Regulation of epithelial sodium channels by cGMP/PKGII

Abstract

Airway and alveolar fluid clearance is mainly governed by vectorial salt movement via apically located rate-limiting Na⁺ channels (ENaC) and basolateral Na⁺/K⁺-ATPases. ENaC is regulated by a spectrum of protein kinases, i.e. protein kinase A (PKA), C (PKC), and G (PKG). However, the molecular mechanisms for the regulation of ENaC by cGMP/PKG remain to be elucidated. In the present study, we studied the pharmacological responses of native epithelial Na⁺ channels in human Clara cells and human αβγδ ENaCs expressed in oocytes to cGMP. 8-pCPT-cGMP increased amiloride-sensitive short-circuit current (I_sc) across H441 monolayers and heterologously expressed αβγδ ENaC activity in a dose-dependent manner. Similarly, 8-pCPT-cGMP (a PKGII activator) but not 8-Br-cGMP (a PKGI activator) increased amiloride-sensitive whole cell currents in H441 cells in the presence of CFTRinh-172 and diltiazem. In all cases, the cGMP-activated Na⁺ channel activity was inhibited by Rp-8-pCPT-cGMP, a specific PKGII inhibitor. This was substantiated by the evidence that PKGII was the sole isoform expressed in H441 cells at the protein level. Importantly, intratracheal instillation of 8-pCPT-cGMP in BALB/c mice increased amiloride-sensitive alveolar fluid clearance by ∼30%, consistent with the in vitro results. We therefore conclude that PKGII is an activator of lung epithelial Na⁺ channels, which may expedite the resolution of oedematous fluid in alveolar sacs.

Expression of epithelial sodium channels (ENaCs) has been detected both biochemically and physiologically in airway and alveolar epithelial cells. In the lung, apical ENaCs, combined with basolaterally located Na⁺/K⁺-ATPase, form a major pathway for the vectorial transport of salt and fluid across the epithelial layer of the lung gas–blood barrier. It has clearly been demonstrated by the fundamental studies that the deletion of α, β and γ ENaC subunits delays and strikingly reduces fluid clearance from the airspace of knockout mice at birth (Hummler et al. 1996; Barker et al. 1998; McDonald et al. 1999; Rossier et al. 2002). On the other hand, a decrease in ENaC expression and activity following influenza virus, respiratory syncytial virus (RSV), and other respiratory virus infections is associated with airspace flooding (Kunzelmann et al. 2000; Chen et al. 2004; Davis et al. 2007). These pioneer preclinical studies confirm that Na⁺ transport via ENaC is the rate-limiting process responsible for maintaining the airspace free of edoematous fluid (Matthay et al. 1982, 2002). Indeed, several phenotypes of pulmonary oedema, for instance, high altitude pulmonary oedema and cardiogenic oedema, together with the aforementioned viral infection-induced oedematous lung injuries, are associated with impaired fluid re-absorption via ENaC.

There has long been interest in the regulation of lung ion transport by second messengers, i.e. cAMP, Ca²⁺ and cGMP. Specifically, a broad spectrum of hormones, neurotransmitters and toxic peptides of pathogens elevates cytosolic cGMP by releasing natriuretic peptides (e.g. guanylins) and nitric oxides (NO), which activates soluble gyanylyl cyclase. The fundamental contribution of physiological levels of NO to lung ENaC activity has well been established using iNOS specific inhibitors and NOS knockout mice (Hardiman et al. 2001, 2004). These well-designed studies demonstrate that the maintenance of amiloride-sensitive Na⁺ channel activity requires a physiological range of NO levels. However, NO interacts with superoxide and thus forms other adducts and reactive species, resulting in variable effects on sodium transport in epithelial cells. This may partially explain the inconsistent observations about the regulation of ENaC by NO donors even in the same model-A549 cells (Xu et al. 1999; Kamosinska et al. 2000; Lazrak et al. 2000).

While increasing evidence indicates that both native and heterologously expressed ENaCs are regulated by protein kinase A, protein kinase C and other kinases, there are no detailed studies of the regulation of lung epithelial ENaCs by cGMP/PKG. This prompted us to explore the specific downstream mechanisms of NO/cGMP/PKG signal pathway that modulate ENaC activity in lung epithelial cells. To address this issue, we applied PDE-resistant, cell permeant cGMP analogues to specifically activate either PKGI or PKGII in human Clara H441 cells. Indeed, ENaC activity was activated by 8-pCPT-cGMP (an activator of PKGII) in a dose-dependent manner. In contrast, a cGMP analogue activating PKGI (8-Br-cGMP) had no effect. These results provide an intrinsic mechanism underlying the up-regulation of transepithelial salt transport by guanosine nucleotides.

Methods

Cell culture

NCI-H441 (H441) cells were obtained from the American Type Culture Collection (ATCC), and grown in RPMI medium (ATCC) containing 10% fetal bovine serum (FBS), 2 mm l-glutamine, 10 mm Hepes, 1 mm sodium pyruvate, 4.5 g l⁻¹ glucose, 1.5 g l⁻¹ sodium bicarbonate and antibiotics (100 U ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin). Dexamethasone at 250 nm was applied to the culture medium to stimulate ENaC expression after splitting on coverslips for patch clamp and on permeable support for Ussing chamber studies. Cells were seeded on permeable support filters (Costar Snapwell culture cup) at a supraconfluent density (∼3 × 10⁶ cells cm⁻²), and incubated in a humidified atmosphere of 5% CO₂–95% O₂ at 37°C. When cells reached confluency (24 h after plating), medium and non-adherent cells in the apical compartment were removed. Air–liquid interface culture was then utilized to obtain highly polarized tight monolayers. Culture medium in the basolateral compartment was replaced every other day and the apical surface was rinsed with phosphate-buffered saline (PBS). The transepithelial resistance was measured with an epithelial tissue volt-ohm-meter (World Precision Instruments, Sarasota, FL, USA) before the culture medium was changed. The confluent filters with a resistance >500 Ω cm², which took 2–3 weeks’ culture, were used for Ussing chamber assays. For patch clamp studies, H441 cells were plated on coverslips at a low density and grown for 24–72 h.

