We investigated the mechanism by which cAMP increases sodium transport in lung epithelial cells. Alveolar type II (ATII) cells have two types of amiloride-sensitive, cation channels: a nonselective cation channel (NSC) and a highly selective channel (HSC). Exposure of ATII cells to cAMP, β-adrenergic agonists, or other agents that increase adenylyl cyclase activity increased activity of both channel types, albeit by different mechanisms. NSC open probability (P o) increased severalfold when exposed to terbutaline, isoproterenol, forskolin, or cAMP analogs without any change in NSC number. In contrast, terbutaline increased HSC number with no significant change in HSC P o. For both channels, the effect of terbutaline was blocked by propranolol and H-89, suggesting a protein kinase A (PKA) requirement for β-adrenergic-induced changes in channel activity. Terbutaline increased cAMP levels in ATII cells, but intracellular calcium also increased. Calcium sequestration with BAPTA blocked β-adrenergic-induced increases in NSC P o but did not alter HSC activity. These observations suggest that β-adrenergic stimulation increases intracellular cAMP and activates PKA. PKA increases HSC number and increases intracellular calcium. The increase in calcium increases NSC P o. Thus increased cAMP levels are likely to increase lung sodium transport regardless of which channel type is present.
- single channel recording
- ion transport
- epithelial sodium channels
- adenosine 3′,5′-cyclic monophosphate
- β-adrenergic agents
- alveolar type II cells
amiloride-sensitive sodium absorption across the alveolar epithelium is believed to be enhanced by cAMP and by agents such as β-adrenergic agonists that stimulate adenylyl cyclase. In the perinatal period, a rise in endogenous catecholamines correlates with enhanced clearance of fetal lung fluid. In postnatal life, β-adrenergic agents have a potential use in limiting pulmonary edema that accompanies a variety of lung conditions. β-Adrenergic agonists increase sodium reabsorption in anesthetized fetal and adult animals (5, 13, 30, 40), isolated rat lungs (10, 20, 37, 38), and cultured type-II cells (9). The effects of β-adrenergic agonists may be mediated by increases in intracellular cAMP, since other agents that increase cAMP produce effects similar to β-adrenergic agonists (9). However, Nakahari and Marunaka (28) have proposed that the action of β-adrenergic agonists is mediated by a cAMP-dependent increase in intracellular calcium that activates calcium-dependent nonselective cation channels (NSC). Despite this difference, the exact mechanism by which cAMP influences sodium transport in vivo is disputed, as is evident from recent editorials and review articles from experts in this area (21, 32, 43). Lazrak et al. (23, 24) have postulated that β-adrenergic agents stimulate sodium absorption by increasing the membrane conductance to sodium (G Na). They based their opinion on single channel studies, which showed that terbutaline increases the open probability (P o) of a 25-pS moderately selective cation channel (21). However, O'Grady et al. (32) propose that the terbutaline-induced increase in sodium transport is secondary to an increase in chloride conductance, which in turn changes the apical membrane potential (V a). In support of their hypothesis, O'Grady et al. (32) cite studies with cultured alveolar epithelial cell monolayers that show increases in apical membrane chloride conductance in the absence of any change in sodium conductance after stimulation by β-adrenergic agonists. However, as pointed out by Widdicombe (42), neither hypothesis is adequately substantiated, and additional studies need to be done to test both hypotheses. Such studies are critical for lung biologists because of the widespread use of β-adrenergic agonists in clinical medicine and the potential therapeutic role for selective adrenergic agents in several pathological states accompanied by pulmonary edema.
One explanation for such disparate results may lie in the differences in the experimental protocols followed by various investigators, especially culture conditions. Our studies have shown that culture conditions have a profound effect on the type of sodium conductance expressed on the apical surface of cultured alveolar epithelial cells (17). Specifically, when alveolar type II (ATII) cells are cultured under conditions when oxygen delivery to the cells is reduced or in the absence of steroid hormones, the predominant channel is a 21-pS NSC with a sodium-to-potassium selectivity ratio of 1:1. However, when ATII cells are grown on permeable supports in the presence of steroids with an air interface, the predominant channel is a low-conductance (6-pS), highly sodium-selective channel (HSC) with an sodium-to-potassium selectivity ratio >80:1. Both channels are amiloride sensitive so that, even though the half-maximal inhibition constant values are different, at a whole tissue level, the two channel types are difficult to distinguish based on amiloride sensitivity alone. Nonetheless, HSC, because of their high selectivity for sodium, should be much more effective as sodium transport pathways than NSC.
We hypothesized that signal transduction pathways mediated by cAMP and protein kinase A (PKA) regulate the expression and activity of both HSC and NSC, and the relative abundance of one channel over the other would determine the net ability of ATII cells to transport sodium. To address these issues, we used single channel studies to evaluate the effect of β-adrenergic agonists and cAMP on both NSC and HSC. One reason to make single channel measurements is to be sure which channels are responsible for transepithelial currents, especially because these agents are known to stimulate chloride channels also. Our results show that β-adrenergic agonists and cAMP increase the activity of both types of channels, but increases in intracellular cAMP appear to increase the P o of the 21-pS NSC, with no change in the density of these channels, whereas cAMP produces no change in the P o of HSC but does increase the density of channels. In fact, cAMP appears to cause insertion of clusters of new HSC in the apical membrane patches.
