A nonselective cation channel in adult alveolar epithelial cells

OBJECTIVE: Efficient gas exchange across the alveolar membrane requires a dry air space. At birth, active Na

Un canal cationique non selectif dans les cellules epitheliales alveolaires de l'adulte OB.JECTIF : Pour etre efficace, !es echanges gazeux a travers la membrane alveolaire exigent un espace aerien sec.A la naissance, le transport actif du Na+ fournit la force motrice pour !'absorption du liquide pulmonaire foetal.A !'age adulte, des mecanismes de transport similaires regulent Jes echanges liquidiens a travers la membrane alveolo-capillaire.Un canal cationique detecte anterieurement dans !'epithelium alveolaire foetal pourrait representer l'une des differentes voies de conduction qui regule le flux apico-basal transepithelial du sodium.L'objet de cette etude etait de determiner si des canaux cationiques etaient aussi presents dans les cellules epitheliales alveolaires de l'adulte.MODELE : Utilisation de la technique du patch-clamp exposant la face cytoplasmique des canaux pour etudier les canaux presents dans les fragments membranaires de !'epithelium alveolaire du rat adulte.POPULATION : Des cellules epitheliales alveolaires ont ete prelevees chez des rats Sprague-Dawley, males, adultes, et etudiees par culture cellulaire primaire.PRINCIPAUX RESULTAlS: On a identifie Lill canal cationique 23 pS non selectif dans 68 % des fragments au potentiel impose de -60 mY.La probabilite de l'ouverture d'un canal n' a pas ete influencee par les variations de potentiel transmembranaire.De l'amiloride ( I 0-5 M) applique sur la membrane extracellulaire a induit une stimulation du canal et a reduit le temps moyen apparent de l'ouverture du canal sans en affecter la conductance.CONCLUSIONS : Ces resultats fournssent la preuve directe de I' existence d' un canal cationique dans la membrane apicale des cellules epitheliales alveolaires de l' adulte.Le canal est identique a un canal cationique non selecti f identifie dans les cellules epitheliales alveolaires foetales et peut jouer un role important dans la reabsorption liquidienne couplee au Na+ et/ou dans la secretion du K+.A 'QUIET REVOLUTION' HAS OCCURRED OVER THE PAST 10 years in our understanding of the forces that regulate fluid movement across the alveolar-capillary membrane ( 1-3).While fluid balance has traditionally been analyzed in terms of hydrostatic and colloid-osmotic pressure gradients, increasing evidence suggests that acti ve ion transport provides the major driving force for absorption of fluid from the alveolar space.The sole importance of 'Starling Forces' was challenged when Matthay et al (3,4) demonstated that proteinaceous fluid instilled into lungs of sheep, was absorbed despite the presence of a large opposing osmotic pressure gradient.The relative contribution of passive and active forces was investigated by Basset et al (5).They determined that human lung depends primarily on active transport mechanisms similar to other fluid absorbing surfaces and estimated that the human lung is capable of absorbing 1.5 L of fluid over a 24 h period (5).
The mechanisms that regulate solute-coupled fluid transport in the lung are of considerable cli nical importance.Several diseases, including respiratory distress syndrome (RDS) in the newborn and pulmonary edema in the mature lung, result from a fluid imbalance across the alveolarcapillary membrane.O'Brodovich et a l (6) reported that an 'RDS-like syndrome' was induced in full-term, newborn guinea-pigs when the Na+ transport inhibitor amiloride was instilled into the lungs before the first breath.This observation raises the possibility that defects in ion transport, as well as surfactant deficiency, contribute to d isease in the premature lung.In adult humans, survival from acute lung injury correlates with the restoration of normal ion transport, suggesting that solute-coupled transport is essential for recovery from alveolar flooding (7).
Determining which cells are responsible for lung fluid clearance has been complicated by the complex anatomy of the lung.Cell culture techniques that allow select populations of cells to be studied in vitro have provided insights into the transport capabi Ii ties of the alveolar epithelium.Isolated type II alveolar epithelial cells, grown on porous supports, spread to form confluent monolayers (8,9).Osmotic forces generated by the active transport of ions from the apical to basolateral membrane cause fluid to accumulate beneath these cells.The resulting dome-like structures are characteristic of fluid absorbing epithelia and result from Na +-coupled transport.The bioelectric properties of al veolar epithelial cells have also been studied in Ussing chambers.Fetal and adult alveolar epithelial cells form 'tight' , high resistance membranes and generate an amiloride-sensitive, Na +-dependent, shortcurcuit current (9)(10)(11).These data indicate that alveolar epithelial cells have the bioelectric characteristics necessary for active ion transport ( 12).
The rate of solute translocation across the alveolar epithelium suggests that the apical site for Na+ entry is an ion channel rather than a cotransport system (13,14). 22Na+ flux measurements from membrane vesicles prepared from adult alveolar epithelium and whole-cell patch-clamp recordings support the presence of amiloride-sensitive Na+ channels (13,14).Recently, an amiloride-sensitive channel which was permeable to Na+ and K+ was identified in fetal alveolar epithelium (15,16).To examine the possibility that fetal and adult alveolar cells use similar apical Na+ transport pathways, we attempted to determine if this channel is also present in adult alveolar epithelial cells.Part of this work was previously published in abstract form ( 17).

