AMP-activated protein kinase (AMPK)–dependent and –independent pathways regulate hypoxic inhibition of transepithelial Na⁺ transport across human airway epithelial cells

Abstract

BACKGROUND AND PURPOSE

Pulmonary transepithelial Na⁺ transport is reduced by hypoxia, but in the airway the regulatory mechanisms remain unclear. We investigated the role of AMPK and ROS in the hypoxic regulation of apical amiloride-sensitive Na⁺ channels and basolateral Na⁺K⁺ ATPase activity.

EXPERIMENTAL APPROACH

H441 human airway epithelial cells were used to examine the effects of hypoxia on Na⁺ transport, AMP : ATP ratio and AMPK activity. Lentiviral constructs were used to modify cellular AMPK abundance and activity; pharmacological agents were used to modify cellular ROS.

KEY RESULTS

AMPK was activated by exposure to 3% or 0.2% O₂ for 60 min in cells grown in submerged culture or when fluid (0.1 mL·cm⁻²) was added to the apical surface of cells grown at the air–liquid interface. Only 0.2% O₂ activated AMPK in cells grown at the air–liquid interface. AMPK activation was associated with elevation of cellular AMP : ATP ratio and activity of the upstream kinase LKB1. Hypoxia inhibited basolateral ouabain-sensitive I_sc (I_ouabain) and apical amiloride-sensitive Na⁺ conductance (G_Na+). Modification of AMPK activity prevented the effect of hypoxia on I_ouabain (Na⁺K⁺ ATPase) but not apical G_Na+. Scavenging of superoxide and inhibition of NADPH oxidase prevented the effect of hypoxia on apical G_Na+ (epithelial Na⁺ channels).

CONCLUSIONS AND IMPLICATIONS

Hypoxia activates AMPK-dependent and -independent pathways in airway epithelial cells. Importantly, these pathways differentially regulate apical Na⁺ channels and basolateral Na⁺K⁺ ATPase activity to decrease transepithelial Na⁺ transport. Luminal fluid potentiated the effect of hypoxia and activated AMPK, which could have important consequences in lung disease conditions.

Introduction

Exposure of the lung to decreased oxygen (e.g. at high altitude or during pathophysiological disorders) is associated with formation of pulmonary oedema (Mairbaurl, 2006). Individuals exposed to high altitude have decreased nasal potential difference, which is indicative of reduced transepithelial transport in the airway (Sartori et al., 2002; Guney et al., 2007). Active vectorial transport of Na⁺ across the lung epithelium drives fluid re-absorption, and this serves an essential role in the maintenance of the pericilliary fluid layer volume in the airway and the prevention and resolution of pulmonary oedema in the alveolar region (Hummler et al., 1996; Widdicombe, 2002; Mairbaurl, 2006; Tarran, 2008; Zhou et al., 2008). This process requires the activity of apical amiloride-sensitive Na⁺ channels and the basolateral Na⁺K⁺ ATPase. In vitro, exposure to hypoxia inhibited amiloride-sensitive ²²Na⁺ influx and Na⁺K⁺ ATPase activity in human alveolar A549 and rat alveolar type II cells and a role for ROS in regulating these processes was identified (Mairbaurl et al., 1997; Planes et al., 1997; 2002; Wodopia et al., 2000; Dada et al., 2003; Dada and Sznajder, 2007).

More recently, adenosine AMP-activated protein kinase (AMPK) has been shown to play a role in sensing and mediating the effects of hypoxia in a number of tissues by responding to changes in cellular energy balance (Evans et al., 2005; Wyatt et al., 2007). It is a heterotrimeric protein complex consisting of a catalytic α subunit and regulatory β and γ subunits (Winder et al., 1997; Hardie and Sakamoto, 2006). An increase in the cellular AMP : ATP ratio stimulates AMPK activity through the binding of AMP to CBS domains present in the γ subunit (Mortimer et al., 2004). This makes AMPK a better substrate for phosphorylation by upstream kinases such as liver kinase B1 (LKB1)(Wilson et al., 1996; Sutherland et al., 2003; Woods et al., 2003), Ca²⁺/calmodulin kinase kinase (CaMKK) (Hawley et al., 2005; Woods et al., 2005) and TGF-β-activated kinase 1 (TAK1) (Momcilovic et al., 2006; Xie et al., 2006), while reducing its dephosphorylation by protein phosphatases PP2A and PP2C (Hawley et al., 1996; Marley et al., 1996; Zhao et al., 2007). It has also been proposed that AMPK can be activated independently of changes in the AMP : ATP ratio by changes in mitochondrial ROS, elevation of intracellular Ca²⁺ and increased phosphorylation by CaMKK (Verhoeven et al., 1995; Woods et al., 2005). Once activated, AMPK inhibits ATP-consuming processes and increases ATP-generating processes.

We have shown that pharmacological activation of AMPK decreases transepithelial Na⁺ transport across H441 airway epithelial cell monolayers via inhibition of apical amiloride-sensitive Na⁺ channels and the activity of Na⁺K⁺ ATPase (Carattino et al., 2005; Woollhead et al., 2005; 2007; Albert et al., 2008; Gusarova et al., 2009). Hypoxia was recently shown to decrease Na⁺K⁺ ATPase abundance in the alveolar cell membrane via a pathway involving AMPK and ROS (Gusarova et al., 2009). However, how hypoxia inhibits Na⁺ transport pathways in the airway has not been explored. In particular, it is unclear whether AMPK mediates the effects of hypoxia on apically localized amiloride-sensitive Na⁺ channels.

We used three different H441 cell culture models for this study as we wanted to compare the responses to hypoxia of air–liquid interface cultures (a physiologically relevant model of the airway) with those of submerged cells cultured on plastic, because much of the published data on the effects of hypoxia use the latter. As transepithelial transport can only be measured in cells grown on permeable supports, we also investigated responses in polarized airway epithelial cells ‘submerged’ by placing fluid on the apical surface (liquid–liquid interface cultures). Using these models, we investigated the role of AMPK in hypoxia-induced functional changes in transepithelial Na⁺ transport across H441 human airway epithelial cell monolayers.