MATERIALS AND METHODS
ATII cell isolation and culture.
ATII cells were isolated by enzymatic digestion of lung tissue from adult Sprague-Dawley rats (200–250 g) using previously described methods (17). Briefly, the rats were anesthetized with pentobarbital sodium and heparinized (100 U/kg). ATII cells were digested by tracheal instillation of elastase (0.4 mg/ml). Lung tissue purification was based on differential adherence of cells to dishes coated with rat IgG. Nonadherent ATII cells were collected, centrifuged, and seeded on permeable supports in a highly enriched medium (3 parts Coon's modification of Ham's F-12 and 7 parts Liebovitz's L-15 with 1.5 μM aldosterone). After isolation, some cells were allowed to grow on glass coverslips, and others were allowed to attach to a specialized culture support (24), which is optimized for patch-clamp recording and allows the cells to grow on a permeable support (Millipore) while they are submerged in medium. After the cells had attached to the culture surface (this usually required 2–4 h), medium was drained from the apical surface, and cells were allowed to grow with medium on the basolateral surface and air on the apical side. Alternatively, cells were cultured in an identical fashion but without draining the medium so that cells remained submerged. Cells were incubated in 95% air and 5% CO2 and were used for patch-clamp studies between 24 and 96 h after plating.
Using this method for isolation of ATII cells, we were able to obtain an ATII cell population with 95% viability (confirmed by Live/Dead Eukolight Viability/Cytotoxicity Kit 1–3224; Molecular Probes, Eugene, OR) and 95% purity (confirmed by staining for surfactant proteins A and B; see Ref. 17). Contamination by macrophages and fibroblasts was <5%. Cells were used for patch-clamp experiments during the first 24–96 h in culture while they maintained ATII cell phenotype (light microscopy) and function (surfactant production using radiolabeled choline incorporation; see Ref. 17). The purity and viability of ATII cells were similar when cells were cultured on glass submerged in media and on a permeable support exposed to an air interface. ATII cells used in these experiments had lamellar bodies and other phenotypic features of type II cells and secreted surfactant (17).
Bath and pipette solutions used in the cell-attached mode contained (in mM) 140 NaCl, 1 MgCl, 1 CaCl2, 5 KCl, and 10 HEPES, pH 7.4, with 2 N NaOH. In the inside-out recordings, pipette solution was the same, but the bath solution was changed to (in mM) 5 NaCl, 140 KCl, 4, CaCl2, 5 EGTA, 1 MgCl2, and 10 HEPES, pH 7.4, with 2 N KOH. The contents of bathing and pipette solutions were varied as appropriate for specific protocols. All chemicals were obtained from Sigma (St. Louis, MO). All treatments (adrenergic agents, cAMP, etc.) were applied to the basolateral side of the cells, i.e., the bottom of the filter support.
Single channel recording.
Patch-clamp experiments were carried out at room temperature. The pipettes were pulled from filamented borosilicate glass capillaries (TW-150; World Precision Instruments) with a two-stage vertical puller (Narishige, Tokyo, Japan). The pipettes were coated with Sylgard (Dow Corning) and fire polished (Narishige). The resistance of these pipettes was 5–8 MΩ when filled with pipette solution. We used the cell-attached configuration for most of our studies since, in this configuration, the cytoplasmic constituents remain intact, thus allowing us to study the role of cytoplasmic second messengers in regulation of ion channel activity. After formation of a high-resistance seal (>50 GΩ) between the pipette and cell membrane, channel currents were sampled at 5 kHz with a patch-clamp amplifier (Axopatch 200A; Axon Instruments, Foster City, CA) and were filtered at 1 kHz with a low-pass Bessel filter. Data were recorded by a computer with pCLAMP 6 software (Axon Instruments). Current amplitude histograms were made from stable, continuously recorded data, and the open and closed current levels were determined from least-square fitted Gaussian distributions. The P o of the channels was calculated using FETCHAN in pCLAMP 6. Single-channel conductance was determined using a linear regression of unitary current amplitudes over the range of applied pipette potentials.
For cell-attached patches, voltages are given as the negative of the patch pipette potential (−V pipette). This potential is the displacement of the patch potential from the resting potential (about −40 mV for ATII cells); positive potentials represent depolarizations, and negative potentials represent hyperpolarizations of the cell membrane away from the resting potential. For a highly selective sodium channel with a sodium concentration gradient of 10 to 1 (outside to inside), the reversal potential would be +60 mV. Therefore, it would require a 100-mV positive voltage displacement (−V pipette = +100 mV) from the resting potential to reach the reversal potential.
Single channel analysis.