Two-electrode voltage clamp studies

Preparation of ENaC cRNAs for human α, β, γ, and δ subunits were conducted as previously described (Ji et al. 2001). Injection of ENaC cRNAs into defolliculated oocytes was performed with a Nanolitre microinjector (World Precision Instruments). Whole-cell currents were recorded as described previously (Ji et al. 2006). Cyclic nucleotides were purchased from AXXORA LLC, San Diago, CA, USA. Other chemicals were from Sigma-Aldrich.

Patch clamp assay

H441 cells grown on coverslips were mounted on the stage of a Leica DMIRB inverted fluorescence microscope (Meyer Instruments Inc., Houston, TX, USA) and continuously perfused with an external solution containing (in mm): 145 sodium gluconate, 2.7 KCl, 1.8 CaCl₂, 2 MgCl₂, 5.5 glucose, and 10 Hepes (pH 7.4, 300 mosmol kg⁻¹). Pipettes were back-filled with internal solution containing (in mm): 135 potassium gluconate, 10 KCl, 6 NaCl, 2 MgCl₂, 2 Mg₂ATP, 2 Na₂GTP and 10 Hepes (pH 7.2, 296 mosmol kg⁻¹). The pipette resistance was 5–10 MΩ when filled with the internal solution. When the whole-cell access was achieved, membrane potential was voltage clamped with an Axopatch 200B patch-clamp amplifier (Molecular Devices, Sunnyvale, CA, USA). Currents were digitized with Digidata 1440A (Molecular Device), filtered through an internal four-pole Bessel filter at 1 kHz, and sampled at 2 kHz. Inward and outward whole-cell currents were monitored by a step-pulse protocol every 10 s before and after perfusion with cGMP and other drugs.

Short-circuit current recordings in H441 monolayers

Measurements of short-circuit current (I_sc) and transepithelial resistance (R_t) in H441 monolayers were performed as described previously (Chen et al. 2006). Briefly, filters containing H441 monolayers were mounted in Ussing chambers (Physiologic Instruments Inc., San Diego, CA, USA) and bathed on both sides with solutions containing (in mm): 120 NaCl, 25 NaHCO₃, 3.3 KH₂PO₄, 0.83 K₂HPO₄, 1.2 CaCl₂, 1.2 MgCl₂, 10 Hepes acid, and either 10 mannitol in the apical compartment or 10 glucose in the basolateral compartment. Osmolality of all solutions, as measured by a freezing depression osmometer (Wescor Inc., Logan, UT, USA), was between 290 and 300 mosmol kg⁻¹. The bath solutions were stirred vigorously by continuous bubbling with 95% O₂ and 5% CO₂ at 37°C (pH 7.4). Monolayers were short-circuited to 0 mV, and I_sc was measured with an epithelial voltage clamp (VCC-MC8, Physiologic Instruments). A 10 mV pulse of 1 s duration was imposed every 10 s to monitor R_t. Data were collected using the Acquire and Analyse program (v. 2.3, Physiologic Instruments). The temperature of vertical Ussing diffusion chambers (37°C) was maintained by using a temperature-controlled water bath for at least 30 min before analysis. Following establishment of steady-state values of I_sc and R_t, cGMP and specific channel inhibitors were added to the apical compartment, and I_sc and R_t were measured continuously until new steady-state values were reached.

To determine whether cGMP elicited an increase in Na⁺ transport across the apical membrane, we established a 145 : 25 mm Na⁺ ionic gradient (apical to basolateral) across monolayers by replacing the 120 mm NaCl with equal-molar N-methyl-d-glucamine, an impermeant cation in the basolateral chamber. After basal I_sc level was recorded, 100 μm amphotericin B was added to the basolateral side of the Ussing chamber. Permeabilized monolayers equilibrate intracellular Na⁺ concentration to that of the bath solution in the basolateral compartment. Under these experimental conditions, I_sc results from a passive electrogenic Na⁺ influx through the apical Na⁺ conductive pathways down the Na⁺ concentration gradient (Guo et al. 1998; Thome et al. 2003). When the I_sc had attained its maximum value following application of cGMP, the amiloride-sensitive current component was determined by adding 100 μm amiloride into the apical compartment.

Reverse transcription-PCR

The use of frozen human lung tissues and isolation of human alveolar type II cells were approved by the Institutional Review Boards of the University of Alabama at Birmingham (UAB) and the University of California–San Francisco (UCSF). To analyse PKG mRNA expression, total RNA was isolated from H441 cells, human lung tissues (Human Tissue Procurement, UAB), human alveolar type II cells, and rabbit and mouse lung tissues using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. The RNA concentration was assessed by spectrophotometry. The ratio of 260 nm/280 nm absorbance was above 1.95. RNA quality was also verified by denatured agarose gel electrophoresis. First-strand cDNA synthesis was performed using random hexamers and SuperScript II RT (Invitrogen). PCR was performed on a mastercycler gradient thermocycler (Eppendorf, Westbury, NY, USA). Primers for PKG isoforms are 5′-AAGATTCTCATGCTCAAGGAGGA-3′ and 5′AGGTCGTGGAAGGACCTG-3′ for PKG1α (NCBI access number: Z92867); 5′-CAAGACGCGGAGCAGCGG-3′ and 5′-GTAGAAGGGCAGGGTCAC-3′ for PKG1β (Z92868); 5′-TGGAGGCCTGCTTAGGTG-3′ and 5′-TTTCCCAGCCTTTCTGTT-3′ for PKGII (NM006259).

Two microlitres of first strand cDNA product was used as template in each 50 μl PCR reaction. PCR conditions were optimized using the PCR Optimizer™ kit (Invitrogen). The cycling parameters consisted of one cycle of 95°C for 3 min and then 30 cycles of 95°C for 30 s, 52–62°C (depending on the primers) for 30 s and 72°C for 30 s followed by a single 10 min extension period at 72°C. RT-PCR products were subsequently separated on a 1.5% agarose gel using 100 bp markers (Promega Corp., Madison, WI, USA) as a standard to identify the product size. PCR products were purified using QIAquick gel extraction kit (Qiagen, Valencia, CA, USA) and subcloned into the pCR-2.1 vector (Invitrogen). Positive clones were selected by blue/white screening and followed by digestion with EcoRI (Promega) to verify incorporation of the correct size insert. Identity of the PCR products was confirmed by sequencing (DNA Sequencing Core, UAB).