Methods

H441 cell culture

The human lung epithelial cell line, H441, was obtained from the American Type Culture Collection (ATCC) (Manassas, VA). Cells were maintained in RPMI-1640 medium with 10% FBS, 2 mM L-glutamine, 1 mM sodium pyruvate (Sigma, Poole, UK), 5 µg·mL⁻¹ insulin, 5 µg·mL⁻¹ transferrin, 10 nM sodium selinite and 100 U·mL⁻¹ penicillin, 100 µg·mL⁻¹ streptomycin and incubated in a humidified atmosphere with 21% O₂+ 5% CO₂ buffered with N₂ (control) at 37°C. Polarized resistive H441 monolayers were cultured on Snapwell clear membranes (Corning, UK) as previously described (Woollhead et al., 2005). Serum was replaced with 4% charcoal stripped FBS, 10 nM tri-iodothyronine (T3) and 200 nM dexamethasone (Sigma) to promote polarization. Cells were maintained for 7–14 days at the air–liquid interface before experimentation. A549 alveolar epithelial cells were cultured in DMEM supplemented with 10% FBS, 2 mM L-glutamine and antibiotics (penicillin/streptomycin). Unless otherwise stated, all cell culture media and supplements were purchased from Invitrogen (Paisley, UK).

Hypoxic treatment

Non-polarized H441 cells grown in submerged culture were seeded onto 6- or 12-well plates. When cells were 90% confluent, the volume of medium overlying the cells was reduced to 0.1 mL·cm⁻² as described by Planes et al. (1997) to decrease the gradient for oxygen diffusion before exposure to a humidified atmosphere containing 0.2% O₂+ 5% CO₂ buffered with N₂ (0.2% O₂) or 3% O₂+ 5% CO₂ buffered with N₂ (3% O₂) at 37°C for 15 or 30 or 60 min. Polarized H441 cells cultured on permeable supports at the air–liquid interface were exposed to a humidified atmosphere containing 0.2% O₂+ 5% CO₂ buffered with N₂ (0.2% O₂) or 3% O₂+ 5% CO₂ buffered with N₂ (3% O₂) at 37°C for 15 or 30 or 60 min. Liquid–liquid interface experiments were conducted with 0.1 mL·cm⁻² PSS added to the apical chamber before exposure to hypoxia as described above. Cells were pretreated with STO609 (Sigma) 25 µM for 30 min, geldenamycin (Enzo Life Sciences, Exeter, UK) 10 µM for 24 h and ionomycin (Merck, Nottingham, UK) 1 µM for 15 min.

Fluo-3 fluorescence imaging

H441 cells were seeded onto glass coverslips forming the bottom of experimental chambers (as submerged cultures). Cells were loaded with 2 µM Ca²⁺-sensitive fluorescent indicator fluo-3AM (Invitrogen) for 1 h at 37°C in culture medium also containing a non-ionic detergent Pluronic F127 (0.002%) and an inhibitor of organic cation extrusion systems probenecid (2.5 mM). The cells accumulated very little of the dye unless two latter compounds were present (McAlroy et al., 2000). Fluo-3 fluorescence was recorded using x–y time series mode of an LSM 510 laser-scanning confocal microscope (Carl Zeiss Ltd., Welwyn Garden City, UK) and analysed as previously described (Harhun et al., 2006).

Ussing chamber experiments

Ussing chamber experiments were performed as described previously (Ramminger et al., 2000; Woollhead et al., 2005). H441 cell monolayers were mounted in Ussing chambers and bathed with PSS (mM): NaCl, 117; NaHCO₃, 25; KCl, 4.7; MgSO₄, 1.2; KH₂PO₄, 1.2; CaCl₂, 2.5; d-glucose, 11; pH 7.3–7.4 and equilibrated with pre-mixed gas 0.2% O₂+ 5% CO₂+ 94.8% N₂ (0.2% O₂) or 3% O₂+ 5% CO₂+ 92% N₂ (3% O₂) or 21% O₂+ 5% CO₂+ 74% N₂ (control) maintained at 37°C and continuously circulated throughout the course of the experiment by airlift. Short circuit current (I_sc) was measured by clamping transepithelial potential (V_t) at 0 using a DVC1000 current /clamp module (WPI, Hitchin, UK). Every 30 s, a 2 mV potential was applied to enable the calculation of transepithelial resistance (R_t). Amiloride-sensitive transepithelial I_sc (I_amiloride) was measured by adding amiloride (10 µM) to the solution perfusing the apical membrane of the monolayer. Ouabain-sensitive basolateral current (I_ouabain) was determined by maintaining the epithelial monolayer under the conditions described above and permeabilizing the apical membrane with nystatin (75 µM) to isolate the basolateral current generated by Na⁺ K⁺ ATPase. Apical permeabilization permits unrestricted movement of monovalent cations across the apical membrane and leads to equilibration of intracellular ion concentrations with external ionic conditions. The increase in intracellular Na⁺ leads to increased basolateral pump current. The decrease in peak basolateral current following the application of 1 mM ouabain is indicative of the activity of the Na⁺ K⁺ ATPase under these experimental conditions. Amiloride-sensitive apical conductance (G_Na+), was determined by bathing monolayers in potassium gluconate solution [composition (mM): potassium gluconate, 121.7; KHCO₃, 25; MgSO₄, 1.2; KH₂PO₄, 1.2; calcium gluconate, 11.5; d-glucose, 11; pH 7.3–7.4]. This was mixed with PSS to give a PSS : potassium gluconate solution ratio of 8.1:91.9 and a final Na⁺ concentration of ∼11.5 mM (equilibrated with premixed gas as described above). Na⁺K⁺ ATPase was inhibited with 1 mM ouabain, and the basolateral membrane was permeabilized with 75 µM nystatin. A Na⁺ gradient across the apical membrane was then established by raising the concentration of Na⁺ in the apical chamber to 55 mM with a sodium gluconate solution [composition (mM): sodium gluconate, 117; NaHCO₃, 25; potassium gluconate, 4.7; MgSO₄, 1.2; KH₂PO₄, 1.2; calcium gluconate, 2.5; d-glucose, 11; pH 7.3–7.4 (91.9:8.1 with PSS) equilibrated with premixed gas as described above]. Amiloride (10 µM) was added to the apical bath to inhibit currents, and amiloride-sensitive G_Na+ was estimated from the amiloride-sensitive apical current (I_ap) using the equation G_Na+=I_ap/V_Na+, where V_Na+ is the driving force for Na⁺ entry where potential difference across the apical membrane (V_a) is clamped at 0 and the equilibrium potential for Na⁺ (E_Na) = 41.8 mV. The value for E_Na+ was calculated using the Nernst equation for the ionic conditions applied. There is no driving force for the passive movement of Cl^- across the apical membrane as the PD is held at 0 mV, and the concentrations of Cl^- in the apical/basolateral baths are the same (Collett et al., 2002; Richard et al., 2004; Woollhead et al., 2005; 2007)