We used the product of the number of channels (N) times the single P o as a measure of channel activity within a patch. This product was calculated without making any assumptions about the total numbers of channels in a patch or theP o of a single channel where n is the channel level, T is the total recording time, N is the maximum level of channel opening, and tn is the open time for thenth channel. The total number of functional channels in a patch was estimated by observing the number of peaks in current-amplitude histograms constructed, when possible, from event records of long enough duration to provide 95% confidence of determining the correct number of channels according to methods we have described previously (22, 26). However, especially in some of the antisense-treated patches, we could not always record long enough to reach a 95% confidence level, and the values for the number of channels in these patches may be an underestimate.
Measurement of intracellular cAMP levels.
The intracellular cAMP level in ATII cells was measured by a Biotrak cAMP enzyme immunoassay (EIA) system (Amersham Life Science). After treatment with experimental agents, ATII cells were placed on ice, scraped, and homogenized and then 65% of ice-cold ethanol was used to extract cAMP from the cells. cAMP EIA is based on the competition between unlabeled cAMP and a fixed quantity of peroxidase-labeled cAMP for a limited number of binding sites on a cAMP-specific antibody. With fixed amounts of antibody and peroxidase-labeled cAMP, the amount of peroxidase-labeled ligand bound by the antibody is inversely proportional to the concentration of added unlabeled ligand. The amount of peroxidase-labeled cAMP bound to the antibody is determined by addition of a tetramethylbenzidine/hydrogen peroxide single pot substrate. The reaction is stopped by addition of an acid solution, and the resultant color is read at 450 nm in a microtiter plate spectrophotometer. A standard curve was made for a range of cAMP concentrations from 12.5 to 3,200 fmol. The concentrations of unknown samples were determined by comparison with the standard curve. The cellular levels of cAMP were normalized to total cellular protein. Protein concentration was measured against a BSA standard using a Bradford dye-binding assay.
Measurement of intracellular calcium.
To measure intracellular calcium levels, indo 1 fluorescence was measured with ATII cells grown on glass coverslips or filter supports. Cells were preincubated with 5 μM indo 1-AM (Molecular Probes) for 30 min at room temperature in darkness and then were washed with saline solution three times. Supports were then placed in a chamber secured to the stage of a Meridian ACAS 570/Ultima laser scanning confocal microscope (Laser Cytofluorimeter Working Station; Meridian Instruments). Cells loaded with indo 1-AM were excited by ultraviolet light at 351–364 nm, and the cytoplasmic free calcium signal was read as the fluorescence ratio at 405- and 530-nm wave lengths. Intracellular calcium concentrations were calculated by using the standard working curves of the fluorescence ratio vs. free calcium in calcium-EGTA buffer.
Statistical analysis for the changes in P o of channels and the biochemical estimations were performed using SPSS or SigmaStat for windows. Statistical significance between two groups was determined by paired or unpaired tests, as appropriate. When the comparison between more than one group was required, statistical significance was determined by one-way ANOVA followed by comparison of treated with untreated cells using Dunnett's test or pairwise comparisons with the Student-Newman-Keul's test to determine significant differences. Values < 0.05 were considered significant.
We have previously shown (15-17) that, when ATII cells are under conditions of reduced oxygen delivery or in the absence of steroids, the predominant sodium-permeant cation channel is a 21-pS NSC with an sodium-to-potassium selectivity of 1:1, a linear current-voltage relationship, and a reversal potential near 0 mV. However, when grown on permeable supports in the presence of steroids with an air interface, the predominant channel is a 6-pS HSC with an sodium-to-potassium selectivity of >80:1, moderate inward rectification, and a reversal potential near +100 mV. Both channels are amiloride sensitive. Details of single channel characteristics have been described previously (17). In this study, we have focused our attention on the effect of β-adrenergic agonists and the cAMP-PKA pathway on these channels.
β-Adrenergic agonists increase cellular cAMP levels in ATII cells.
We measured the intracellular cAMP levels in ATII cells before and after 15 min of exposure to agents known to increase cAMP levels in other epithelial cells and found, as expected, that β-adrenergic agents and forskolin increase cAMP levels in ATII cells from 111.8 ± 51.3 fmol/mg protein in untreated cells to 630.8 ± 275.2 fmol/mg protein in 20 μM terbutaline, 358.4 ± 121.6 fmol/mg protein in 20 μM isoproterenol, and 323.6 ± 108 fmol/mg protein in 20 μM forskolin (cAMP levels for each of the treatments was significantly different from untreated cells; mean ± SD ,n = 6 for all conditions). We added no phosphodiesterase inhibitors in any of the experiments described in this study, although addition almost surely would have increased the magnitude of our responses. Thus the levels of cAMP and the responses of sodium channels are what might be expected from normal physiological responses.
β-Adrenergic agonists increase intracellular calcium levels in ATII cells.
Although β-adrenergic agents are usually thought to act by increasing cAMP with subsequent activation of PKA, there are several reports that suggest that an additional effect of β-adrenergic agents is to increase intracellular calcium in a variety of target cells (27,28, 30). Therefore, we measured changes in intracellular calcium in ATII cells in response to the addition of terbutaline (20 μM) and found that terbutaline produced a significant increase in intracellular calcium from basal levels of 118 ± 49.7 nM to a maximum stimulated level of 909 ± 97.9 nM (peak calcium level after terbutaline was significantly different from the basal level; mean ± SD, n = 6). The steady-state level in the continued presence of terbutaline was 2.5- to 3-fold higher than baseline levels (293 ± 60.5 nM; Fig. 1).
