tested for their cytokine levels using a protein array approach. Although no cytokine could be detected in fresh, unconditioned medium, a few cytokines, including monocyte chemoattractant protein-1 (MCP-1), were found to be secreted in the apical conditioned medium from control alveolar epithelial cells (Table 1). TNF increased the secretion in apical medium of macrophage inflammatory protein-3α (MIP-3α) and cytokine-induced neutrophil chemoattractant-3 (CINC-3) compared with untreated cells (Table 1). Dex had no effect on the apical cytokine expression profile of TNF-conditioned medium (Table 1). The cytokine profile in basolateral-conditioned medium was slightly different, since several cytokines present at the apical side could not be detected (Table 1). Compared with untreated cells, there was an increase of MIP-3α, fractalkine, CINC-3, and LPS-induced C-X-C chemokine (LIX) with TNF treatment (Table 1). While Dex reduced fractalkine secretion and slightly decreased CINC-3 and LIX in the basolateral medium of TNF-treated cells, the steroid had no other significant effect on the cytokine profile of TNF-conditioned medium.
Amiloride-sensitive current at the apical membrane after Dex and TNF treatment.
The presence of active ENaC at the apical membrane was evaluated in an Ussing chamber after amphotericin B permeabilization of the basolateral membrane (Fig. 5). In the conditions tested, NMDG-Cl replaced ∼82% of Na+ in the basolateral solution, creating an apical-to-basolateral Na+ gradient. In these conditions, TNF decreased amiloride-sensitive current (ENaC current) to 22% of the controls (P < 0.05) (Fig. 6). In Dex-treated cells, ENaC current was increased to 198% (P < 0.05), whereas in TNF+Dex-treated cells, it was augmented to 143% (Fig. 6). Clearly, Dex treatment upregulates Na+ channel activity at the apical membrane and can overcome the reduction evoked by TNF.
Open circuit measurement and liquid clearance ability of TNF+Dex-treated cells.
The effects of TNF and Dex treatments were also evaluated in open circuit configuration to study the impact of these conditions on the transepithelial characteristics of the monolayers (Table 2). In control cells, 1 μM amiloride-sensitive current represents 57% of total Ite (Table 2). TNF treatment decreased amiloride-sensitive current to 27% of the controls (P < 0.05) (Fig. 7A) and also reduced Rte from 1,601 Ω/cm2 to 1,313 Ω/cm2 (P < 0.05) (Table 2). Although Dex treatment increased the current only very modestly to 116% of the controls (P < 0.05) (Fig. 7A), in TNF+Dex cotreatment it inhibited the effect of TNF, allowing the current to be maintained at 112% of the controls (P < 0.05) (Fig. 7A). Because transepithelial Na+ transport is the main force involved in liquid absorption in the distal lung, we also tested if these treatments could affect the liquid clearance ability of alveolar epithelial cells. Alveolar epithelial cells were cultured at the ALI, and liquid clearance was measured by the Fang et al. (26) method. Fluid transport was 0.678 μl·cm−2·h−1 in control cells (Fig. 7B) and was reduced to 0.349 μl·cm−2·h−1 (51%) after 24-h TNF treatment (P < 0.05). This is in the range (53%) of the fluid absorption measured in the presence of 10 μM amiloride (Fig. 7B). Dex had no impact on the fluid absorption rate; however, in cotreatment with TNF, it inhibited the decrease evoked by the cytokine (Fig. 7B). The permeability of alveolar epithelial cells was also investigated by measuring 125I-albumin leakage into the basolateral compartment. TNF treatment increased 125I-albumin permeability by 25% (P < 0.05) compared with untreated cells. Dex treatment had no impact on paracellular permeability and could not decrease the permeability induced by TNF (data not included).
In this study, we tested if Dex could inhibit the effect of TNF on ENaC expression and activity in alveolar epithelial cells. We found that Dex in cotreatment could indeed increase αENaC promoter activity and mRNA expression in TNF-treated cells and was able to normalize ENaC transepithelial current as well as the liquid clearance ability of the cell monolayers.