Lentiviral vector construction and cell transduction

cDNA encoding for a constitutively active mutant of α-AMPK (CA) was amplified from pcDNA3/AMPKaCA (kind gift of Prof David Carling, Imperial College London, UK) by PCR and subcloned under the control of the CMV promoter into XhoI/SpeI of the self-inactivating (SIN) (Zufferey et al., 1998) lentiviral transfer plasmid pRRLsc_SFFV_eGFP_CMV_NCS1_WPRE, replacing the NCS1 cDNA. AMPKα (CGCAGCAATAAGCATGCATA) (Myerburg et al., 2010) and non-targeting shRNA sequences (Sh-neg: ACTACCGTTGTTATAGGTG) were cloned under the control of the human H1 promoter into MluI/PstI sites of pRRLsc_CMV_eGFP_WPRE_H1 (pshLENTImax), a SIN lentiviral transfer plasmid for RNAi. Lentiviral transfer plasmids were packaged into third-generation, VSV-G pseudotyped vectors. Briefly, vectors were produced by calcium phosphate-mediated transient transfection of HEK293T. The packaging plasmids were pMDLg/pRRE and pRSV-rev, while pMD2.G provided the VSV-G envelope gene (Dull et al., 1998). Cell culture supernatant was harvested 2–3 days post transfection, and vector particles were concentrated by ultracentrifugation. Vectors were titrated in HeLa cells by scoring eGFP fluorescence units using flow cytometry (Yanez-Munoz et al., 2006). H441 cells were seeded onto permeable supports. Following overnight incubation, the cells were transduced with lentiviral vectors at an approximate multiplicity of infection (MOI) of 10 for 4 h. The cells were incubated in fresh growth medium overnight, cultured at the air–liquid interface and used for functional studies 96 h after transduction.

Measurement of intracellular adenine nucleotide concentration

The adenine nucleotides ATP, ADP, AMP and TAN (total adenine nucleotides) were measured in untreated monolayers or those exposed to 0.2% O₂ or 3% O₂ for 60 min. Cells were washed with PBS on ice, and then (0.4 M) perchloric acid was added to extract the nucleotides. Cell extracts were neutralized with 3 M K₃PO₄ and analysed by reverse-phase HPLC (Hewlett-Packard 1100 series linked to a diode array detector, ISS Ltd., Dartford, UK) as described previously in detail (Woollhead et al., 2007; Nofziger et al., 2009). Protein content of the perchlorate precipitate was analysed using the Bradford assay (Fisher Scientific UK Ltd., Loughborough, UK).

Western blotting

Cells were lysed in lysis buffer [50 mM Tris–HCl, pH 7.4; 1% NP-40; 0.25% sodium deoxycholate; 150 mM NaCl; 1 mM EDTA; 1 mM EGTA, 1 mM PMSF, 1 mM Na₃VO₄, 50 mM NaF, 5 mM sodium pyrophosphate and 1% (w v^-1) Protease inhibitor cocktail]. Cellular debris was removed by centrifugation at 13 000 ×g for 10 min, and protein concentrations were determined using the Bradford assay (as above). Approximately 30–50 µg of proteins were fractionated on 4–12% Bis–Tris gel (Invitrogen) alongside pre-stained protein standards (Santa Cruz, CA) and transferred onto Hybond-P PVDF Membrane (GE Healthcare, Amersham, UK). Membranes were incubated in TBS-T containing 5% (w v^-1) nonfat milk powder for 1 h at room temperature before incubation with primary antibodies; anti-phospho-AMPKα (Thr¹⁷²), anti-AMPKα, anti-phospho-ACC(Ser⁷⁹), anti-ACC (Cell Signaling Technology, Danvers, MA) (all diluted 1:1000) or anti-LKB1 (Abcam, Cambridge, UK) (1:500) or anti-β-actin (Abcam) (1:2000), followed by incubation with species-specific HRP-conjugated secondary antisera (Sigma). Immunostained proteins were visualized with SuperSignal west pico chemiluminescent substrate (GE Healthcare). Densitometric quantification was performed using Scion Image (NIH).

Statistical analysis

Statistical analysis was carried out using anova or Student's unpaired or paired t-tests where applicable. P-values of <0.05 were considered significant. Results are presented as mean ± SEM.

Results

Effect of hypoxia on the intracellular adenine nucleotide concentration

As the cellular AMP : ATP ratio is an important activator of AMPK, we determined whether the AMP : ATP ratio was altered in cells exposed to hypoxia. Cellular extracts from H441 cells monolayers were analysed by HPLC. Exposure of submerged cultures to 3% O₂ for 60 min resulted in an 8 fold increase in the AMP : ATP ratio compared with control (P < 0.001, n= 3). This was the result of both a decrease in ATP concentration and an increase in AMP concentration (P < 0.01, n= 3 respectively) (Table 1). In H441 cells grown at the air–liquid interface, exposure to 3% O₂ for an hour did not affect the AMP : ATP ratio (n= 3). However, when exposed to 0.2% O₂ the AMP : ATP ratio was approximately fivefold higher than control (P < 0.001, n= 3) (Table 1).

Table 1.

Cellular adenine nucleotide concentration (nmol mg⁻¹ cell protein) determined in H441 cells grown in submerged culture or at the air–liquid interface and exposed to 3% or 0.2% O₂ for 60 min

	ATP	ADP	AMP	TAN	ADP : ATP	AMP : ATP
Submerged cultures
Control	13.8 ± 0.4	0.57 ± 0.05	0.12 ± 0.02	14.5 ± 0.4	0.04 ± 0.002	0.01 ± 0.002
3% O2	7.2 ± 0.04**	1.37 ± 0.08*	0.56 ± 0.06**	9.1 ± 0.2*	0.19 ± 0.01**	0.08 ± 0.01***
Air–liquid interface cultures
Control	30.2 ± 0.5	1.38 ± 0.12	0.18 ± 0.01	31.8 ± 0.6	0.046 ± 0.004	0.006 ± 0.001
3% O2	31.2 ± 0.7	1.38 ± 0.05	0.30 ± 0.01	33.0 ± 0.7	0.044 ± 0.002	0.009 ± 0.001
0.2% O2	31.1 ± 0.2	1.33 ± 0.12	0.91 ± 0.06***	33.3 ± 0.1	0.043 ± 0.004	0.029 ± 0.001***

Results are shown as mean ± SEM. Significantly different from control: *P < 0.05, **P < 0.01, ***P < 0.001 (n= 3). TAN, total adenine nucleotides.

Effect of hypoxia on the activity of AMPK

In human H441 airway epithelial cells grown in submerged culture, exposure to 3% O₂ significantly increased the phosphorylation of Thr¹⁷² of the AMPK α-subunit in a time-dependent fashion (pAMPK/tAMPK, P < 0.05, n= 4). Phosphorylation of Ser⁷⁹ of ACC was also observed to increase with time (Figure 1A and B). Treatment with AICAR, an AMP-mimetic activator of AMPK (2 mM for 60 min), significantly increased phosphorylation of AMPKα in H441 cells (P < 0.05, n= 4). Phosphorylation of ACC followed a similar pattern (Figure 1A and B).