Immunoassays of δ ENaC expression

Human lung tissue slides (Human Tissue Procurement, UAB) were fixed in 3% buffered formaldehyde (Electron Microscopy Sciences, Hatfield, PA, USA) for 45 min; cells were permeabilized in 0.5% Triton X-100 in PBS for 3 min and non-specific protein binding sites were blocked using 5% BSA in PBS for 30 min. Anti-δ ENaC (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) was used at 5∼20 μg ml⁻¹ for 2 h at room temperature to detect ENaC proteins. Cells incubated with non-immune IgG serve as a control. Nuclei were stained with Vectashield-DAPI (Vector Laboratories, Inc., Burlingame, CA, USA). Images were captured on an Olympus IX170 (Olympus) inverted epifluorescence microscope using SenSys-cooled charge-coupled high-resolution camera (Photometrics, Tucson, AZ, USA). Images were finally processed and assembled using Adobe Photoshop.

Immunoblotting assays

For detection of PKGI and II, homogenized proteins were dissolved in a solution containing (in mm): 1% Triton X-100, 150 NaCl, 2 EDTA, 2 EGTA, 50 Tris, pH 7.4 (adjusted with HCl), and complete protease inhibitors (Roche Applied Science). Protein concentration was assayed by the Lowry method. Thirty to 60 μg of protein was redissolved for SDS-PAGE on a 7.5% acrylamide gel. Twenty micrograms of mouse brain tissue extract (Stressgen, Ann Arbor, MI, USA) and 0.075 μg rat protein kinase G II purified protein expressed in Sf9 cells (Sigma) were used as positive controls for PKGI and II, respectively. Proteins were subsequently transferred electrophoretically onto PVDF membranes (Millipore), blocked with 0.2% I-blocker (Tropix, Applied Biosystem, Foster City, CA, USA) and probed with anti-PKGI (Stressgen, 1 : 2 000) or anti-PKGII antibodies (Santa Cruz, 1 : 2 000). Blots were then incubated with horseradish peroxidase conjugated goat anti-rabbit secondary antibody (1 : 10 000; Pierce Biotechnology, Inc., Rockford, IL, USA), and visualized by enhanced chemiluminescence detection (Pierce).

In vivo alveolar fluid clearance

Animals were kept under pathogen-free conditions, and all procedures performed were approved by the Institutional Animal Care and Use Committee of the University of Texas Health Science Center at Tyler and the University of Alabama at Birmingham. Alveolar fluid clearance (AFC) was measured as previously described (Hardiman et al. 2004). In brief, BALB/c mice (NCI) were anaesthetized with diazepam (17.5 mg kg⁻¹, intraperitoneally) followed 6 min later by ketamine (450 mg kg⁻¹, intraperitoneally) and were placed on a heating pad (Braintree Scientific, Inc., Braintree, MA, USA). The trachea was exposed and cannulated with a trimmed 18-gauge intravenous catheter, which was then connected to a mouse respirator (model 687; Harvard Apparatus, Holliston, MA, USA). Mice were ventilated with 100% O₂ with a 200 μl tidal volume (8–10 ml kg⁻¹) at a 160 breaths min⁻¹. Once stable anaesthesia was obtained, mice were positioned in the left decubitus position, and 300 μl of isosmolar NaCl containing 5% fatty acid-free bovine serum albumin (BSA) was instilled via the tracheal cannula, followed by 100 μl of room air to clear dead space. After instillation, mice were ventilated for a 15 min period, and then the alveolar fluid was aspirated. AFC was calculated from the ratio between the protein concentration of the instillate before instillation and of the alveolar sample. All reagents were added to the AFC instillate from stock solutions directly before instillation, in a minimal volume of solvent (1–10 μl ml⁻¹). Mice were killed with a dose of ketamine (425 mg kg⁻¹) and xylazine (75 mg kg⁻¹) and then exsanguinated.

Data analysis

Electrophysiological data from patch clamp and Ussing chamber studies were primarily analysed with the Clampfit 10.1 (Molecular Devices) and Acquire and Analyse 2.3 (Physiological Instruments), respectively. Furthermore, the measurements were imported into OriginPro 8.0 (OriginLab Corp., Northampton, MA, USA) for statistically computation and graphic plotting. The EC₅₀ values of cGMP activation were calculated by fitting the dose–response curves with the Hill equation. The whole-cell currents between each single H441 cell and each laboratory were divergent due to the variant cell size (capacitance), passage, and culture reagents. The whole-cell channel activity was thus presented as amiloride-sensitive whole-cell current density (pA pF⁻¹) or normalized currents to the one at −120 mV of control group.

All results are presented as mean ±s.e.m. Student's paired-sample t-test was used to analyse the difference in the paired data obtained before and after addition of drugs. Both null (mean₁– mean₂= 0) and alternative hypotheses (mean₁– mean₂ <> 0) were adapted to compare the average current levels. Otherwise, one-way ANOVA computations with either Turkey's (for pairwise data) or Bonferroni's test (for unpairwise data) was used for means comparison. A probability level of <0.05 was considered statistically significant. The power analyses of sample sizes were simultaneously performed with this significant level and the actual power >0.95 was accepted.

Results

Cyclic GMP improves alveolar fluid clearance in vivo

Cyclic GMP has been reported to alter transepithelial salt/fluid transport by up-regulating cystic fibrosis transmembrane conductance regulator (CFTR) (Fang et al. 2006) and amiloride-sensitive cation channels (Schwiebert et al. 1997; Kemp et al. 2001; Norlin et al. 2001) in respiratory epithelial cells. To examine the potential effects of cGMP on alveolar fluid clearance in vivo, mice were intratracheally instillated with a 5% BSA solution with or without 1 mm 8-pCPT-cGMP, a concentration used by Jain and colleagues to study the effects of Br-cGMP on native ENaC activity in primary rat alveolar type II cells (Jain et al. 1998). As shown in Fig. 1, cGMP increased the total absorption rate of the instilled BSA solution by approximately 34.8% in 15 min (19.8 ± 1.1%vs. 14.7 ± 0.9% for control, P= 0.001, one-way ANOVA, n= 16) compared with the control group. The amiloride-sensitive AFC fraction, which is contributed by apical Na⁺ channels, pronouncedly increased by 46.8% in cGMP-instilled lungs. These in vivo results suggest that cGMP may up-regulate ENaC activity and accelerate fluid re-absorption in distal lung airspace under physiological conditions. A new question raised from these preclinical studies is what is the molecular basis of the cGMP-activated cation channels?