β-Adrenergic agents increase the number of HSC without changing the Po.
As mentioned above, when ATII cells are cultured on permeable supports with an apical air interface, they express HSC in their apical membranes. Under these conditions, examination of single channel records from untreated ATII cells and cells treated with the β-agonist terbutaline showed that 20 μM terbutaline produced a large increase in the number of channels observable in a typical cell-attached patch (Fig. 2). In 45 untreated cell-attached patches, we observed 2.6 ± 0.18 channels/patch (mean ± SE); after 10 min of exposure to 20 μM terbutaline the mean number of channels per patch in 43 patches increased to 6.0 ± 0.34. After 1 h of exposure, the mean number of channels was lower but still elevated (4.7 ± 0.36,n = 45). Both values were significantly larger than the number of channels in untreated patches. In contrast,P o of the channels in terbutaline-treated cells (0.224 ± 0.0269 after 10 min and 0.188 ± 0.0268 after 1 h) did not change significantly from theP o of untreated cells (0.182 ± 0.0202). These results are summarized in Fig. 3.
Effect of terbutaline on HSC is mediated by β-adrenergic receptors.
Terbutaline is usually considered to be a β-adrenergic agonist. To test this, we applied 20 μM propranolol (a β-adrenergic receptor antagonist) in the bath solution. Propranolol blocked the terbutaline-mediated increase in channel number with no significant effect on channel P o. [The number of channels per patch in 45 patches on untreated cells was 2.6 ± 0.175 and increased to 6.0 ± 0.340 in 45 patches after 20 μM terbutaline but was only 2.5 ± 0.143 when 25 patches were pretreated for 30 min with 20 μM propranolol and then treated with 20 μM terbutaline (Table 1). The number of channels in propranolol-treated patches is not different from untreated patches (P > 0.05).]
Increases in intracellular cAMP produce the same effect on HSC as β-adrenergic agonists.
One of the ways in which β-agonists produce their effect is through activation of adenylyl cyclase with a subsequent increase in intracellular cAMP. Indeed, we demonstrated that β-adrenergic agonists increase intracellular cAMP and that the same agonists increase channel density (Fig. 1). Therefore, we examined the effect of cAMP on HSC activity. Application of a membrane-permeable analog of cAMP, 8-(4-chlorophenylthio)cAMP (cpt-cAMP), produces the same increase in HSC density as terbutaline, with no significant change inP o. [The number of channels per patch in 26 patches on cells treated with 100 μM cpt-cAMP for 15 min was 5.5 ± 0.267 compared with 6.0 ± 0.340 in 45 patches after 10 min exposure to 20 μM terbutaline (Table 1). The number of channels in cpt-cAMP-treated patches is not different from terbutaline-treated patches (P > 0.05).] These results imply that the increase in HSC density depends on the β-adrenergic-mediated increase in intracellular cAMP.
Increases in intracellular calcium are not associated with the β-adrenergic-induced increase in HSC density.
We show that β-adrenergic agonists increase intracellular cAMP. On the other hand, β-adrenergic agonists also increase intracellular calcium (Fig. 1) so that the increase in channel density could also depend on the increase in calcium. We tested this by adding the membrane-permeable calcium chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA)-AM to inhibit the increase in intracellular calcium. Incubation with 1 mM BAPTA-AM for 30 min reduced intracellular calcium in 30 cells to 78 ± 23.4 nM (compared with 118 ± 49.7 nM in untreated cells), but addition of terbutaline increased calcium only to 108 ± 38.5 nM (compared with 909 ± 97.9 nM in untreated cells). Despite the effect of BAPTA on intracellular calcium, incubation with BAPTA-AM did not change the terbutaline-induced increase in HSC density [in 22 patches on cells pretreated with 1 mM BAPTA-AM for 30 min, terbutaline still increased channel density to 5.8 ± 0.45 channels/patch, comparable to patches treated with terbutaline alone and significantly greater than patches from untreated cells (Table 1)].
PKA activity is necessary for the β-adrenergic-induced increase in HSC density.
The effects of increases in intracellular cAMP are often associated with an activation of PKA. Therefore, we examined whether blocking PKA activity reduced the effect of terbutaline on HSC density. If PKA was blocked with the A kinase blocker H-89, terbutaline was no longer capable of producing an increase in the density of HSC [in 19 patches on cells pretreated with 5 μM H-89 for 30 min, terbutaline treatment produced only 2.6 ± 0.16 channels/patch, which is not significantly different from 45 untreated patches that had 2.6 ± 0.18 channels/patch (Table 1)]. These results taken together imply that terbutaline increases the activity of HSC by increasing the number of channels per unit area of membrane through a PKA-dependent process.
β-Adrenergic agents increase the Po of NSC but do not change the channel number.
