Evidence for apical sodium channels in frog lung epithelial cells

Evidence for apical sodium channels in frog lung epithelial cells HORST

FISCHER,

WILLY

VAN DRIESSCHE,

AND

WOLFGANG

CLAUSS

Institut fiir Veterinkrphysiologie, Freie Universitkt Berlin, 1000 Berlin 33, Federal Republic of Germany; and Laboratorium voor Fysiologie, K.U.L., Campus Gasthuisberg, 3000 Louvain, Belgium

FISCHER, HORST, CLAUSS. Evidence

WILLY

VAN

DRIESSCHE,

AND WOLFGANG

for apical sodium channels

in frog lung epithelial cells.Am. J. Physiol. 256 (Cell Physiol. 25): C764-C771, 1989.-To reveal the mechanismof Na+ transport acrossXenopus lung epithelium, we recorded short-circuit current (I&

transepithelial resistance(I&), and current noisespectra while the isolatedlung tissuesweremounted in an Ussing-typechamber. Mean values of Isc and Rt obtained while the tissue was bilaterally incubatedwith NaCl-Ringer solution wereIsc= 11.57 t 1.19PA. cmB2and R, = 0.82 t 0.07 kQ. cm2.Amiloride added to the mucosal(apical) sidedepressedIScby 61 to 99%. Ouabain abolishedI,, totally when added to the basolateral compartment. Adenosine 3’,5’-cyclic monophosphate (CAMP), epinephrine, and a variety of other compoundsdid not alter I,, significantly. Transepithelial depolarization with serosalKC1 solution reducedI,, to 6.22 t 1.37PA. cmD2.Amiloride-sensitive current and the kinetics of amiloride interaction were not significantly affected by depolarization. Fluctuation analysisof I,, in the presenceof amiloride revealed a Lorentzian component in the power density spectrum indicating apical Na+ channels.Assumingpseudo-first order kinetics, we calculated single channel currents (i& and channel density (M): iNa = 0.29 t 0.04pA and M = 0.24 t 0.04 pm2.Our results showthat the route for Na’ transport through lung epithelial cellsfollows the classicalKoefoed-Johnson-Ussingmodelfor tight epithelia. kzevis; apicalmembrane;noiseanalysis;sodium;transport; amiloride; ouabain Xenopus

TRANSPORT through lung epithelium has so far been investigated mainly by means of isotope flux measurements (10, 24). By these methods it is merely possible to differentiate between ion movement occuring paracellular or transcellular and to distinguish between active carrier-mediated ion transport and passive ion movement through channels in the latter path. Transport of Na+ through lung epithelium is of special interest because of its putative role in the processes responsible for maintaining a nearly liquid-free alveolar airspace. There is evidence that Na+ transport is closely coupled to fluid movement (1, 2, 23). Inhibition of Na+ transport with the specific Na+ channel blocker amiloride was already investigated in lung tissues of different species and cell cultures of epithelial lung cells (1,2, 19,23,26). All these studies provided only indirect evidence for the existence of Na’ channels as entry sites in the apical membrane of lung epithelium. However, direct evidence demonstrating passive uptake of Na+ through channels, and the

ION

C764

0363-6143/89

$1.50 Copyright

characterization of this uptake process is still missing. The driving force for Na+ movement through such channels may be generated by a ouabain-sensitive basolateral located Na+-K+-ATPase that has been already described in lung epithelial cells of different species (2, 9, 19). The luminal surface of the frog lung is enlarged and partitioned by septa. Because of the complex anatomy of the mammalian lung, we examined the frog lung as a simpler model that may share apical membrane properties with mammalian alveolar pneumocytes. The relatively flat structure of this tissue has the advantage that investigations of transepithelial Na+ transport can be performed in an Ussing chamber. Transport processes in bullfrog lung are already intensively studied. Because of its simple anatomy it was a useful tissue for the investigations of ion transport. However, Na+ transport was not detected in bullfrog lung. The observed short-circuit current was mainly Cl- secretion (10). In the adult fluid-filled mammalian lung epithelium, an amiloride-blockable Na+ current occurs (1, 2). We found evidence for Na+ transport in the lung of Xenopus directed from the airspace to the interstitium. In this study, Xenopus lung epithelium was investigated with electrophysiological methods to reveal macroscopic data about ion transport properties and microscopic data about Na+ channel distribution, single Na+ channel current, and amiloride-binding kinetics of single Na+ channels in the apical cell membrane. In a morphological approach, we investigated the microscopic structure of Xenopus lung epithelial cells. We found it structured suitable for normal gas exchange. MATERIALS