Figure 1.

Time course of AMPK activation in H441 cells grown in submerged culture. H441 cells were treated with 2 mM AICAR for 60 min or exposed to 3% O₂ for 5, 15 or 60 min. (A) Typical Western blot of cell extracts immunoblotted for phospho-Ser⁷⁹ of ACC (pACC), total ACC (tACC), phospho-Thr¹⁷² of AMPKα (pAMPK) or total AMPKα (tAMPK). (B) Densitometry analysis of protein abundance pAMPK/ tAMPK and pACC/tACC from Western blots as shown in (A). Results are shown as mean ± SEM. Significantly different from control: *P < 0.05, **P < 0.01, ***P < 0.001 (n= 4).

In monolayers grown at the air–liquid interface, exposure to 3% O₂ for 60 min did not increase the phosphorylation of AMPKα and ACC (phosphoAMPK /totalAMPK and phosphoACC/totalACC) (Figure 2A and C). However, consistent with the observed increase in the AMP : ATP ratio, exposure to 0.2% O₂ for 15 or 60 min significantly elevated phosphorylation of AMPKα (P < 0.01, n= 4) and (P < 0.05, n= 4), respectively (Figure 2A–D).

Figure 2.

AMPK activation in H441 cells grown at the air–liquid interface. (A) Typical Western blot of extracts from cells grown at the air–liquid interface and maintained at 21% O₂ (control) or exposed to 3% or 0.2% O₂ for 60 min and immunoblotted for phosphor-Ser⁷⁹ of ACC (pACC), total ACC (tACC), phospho-Thr¹⁷² of AMPKα (pAMPK) or total AMPKα (tAMPK). (B) Typical Western blot of extracts from cells exposed to 0.2% O₂ for 15 or 60 min and immunoblotted for phospho-Ser⁷⁹ of ACC (pACC), total ACC (tACC), phospho-Thr¹⁷² of AMPKα (pAMPK) or total AMPKα (tAMPK). (C and D) Densitometry analysis of protein abundance pAMPK/ tAMPK and pACC/tACC from Western blots as shown in (A) and (B), respectively. Results are shown as mean ± SEM. Significantly different from control: *P < 0.05, **P < 0.01 (n= 4).

Addition of 0.1 mL·cm⁻² of PSS to the apical surface increased phosphorylation of AMPK and ACC (P < 0.05, n= 3, respectively) in cells exposed to 3% O₂. Furthermore, we observed that addition of 0.1 mL·cm⁻² of fluid to the apical compartment mildly activated AMPK in cells exposed to 21% O₂ (Figure 3A–C). These data indicate that fluid submersion of cells increases sensitivity to hypoxia.

Figure 3.

AMPK activation in H441 cell monolayers grown at the air–liquid interface or liquid–liquid interface. (A) Monolayers grown at the air–liquid interface (A) or with 0.1 mL·cm⁻² added to the apical surface (liquid–liquid interface, L) were maintained at 21% O₂ (control) or exposed to 3% or 0.2% O₂ for 60 min. Typical Western blot of cell extracts immunoblotted for phospho-Ser⁷⁹ of ACC (pACC), total ACC (tACC), phospho-Thr¹⁷² of AMPKα (pAMPK) or total AMPKα (tAMPK). (B) Densitometry analysis of protein abundance pACC/ tACC. (C) Densitometry analysis of protein abundance pAMPK/tAMPK from Western blots as shown in panel A. Results are shown as mean ± SEM. ^§Significantly different from control at the air–liquid interface, P < 0.05 (n= 3); ^†significantly different from control at the liquid–liquid interface, P < 0.05 (n= 3); *significantly different from air–liquid interface at the same % O₂, P < 0.05 (n= 3).

As we wished to investigate the regulation of AMPK activity and whether AMPK mediated effects on Na⁺K⁺ ATPase and/or apical amiloride-sensitive Na⁺ conductance, in all subsequent experiments we only used hypoxic situations in which AMPK was activated. As some drugs needed to be administered to the apical surface for efficacy, these were either cells grown at the air–liquid interface and exposed to 0.2% O₂ or cells exposed to 3% or 0.2% O₂ at the liquid–liquid interface.

Effect of ROS inhibitors on AMPK activity

Under the conditions tested, our data indicated that AMPK was only activated in cells where there was an increase in the AMP : ATP ratio. However, hypoxic elevation of ROS may also contribute to the activation of AMPK. We therefore investigated whether N-acetylcysteine (NAC; a scavenger of H₂O₂) and TEMPOL (a scavenger of superoxide, O₂^-) had any effect on AMPK activity. Pretreatment with NAC (10 µM) had no effect on phosphorylation of AMPKα and ACC (n= 3) in cells exposed to 3% or 0.2% O₂ at the liquid–liquid interface (Figure 4A and B). Similarly, TEMPOL (3 mM) had no observable effect on hypoxia-induced phosphorylation of ACC (n= 2) (Figure 4C). These data indicate that in H441 cells, activation of AMPK is not dependent on changes in intracellular H₂O₂ or superoxide.

Figure 4.

Effect of NAC and TEMPOL on hypoxia-induced AMPK activation. (A) Cells grown at the liquid–liquid interface were maintained at 21% O₂ (control) or exposed to 3% or 0.2% O₂ for 60 min in the presence or absence of NAC. Typical Western blot of cell extracts immunoblotted for phospho-Ser⁷⁹ of ACC (pACC), total ACC (tACC), phospho-Thr¹⁷² of AMPKα (pAMPK) or total AMPKα (tAMPK). (B) Densitometry analysis of protein abundance pAMPK/ tAMPK and pACC/tACC from Western blots as shown in (A). Results are shown as mean ± SEM. Significantly different from control: *P < 0.05, **P < 0.01 (n= 3). (C) Cells grown at the liquid–liquid interface were maintained at 21% O₂ (control) or exposed to 3% or 0.2% O₂ for 60 min in the presence or absence of TEMPOL. Typical Western blot of cell extracts immunoblotted for phospho-Ser⁷⁹ of ACC (pACC), total ACC (tACC). This blot was repeated with similar results.