Dex at a 100-nM concentration could heighten αENaC mRNA expression by ∼4.5-fold in alveolar epithelial cells. This is in the range of what we reported previously (17). In Dex and TNF cotreatment, although TNF decreased the Dex response, αENaC mRNA expression was still 3.2-fold higher than in the untreated controls. When Dex was added after 24-h pretreatment with TNF, the steroid response was smaller but still 1.8-fold higher than in the controls. This indicates that even when TNF had time to promote the expression of proinflammatory factors, Dex was still able to reverse the effects of the cytokine on αENaC mRNA expression. The modulation of β and γENaC mRNA expression was also investigated in epithelial cells treated with TNF and Dex. Although we found Dex to be a less potent activator of β and γENaC mRNA expression compared with αENaC, in TNF and Dex cotreatment it could maintain βENaC mRNA expression to 1.71-fold of untreated controls while γENaC mRNA transcripts were normalized to 1.1-fold expression. When Dex was added after 24-h TNF pretreatment, the steroid was unable to normalize β and γENaC mRNA expression to the level of the untreated controls. βENaC mRNA expression was downregulated to 47% of the controls while γENaC mRNA expression was very low (14% of the controls). The low response of the β and γENaC subunits to this treatment, compared with αENaC, could be explained by a greater sensitivity to TNF treatment for these subunits and a smaller stimulation of their expression by Dex. Because the ratio between the three ENaC subunits could change, there could be a shift in the nature of Na+ channels present in the cells. Two different types of amiloride-sensitive Na+ channels have been detected in alveolar epithelial cells: ENaC, a 4-pS low-conductance, highly selective channel (HSC) for Na+ that is made of α-, β-, and γENaC subunits (41, 81), and a nonselective channel (NSC) with ∼20 pS conductance that does not discriminate between Na+ and K+ (28, 40, 84). Studies with antisense oligonucleotides against each of the three ENaC subunits have shown that only αENaC inhibition decreases the density of NSC (40, 41). The presence of NSCs and HSCs has been reported to be modulated by culture conditions of alveolar epithelial cells (41). While TNF, by affecting the modulation of the three ENaC subunits, could affect both types of channels, the low expression of the β- and γ-subunits after TNF pretreatment, even in the presence of Dex, suggests that in these conditions there could be a shift in alveolar epithelial cells from HSC to NSC. The consequences of such changes on the Na+ transport ability of the monolayers will need further investigation.
To establish if TNF was decreasing αENaC transcription and to determine how Dex could modulate this process, αENaC promoter activity was tested in A549 cells, an alveolar epithelial cell line with characteristics of alveolar type II cells (50). These cells were transfected with αENaC-Luc, a construct where the mouse αENaC promoter, including a GRE sensitive to the steroid, controlled Luc expression (31). TNF decreased αENaC promoter activity by ∼25%. This is a modest reduction compared with the 70% decline of αENaC mRNA evoked by the cytokine (18). These results show that a decrease of αENaC transcription is not a major mechanism allowing the strong diminution of αENaC transcript by TNF and indicate that posttranscriptional events could be involved. We reported previously that TNF reduces the half-life of αENaC mRNA (18). The strong decline of αENaC expression induced by TNF in these cells could, therefore, be a combined decrease in the transcription and stability of αENaC mRNA.
Some studies have described an effect of TNF on ENaC promoter activity in the distal colon, where TNF has been found to induce a significant reduction of β and γENaC promoter activity (2, 3). αENaC promoter activity was not investigated in these experiments because in the distal colon, TNF does not modulate αENaC expression (2, 3). Although we do not know if TNF modulated β and γENaC promoter activity in our cells, the modulation process involved seemed very different in the two systems, since there was strong transcriptional downregulation of the β and γENaC genes in the distal colon, whereas we found only weak inhibition of αENaC promoter activity in alveolar epithelial cells.