Figure 1. Up-regulation of mouse alveolar fluid clearance (AFC) in vivo by 8-pCPT-cGMP.

Anaesthetized, ventilated mice were intratracheally instilled with 5% BSA and the instillate was collected after 15 min. Reabsorption rate of instillate was computed as the percentage of instilled volume (% AFC₁₅). Data are presented as means ±s.e.m. Numbers within the brackets are total mice examined for control, amiloride (1 mm), 8-pCPT-cGMP (cGMP, 1 mm), and cGMP+amiloride (+amiloride). P values were computed with one-way ANOVA.

Expression of δ ENaC in human, rabbit, and mouse lung tissues and primary human alveolar type II cells

In addition to well-characterized α, β and γ ENaC subunits, another ENaC subunit previously identified only in mammalian tissues, namely, δ ENaC has recently been found to be expressed in H441 cells, a Clara cell line of human origin (Ji et al. 2006). To address the aforementioned question, we first examined the expression of δ ENaC in human, rabbit, and mouse lung tissues as well as primary human alveolar type II cells (hAT2) at the mRNA level. Indeed, two mRNA products for δ ENaC were detected in hAT2 cells and human lung tissues: 419 bp and 325 bp (Fig. 2A). These two products encode a δ ENaC splicing variant with a longer amino terminal tail (714 amino acids) and the originally reported δ ENaC (638 amino acids), respectively. Interestingly, the mRNA encoding the 419 bp δ ENaC splicing variant was abundantly expressed in H441 cells and rabbit and mouse lung tissues.

Figure 2. Expression of ENaC δ subunit in lung epithelial tissues.

A, reverse transcription-PCR analysis of δ ENaC mRNA expression in human lung. This RT-PCR represents 3 independent experiments with similar results. H441 cells were used as a positive control (Ji et al. 2006). Lanes from left to right represent human alveolar type II cells (hATII), human lung tissue (HLT), H441 cells, rabbit lung tissue (RLT), and mouse lung tissue (MLT). B, indirect immunofluorescence detection of δ ENaC in human lung tissue. Human lung sections were incubated with specific antibody against δ ENaC. ENaC δ protein was detected (green channel) with FITC conjugated secondary antibody. Left panel is a negative control incubated with non-immune IgG substituted for the anti-δ ENaC antibody (control). Cell nuclei were stained with DAPI (blue channel). Bar equals 20 μm; n= 5. C, immunoblotting assay of δ ENaC protein. A polypeptide of approximately 85 kDa was detected in the lysate of human lung tissue (HLT). H441 cell lysate serves as a positive control (Ji et al. 2006). n= 3.

To extend these observations, we then detected the expression of δ ENaC in pneumocytes at the protein level using immunocytochemical approaches. As shown in Fig. 2B, δ ENaC proteins (green channel) were observed in normal human lung slices incubated with anti-δ ENaC antibody. In contrast, only nuclei were seen in control incubated with non-immune IgG. In addition, Western blots, as shown in Fig. 2C, revealed that a δ ENaC polypeptide of 85 kDa in human lung tissues and H441 cells was detected, consistent with our previous findings (Ji et al. 2006). These immunoassays, together with the RT-PCR data, suggest that δ ENaC subunit is abundantly co-expressed with conventional α, β and γ counterparts in alveolar epithelial cells. Even though the architecture regarding the αβγ-containing channels has not been completely known, co-expression and subunit-subunit physical interaction of these four mammalian ENaC subunits in lung epithelial cells implicate the existence of the αβγδ-containing ENaCs (Ji et al. 2006).

Stimulation of heterologously expressed ENaC activity in Xenopus oocytes

Based on the physical intermolecular regulation between δ ENaC and other counterparts as well as the co-expression of four ENaC subunits in lung epithelial cells as described above and previously (Ji et al. 2006), we hypothesized that cGMP might stimulate heterologously expressed αβγδ ENaCs. To address this issue, we co-injected cRNAs for human α, β, γ and δ ENaC into Xenopus oocytes, a well-established expression model for studying epithelial cation and anion channels. Whole-cell Na⁺ currents in oocytes expressing αβγδ ENaC were continuously monitored before and after cGMP and amiloride application. Typical current traces recorded at −100 mV and +40 mV were plotted in Fig. 3A for characterizing the concentration dependence of cGMP. A cell permeant cGMP, 8-pCPT-cGMP, even at micromolar levels considerably increased both inward and outward ENaC currents. Comparably, the inward currents, which reflect Na⁺ influx in oocytes and transapical Na⁺ re-absorption in polarized epithelial cells, were much more sensitive to cGMP. A saturated concentration of 1 mm, as revealed by the inward current, was observed for cGMP. This is consistent with the dose of cGMP we used for the in vivo alveolar fluid clearance. Addition of amiloride in the perfusate reversibly blocked both the cGMP-elevated and basal currents associated with the expressed αβγδ ENaCs. The values of EC₅₀ for cGMP were 5 μm and 225 μm, respectively, for inward (at −100 mV) and outward (at +40 mV) currents, as retrieved by fitting the experimental data with the Hill equation (Fig. 3B). On average, 8-pCPT-cGMP (0.2 mm) increased αβγδ ENaC activity over 2-fold in voltage clamped oocytes (Fig. 3C). Amiloride-sensitive Na⁺ currents at −100 mV reduced to −14 820 ± 2134 nA from −6171 ± 813 nA (n= 10, P < 0.001). The activation of ENaC by cGMP was reversible as the increased currents returned to the basal levels subsequent to washout of cGMP (data not shown). These results clearly demonstrate that αβγδ ENaCs can be activated by cell permeant cGMP in a dose-dependent manner.