Effect of inhibition of AMPK upstream kinases on hypoxia-induced activation of AMPK

LKB1 and CaMKK are major upstream kinases of AMPK. In some cell types, hypoxia can elevate intracellular Ca²⁺ and activate CaMKK (Gusarova et al., 2011). Therefore, we tested whether CaMKK played a role in the phosphorylation of AMPK in H441 cells after exposure to hypoxia. A brief (2 min) exposure to the Ca²⁺ ionophore ionomycin (1 µM) evoked a robust increase in intracellular Ca²⁺ in H441 cells preloaded with the Ca²⁺ indicator fluo-3AM (Liu and Hermann, 1978). Intracellular Ca²⁺ concentration was restored following washout of ionomycin from extracellular media consistent with findings in other cell types (Clair et al., 2001; Abramov and Duchen, 2003). This indicated that Ca²⁺ uptake by the intracellular stores and Ca²⁺ extrusion from the cells remained viable (Figure 5A and B). However, we were unable to show any changes in intracellular Ca²⁺ concentration after exposure to hypoxia. Treatment with ionomycin for 15 min increased the phosphorylation of AMPKα and ACC in H441 cells grown at the air–liquid interface. The phosphorylation of AMPKα was attenuated by pretreatment with the CaMKK inhibitor STO-609 (Figure 5C). These data indicate that elevation of Ca²⁺ and activation of CaMKK increased AMPK phosphorylation in H441 cells in the absence of metabolic conditions that would alter the AMP : ATP ratio. However, STO609 did not prevent phosphorylation of AMPKα and ACC in cells grown in submerged culture and exposed to 3% O₂ for 60 min (Figure 6A, B and F) or in cells grown at the air–liquid interface and exposed to 0.2% O₂ for 60 min (Figure 6C and D). As a positive control, STO609 inhibited the elevation of phosphorylated AMPKα in human alveolar A549 cells grown in submerged culture and exposed to 3% O₂ (Figure 6E) consistent with the findings of (Gusarova et al., 2011).

Figure 5.

Effect of ionomycin on intracellular Ca²⁺ and activation of AMPK in H441 cells. (A) Representative fluorescent confocal x–y images of H441 cells taken before, during and after exposure to 1 µM ionomycin. (B) Plot of the time course of the normalized fluorescence averaged from a group of ∼20 H441 cells briefly exposed to ionomycin (1 µM). (C) H441 cells grown at the air–liquid interface were treated with ionomycin (15 min) with or without pretreatment (30 min) with the CaMKK inhibitor STO609 (STO609). Typical Western blots of cell extracts immunoblotted for phospho-Ser⁷⁹ of ACC (pACC), total ACC (tACC), phospho-Thr¹⁷² of AMPKα (pAMPK) or total AMPKα (tAMPK). These blots were repeated with similar results.

Figure 6.

Effect of STO609 on activation of AMPK in H441 cells exposed to hypoxia. (A) Typical Western blots of cell extracts from H441 cells grown at the liquid–liquid interface, with or without pretreatment with the CaMKK inhibitor STO609 and maintained at 21% O₂ (control) or exposed to 3% for 60 min and immunoblotted for phospho-Ser⁷⁹ of ACC (pACC), total ACC (tACC), phospho-Thr¹⁷² of AMPKα (pAMPK) or total AMPKα (tAMPK). (B) Typical Western blots of cell extracts from H441 cells grown at the air–liquid interface, with or without pretreatment with STO609 and maintained at 21% O₂ (control) or exposed to 0.2% O₂ for 60 min and immunoblotted for phospho-Ser⁷⁹ of ACC (pACC), total ACC (tACC), phospho-Thr¹⁷² of AMPKα (pAMPK) or total AMPKα (tAMPK). (C and D) Densitometry analysis of protein abundance pAMPK/ tAMPK and pACC/tACC from Western blots shown in (A) and (B). Results are shown as mean ± SEM. Significantly different from control: *P < 0.05 (n= 3). (E and F) A549 and H441cells grown in submerged culture maintained at 21% O₂ (control), or exposed to 3% O₂ for 60 min before or after pretreatment with STO609 prior to exposure to 3% O₂, fixed and immunostained for pAMPK. pAMPK was visualized with a Texas Red secondary antiserum. Nuclei are stained blue using DAPI.

There are no specific inhibitors of LKB1. However, LKB1 is associated with Heat shock protein 90 (HSP90) and Cdc37 to prevent degradation by the proteosome (Nony et al., 2003; Sutherland et al., 2003). In H441 cells, geldanamycin (10 µM, 24 h), a pharmacological inhibitor of HSP90, markedly decreased LKB1 protein abundance to 40% of control (P < 0.05, n= 3) (Figure 7A and B) but did not affect total cellular AMPKα or ACC abundance (Figure 7C and D). A decreased LKB1 abundance reduced the phosphorylation of AMPKα in cells grown at the air–liquid interface and exposed to 0.2% O₂ for 60 min when compared with non-geldanamycin-treated cells (P < 0.05, n= 3). The phosphorylation of ACC was also concomitantly reduced (P < 0.05, n= 3) (Figure 7C and D). Thus, these data indicate that hypoxia-induced phosphorylation of AMPK in polarised H441 cells is associated with LKB1-mediated phosphorylation.

Figure 7.

Effect of geldenamycin on activation of AMPK in H441 cells exposed to hypoxia. (A) Typical Western blot of cell extracts from H441 monolayers grown at the air–liquid interface before and after treatment with geldnamycin (10 µM for 24 h) immunoblotted for LKB1 or β-actin. A lane of cell extract from A549 cells (which do not contain LKB1) is shown to the right-hand side as a negative control. (B) Densitometry analysis of protein abundance LKB1/β-actin from Western blots shown in (A). *Significantly different from control: P < 0.05 (n= 3). (C) Typical Western blot of extracts from cells grown at the air–liquid interface pretreated with geldnamycin before exposure to 0.2% O₂ for 60 min and immunoblotted for phospho-Ser⁷⁹ of ACC (pACC), total ACC (tACC), phospho-Thr¹⁷² of AMPKα (pAMPK) or total AMPKα (tAMPK). (These lanes were contained on the same gel shown in Figure 6C and utilize the same controls but have been presented as an independent figure for clarity). (D) Densitometry analysis of protein abundance pAMPK/tAMPK and pACC/tACC from Western blots as shown in (C). *Significantly different from control: P < 0.05 (n= 4). †Significantly different from untreated cells: P < 0.05 (n= 4).