Dex upregulates ENaC expression via a very conserved GRE in the αENaC promoter (17, 63, 74). As with mRNA expression, Dex treatment increased Luc activity by fivefold in the cells, and this increase was sensitive to TNF. To test if the 25% decrease in promoter activity detected in TNF-treated cells could be explained by the inhibition of endogenous glucocorticoid stimulation, the same experimental conditions were repeated in the presence of RU-486, a steroid receptor antagonist, or with transfection of αENaC Δ GRE-Luc, a GRE deletion mutant of the αENaC promoter. In both cases, Dex could not stimulate αENaC promoter activity any longer. Still, Luc activity was decreased when the cells were treated with TNF alone. Together, these data reveal that TNF modulates ENaC transcription via two independent pathways, a steroid-independent pathway when cells are treated with TNF alone, and partial inhibition of steroid action with TNF+Dex exposure.
Different mechanisms were investigated in this work to explain how TNF could decrease the stimulation by Dex of αENaC promoter activity. The ERK pathway was the first to be examined since it was known to inhibit the Dex action on αENaC promoter activity after Ras activation (51), TGF-β1 treatment (31), or oxidative stress (82). We found that PD-98059, an ERK pathway inhibitor, could not suppress the effect of TNF on Dex stimulation (data not included). It has been demonstrated that the GR undergoes direct protein-protein interaction with NF-κB or AP-1 in the cytoplasm, a mechanism that could repress the steroid response (35, 67, 75, 83). For this reason, we tested if GR overexpression could alleviate the effect of TNF on αENaC promoter activity in Dex-treated cells. Cotransfection of a plasmid coding for the GR along with αENaC-Luc had no impact on Luc activity. TNF could still decrease the Dex response even in cells overexpressing the GR (data not presented). We, therefore, tested if the effect of TNF on Dex stimulation was the consequence of transrepression of the steroid-transducing pathway by TNF in alveolar epithelial cells or was something specific to the αENaC promoter. The cells were transfected with pHH-Luc, a plasmid where the Luc reporter gene was regulated by the MMTV promoter, a system very sensitive to Dex stimulation (38, 57). Whereas Dex allowed strong Luc expression upon pHH-Luc transfection, TNF in cotreatment could not inhibit the steroid response. This result shows that TNF does not block the steroid transduction pathway in A549 cells. The partial repression of the steroid response by TNF for αENaC promoter activity is thus not related to a lack of steroid stimulation in these cells, but is rather a specific response of the αENaC promoter.
A dose response curve to Dex in presence or absence of TNF was performed to determine how TNF modulates the αENaC promoter response to Dex. A similar dose response curve was also tested in alveolar epithelial cells to study how the treatment affects αENaC mRNA level. Dex induced a dose-dependent increase of Luc activity and of αENaC transcript in cells. The dose response curves to Dex showed a similar profile for αENaC promoter activity and αENaC transcript level. It confirms the results reported elsewhere (46, 74, 80) that Dex is an important modulator of αENaC expression. It also confirms that the main effect of Dex on αENaC expression is evoked by an increase in transcriptional activity of the promoter. In the presence of TNF, however, stimulation of αENaC promoter activity and mRNA level by Dex was markedly reduced at all concentrations tested. Interestingly, a 1.35-nM concentration normalized αENaC promoter transcriptional activity, and a 1.65-nM concentration normalized the αENaC transcript level in TNF-treated alveolar cells. The work by Renard et al. (68) has shown that the lung is a tissue that is very sensitive to steroid action. The results presented here confirm that Dex has a strong impact on αENaC promoter transcriptional activity and mRNA level in alveolar epithelial cells. In addition, they reveal that a low level of Dex is needed to inhibit the negative effect of TNF on αENaC expression.