Figure 3. Stimulation of heterologously expressed human ENaCs in Xenopus oocytes by cGMP.

A, representative current traces recorded at +40 mV and −100 mV. Oocytes were perfused with 8-pCPT-cGMP ranging from 1 μm to 1 mm as indicated by arrows. Amiloride was added to the bath before the recordings were terminated. B, dose–response curve. Average cGMP-activated current fraction in the presence of cGMP (ΔI) was normalized to the maximal cGMP-elevated current (ΔI_max) and plotted as a function of cGMP contents. Smooth lines and the EC₅₀ value was created by fitting the raw data with the Hill equation. n= 9. C, average amiloride-sensitive (AS) αβγδ ENaC activities before (control) and after cGMP perfusion. Holding potential, −100 mV. n= 10.

Dose-dependent up-regulation of amiloride-sensitive short-circuit currents (I_sc) by cGMP in H441 monolayers

ENaC expression in H441 cells, a human Clara cell line, has been well characterized electrophysiologically by several groups for studying native ENaC (Itani et al. 2002; Lazrak & Matalon, 2003; Thomas et al. 2004; Woollhead et al. 2007). The properties of amiloride-sensitive Na⁺ channels in H441 cells are similar to those in primary alveolar type II cells. To examine the sensitivity of epithelial Na⁺ channels to cGMP in polarized H441 cells, we applied 8-pCPT-cGMP at a variety of concentrations (0–1 000 μm) into the apical compartment of the Ussing chamber housing confluent monolayers. 8-pCPT-cGMP moderately increased I_sc level within minutes in a dose-dependent manner (Fig. 4A). The cGMP-elevated I_sc fraction in addition to the basal current was inhibited by the specific ENaC inhibitor amiloride (0.1 mm). The corresponding dose–response curve (Fig. 4B) showed that the EC₅₀ value for 8-pCPT-cGMP was 147 μm with a Hill co-efficient of 0.7 (n= 5). This EC₅₀ value at 10 mV may be quite close to the predicted one, which shall be definitely between 5 μm at −100 mV and 225 μm at +100 mV for αβγδ ENaCs in oocytes. In addition to expressed αβγδ ENaCs, amiloride-sensitive Na⁺ channels in human lung epithelial cells can also be activated by 8-pCPT-cGMP.

Figure 4. Activation of epithelial Na⁺ channels in confluent H441 monolayers.

A, representative current trace showing elevated Na⁺ current (I_sc) in the presence of various concentrations of 8-pCPT-cGMP (cGMP). B, dose–response curve. The cGMP-activated I_sc fraction (ΔI) was normalized to the maximal cGMP-activated I_sc level (ΔI_max). This plot was fitted with the Hill equation to compute the EC₅₀ value. n= 5.

Activation of whole-cell Na⁺ currents by 8-pCPT-cGMP in patch clamped H441 cells

To corroborate our aforementioned findings in oocytes and alveolar fluid clearance studies, we recorded whole-cell Na⁺ currents in H441 cells using the whole-cell patch clamp technique. Based on our dose–response studies on αβγδ ENaC expressed in oocytes (Fig. 3) and on native epithelial Na⁺ channels in polarized confluent H441 monolayers (Fig. 4), 1 mm 8-pCPT-cGMP was supposed to maximally activate amiloride-sensitive channel activity. As shown in Fig. 5A and B, cGMP increased inward Na⁺ currents, which were abolished by amiloride (10 μm). The corresponding amiloride-sensitive Na⁺ currents (ASIs) at each holding potential were plotted as current–voltage curves (Fig. 5C). The actual reversal potential of over +40 mV is consistent with the predicted one for highly Na⁺ selective channels. Besides, the inward rectified I–V curves and highly amiloride sensitivity are in good agreement with those of heterologously expressed ENaCs previously reported by several independent groups (Canessa et al. 1994; Snyder et al. 1998; Ji et al. 2001, 2004; Sheng et al. 2004). Average ASIs were normalized to controls and graphed in Fig. 5D. 8-pCPT-cGMP significantly increased Na⁺ currents by 43.5 ± 14.5% (P <0.05, n= 11–19). This stimulatory effect was not affected by addition of diltiazem (50 μm), a cyclic nucleotide-gated (CNG) channel inhibitor (45.2 ± 11%, P > 0.05 vs. cGMP group, n= 4), excluding the possible contributions of amiloride-sensitive CNG channels.

Figure 5. Activation of amiloride-sensitive Na⁺ currents by cGMP in patch clamped H441 cells.

A, representative time course showing 8-pCPT-cGMP (cGMP)-activated whole cell current trace at −40 mV. B, amiloride-sensitive current traces recorded in a cell before (control) and after delivery of 8-pCPT-cGMP (cGMP) to the bath. Dashed lines indicate the zero current levels. C, I–V curves for average control currents (control), in the presence of 8-pCPT-cGMP (cGMP), and amiloride plus 8-pCPT-cGMP (amil+cGMP). Amiloride-sensitive currents (ASIs) were normalized to the current density at −120 mV of control. D, normalized whole-cell current densities (I/I_control). Whole-cell currents in the presence of 8-pCPT-cGMP (cGMP), diltiazem (50 μm) and cGMP (diltia + cGMP), amiloride (10 μm) + cGMP, and amiloride alone (amil) were normalized to the control. Data were pooled from five independent experiments performed by three patch clampers. Numbers in brackets are patched cells.