Effect of hypoxia on transepithelial ion transport processes across H441 cell monolayers

We wanted to determine whether AMPK mediated the effects of hypoxia on transepithelial Na⁺ transport processes in H441 cells. Therefore, we investigated monolayers under conditions where AMPK was activated with Ussing chamber experiments performed on polarized H441 cell monolayers at the liquid–liquid interface exposed to 3% or 0.2% O₂ at for 60 min. Cells exposed to 3% and 0.2% O₂ exhibited a significant decrease in transepithelial I_amiloride from 20.9 ± 1.7 to 16.6 ± 1.4 µA·cm⁻² and 12.5 ± 1.4 µA·cm⁻² (P < 0.01, n= 4, respectively). I_ouabain was reduced from 102.9 ± 10.6 to 82.4 ± 11.8 and 66.4 ± 13.6 µA·cm⁻² (P < 0.05, n= 4, respectively) (Figure 8A–C). Amiloride-sensitive G_Na+ was also significantly reduced from 197 ± 37.4 to 131.3 ± 23.0 and 120 ± 9.8 µS·cm⁻² (P < 0.05, n= 4, respectively) (Figure 8D and E). Taken together, these data indicate that under these conditions, exposure to hypoxia inhibits both the apical and basolateral components of transepithelial Na⁺ transport. Interestingly, exposure to 0.2% O₂, but not 3% O₂, resulted in a significant increase in the transepithelial resistance from 467.3 ± 39.7 to 536.2 ± 49.2 Ω·cm² (P < 0.01, n= 4).

Figure 8.

Effect of hypoxia on transepithelial I_amiloride, basolateral I_ouabain and apical G_Na+. Cells grown at the liquid–liquid interface were maintained under 21% O₂ (control) or exposed to 3% or 0.2% O₂ for 60 min before being mounted in Ussing chambers. (A) Amiloride-sensitive transepithelial I_sc (I_amiloride). (B) Basolateral ouabain-sensitive I_sc (I_ouabain). (C) Typical current traces from which values for I_amiloride and I_ouabain were obtained. Spontaneous transepithelial I_sc was measured before the addition of amiloride (10 µM) to the apical bath over the time course shown in the box above the trace. Under the same recording conditions, nystatin (75 µM) was added to the apical chamber. When the current reached peak values, ouabain (1 mM) was added to the basolateral chamber. (D) Apical amiloride-sensitive Na⁺ conductance (G_Na+). (E) Typical current traces showing spontaneous I_sc (in potassium gluconate solution) before the addition of nystatin to the basolateral bath over the time course shown in the box above the trace. A Na⁺ gradient was then applied across the apical membrane as indicated, followed by the addition of amiloride (10 µM). G_Na+ was calculated from the amiloride-sensitive apical current (I_ap). Significantly different from control: *P < 0.05, **P < 0.01 (n= 4).

Effect of AMPK lentiviral vectors on transepithelial Na⁺ transport

Transduction of H441 cells grown at the air–liquid interface with lentivector containing a short hairpin sequence specific to AMPKα1/2 (sh-AMPK) resulted in a reduction of total AMPK abundance compared to control lentivector (Sh-negative) (Figure 9A). Exposure to 0.2% O₂ for 60 min resulted in a significant inhibition of I_ouabain to 76 ± 7% of control in Sh-negative transfected cells (P < 0.05, n= 4) but had no effect in the presence of Sh-AMPK (Figure 9B). Exposure to 0.2% O₂ inhibited G_Na+ to 60 ± 4 % (P < 0.01, n= 4) and to 69 ± 9 % (P < 0.05, n= 4) of control in Sh-negative and Sh-AMPK-treated cells (Figure 9C). Transduction with lentivector containing constitutively active AMPK (CA-AMPK) resulted in expression of CA-AMPK protein in addition to endogenous AMPK in H441 cells (Figure 10A). Under non-hypoxic control conditions, expression of CA-AMPK reduced I_ouabain to 70 ± 7% (P < 0.05, n= 4) of that determined in Sh-negative transfected cells (Figure 10B). Exposure to 0.2% O₂ for 60 min in the presence of CA-AMPK had no further effect (n= 4) (Figure 10C). These data support a role for AMPK in mediating the effects of hypoxia on Na⁺K⁺ ATPase but not apically localized amiloride-sensitive Na⁺ channels in H441 cells.

Figure 9.

Effect of expression of Sh-neg and Sh-AMPK lentiviral vectors on I_ouabain and G_Na+ in H441 cells. (A) Western blot of total AMPK in cells grown at the air–liquid interface and transduced with Sh-neg or Sh-AMPK lentivector. (B) I_ouabain in cells transduced with Sh-neg or Sh-AMPK lentivector maintained under 21% O₂ (control) or exposed to 0.2% O₂ for 60 min. (C) G_Na+ in cells transduced with Sh-neg or Sh-AMPK lentivector maintained under 21% O₂ (control) or exposed to 0.2% O₂ for 60 min. Data are shown as % control for comparative purposes. Significantly different from control: *P < 0.05, **P < 0.01 (n= 4).

Figure 10.

Effect of expression of CA-AMPK lentiviral vector on I_ouabain in H441 cells. (A) Western blot of total AMPK in cells grown at the air–liquid interface and transduced with Sh-neg or CA-AMPK lentivector. CA-AMPK is a smaller protein of 36 kDa, which runs below the endogenous protein. (B) Transepithelial I_ouabain in cells maintained at 21% O₂ and transduced with CA-AMPK compared to that of cells transduced with Sh-neg lentivector. Data are shown as % Sh-neg for comparative purposes. Significantly different from control *P < 0.05 (n= 3). (C) I_ouabain in cells transduced with CA-AMPK lentivector maintained at 21% O₂ (control) or exposed to 0.2% O₂ for 60 min (for reference, the effect of 0.2% O₂ for 60 min on I_ouabain in cells transduced with Sh-neg is shown in Figure 8C). Data are shown as % control for comparative purposes. *Significantly different from control: P < 0.05 (n= 4).

Effect of NAC, TEMPOL and apocynin on apical G_Na+

H441 cells grown at the liquid–liquid interface were pretreated with NAC (10 mM) or TEMPOL (3 mM) or apocynin (300 µM) and exposed to 3% or 0.2% O₂ for 60 min. The transepithelial resistance (R_t) in NAC treated cells was 395.5 ± 36.0 Ω cm², and this was not significantly affected by exposure to 3% or 0.2% O₂ (487.9 ± 81.5 and 512.5 ± 68.0 Ω cm², n= 3, respectively). However, in the presence of NAC, exposure to 3% or 0.2% O₂ inhibited amiloride-sensitive G_Na+ from 170.4 ± 7.2 to 110.0 ± 10.8 (P < 0.01, n= 3) and 78.0 ± 2.7 µS cm⁻² (P < 0.001, n= 3), respectively (Figure 11A). The R_t in TEMPOL-treated cells was 403.0 ± 96.0 Ω cm² and was also not affected by exposure to 3% or 0.2% O₂ (442.5 ± 91.6 and 430.8 ± 125.8 Ω cm², n= 3 respectively). The apical amiloride-sensitive G_Na+ was reduced by TEMPOL to 50.4 ± 4.7 µS cm⁻² compared with untreated cells and was not further inhibited by exposure to 3% or 0.2% O₂ for 60 min (30.2 ± 7.7 and 47.9 ± 4.2 µS cm⁻², n= 3 respectively) (Figure 11B). In apocynin-treated monolayers, amiloride-sensitive G_Na+ was 255.0 ± 33.3 18.3 µS cm⁻², and exposure to 3% or 0.2% O₂ had no significant effect (251.9 ± 2.1 and 216.8 ± 18.3 µS cm⁻², n= 3 respectively) (Figure 11C). Taken together, these data indicate that TEMPOL and apocynin prevented the effect of hypoxia on apical G_Na+.