Primary culture of adult epithelial cells
Alveolar epithelial cells were obtained from the lungs of adult male Sprague-Dawley rats using the method of Dobbs et al (18).In brief, alveolar cells were separated from the basement membrane by incubation with porcine pancreatic elastase followed by removal of macrophages by differential adherence.Cells were seeded at a plating density of 5xl0 5 cells/en/ on collagen coated plastic petri dishes and cultured in a humidified incubator (95% air, 5% carbon dioxide) at 37°C for 24 to 48 h.Alveolar epithelial cells grown under these culture conditions develop the morphological characteristics of polarized epithelia ( JO, 19).It was therefore assumed that a patch excised from the surface of these cells was obtained from the apical membrane.The purity of alveolar epithelial cell cultures was measured by histochemical staining for alkaline phosphatase (20) and was consistently greater than 93 %.

Single channel recording and analysis techniques
The composition of solutions used in the experiments is indicated in Table I.All sol utions contained 1.5 mM Ca 2 + as it was previously observed that the stability of the patch improved when high concentrations of Ca 2 + were present in the recording media (15).All solutions were fi ltered (0.22 µM) (Millipore Products, Massachusetts) before use.
Electrodes were constructed from borosilicate glass ( 1.5 mm outer diameter) (World Precision Instruments Inc, Florida) coated with Sylgard (Dow Corning, Michigan) and fire polished (Narishige, Tokyo, Japan).Conventional techniques were used to acquire inside-out membrane patches and all experiments were performed at room temperature.Cells were visualized using an inverted microscope (Zeiss, Germany) and currents were recorded using an Axo-patch ampl ifier (Axon Instruments, California).
The connection of the perfusion chamber to the ground was established using a silver-silver chloride pellet electrode, whereas the pipette was coupled to the headstage using a silver-silver chloride wire.At the beginning of each experiment, the recording pipette was immersed in a solution containing 140 mM sodium chloride (solution A) and the amplifier potential was set to zero.It was assumed that no junctional potentials developed when the bath solution was exchanged with a solution containing 140 mM potassium chloride (solution B).However, when the bathing solution was replaced with one containing 47 mM sodium chloride (solution C), importantjunctional potentials were expected to occur.These potentials could influence the accuracy of the measured zero cunent potentials and hence the calcu lation of relative permeability of the channel to anions and cations.
We measured the value of the junctional potentials (Vu) using membrane-free electrodes filled with the standard pipette solution (2 l ).The pipette was immersed in the bathing solution (solution A) and the amplifier zeroed.The bathing solution was then exchanged with solution C and the amplifier potential was set to zero several times over the nex t 5 mins .The amplitude of the compensating potential was noted.When the variation in the compensating voltage was less than I mV over 2 to 3 mins, the signal was considered to be stable.Junctional potentials were measured for seven electrodes and the mean Vu was added to the measured zero current potential (VMeasured) in order to obtain the reversal potential (V Rev) The permeabi lity of the channel for Cl-anions was expressed as the C l-permeability to Na+ permeability ratio (Po/PNa) and was analyzed using the Goldman-Hogkin-Katz (GHK) equation for the reversal potential.Activities rather than concentrations were used for all calculations of ion permeability .
Cunent signals were filtered at I kHz and recorded onto an FM tape.For analysis, data records were played back off the tape, sampled at I 00 to 200 µs and stored on a personal computer.Records were analyzed using pCLAMP program (Axon Instruments) and single-channel events were detected using the 50% threshold crossing method .Dwell times were reported as the arithmetic mean of channel open times .Single channel conductance was determined using amplitude histograms and the slope of the current to voltage (i-v) curve.Membrane potential refers to the intracellular or cytoplasmic potential relative to extracellular or pipette potential, which is assumed to be zero .Inward currents arc depicted as downward deflections in all current tracings.The probability of channel opening (Popen) was determined as previously reported using the relative areas of the all-points histograms ( 15).A one-way analysis of variance (ANOV A) was used to V (Vm)