To narrow down the involvements of cGMP-regulated non-Na⁺ channels, for instance, K⁺ channels, and anion channels, we repeated these patch clamp experiments in the presence of 10 μm amiloride, enough to completely inhibit epithelial Na⁺ channels. As shown in Fig. 5C, the same amount of cGMP did not alter amiloride-resistant currents within the range of tested membrane potentials (from −120 mV to +40 mV). Taken together, the possible contribution of amiloride-insensitive, non-Na⁺ channels to the cGMP-activated fraction in H441 cells can be definitely ruled out.

cGMP increases transapical I_sc level in basolaterally permeabilized H441 monolayers

Ouabain-inhibitable Na⁺/K⁺-ATPase is the major exit for accumulated intracellular Na⁺ ions subsequent to salt entry through the apically located Na⁺ channels (Matthay et al. 2002). To eliminate the transbasolateral conductance contributed by Na⁺/K⁺-ATPase and any other cGMP-sensitive transport pathways, we then basolaterally permeabilized H441 monolayers with amphotericin B as previously described (Guo et al. 1998; Thome et al. 2003). In contrast to the intact monolayers, a specific Na⁺/K⁺-ATPase blocker, ouabain (1 mm), when applied basolaterally, did not reduce the I_sc levels in permeabilized monolayers. This indicates that the entire basolateral membrane was completely permeabilized (data not shown). 8-pCPT-cGMP (0.2 mm) elevated the transapical membrane I_sc levels from 31.5 ± 3 to 36.3 ± 3 μA cm⁻² (Fig. 6B, P <0.01, paired-sample t-test, n= 8), in agreement with our findings in intact H441 monolayers (Fig. 4A). In contrast, when the same dose of 8-pCPT-cGMP was added in the presence of amiloride, no increase in amiloride-resistant transapical I_sc level (from 25.1 ± 1 to 25.2 ± 2 μA cm⁻²) was observed in the permeabilized H441 monolayers (Fig. 6C and D, P > 0.05, n= 7). Clearly, the cGMP-stimulated transapical I_sc levels are associated with apically located amiloride-sensitive Na⁺ channels at least under these experimental conditions.

Figure 6. Activation of epithelial Na⁺ channels by cGMP in basolaterally permeabilized H441 monolayers.

A, representative short-circuit current (I_sc) trace. Amphotericin B (am-B, 100 μm) was added to the basolateral bath as indicated by an arrow. 8-pCPT-cGMP (0.2 mm cGMP) was applied to the apical bath followed by amiloride (100 μm). B, summary of I_sc levels. Mean amiloride-sensitive (AS) I_sc values before and after addition 8-pCPT-cGMP (cGMP) were averaged. n= 8. C and D, representative trace and average I_sc levels in amiloride-pretreated H441 monolayers. n= 7.

PKGII mediates cGMP-stimulated Na⁺ channel activity in confluent H441 monolayers

Three PKG isoenzymes, namely, PKGIα, PKGIβ, and PKGII, have been isolated to date. Albeit 8-pCPT-cGMP predominately activates PKGII, it also stimulates PKGI enzymatic activity at higher concentrations. To identify which PKG isoform is involved, we applied the isoform specific antagonists for PKGI and II following 8-pCPT-cGMP. Rp-8-Br-PET-cGMP specifically inactivates PKGIα and Iβ with similar IC₅₀ values of 35 and 30 nm, respectively (Jernigan et al. 2003), whereas Rp-8-pCPT-cGMP is a specific PKGII inactivator with a larger IC₅₀ value of 160 nm (Gamm et al. 1995). Additionally, Rp-8-Br-PET-cGMP is much more cell permeant than Rp-8-pCPT-cGMP, as predicted by their respective lipophilicities (approximately 130-fold and 60-fold greater than that of standard cGMP for Rp-8-Br-PET-cGMP and Rp-8-pCPT-cGMP, respectively). We therefore applied 100 nm Rp-8-Br-PET-cGMP to inhibit PKGI activity only. A higher dose of Rp-8-pCPT-cGMP (10 μm) was then added based on its IC₅₀ value and less cell permeability. As shown in Fig. 7A, the I_sc level was activated by cGMP (0.2 mm) within minutes. On average 8-pCPT-cGMP stimulated I_sc currents from 20.7 ± 0.8 to 22.1 ± 0.8 μA cm⁻² (Fig. 7B, P < 0.001, Paired-sample t-test, n= 10). Subsequently addition of Rp-8-Br-PET-cGMP caused a slight reduction of I_sc level without significant statistical difference (from 22.1 ± 0.8 to 22.0 ± 0.8 μA cm⁻², P > 0.05). In contrast, Rp-8-pCPT-cGMP completely eliminated the cGMP-stimulated I_sc fraction progressively (Fig. 7A, from 22.0 ± 0.8 to 20.7 ± 0.7 μA cm⁻², P < 0.05, n= 10). These observations strongly suggest that the PKGII isoenzyme is the predominant mediator for the cGMP-activated epithelial Na⁺ channels.

Figure 7. Inhibition of cGMP-activated epithelial Na⁺ channel activity in H441 monolayers by isoform specific protein kinase G inhibitors.

A, representative I_sc trace showing the inhibitory effects of Rp-8-Br-PET-cGMP (PKGI Inh) and Rp-8-pCPT-cGMP (PKGII Inh) on the 8-pCPT-cGMP-elevated I_sc. B, amiloride-sensitive (AS) I_sc levels before (control) and after addition of 8-pCPT-cGMP (cGMP), Rp-8-Br-PET-cGMP (PKGI Inh), and Rp-8-pCPT-cGMP (PKGII Inh). n= 10.

Isoform specific PKGII inhibitor abolishes cGMP-induced activation of Na⁺ channels in patch clamped H441 cells

To further verify the cGK isoform specificity for the cGMP-mediated activation, the mixture of 8-pCPT-cGMP and Rp-8-pCPT-cGMP was applied to the patch clamped H441 cells in the whole-cell mode. 8-pCPT-cGMP did not result in an increment in whole-cell Na⁺ currents in the presence of Rp-8-pCPT-cGMP, a specific inhibitor of PKGII isoenzyme (Fig. 8A). Average whole-cell Na⁺ currents at −100 mV even decreased by 82 ± 19% (n= 5, P > 0.05, Fig. 8B). Taken together with the results of the Ussing chamber studies, the cGKII but not the cGKI isoform is a critical mediator of the cGMP-activated Na⁺ currents in both polarized and non-polarized H441 cells.

Figure 8. PKGII inhibitor eliminates cGMP-activated whole-cell Na⁺ currents in voltage clamped H441 cells.

A, amiloride-sensitive step I–V traces before (control) and after addition of the mixture of 8-pCPT-cGMP (1 mm) and PKGII inhibitor, Rp-8-pCPT-cGMP (10 μm, +Rp-8-pCPT-cGMP). B, normalized whole-cell current densities (I/I_control). See Fig. 5 legend for additional details.