Figure 11.

Effect of NAC, TEMPOL and apocynin on apical amiloride-sensitive G_Na+. Cells grown at the liquid–liquid interface were pretreated with NAC, TEMPOL or apocynin and maintained at 21% O₂ (control) or exposed to 3% or 0.2% hypoxia for 60 min. G_Na+ was measured in basolaterally permeabilized monolayers after apical application of 10 µM amiloride. Significantly different from control: *P < 0.05, **P < 0.01, ***P < 0.001 (n= 3).

Discussion and conclusions

Exposure to hypoxia activated AMPK in H441 airway and A549 distal lung epithelial cells. However, we found that there was a difference in the response to hypoxia between H441 cells grown in submerged culture or at the liquid–liquid interface and those grown at the air–liquid interface. In cells grown in submerged culture, exposure to 3% O₂ was sufficient to increase the AMP : ATP ratio and activate AMPK, consistent with findings that 3% O₂ activated AMPK in rat alveolar type II cells (Gusarova et al., 2009). In contrast, in cells grown at the air–liquid interface, only 0.2% O₂ increased the AMP : ATP ratio and activated AMPK. Moreover, after exposure to 0.2% O₂, the adenine nucleotide concentrations were comparable with those we have previously published for pharmacological activation of AMPK with phenformin (a mitochondrial inhibitor). Given that mitochondrial PO₂ is approximately 0.4% O₂, it would seem reasonable that exposure to 0.2% O₂ would disrupt mitochondrial function and energy generation. Interestingly, whilst phenformin robustly activated AMPK at 60 min (similar to 0.2% O₂), a 4 h exposure to metformin was required to activate AMPK (Woollhead et al., 2007). Thus, a longer exposure to 3% O₂ may be necessary to elicit changes in the AMP : ATP ratio and activation of AMPK in cells grown at the air–liquid interface.

That apical fluid modified the response to hypoxia was further supported by our finding that the addition of fluid to the apical surface (0.1 mL·cm⁻²) of cells grown at the air–liquid interface resulted in activation of AMPK by 3% O₂. Thus, the presence of fluid ‘sensitizes’ the response to reduced PO₂. O₂ is poorly soluble (approx 0.231 mL L⁻¹ kPa⁻¹) in aqueous solution, and Fick's law of diffusion predicts that medium between the surface of the cells and the gaseous environment produces a barrier to the diffusion of O₂ to the cell surface and subsequently the mitochondria (Allen et al., 2001). A number of studies have investigated how fluid influences the PO₂ of oxygen near the cell membrane using oxygen electrodes (Metzen et al., 1995; Allen et al., 2001; Bambrick et al., 2011; Lynn et al., 2011). These studies indicate that pericellular PO₂ is lower than the environmental gas PO₂, even in hypoxia. The level of pericellular PO₂ is dependent on fluid depth, the number and type of cells. Thus, if O₂ demand exceeds O₂ diffusion, the pericellular PO₂ falls. It is likely that both proliferating H441 cells and those at the liquid–liquid interface, which are actively transporting ions, have high O₂ demand (Metzen et al., 1995). Therefore, the apical pericellular PO₂ of cells in submerged culture could be lower than those at the air–liquid interface. This would explain why 3% O₂ activated AMPK in H441 cells grown in submerged culture and at the liquid–liquid interface. It would also explain our observation that AMPK was mildly activated when polarized epithelial cells at the liquid–liquid interface were exposed to normoxia (∼21% O₂). In support of the latter findings, the addition of 5–10 mm depth of apical fluid was shown to elevate a number of hypoxic markers, including HIF1α, in human colonic HT29 cells grown as monolayers in normoxia (Guimbellot et al., 2008). The changes in the pericellular PO₂ of submerged cultures are also likely to be more rapid in the face of a reduction rather than an increase in gaseous PO₂ (Bambrick et al., 2011). For example, the pericellular PO₂ of cultured A549 lung epithelial cells exposed to hypoxia decreased within 15 min (Allen et al., 2001), consistent with the time course of activation we determined for AMPK. We cannot rule out the possibility that increased apical fluid volume may also alter other factors that could contribute to increased cellular stress [i.e. apical pressure, rapid activation of ENaC (Tan et al., 2011)]. However, we suggest that apical fluid restricted diffusion of O₂ to the cell surface resulting in a lowering of PO₂ at the cell membrane and potentiation of the hypoxic effect, which led to energy restriction and an alteration of the cellular AMP : ATP ratio. A future study, utilizing oxygen electrodes to determine exactly the PO₂ at the cell membrane required to activate AMPK, would now be useful.

These findings are important because much work on respiratory cells has utilized cells grown in submerged culture to study effects of altered PO₂. We are proposing that this may not accurately reflect the normal in vivo conditions, which are more like those of cells grown at the air–liquid interface. Thus, in situations where lung apical fluid volume is increased (pulmonary oedema), more physiologically relevant levels of hypoxia could activate AMPK in lung epithelial cells.

There is mounting evidence that AMPK can be activated by hypoxic elevation of ROS in the absence of changes in the AMP : ATP ratio (Emerling et al., 2009; Gusarova et al., 2011). In addition, an elevation of intracellular Ca²⁺ leading to activation of CaMKKβ and phosphorylation of AMPK was shown to mediate the effects of hypoxia in A549 distal lung epithelial cells (Gusarova et al., 2011). However, our data do not currently support these pathways in H441 airway cells. Under the conditions investigated, AMPK was not activated in the absence of a change in the AMP : ATP ratio. Furthermore, in monolayers grown at the liquid–liquid interface, AMPK was activated by both mild and extreme hypoxia, but neither of the ROS scavengers NAC or TEMPOL had any effect on the phosphorylation of AMPK or ACC. Moreover, although elevation of intracellular Ca²⁺ with ionomycin resulted in activation of AMPK via a STO609-sensitive pathway in H441 cells, hypoxia-induced phosphorylation of AMPK was not affected by this inhibitor. Conversely, hypoxic activation of AMPK in A549 cells was inhibited by STO609, consistent with the findings of Gusarova et al. (2011). Thus, these data indicate that ROS and CaMKK do not play a critical role in hypoxic activation of AMPK in H441 cells.