Several groups have reported that Dex is a strong activator of ENaC expression in alveolar epithelial cells (14, 17, 39, 46, 74). The Northern blots and αENaC promoter study included here confirm this point. At the same time, Dex has important anti-inflammatory properties as it has been found to block the TNF stimulation of proinflammatory genes in the lung by its inhibition of the NF-κB pathway (49, 66). Using a protein array, we tested if Dex could modulate the cytokine profile of TNF-treated cells. We observed that with a few exceptions, Dex did not significantly affect the cytokine profile in the conditioned medium of TNF-treated cells. The protective action of Dex on ENaC expression in TNF-treated cells does not seem, therefore, to rely on the anti-inflammatory action of the steroid, but rather on its activation of ENaC synthesis. Dex has been found to suppress the surfactant protein B (SP-B) mRNA downregulation induced by TNF in alveolar epithelial cells (65) as well as the inhibition of cadherin and catenins by the cytokine in bronchial epithelial cells (13). These findings are similar to what we reported here for ENaC. They suggest that in TNF-treated cells, Dex reverses the effects of TNF on genes that are not involved in inflammatory response but are important for lung epithelial functions.
In addition to its modulation of αENaC promoter activity and ENaC mRNA expression in TNF-treated cells, we tested how Dex could alleviate the effect of the cytokine on the Na+ transport abilities of alveolar epithelial cells. The proportion of active ENaC at the apical membrane was evaluated in Ussing chamber after amphotericin B permeabilization of the basolateral membrane in the presence of a Na+ gradient. In these conditions, we found that TNF reduced amiloride-sensitive Isc at the membrane by ∼78%. We reported previously that TNF had no significant impact on Na+-K+-ATPase activity (18). The results presented here confirm that TNF greatly reduces the ENaC current at the apical membrane. Although the amount of membrane-associated ENaC subunits was not measured, the amphotericin B experiment allowed us to determine the current generated by ENaC at the apical membrane. TNF, by decreasing ENaC protein and mRNA expression (18), could affect the amount of active channel at the membrane. The cytokine could also affect insertion and retrieval of the channel at the membrane as well as the activation of silent channels. Although the experimental design used in permeabilization experiments does not allow us to determine the exact mechanism leading to such decrease, it is possible that low expression of ENaC at the mRNA and protein levels could be involved in the reduction of ENaC current in TNF-treated cells.
Dex augmented ENaC activity at the apical membrane by ∼2-fold compared with untreated cells. This increase was smaller than the rise in ENaC mRNA, probably because the number of active channels at the membrane is not only the result of transcription, but also of translation (62, 64) as well as membrane insertion and retrieval of ENaC (70). In addition, ENaC activity at the membrane depends on other factors, such as CAP-1, a protease that activates ENaC (71, 79). The effect of Dex on the protease activity is unknown, however. Dex is known to increase ENaC activity by promoting the synthesis of new channels and also by its activation of sgk, a kinase that has been shown to inhibit ENaC retrieval and degradation (39) to activate silent ENaC at the membrane (20) and to augment ENaC transcription (7). In cotreatment with TNF, Dex was able to increase ENaC activity at the membrane to 143%. The decrease of ENaC current at the apical membrane in TNF-treated cells was thus corrected in presence of Dex. As discussed above, this could be related to the compensatory effect of Dex on ENaC current generated at the apical membrane. These data demonstrate that, as with ENaC mRNA, TNF and Dex have opposite effects on ENaC activity at the membrane.
The impact of TNF and Dex on amiloride-sensitive transepithelial Na+ transport in alveolar epithelial cells was also measured. First, there was an 18% decrease in the Rte of TNF-treated cells. This has been reported in alveolar epithelial cells treated with TNF (85) as well as in human airway cells (16) and could reflect a small modification in the barrier function of the cells under proinflammatory conditions. ENaC amiloride-sensitive Ite was decreased by 73% in TNF-treated cells, similar to what we reported previously (18), and in accordance with the permeabilization results presented above. In TNF-treated cells, because much less active channels occur at the membrane, ENaC transepithelial Na+ transport is decreased. This suggests that in alveolar epithelial cells, a reduction of ENaC is effectively limiting Na+ transport. In TNF+Dex cotreatment, there was no increase of current, but the steroid was able to maintain the current in the normal range. The data presented suggest that amiloride-sensitive current was maintained within the normal range because steroid treatment was increasing the amount of active ENaC at the membrane as reported also in the amphotericin B permeabilization. However, there was no further increase in current, probably because there was no increased expression of γENaC mRNA compared with α and βENaC transcripts in TNF+Dex-treated cells (Fig. 1).