RESULTS
Adult alveolar epithelial cells grown on a collagen matrix adopted a typical 'cobblestone' appearance when examined using inverted light microscopy .Patches were excised from cells located in the centre of the clusters.High resistance seals (giga-Q) were difficult to form on the surface of the cells and excised patches frequently became unstable during exchanges of the bathing solution or following membrane depolarization.In contrast to alveolar epithelium, the present authors have studied other cell types includ ing cultured neurons and transfected epithelia without difficulty using the same techniques and recording equ ipment.
With symmetrical concentrations of sodium chloride ( 140 mM, solution A) in the pipette and bath and the membrane voltage clamped to O m V, no current events were observed.Membrane hyperpolarization evoked inward current events and membrane depolarization induced outward current, respectively.Channel openings were observed in 68% (24 of 35) of patches voltage clamped to -60 mV .In most recordings, a single conductance state was evident suggesting the presence of a single channel.In 2 1 % of active patches, two or more conductance level s of equal ampl itude were observed.The relationship between current (i) and voltage (v) was plotted and the slope of the line of linear regression used to estimate channel conductance.The i-v curve was linear with a reversal potential close to O m V (Figure 1).The

V(mV)
.,..  E is the membrane potenlia/, Na+i and Na+ e are 47 111M and 140 mM, respectively.PNa was eslimated to be 7. 2 10-14 cm 3 !vpletely selective for cations.There are several possible reasons why these data deviated slightly from the theoretical curve.The GHK current equation, similar to the GHK voltage equation, depends on two important assumptions: ions move independently through the channel pore and the electrical field in the membrane is constant with membrane potential decreasing linearly across the cell membrane (22).In reality, biological membranes are not homogeneous slabs, and ion fluxes are not linearly proportional to ion concentrations.Interactions between ions and energy barriers within the channel pore may have reduced the movement of ions across the membrane.Voltage sensitivity of channel gating was investigated by examining the relationship between membrane potential and the probability of channel opening (Papen).The probabi lity of channel opening was not significantly influenced by changes in membrane potential (Figure 4).For example, at a holding potential of -40 mV, the probability of channel opening was 0.39±0.07(n=4) whereas Papen was 0.34±0.12(n=4) when the membrane was depolarized to +40 m V.
The addition of amiloride ( I o-5 M) to the pipette solution (solution D) consistently altered channel openings from long rectangular events to brief flickering events (Figure 5).The duration of channel opening was significantly reduced in amiloride versus control patches when measured at a mem-

DISCUSSION
The patch-clamp method has recently provided significant insights into the role ion channels play in respiratory disease (23)(24)(25)(26).Despite the importance of the alveolar membrane for normal lung function, studies of channels present in alveolar epithelial cells have been limited.T his is due, in part, to difficulties experienced by ourselves and others in obtaining high quality recording conditions with primary epithelial cells grown in dissociated cell cultures (27).This report describes a 23 pS nonselective cation channel present in the apical membrane of adult alveolar epithel ial cells.The channel was permeable to Na+ and K+ but relatively impermeable to anions.The probability of channel opening was insensitive to changes in membrane potential and amiloride-induced channel flickering and reduced channel open time.
The single channel characteristics of the nonselective cation channel present in fetal and adult alveolar epithelial cells are remarkably similar.The 23 to 25 pS fetal channel is also voltage-insensitive, selective for K\ impermeable to Cl-and blocked by amiloride (3,4 ).In contrast, in our findings, the selectivity and kinetic properties of several other channels studied in fetal and adult cells change dramatically during the first few weeks of postnatal life (28)(29)(30).For example, the acetylcholine receptor/ionophore is a nonselective cation channel present in the neuromuscular junction.During early postnatal development, channel conductance increases by approximately 50% whereas the average duration of channel opening is reduced.These changes result from a substitution in one of the protein constituents that form the acetylcholine channel pore (31 ) .Our results indicate that the nonselective cation channel in fetal alveolar epithelial cells does not undergo dramatic postnatal modifications.In addition, channel behaviour did not appear to be influenced by the different enzymes and procedures used to dissociate fetal and adult alveolar epithelia.

