INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE PURINE NUCLEOSIDE ADENOSINE is an important mediator of many physiological functions, including transepithelial electrolyte secretion in human airways. Adenosine mediates its effects through the activation of high-affinity receptors, A₁ and A_2A, that are physiologically relevant and low-affinity receptors, A_2B and A₃, that could play a crucial role under inflammatory conditions (34). Stimulation of adenosine receptors activates anion conduction via both cystic fibrosis transmembrane conductance regulator (CFTR)-dependent and -independent pathways (11, 26, 39), although the underlying mechanisms are not fully understood.

The magnitude of the effect of adenosine on ion transport is related to its concentration in the vicinity of its cell surface receptors. In healthy subjects the average adenosine concentration in airway surface liquid is ~60 µM, whereas in patients with asthma it is ~200 µM (14). The increased adenosine concentration in asthmatic patients raises two important questions: What is the source of adenosine and how are its levels controlled? One factor that is believed to control adenosine concentrations is the balance between the activities of enzymes that catalyze its synthesis and those that catalyze its metabolism. Adenosine is produced by the action of membrane-bound 5'-nucleotidase on extracellular AMP, which is itself produced by the action of nonspecific (alkaline or acidic) phosphatases on ADP and ATP (35). Conversely, the metabolism of adenosine is mediated by the action of either adenosine kinase or adenosine deaminase, resulting in the conversion of adenosine to AMP or inosine, respectively (35). Because the majority of adenosine synthesis occurs extracellularly whereas most of its metabolism occurs intracellularly, inwardly directed transport of adenosine across the plasma membrane is also an important determinant of its extracellular concentration (8).

Because adenosine and other nucleosides are relatively hydrophilic, their uptake and release from cells depend on specialized nucleoside transporter proteins present in the plasma membrane (8). These are members of the concentrative (Na⁺ dependent) nucleoside transporter (CNT) and equilibrative (Na⁺ independent) nucleoside transporter (ENT) families (1, 7, 46). Molecular cloning studies in humans and rodents have identified three distinct members of the concentrative family (CNT1, CNT2, and CNT3) and two members of the equilibrative family (ENT1 and ENT2). Human (h)CNT1 and hCNT2 both transport uridine and certain uridine analogs but are otherwise selective for either pyrimidine (hCNT1) or purine (hCNT2) nucleosides, except for modest transport of adenosine by hCNT1 (37, 38, 41). In contrast, hCNT3 transports both purine and pyrimidine nucleosides (36). hENT1 and hENT2 also transport both purine and pyrimidine nucleosides and are distinguished functionally by a difference in sensitivity to inhibition by nitrobenzylthioinosine (NBTI), hENT2 being NBTI insensitive (12, 18, 19). They also differ in sensitivity to inhibition by the coronary vasodilators dipyridamole, dilazep, and draflazine (hENT1 > hENT2) and in the ability of hENT2 to transport nucleobases as well as nucleosides.

The concentrative (inwardly directed) nucleoside transporters of rodents and humans are expressed in specialized cells such as intestinal and renal epithelia, liver, choroid plexus, splenocytes, macrophages, and leukemic cells (1, 7, 46). The equilibrative (bidirectional) nucleoside transporters have generally lower substrate affinities than the concentrative transporters and occur in most, possibly all, human and rodent cell types. Although the role of nucleoside transport in the control of transepithelial anion secretion is unknown, it has been recently suggested that CNT2 could play an important role in the regulation of Na⁺ reabsorption in cultured rat epididymal epithelium (28).