In apparent contradiction to the amphotericin B data, however, the effect of Dex treatment, although significant (P < 0.05), was very modest, with only a 116% increase of ENaC current compared with the untreated controls. The Na+ transport process in alveolar epithelial cells is the result of entry of Na+ at the apical membrane mainly through ENaC and Na+ extrusion by Na+-K+-ATPase at the basolateral side (5, 73). When a normal amount of active ENaC is present at the membrane, however, the basolateral membrane could behave as a rate-limiting step for transepithelial Na+ transport. This could explain why in Dex-treated cells, although amphotericin B showed increased ENaC current at the apical membrane, the treatment had no significant impact on Na+ transepithelial current. This lack of response to Dex could be potentially explained by a lack of stimulation of the different channels, cotransporters, or the Na+-K+-ATPase expressed at the basolateral membrane (60) that are essential to Na+ absorption. Na+-K+-ATPase activity is important for Na+ absorption in alveolar epithelial cells. Increasing Na+-K+-ATPase subunit expression by gene targeting has been shown to raise the amiloride-sensitive current in alveolar cells (24, 25, 78) and enhances lung liquid clearance (24, 76). Dex has been shown to augment Na+-K+-ATPase activity in alveolar epithelial cells via heightened expression of the β1-subunit (4). Ouabain-sensitive K+ uptake is also increased by Dex at 12 h but not at 24 h (4). The absence of Dex stimulation on Na+ current reported here could, therefore, also be related to the fact that the measurements were taken after 24 h of treatment. This lack of response to Dex could also be secondary to the lack of stimulation of other channels, like K+ channels, which are important for Na+ extrusion at the basolateral membrane. In fact, we reported recently that inhibition of KATP, KvLQT1, and KCa channels, present at the basolateral membrane, greatly downregulated amiloride-sensitive transepithelial Na+ transport (47). Although Dex had a limited impact on the transepithelial Na+ transport, the treatment was effective to normalize amiloride-sensitive Na+ current in alveolar cells treated with TNF.
Na+ transport is essential for the process of liquid absorption in alveolar epithelial cells. For this reason, we tested if TNF and Dex could modulate the fluid clearance ability of alveolar cells, using an in vitro model developed recently (26). The basal fluid transport rate we found (0.678 μl·cm−2·h−1) was in the range of what has been reported (26). Amiloride inhibited 47% of fluid clearance, showing that ENaC is involved in this process. The amiloride-sensitive/insensitive ratio of lung liquid clearance is in the range of what has been presented in Table 2 for the ratio of amiloride-sensitive/insensitive Ite. It is also in the range of what has been obtained for the amiloride-sensitive/insensitive ratio of rat lung liquid clearance (42, 59). It reveals that amiloride-insensitive channels are also involved in the liquid clearance process. Besides ENaC, several amiloride-insensitive channels have been detected in alveolar cells such as the cGMP-gated channel (CNG), Na-glucose, and other Na transporters (44, 58). CNG-1 channel expression has been detected in the lungs (21), and channel activity was recorded by patch clamp in alveolar cells (45). This type of channel could also be involved in fluid absorption as it has been demonstrated to be important for lung liquid clearance in sheep (43) and in terbutaline- or DBcGMP-stimulated clearance in rat lungs (59). TNF decreased the liquid clearance ability of the monolayers to the same extent as did amiloride, with a 49% decline. As with ENaC expression and activity, the cytokine had a significant impact on the fluid clearance ability of the cells. Whereas Dex alone did not affect fluid clearance, in cotreatment with TNF, it could inhibit the decrease evoked by the cytokine.