β-Adrenergic agonists also increased the activity of NSC in ATII cells, but the mechanism was quite different from the effect on HSC. Figure 4 shows, in a continuous recording from a single cell-attached patch, that application of 20 μM terbutaline dramatically alters the P o of NSC with no change in the number of channels. In 17 cell-attached patches, the P o before addition of terbutaline was 0.207 ± 0.0281 (mean ± SE), but after 15 min of exposure to 20 μM terbutaline, P o increased to 0.455 ± 0.0667, which is significantly larger than theP o before terbutaline. In contrast, the number of channels per patch in terbutaline-treated cells (3.1 ± 0.90) did not change significantly from the number of untreated cells (3.3 ± 0.70). The results of these experiments are summarized in Fig. 5. As expected, isoproterenol, a drug known to activate both β1- and β2-receptors, also increased channel activity by increasing P o from 0.269 ± 0.0302 before application to 0.515 ± 0.111 (n = 17,P < 0.01; Table 2). There was no change in the number of channels in cell-attached patches treated with isoproterenol (2.8 ± 0.29 in untreated cells vs. 2.7 ± 0.56 after isoproterenol), and single channel conductance remained unchanged.
Effect of terbutaline on NSC is mediated by β-adrenergic receptors.
Like HSC, the effect of terbutaline on NSC was also mediated through β-adrenergic receptors. Propranolol (20 μM) added to the basolateral surface of a cell on which a cell-attached patch containing NSC blocked the effect of terbutaline, but it did not change the basal channel activity (the P o of untreated channels in nine patches is 0.248 ± 0.0160, 0.243 ± 0.0217 after 20 μM propranolol, and 0.235 ± 0.0164 after 20 μM propranolol and 20 μM terbutaline together). TheP o of all of the treatments are not significantly different, and the number of channels per patch are also not different.
cAMP increases Po of NSC in ATII cells.
In a cell-attached patch with NSC activity, addition of 100 μM cpt-cAMP caused a significant increase in NSC P o[increase from 0.139 ± 0.0411 to 0.5826 ± 0.0471,n = 10, P < 0.01 (Table 2)] with no significant change in channel density (untreated 2.9 ± 0.433 vs. 2.7 ± 0.47 after cpt-cAMP). Similar results were obtained with 1 mM 8-bromo-cAMP [8-Br-cAMP; P o significantly increased from 0.130 ± 0.0339 to 0.554 ± 0.0574,n = 9, P < 0.005 (Table 2)]. In another set of experiments, we pretreated ATII cells with 1 mM 8-Br-cAMP for 5 min before patch-clamp recordings, with a group of untreated cells serving as a control. Here also, pretreatment of cells with 8-Br-cAMP resulted in a significant increase in the number of channels times the P o(NP o; Table 2), with all of the increase resulting from an increase in P o (controlNP o 1.053 ± 0.157 vs. 1 mM 8-Br-cAMP 1.819 ± 0.185, +73%, n = 24, P< 0.005). Further confirmation of the cAMP effect was obtained by the use of forskolin, an agent known to increase cellular cAMP by direct stimulation of adenylate cyclase. Treatment of cells with 20 μM forskolin added to the bath after stable cell-attached patches were obtained resulted in a significant increase in NSCP o [untreated 0.167 ± 0.0324 to 0.564 ± 0.0455 after treatment with 20 μM forskolin,n = 16, P < 0.001 (Table 2)]. These results suggest that an increase in cellular cAMP level stimulates NSC activity in apical membranes of ATII cells.
cAMP stimulation of NSC is mediated by PKA.
To determine whether cAMP action on NSC is mediated via PKA, we pretreated cells with 5 μM H-89 for 5 min after we acquired control channel activity but before using any reagents to induce an increase in cAMP. We found (Table 2) that H-89 blocked the effect of 20 μM terbutaline to increase NSC P o. H-89 itself did not cause any significant change in the basal channel activity (untreated P o was 0.239 ± 0.0138 vs. H-89 0.227 ± 0.00919 vs. H-89 + terbutaline 0.213 ± 0.00635,n = 14, no significant differences between either treatments or control). In addition, 5 μM H-89 blocked the effect of 20 μM forskolin on NSC (untreated P o was 0.232 ± 0.0188 vs. H-89 0.228 ± 0.0132 vs. H-89 + forskolin 0.233 ± 0.0115, n = 14, no significant differences between either treatments or control). These results suggest that the effect of cAMP (and agents that increase cAMP) on NSC is mediated by PKA.
Increases in intracellular calcium are necessary for the β-adrenergic-induced increase in NSC Po.