Expression of protein kinase G isoforms in human lung tissues and H441 cells

To detect the expression of three PKG isoforms (i.e., PKGIα, PKGIβ and PKGII) in H441 cells as well as whole lung tissue at the transcriptional and protein levels, we conducted RT-PCR and Western blot assays. Our RT-PCR results showed that all three PKG isoforms were expressed in both H441 cells (Fig. 9A) and human lung tissue (Fig. 9C) at the mRNA level. The product sizes were 203 bp for PKGIα, 346 bp for PKGIβ, and 502 bp for PKGII isoform.

Figure 9. Expression of protein kinase G isoforms in H441 cells and human lung tissues (HLT).

A and C, reverse transcription-PCR of protein kinase G mRNA expression. The expression of three protein kinase G isoforms, PKGIα (Iα), PKGIβ (Iβ) and PKGII (II), was examined in H441 cells (A) and human lung tissues (C). M and N denote molecular mass markers and the running buffer only without RNA. Arrows indicate the 500 bp marker. n= 3. B and D, Western blots of protein kinase G isoform I and II. H441 cell and human lung tissue lysates were probed with isoform specific anti-PKG antibodies. Mouse brain extract (mBE) was used as a positive control for PKGI and purified cGKII protein for PKGII. These experiments were repeated 3–5 times.

We then analysed the expression of PKG isoforms at the protein level. Mouse brain extract (mBE) was used as a positive control for PKGI and purified PKGII protein for PKGII. A 75 kDa polypeptide, corresponding to PKGI, was detected in both human lung tissue (HLT) and the lane for positive control. Nevertheless, PKGI signal was not seen in the lane loaded with the H441 protein extracts (Fig. 9B). When we probed both the human lung tissue and H441 cell lysates with a specific PKGII antibody, a protein band at 86 kDa was detected (Fig. 9D). This is consistent with the size of purified PKGII protein (PKGII) and supportive of a previous report (Vaandrager et al. 1998). Taken together, these RT-PCR and immunoblotting results provide direct evidence that PKG isoforms are expressed in both human lung tissue and Clara cells but with inconsistent protein expression patterns.

8-Br-cGMP does not activate amiloride-sensitive Na⁺ currents in patch clamped H441 cells

Another apically located Cl⁻ channel, CFTR, is also a cGKII-activated pathway for alveolar fluid clearance (French et al. 1995; Fang et al. 2006). To rule out the potential contribution of CFTR to the cGMP-activated whole-cell currents in H441 cells, we repeatedly examined the effects of cGMP in the presence of a specific CFTR inhibitor, CFTRinh-172, in addition to diltiazem to inhibit CNG channels (Fig. 10). As anticipated, CFTRinh-172 and diltiazem slightly reduced whole-cell currents (Fig. 10A and B). Subsequently, 8-pCPT-cGMP stimulated the whole-cell Na⁺ current density from −2.3 ± 0.5 to −5.0 ± 1.1 pA pF⁻¹ (P < 0.05, n= 10). In contrast, the same dose of 8-Br-cGMP, a PKGI activator, did not significantly increase the currents (from −2.9 ± 0.5 to −3.0 ± 0.6 pA pF⁻¹, P > 0.05, n= 15), which is consistent with other reports (Jain et al. 1998; Albert et al. 2008). These patch clamp data are supportive of our observations of in vivo alveolar fluid clearance.

Figure 10. Effects of cGMP on ENaC activity in H441 cells pre-exposed to CFTR and CNG channel inhibitors.

H441 cells were first perfused with CFTRinh-172 (20 μm) and diltiazem (50 μm) to inhibit CFTR and CNG channels. Addition of either 0.5 mm 8-pCPT-cGMP (A) or 8-Br-cGMP (B) to the bath is indicated by arrows. Holding potential was −100 mV. C and D, average amiloride-sensitive whole-cell current (ASI) density.

Discussion

The regulation of amiloride-sensitive salt transport across epithelial cells by nitric oxide (NO) prompted us to hypothesize that cGMP, a key downstream effector of the NO/cGMP/PKG signal pathway, may influence ENaC activity acutely. Our Ussing chamber and patch clamp results showed that native amiloride-sensitive Na⁺ channels were up-regulated by the cell permeant PKGII activator 8-pCPT-cGMP in human lung Clara cells. Activation of native ENaCs by cGMP was dose dependent. PKGII blocker completely removed 8-pCPT-cGMP-activated ENaC activity, both I_sc and whole-cell currents. These were further confirmed by the observation that PKGII was the only isoform expressed in H441 cells at the protein level. The potential research and pharmaceutical implications are validated by the improved distal lung fluid clearance in mice intratracheally instilled with 8-pCPT-cGMP.

Diverse expression patterns of PKG isoforms in different epithelial cells may result in inconsistent responses to cGMP. The expression of PKGII as a sole isoform in Clara cells explains why PKGI modulators (8-Br-cGMP and Rp-8-Br-PET-cGMP) did not alter ENaC activity. Interestingly, another apically located anion channel, CFTR, is activated by cGMP via PKGII but not PKGI (French et al. 1995; Vaandrager et al. 1997, 1998). Moreover, the inhibition of the Na⁺/H⁺ antiporter in intestinal and renal epithelial cells is also mediated by PKGII (Cha et al. 2005). Notably, PKGI is mainly expressed in non-epithelial tissues, for example smooth muscle and endothelium, which explains why PKGI was detected in whole lung protein extracts.

Although crosstalk between PKG and other kinases, for instance protein kinase A, has been reported (Schlossmann & Hofmann, 2005), it is most unlikely that cGMP activates other protein kinases based on the following evidence. As shown in Fig. 7A, the isoform-specific PKGII inhibitor, when applied at a micromolar concentration that specifically inhibits only PKGII enzymatic activity, completely blocked the cGMP-elevated current fraction. The second line of evidence is from the lack of acute response of cloned ENaC to protein kinase A when expressed in oocytes. Third, pretreatment with an inhibitor cocktail to inhibit protein kinase A, protein kinase C, mitogen-activated protein kinase (p38 MAP), and MAP-ERK kinase (MEK), which are all cGMP-regulated kinases which alter ENaC activity, did not affect cGMP-mediated activation (H.-G. Nie, unpublished data). Conclusively, it is the PKGII isoenzyme that plays a significant role in the response to 8-pCPT-cGMP of native epithelial Na⁺ channels in Clara cells and heterologously expressed αβγδ ENaCs in Xenopus oocytes. However, we cannot completely rule out other untested kinases downstream of PKGII to regulate ENaC activity.