TNF treatment also induces a 25% increment of paracellular permeability. This increase is in agreement, and of similar magnitude, with the drop in Rte reported above. It suggests that besides altering transepithelial Na+ transport, TNF could modulate the permeability properties of alveolar epithelial cells at the tight junctions. In human airways, it has been suggested that the change in paracellular permeability, and in tight junctions brought by TNF and IFN-γ, could be involved in facilitation of neutrophil passage across epithelial cells (16). TNF clearly modulates the ionic transport and permeability properties of alveolar epithelial cells.
There are some conflicting results concerning how TNF modulates lung liquid clearance. Whereas in bacterial pneumonia (69) and intestinal ischemia-reperfusion injury (6) TNF increases lung liquid clearance by a TNF-dependent mechanism, our previous report (18) as well as the work presented here show that chronic stimulation with TNF decreases ENaC expression and Na+ transport ability in alveolar epithelial cells. Although the time of TNF stimulation, acute vs. chronic, could explain these differences in part, recent reports suggest that TNF could promote both activities. Whereas the lectin-binding domain of TNF (Ltip) has been demonstrated to promote increased lung liquid clearance in tumor necrosis factor receptor I (TNF-RI) and TNF-RII KO mice (23) and in the rat lung (8), TNF instilled in the rat lung has been found recently to decrease lung liquid clearance (8). In this latter study, inactivation of the TNF receptor-binding domain by soluble TNF-RI or inactivating antibodies enabled TNF, via its Ltip domain, to increase lung liquid clearance (8). It has been shown, by patch-clamp experiments in A549 cells, that infusion of TNF was able to increase amiloride-sensitive current within a few minutes (32). We tried to determine if TNF as acute treatment could heighten Isc of alveolar epithelial cells. After TNF perfusion, there was no modulation of the Ite detected, except for a 27% decrease after 60 min of treatment (P < 0.05) (data not shown). The different electrophysiological methods used as well as the different cell types could explain this divergence. In A549 cells, for instance, basal ENaC current is very low (32) compared with alveolar epithelial cells where it is the principal current detected. This could explain why TNF seems to be less active in its stimulating activities in alveolar epithelial cells.
In summary, the data presented here reveal that Dex and TNF have opposite effects on ENaC expression and activity. Dex increases the amount of α, β, and γENaC transcripts and reverses the effect of TNF on ENaC current and the fluid absorption ability of cultured alveolar epithelial cells. We found that Dex is a strong activator of αENaC promoter activity but had a limited ability to reduce in alveolar epithelial cells the proinflammatory cytokines induced by TNF. Alveolar epithelial cells are very sensitive to Dex, since low concentrations are needed to normalize αENaC promoter activity and transcript level in the presence of TNF. The results presented here disclose that besides its anti-inflammatory properties, Dex could also be beneficial for its impact on ENaC expression and activity when alveolar epithelial cells are submitted to a proinflammatory cytokine such as TNF. Another interesting observation from our study is that the level of ENaC mRNA expression in cells cotreated with TNF and Dex seems to be the result of an antagonistic balance between Dex upregulation and TNF downregulation of the transcript level. These data indicate that different pathways acting simultaneously, even in opposite directions, can regulate ENaC transcript level and allow very fine tuning of ENaC expression and activity.
E. Brochiero is a scholar of Fonds de la Recherche en Santé du Québec (FRSQ). R. Fréchette was supported by a studentship from the Canadian Cystic Fibrosis Foundation, and M.-E. Clermont was the recipient of a studentship from the FRSQ. This work was funded in part by the Canadian Cystic Fibrosis Foundation and the Canadian Institutes of Health Research.
We acknowledge the editorial work done on this manuscript by Ovid Da Silva, Research Support Office, Research Centre, Centre hospitalier de l'Université de Montréal (CHUM).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society