We previously showed that β-adrenergic agonists increase intracellular cAMP and that β-adrenergic agonists also increase intracellular calcium (Fig. 1). However, an increase in the number of HSC only requires an increase in cAMP and not an increase in calcium. The increase in NSC activity is different; the increase in NSCP o is associated with an increase in cAMP but also depends on an increase in intracellular calcium. Incubation of with 1 mM BAPTA-AM for 30 min (Fig. 6) prevented the terbutaline-induced increase in NSCP o (in 12 patches on cells pretreated with 1 mM BAPTA-AM for 30 min the P o was 0.0926 ± 0.0333 and after application of terbutaline P owas 0.115 ± 0.0519, which is not significantly different from BAPTA alone). The terbutaline-induced increase in NSCP o could require both an increase in intracellular cAMP and calcium, or the increase in calcium could be a response to the increase in cAMP, and the change inP o could finally only require a change in calcium. To investigate this question, we excised patches of membranes containing NSC in a saline that mimicked the intracellular composition (Fig. 6). In particular, it had 100 nM calcium but no cAMP. We then increased the calcium to 1,000 nM and found that theP o of the NSC increased substantially (in 19 cell-free, excised patches exposed to 100 nM calcium on their cytosolic face the P o was 0.112 ± 0.0304 and after application of 1,000 nM calcium P o was 0.459 ± 0.0504, which is significantly different from 100 nM calcium, P < 0.001). We interpret these results to mean that NSC are activated by intracellular calcium but that the increase in intracellular calcium is dependent on an increase in intracellular cAMP.
In vivo and in vitro studies that use whole tissue measurements of sodium transport indicate that β-adrenergic agonists and cAMP increase alveolar salt and water reabsorption. However, the underlying mechanism is unclear and is the subject of an ongoing debate. In this study, we have used single channel measurements to evaluate the effect of β-adrenergic agonists and cAMP on the sodium channels in ATII cells. This approach allows us to unambiguously identify the effect of β-adrenergic agonists on specific channels in ATII cells. As previously reported, we have identified two different types of amiloride-sensitive sodium-permeable channels in ATII cells (15-17), a 20-pS NSC and a 6-pS HSC. The single channel characteristics of the NSC are comparable to those described in both adult and fetal lung cells by several other research groups (11, 25, 35, 38). The HSC has single channel characteristics that are similar to those of channels seen when α-, β-, and γ-epithelial sodium channel subunits are reconstituted inXenopus oocytes (6) and that are observed in other sodium-transporting epithelial cells (for a review, see Ref.12). Our studies show that the β-adrenergic-cAMP-PKA pathway has a direct regulatory role for sodium channels in ATII cells, although the specific mechanism by which channel activity is increased varies with the channel type. An increase in intracellular cAMP causes an increase in P o of individual NSC in ATII cell apical membranes without affecting their numbers; on the other hand, cAMP increases the channel density or number of HSC without affecting their P o. The exact role of these two channels in different physiological and pathological states has yet to be completely understood. However, our studies show that agents that increase intracellular cAMP are likely to be effective in increasing alveolar sodium transport regardless of which of these two channels is present. Our results also clearly differ from those reported by Jiang et al. (20), who could find no effect of β-adrenergic agents on sodium transport in ATII cells and, indeed, could find no sodium channels even though their cells were grown under conditions similar to ours. We are somewhat at a loss to reconcile their results with ours.
β-Adrenergic agents and cAMP enhance active sodium absorption across alveolar epithelium.
Vectorial transport of solutes between the alveolar surface and the interstitial space is believed to play a key role in regulation of lung salt and fluid balance. Several investigators, using different experimental approaches, have shown that this process can be stimulated by agents that increase cellular cAMP. Using an in vivo sheep model, Berthiaume et al. (4) have shown that β-adrenergic agents or cAMP analogs stimulate the removal of fluid from saline-filled lungs, and this increase in fluid clearance can be inhibited by amiloride. Similar results were obtained by Olver et al. (33, 34) in the fetal lamb model. In monolayers of alveolar epithelial cells, terbutaline has similarly been shown to enhance amiloride-sensitive short-circuit current (I sc), and an increase in transepithelial fluxes of 22Na (9, 31). Even though the first of these studies was published more than a decade ago, the mechanism of action of β-agonists continues to be a subject of debate (23,32). There are at least two broad mechanisms that have been considered (21, 32, 43). In general, because the net inward movement of sodium across the apical membrane is dependent onG Na and the driving force for entry of sodium into the cell (V a −E Na, where E Na is the equilibrium potential for sodium; see Ref. 42), cAMP could be working by affecting G Na (24) orV a. O'Grady et al. (32) have proposed that, instead of a direct effect on sodium channels, cAMP works by increasing chloride conductance. This increase in cAMP-dependent inward movement of chloride leads to hyperpolarization of the apical membrane and generates a driving force (V a) for movement of sodium from the alveolar space into the cell. Their conclusion is based on studies involvingI sc and whole cell patch-clamp measurements (18-20) in which they permeabilized basolateral membranes of alveolar epithelial cells in monolayers using amphotericin B. In this preparation, in the presence of a chloride gradient, terbutaline was found to increase I sc, butI sc was not increased by terbutaline when a sodium gradient was imposed across the apical membrane. However, the authors were unable to show an increase in amiloride-sensitiveI sc with terbutaline and did not conduct any single channel studies.