The reduction of LKB1 protein with geldenamycin suppressed the hypoxia-induced AMPK activity indicating that LKB1-mediated phosphorylation is the primary route for activation of AMPK. We were unable to completely inhibit hypoxic activation of AMPK using this method, probably because of the residual activity of the remaining LKB1 protein. We therefore propose that hypoxic activation of AMPK in H441 epithelial cells requires a change in the cellular AMP : ATP ratio and LKB1 activity. These results contrast with findings in A549 and type II alveolar epithelial cells (Gusarova et al., 2011). This may reflect differences between the cell types, sensitivity to changes in PO₂ and the signalling pathways utilized in the distal lung versus the airway. It is important to note that A549 cells do not contain LKB1 and may therefore preferentially exploit alternative hypoxia-sensing pathways.

We showed for the first time that, under conditions where AMPK is activated, hypoxia (3% and 0.2% O₂) inhibits transepithelial I_amiloride, I_ouabain and G_Na+ across H441 airway epithelial cells. These findings are consistent with studies in human A549 and rat alveolar type II cells where exposure to hypoxia (1.5–3%) inhibited Na⁺ transport processes (Heberlein et al., 2000; Wodopia et al., 2000; Mairbaeurl et al., 2002). Interestingly, transepithelial resistance was also increased in response to hypoxia. Similar changes in resistance have been associated with pharmacological activation of AMPK in H441 cells (Woollhead et al., 2005; 2007; Mace et al., 2008).

The effect of hypoxia on I_ouabain was disrupted by the expression of a shRNA construct that reduced AMPK protein abundance in the cells. Furthermore, the expression of constitutively active α-AMPK inhibited I_ouabain and prevented any further inhibition by hypoxia. These data indicate that hypoxic activation of AMPK inhibited Na⁺K⁺ ATPase in H441 cell monolayers. These findings are consistent with our previous observations that pharmacological activation of AMPK inhibits Na⁺K⁺ ATPase in H441 cells (Woollhead et al., 2005; 2007; Albert et al., 2008). Exposure to hypoxia (1.5% O₂) has also been shown to inhibit Na⁺K⁺ ATPase function in human A549 and rat alveolar type II cells via AMPK (Gusarova et al., 2009). However, we were unable to inhibit the effect of hypoxia on apical G_Na+ using such constructs, even though there is evidence that activation of AMPK inhibits apical G_Na+ and ENaC channels in these and other epithelial cells (Carattino et al., 2005; Woollhead et al., 2005; 2007; Bhalla et al., 2006; Albert et al., 2008). We cannot rule out the possibility that more prolonged exposure to hypoxia could elicit AMPK-mediated effects on apical channels. However, in renal collecting duct cells, pre-activation of AMPK with AICAR did not mitigate the inhibitory effect of hypoxia (8% O₂ for 24 h) on Na⁺ transport, indicating that the apical response to hypoxia involves other AMPK-independent mechanisms (Husted et al., 2011).

Hypoxia modifies intracellular ROS, and there is evidence that amiloride-sensitive ENaC channels are acutely regulated by ROS (Helms et al., 2008; Althaus et al., 2009; Yip et al., 2010). That NAC did not prevent the hypoxic inhibition of apical G_Na+ indicates that changes in H₂O₂ were not involved. However, both TEMPOL and apocynin prevented the effect of hypoxia on apical G_Na+, indicating a potential role for superoxide (O₂^-) in the hypoxic suppression of apical ENaC channel activity. TEMPOL is a scavenger of O₂^- that can be generated by a number of cellular sources including NADPH oxidase. In A6 renal cells, aldosterone elevation of intracellular O₂^- increased ENaC activity, which was abolished in the presence of TEMPOL (Yu et al., 2007). This would be consistent with our finding that TEMPOL reduced G_Na+ and prevented any further suppression of ENaC activity by hypoxia. Apocynin inhibits NADPH oxidase (NOX) activity in the plasma membrane, which is responsible for the generation of O₂^- from O₂, potentially providing a localized source of O₂^- proximal to the Na⁺ channel. Our data support the idea that when O₂ availability is reduced, the localized production of O₂^- by NOX is compromised, leading to a reduction in Na⁺ channel activity. More work is now required to identify how hypoxia affects different sources of intracellular O₂^- and how these contribute to the regulation of apical Na⁺ channels. In addition, O₂^- may not regulate ENaC channels directly. For example, O₂^- can sequester NO, which has been shown to inhibit ENaC channels (Althaus et al., 2011; Husted et al., 2011).

In conclusion, we have shown that a 60 min exposure to hypoxia activated AMPK in H441 airway epithelial cells. The sensitivity to hypoxia was increased by the addition of fluid to the apical surface of cell monolayers grown at the air–liquid interface, which may have important consequences in lung disease conditions. Activation of AMPK was associated with an elevation of the cellular AMP : ATP ratio and the presence of the upstream kinase LKB1. Hypoxia inhibited amiloride-sensitive transepithelial Na⁺ transport by inhibiting both basolateral Na⁺K⁺ ATPase and apical amiloride-sensitive Na⁺ channel activity. However, the signalling pathways regulating apical and basolateral responses to hypoxia were different. AMPK mediated the effect of hypoxia on Na⁺K⁺ ATPase, but the effect on apical Na⁺ channels involved the action of superoxide O₂^-. Therefore, we propose that hypoxia simultaneously activates AMPK-dependent and -independent pathways to decrease transepithelial Na⁺ transport in polarized H441 airway epithelial cells.

Glossary

AICAR

5-aminoimidazole-4-carboxamide-1-β-d-ribofuranoside

AMPK

AMP-activated protein kinase

CaMKK

Ca²⁺/calmodulin kinase kinase

ENaC

epithelial sodium channel

G_Na+

amiloride-sensitive apical conductance

I_amiloride

amiloride-sensitive transepithelial I_sc

I_ouabain

ouabain-sensitive basolateral current

I_sc

short circuit current

LKB1

liver kinase B1

NAC

N-acetylcysteine

NOX

NADPH oxidase

R_t

transepithelial resistance

TAK1

TGF-β-activated kinase 1

TEMPOL

nitrox 2,2,6,6-tetramethylpiperdin-l-yloxy

V_t

transepithelial potential

Conflicts of Interest

None.