It has been found that PKGII-anchoring proteins were required to form a cGMP-anchoring protein–Na⁺/H⁺ exchanger (NHE) complex to regulate NHE activity (Cha et al. 2005). In addition to myosin heavy chain, NHERF2 is the only PKGII anchoring protein identified to date. In this process, myristoylation of PKGII attaches it to the apical membrane of polarized epithelial cells. Deletion of the myristoylation motif in the N terminus of PKGII has been shown to prevent the translocation of PKGII from the cytosol to the plasma membrane (Cha et al. 2005). Whether it is also the scenario for the regulation of ENaC by cGMP awaits follow-up studies.

CFTR has recently been found to be one of major contributors to the alveolar fluid clearance (Fang et al. 2006), which is activated by cGMP/PKGII (French et al. 1995). It is conceivable that the cGMP-activated transepithelial I_sc currents were at least partially carried by CFTR. The Ussing chamber results from the amiloride-pretreated H441 monolayers (Fig. 6C) and whole-cell patch clamp observations in the presence of CFTRinh-172 (Fig. 10) confirm that ENaC is the charge carrier of the cGMP-activated currents. This is further substantiated by the in vivo alveolar fluid clearance and oocytes only expressing human ENaCs. Taken together, our results clearly indicate that cGMP-activated Na⁺ channels is a novel pathway for resolving oedematous fluid in injured lungs.

Albeit we cannot exclude the possible contributions of epithelial CNG channels to alveolar fluid clearance in vivo, at least under our experimental conditions, native epithelial Na⁺ channels can be activated by cGMP to the same extent, either in the presence or absence of CNG inhibitors. Reminiscently, cGMP may directly activate ENaCs through a mechanism analogous to that for the cGMP-mediated activation of CNG channels. Cyclic nucleotides directly bind to cyclic nucleotide binding domains (CNBD) and alter the channel gating kinetics. In addition, blocking of the cGMP-elevated I_sc by specific PKGII isoform inhibitor demonstrates PKGII is the necessity for the cGMP-mediated activation of ENaC.

The molecular basis of the cGMP-activated ENaC in H441 cells remains to be elucidated. Three subtypes of amiloride-sensitive Na⁺ channels have been reported in H441 cells grown on coverslips: 4 pS, 11 pS, and 20 pS channels (Lazrak & Matalon, 2003; Thomas et al. 2004). Similarly, three types of native amiloride-sensitive Na⁺ channels have also been functionally detected in other pneumocytes. These results imply there are at least three groups of ENaCs in the unpolarized H441 cells. The highly Na⁺ selective, more amiloride sensitive, 4 pS channels have been attributed to αβγ channels. However, the stoichiometry of the αβγ ENaCs is still inconclusive. So far, three schools of thought have been published regarding the architecture of the αβγ ENaCs: 1α+1β+1γ, 2α+1β+1γ and 3α+3β+3γ (Firsov et al. 1998; Kosari et al. 1998; Snyder et al. 1998; Jasti et al. 2007). In oocytes co-expressing δ ENaC with the other three subunits, the only detected unitary Na⁺ conductance is 8 pS, which is much close to the 11 pS channels in H441 cells (Ji et al. 2006). In addition, δ ENaC physically interacts with other counterparts and confers the biophysical features on αβγ ENaC. Taken together, the 11 pS channels most likely consists of αβγδ ENaC subunits and is sensitive to cGMP/PKGII. However, the stoichiometry of the αβγδ channels awaits further studies following future conclusive studies with regard to the architecture of αβγ ENaCs.

Up-regulation of vectorial salt and fluid transport across the distal lung epithelial layer has been a critical therapeutic target to ameliorate oedematous respiratory diseases for a long time. Cyclic GMP has been observed to be increased in murine and rat lungs both in vivo and in vitro following NO application (Tsubochi et al. 2003; Hardiman et al. 2004). Increased cGMP may augment the cGMP-sensitive pathway for lung fluid removal from alveolar sacs (Junor et al. 2000; Norlin et al. 2001; Sakuma et al. 2004). Herein to our best knowledge, PKGII is identified as an ENaC activator for the first time in Clara cells. In addition to the regulation of CFTR by PKGII, which is another critical pathway for alveolar fluid clearance (Fang et al. 2006), up-regulation of the rate-limiting ENaCs in respiratory epithelial cells by specific PKGII activators may be a potent clinicopharmaceutical strategy for alleviating airspace flooding in fatal oedematous lung diseases.

In toto, our observations suggest that cell permeant 8-pCPT-cGMP, a PKGII activator, stimulates native lung ENaC activity. These results indicate 8-pCPT-cGMP most possibly stimulates PKGII enzymatic activity and subsequently activates channel function.

Glossary

Abbreviations

AFC

alveolar fluid clearance

ASI

amiloride-sensitive Na⁺ current

CFTR

cystic fibrosis transmembrane conductance regulator

ENaC

epithelial Na⁺ channel

I_sc

short-circuit current

NOS

nitric oxide synthase

PKA

PKC, PKG, protein kinase A/C/G

RSV

respiratory syncytial virus

R_t

transepithelial resistance

Author contributions

All of experiments were performed at the University of Texas Health Science Center at Tyler and the University of Alabama at Birmingham. H.G.N. and L.C. contributed to the Ussing chamber studies, D.Y.H., W.F.S. and S.P.W. to patch clamp assays, J.L. and X.H.F. to immunoassays, X.G. and S.P.W. to in vivo A.F.C., S.M. and H.L.J. were responsible for conception and design, interpretation of data, and preparation of the finally version.

Author's present address

Hong-Guang Nie: School of Pharmaceutical Sciences, China Medical University, Liaoning, Shenyang 110001, China.