On the other hand, Lazrak et al. (23) proposed that cAMP increases the overall G Na and not by increasing chloride conductance. This conclusion is based on patch-clamp studies (26, 26, 40, 44) that show that β-adrenergic agents and cAMP increase the P o of NSC recorded from apical membranes of ATII cells. They also citeI sc measurements by several laboratories, which show that cAMP increases sodium currents (8, 9, 29).
Our work implies that all types of sodium channels in ATII cells are activated by β-adrenergic agents and cAMP (although the activation of NSC requires an increase in intracellular calcium subsequent to the increase in cAMP). Because we are examining the properties of single channels, there can be no ambiguity about which channels we are activating. Our results do not preclude a role for changes in apical sodium entry resulting from changes in driving force caused by activation of chloride channels in the apical membrane. Indeed, on occasion, we do observe cystic fibrosis transmembrane conductance regulator-like channels that are likely activated by cAMP as they are in many other epithelial tissues. Nonetheless, the change in driving force cannot be very large, since a large change in the apical potential would produce a large change in the unit current observed in cell-attached patches. In fact, the change in current is small; for HSC 6-pS channels, the single channel current in 91 patches is 0.139 ± 0.00572 pA in untreated cells and 0.126 ± 0.00484 pA after 20 μM terbutaline. This difference is not statistically significant, but, even if we presumed it were, the difference would only imply a difference in driving force of 2 mV. Similarly, for NSC in 11 patches, the single channel current is 0.666 ± 0.0877 pA in untreated cells and 0.630 ± 0.0733 pA in terbutaline-treated channels, which also implies a difference in driving force of only 2 mV. These results are interesting since, if only G Na or chloride conductance (G Cl) were increasing, then there should be a large change in driving force as the apical potential moved toward E Na or the equilibrium potential for chloride. That there is no change in driving force in the face of up to a sixfold change in sodium channel activity strongly suggests that both G Na and G Cl are increasing in such a way as to keep the driving force for each ion unchanged. This implies that there is coordinate regulation of bothG Na and G Cl but that the amount of transport depends primarily on NP orather than the driving force.
Regulation of sodium-permeant channels by β-adrenergic agonists varies, depending on the type of channel being studied.
We have found that the effect of cAMP on cation channels varies with the type of channel. Any treatment that increases intracellular cAMP appears to increase the P o of 20-pS NSC with little or no change in the density of channels on the apical surface of epithelial cells. Similar regulation of NSC activity by β-agonists has also been reported by other investigators. Yue et al. (44) found that 10 μM terbutaline or PKA and ATP increased P o and the mean open time without affecting single channel conductance of the 27-pS NSC in ATII cells. Marunaka et al. (27) and Tohda and Marunaka (41) also demonstrated that terbutaline increasesP o of an amiloride-sensitive, calcium-activated, chloride-inhibitable NSC in fetal rat alveolar epithelium by increasing the mean open time without any significant change in the mean closed time or in the single channel conductance. Senyk et al. (39) and Berdiev et al. (3) showed that addition of the catalytic subunit of PKA plus ATP to the presumed cytoplasmic side of the lipid bilayers significantly increasedP o of NSC reconstituted into planar lipid bilayers. Virtually all reports of the effects of agents that increase intracellular cAMP report an increase in the P oof NSC. Occasionally, there have been reports that the number of channels also increases (24, 44). The problem with these observations is that, even if there were apparent changes in the number of channels, these may be an experimental artifact, since counting current levels tends to underestimate the number. This is a particular problem if the P o is low (since channel openings from multiple channels are unlikely to overlap). If theP o increases significantly, then the probability that multiple channels will open simultaneously increases substantially, thus making it appear that the patch now has more channels than before treatment. As one of us has established in published work (22, 26), there are relatively straightforward statistical tests that can establish the confidence that one can place in estimates of the number of channels. We have used them in this work, but such tests have not been commonly used previously in the lung cell single channel literature, making any conclusion that both the number of channels andP o change unclear.
In contrast to NSC, HSC show little if any change inP o in response to changes in cellular cAMP. Rather, cAMP appears to cause insertion of new HSC in the apical membrane of epithelial cells. Similar results have been reported from renal epithelial (A6) cells with an increase in density of HSC after treatment with terbutaline, but no increase inP o (8, 26). Marunaka and Eaton (26) found that pretreatment of renal epithelial A6 cells for 1 h with 1 mM dibutyryl-cAMP increasedNP o, which was the result of the increase in the number of conductive sodium channels in the membrane with no change in the P o of individual sodium channels. More recently, Baxendale-Cox (2) showed that terbutaline increases the number of open epithelial sodium channels inXenopus laevis distal lung epithelia. However, Awayda et al. (1) were unable to show PKA-mediated activation of the cloned α-, β-, and γ-rat epithelial sodium channel expressed in Xenopus oocytes and in planar bilayers.
Molecular mechanisms underlying regulation of sodium channels by β-agonists.