Physiological role of a nonselective cation channel
According to the Koefoed-Johnsen-Ussing model of ion transport, epithelial cells maintain a low intracell ular concentrations of Na+ and high intracell ular concentrations of K+ by actively exchanging ions across the basolateral cell mem-brane ( 14).The enzyme Na +,K+ -ATPase ensures the maintenance of a favourable electrochemical gradient for Na+ influx, and Na+ enters the cell by diffusing through channels or by coupling to transport proteins located in the apical membrane.The net transport of Na+ from the apical to basolateral membrane generates the osmotic force that drives fluid movement across the epithelial ba1Tier.The rate of transep1thelial Na+ flux, and hence fluid absorption, is primarily governed by the ion channels present in the apical membrane.The nonselective cation channel we described could be the apical Na+ pathway and would therefore regulate Na +-coupled fluid absorption.
Our results also demonstrated that the channel was highly permeable to K+ .Under normal physiological conditions, channel opening would allow K+ efflux into the alveolar space.Such a mechanism is supported by the high concentration of K+ in the alveolar subphase.In humans and animals.the concentration of K+ in the fluid lining the alveolus is approximately twice the concentration of K+ in the plasma (32,33).Channels that are highly selective for K+ and K+ -dependent cotransport systems are also present in the apical membrane of alveolar epithelial cells and may also contribute to K+ secretion (34)(35)(36)(37).
Nonselective cation channels have been described in several other epithel ial and nonepithelial cell types (38)(39)(40)(41)(42)(43)(44).Physiological functions attributed to these channels include: electrogenic Na+ absorption, stimulus-secretion coupling, excitation-contraction coupling, and control of resting membrane potential.Interestingly, an amiloride-insensitive nonselective cation channel has recently been identified in human nasal epitheli um (45).This channel is also present in patients with cystic fibrosis and may contribute to the pathogenesis of the disease ( 45).
In our preparation, the activity of the nonselective cation channel was examined using the inside-out patch configuration, in the presence of high concentrations of Ca 2 +.The concentration of Ca 2 + in the cytosol may have influenced the single channel characteristics.Recently, Marunaka et al ( 16) reported for the nonselective channel in fetal alveolar epithelial cells that lowering the concentration of Ca 2 + in the bathing solution decreased the probability of channel opening, increased amiloride sensitivity and increased the selectivity of the channel for Na+ relative to K+.These findings suggest that, in vivo, the 'nonselective' channel may be a selective Na+ conduit.The authors also reported that the beta-adrenergic agonist terbutaline increased the probability of channel opening.This observation is important because adrenergic stimulation accelerates the clearance of flu id from the air spaces and stimulates active ion transport in alveolar epithelium ( 1, [46][47][48][49].Further studies are required to determine whether channel activity is modulated by other factors that influence alveolar flu id absorption.
In summary, the 23 pS nonselective channel identified in adult alveolar epithelial cells is similar to a cation channel present in fetal alveolar cells.This channel may participate in absorption of alveolar fluid, and we speculate that impaired channe l function contributes to the reduction of fluid clearance observed in acute lung injury.Our data provide the basis for further investigation of channel activity under various physiological and pathological conditions.

I
examine differences in open channel probabili ty at the various holding potentials.A Student' s t test (unpaired) was used to analyze differences in open dwell times between control and amiloride treated patches.P<0.05 was considered statis-Can Respir J Vol 1 No 2 Summer 1994

Figure 3 )
Figure 3) Anion selectivity of the channel.The current-voltage ( i-1•) relationship for three patches recorded with 47 mM sodium chloride in the bath and 140 mM sodium chloride in the pipette.The solid line indicates the line of linear regression through the data poinls.Under these recording condition.1•the i-v curve is sh(fled to the right with an x intercept of 22±3 111 V.The dashed line represenls the i-v curve predictedfrom the GHK cons/ant.fieldequation.fora channel completely selective for cations where: i=EFIRT * PNa + * [Ne/ Je -I Ne/Ji exp ( EFIRT)I I-exp( EFIRT)

Figure 4 )
Figure 4) Effect of111e111bra11e potential 011 the probability of channel opening.Currents H'ere recorded H'ith sodium chloride ( 140 111M) in the bath and recording electrode.No significant difference H'C/S observed for open probabilit_ 1• of'the channel when measured at hrpe1polari::,i11g and depolari:ing membrane potentials (11=4)

Figure 5 )
Figure 5) Amiloride-sensitivit_ v of the nonselective cation channel.Single-channel currents recorded ji-0 111 Mo inside-out patches voltage clamped to -60 mV.A Channel openings/ram a patch rerorded il'ith sodium chloride ( 140 mM) in the bath a,1d pipette.B When amiloride 10-5 M ( solution D) was added to the pipette solution, channel.f]ickering was consistently obsen•ed.Both current records were filtered at I kH~ and sampled at JOO µs!point