AND

METHODS

Adult frogs of both sexes (Xenopus laevis, obtained from H. Kahler, Hamburg) were kept in tap water at room temperature. Once a week they were fed with commercial cat food ad libitum. The animals were killed by decapitation. The spinal chord was pithed, and the lungs were removed. For preparation of a planar sheet of tissue, the lung was incised along the pathway of the large pulmonary artery and, if necessary, connecting septa were cut to unfold the lung sac. The tissue was glued with tissue adhesive (Histoacryl blau, B. Braun Melsungen) on its pleural side to a Lucite ring and was mounted with minimal edge damage in an Ussing-type chamber (for details see Ref. 7). Effective aperture was 0.5 cm2. Serosal and mucosal compartments were contin-

0 1989 the American

Physiological

Society

NA+

CHANNELS IN LUNG EPITHELIAL

uously perfused with solution with a flow rate of -5 ml/ min. The transepithelial potential was clamped to zero with a voltage-clamp unit (Elke Nagel Biomedical Instruments, Munich, FRG). Short-circuit current (1& and the tissue conductance were recorded continuously on a strip chart recorder. Conductance was calculated from superimposed unipolar pulses of IO-mV amplitude, ZOO-ms duration, and a repetition time of 3 s. Polarity of pulses had no effect on the calculated conductance. Transepithelial capacitance (CJ was measured under open-circuit conditions. Rectangular current pulses were applied, and the time course of the transepithelial potential (VJ was analyzed. Ct was calculated as T/transepithelial resistance (RJ, where 7 denotes the time after Vt reached 63% of its maximal value. There was no dependency of both amplitude and polarity of the applied current pulses (+5, HO, t15 PA) on calculation of Ct. The values were pooled and each Ct measurement represents the mean of six different current pulses across the tissue. For the noise analysis experiments, the transepithelial voltage was clamped to zero with a low-noise voltage clamp (29, 30). I,, was recorded continuously. Rt was calculated from the deflections in I,, caused by single voltage pulses of -1-s duration. The stepwise depression of Isc by increasing concentrations of mucosal amiloride ([A],) was analyzed with the direct linear plot method by Eisenthal and Cornish-Bowden (8). They rearranged the Michaelis-Menten equation as follows

mtaaxl~sc)- uw[Al,)

=1

(1)

where Gzx is the total amount of amiloride-blockable current, AIs, is the depression in I,, caused by a given constant. Each [Al m, and Kt is the Michaelis-Menten [Aim and its corresponding AIs, reveals one straight line described by Eq. 1 in the direct linear plot, where I,, is the ordinate intercept and [Aim is the intercept on the abscissa. The intersection of each two straight lines derived from different [Aim revealed one value of KA, and GIx in the direct linear plot. The median of the K”, and &,““values is used as the best estimate of the true values describing the dose-response curves. Perfect MichaelisMenten kinetics of the blocking reaction would reveal exactly one intersection point of all straight lines. Michaelis-Menten kinetics were additionally checked by determination of the Hill-coefficient (nn). ?2His the slope of a regression line of (E,““/&,) - 1 vs. [Aim in a double logarithmic plot, where lNa is the amiloride-dependent Na’ current. The fluctuation in I,, was recorded and processed as described previously (28). The power spectra recorded in the presence of amiloride in the apical solutions consisted of three noise components (see Fig. 5 in RESULTS): 1) amplifier noise at the higher frequency end of the spectrum; 2) a low-frequency noise component described by the empirical equation S(f) = KB/f* and 3) a blocker-induced

Lorentzian

(2) component

that is

C765

CELLS

associated with random opening and closing of the channel, described by S(f) = So/[1 +