As a first step toward elucidation of the mechanism of action of β-adrenergic agents, we looked at the effect of cAMP on sodium channels, since β-agonists are known activators of adenylyl cyclase. Commercially available analogs of cAMP had a similar effect on NSC as terbutaline. Furthermore, forskolin, a direct stimulator of adenylyl cyclase, increased the P o of NSC by sevenfold. One of the ways by which cAMP produces its effects is by stimulating PKA. H-89, a commercially available PKA inhibitor, blocked the effect of terbutaline, implying a cAMP- and PKA-dependent mechanism of action. H-89 also blocked the effect of forskolin. In our studies, propranolol (a β-antagonist) abolished the terbutaline effect, implying that the effect was mediated via activation of β-adrenergic receptors. The presence of both β1 and β2 receptors has been reported in vivo by autoradiographic techniques (7). Taken together, these findings point to a regulatory role for β-adrenergic agents in regulation of lung sodium and water transport.
The final step in modification of channels in this pathway is still unclear. Our studies show that this step may be different for NSC and HSC. For NSC, changes in intracellular calcium are required for activation. Because the effect can be mimicked in excised, cell-free patches of membrane, the simplest explanation would be that calcium is interacting directly with the channel. There are numerous examples of channels directly activated by calcium (e.g., see Refs. 13and 21). Because the cytosolic solution does not contain any ATP, the effect of calcium is unlikely to be mediated by a calcium-dependent protein kinase. On the other hand, calcium can activate membrane-associated enzymes (like phospholipase A2 and C) that remain active in excised patches (10). At least one lipid product of these enzymes, arachidonic acid, can activate sodium channels (43). Thus calcium-dependent production of a lipid second messenger could be an alternative mechanism to activate NSC rather than direct activation by calcium.
The mechanism that produces an increase in the density of HSC is also unclear. In A6 cells where increases in intracellular cAMP also cause an increase in HSC number, Marunaka and Eaton (26) showed that the effect of increased intracellular cAMP was probably mediated by incorporation of new channel proteins from the cytosol (probably in vesicles) into the cell membrane. More recently Butterworth et al. (5) have demonstrated in A6 cells that increases in intracellular cAMP dramatically increase the rate of endo- and, by inference, exocytosis, concomitant with increases in HSC density. A similar mechanism of action was proposed by Ito et al. (14) in fetal lung cells, who found thatI sc was increased 2.5-fold 50 min after application of 10 μM terbutaline from the basolateral side, but this response was abolished by pretreatment with 1 μg/ml brefeldin A, which inhibits the intracellular trafficking of membrane proteins from the cytosolic pool to the cell surface.
An examination of all of our single channel data suggests that vesicle insertion may also be the operative mechanism for HSC in ATII cells. In Fig. 7, we have replotted all of the values for the number of channels under untreated and stimulated conditions for both NSC and HSC. As expected, the histograms that describe the distribution of the number of NSC per patch before and after terbutaline or cAMP treatment look the same (i.e., no change in the number of channels). In contrast, the distributions of HSC look quite different. Untreated cells have a distribution of the number of channels per patch that is described by a single Gaussian distribution with a mean of 2.1 channels/patch. After 10 min and 1 h of treatment, the distributions cannot reasonably be fitted by any less than two Gaussians. After only 10 min of exposure, one Gaussian has a mean near that of untreated cells (∼2.8 channels/patch) and one with a mean of 7.9 channels/patch. One interpretation of these results is that, as we make patches at random places on the cells, we find areas into which a vesicle has fused and that have a high density of channels and other areas with channel densities like untreated cells. In general, membrane proteins are mobile in the plane of the membrane, and, after 1 h, the main distribution is broader with a mean value of 3.9 channels/patch, although a much smaller distribution with a mean of 10.4 channels/patch is still visible. We interpret these observations to mean that HSC are inserted in packets of seven or eight channels and gradually diffuse to form an approximately uniform distribution with about two times as many channels per unit area of membrane as untreated cells.
There is one report by Palmer and Frindt (36) that theP o of channels in renal principal cells, which appear identical to lung HSC, is reduced by intracellular calcium. This is apparently inconsistent with our result in which there is little if any effect on P o of HSC when intracellular calcium increases in response to terbutaline (although theP o after cAMP treatment may be slightly less than untreated cells; Table 1). However, in this previous work of Palmer's (36), intracellular calcium was increased by use of the calcium ionophore ionomycin at concentrations that are likely to cause a sustained level of intracellular calcium in excess of several micromolar. Such a large sustained increase could likely produce a different response than β-adrenergic stimulation in which the sustained level of calcium is much less, ∼250 nM (Fig. 1).
In summary, our studies show that β-agonists regulate the activity of both NSC and HSC. Although the mechanism of action is different for the two channel types, the net effect would be to increase reabsorption of alveolar sodium and water. Studies have also shown that β-agonists increase the activity of sodium-potassium-ATPase in the basolateral membrane of ATII cells, thus complementing the increased influx of sodium in the cell (37). Taken together, these studies point to a role for β-agonists and other agents that increase intracellular cAMP in regulation of lung sodium reabsorption.
Address for reprint requests and other correspondence: L. Jain, Dept. of Pediatrics, Emory Univ. School of Medicine, 2040 Ridgewood Dr. NE, Atlanta, GA 30322 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2002 the American Physiological Society