The aim of the present study was to characterize the role of adenosine receptors and transporters in the regulation of ion transport in human airway epithelial cell line A549. These cells exhibit metabolic and transport properties consistent with type II pneumocytes and, because they do not express CFTR (16), constitute a convenient model for studying CFTR-independent regulation of ion transport by adenosine. Our data show that regulation of adenosine receptor function by hENTs controls ion transport in A549 cells.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell culture. A549 cells were obtained from the American Type Culture Collection (Rockville, MD) and grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 50 µg/ml gentamicin sulfate, 60 µg/ml penicillin G, and 100 µg/ml streptomycin. Cells were maintained in T25 tissue-culture flasks (Costar, Cambridge, MA) at 37°C in a humidified atmosphere of 5% CO₂ in air. Confluent cell layers were passaged using saline solution containing 0.05% trypsin and 0.02% EDTA. For adenosine transport experiments, cells were seeded at a density of 10⁶ cells/cm² onto Costar Snapwell inserts (0.45-µm pore size, 1-cm² surface area) coated with type VI human placental collagen (Sigma, St. Louis, MO). For the first 6 days, cells were grown submerged in culture medium that was changed every 2-3 days. Subsequently, air interface culturing was used, in which the medium was added only to the basolateral side of the inserts. For whole cell patch-clamp studies, cells were seeded onto 35-mm plates (Becton Dickinson, Franklin Lakes, NJ) at a density of 1.5 × 10³ cells/cm² at least 4 h before experiments.

Whole cell patch-clamp recordings. Pipette electrodes were made from thin-walled borosilicate glass (A-M Systems, Everett, WA) with a two-stage vertical puller (Nirashige). Electrode tips were fire polished to a final resistance of 3-6 M immediately before experiments. The composition of pipette and bath solutions is given in Table 1. Cultured cells were rinsed three times in bath solution immediately before being mounted into a holder fixed to the stage of an Olympus IMT-2 inverted research microscope (Lake Success, NY). The holder maintained the bath solution at 37°C by means of a heat-exchange perfusion system. After the pipette had been immersed in the bath solution, offset potentials were compensated before a gigaohm seal was formed. Once sealed, the whole cell configuration was obtained mechanically, by suction, and the cell was immediately clamped to 40 mV. Currents were recorded at 1-min intervals with an Axopatch 200A amplifier and Clampex 8.0 software, both from Axon Instruments (Foster City, CA), in response to the voltage protocol shown in Fig. 1C. All currents were reported with reference to the ground electrode in the bath. The access resistance of the patch and cell capacitance were measured directly by the compensation circuitry of the patch-clamp amplifier and by Clampex 8.0 software. The whole cell capacitance in these experiments, expressed as mean ± SD, was 27 ± 6 pF (n = 125), and only seals with a series resistance of <20 M were used. All data were analyzed by Clampfit 8.0 (Axon Instruments), Microsoft Excel 97 (Seattle, WA), and Micrococal Origin 5.0 (Northampton, MA) software. Traces were first normalized to 1 pF to remove variability due to cell size. The current-voltage relationship was obtained from the mean current during the central 140 ms of the recording. The whole cell current chord conductance () was computed from the equation I = (V  E_rev), where I is the whole cell current, V is the applied voltage, and E_rev is the whole cell current reversal potential. Calculations of chord conductance were performed at V = 40 mV.

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Stimulatory effect of adenosine and nitrobenzylthioinosine (NBTI) on whole cell current. A: representative traces showing the activation of whole cell current by 100 µM adenosine (Ado, n = 8). B: representative traces showing the activation of whole cell current by 10 µM NBTI (n = 15). C: voltage protocol used for all experiments performed in this study. D: summary of experiments done with adenosine and NBTI showing the change in membrane chord conductance on addition of nucleoside or nucleoside transport inhibitor. All values are means ± SE of chord conductance density. *Statistical significance (P < 0.01).

RT-PCR. Total RNA was isolated from 2 × 10⁶ cells using the Qiagen RNeasy kit. The average amount of RNA obtained from 2 × 10⁶ cells was ~400 ng. One-fourth of the RNA was reverse transcribed with the use of Superscript II reverse transcriptase (GIBCO BRL) and either oligo(dT) or random hexamers (50 A₂₆₀ units; Boehringer Mannheim) as primers. Thereafter, PCR was performed in 20-µl reactions with the primer pairs (25 µM) described in Table 2. In addition to the primers designed to amplify sequences of interest, reactions with glyceraldehyde-3-phosphate dehydrogenase (GAPDH)- specific primers were run in all rounds of PCR reactions to serve as internal positive controls. One-tenth of the cDNA was used in PCR experiments, and amplification proceeded by annealing for 30 s at the temperatures indicated in Table 2, followed by an elongation step at 72°C for 1 min. Sequences were amplified over 30 or 38 cycles, and PCR products with the expected sizes, shown in Table 2, were resolved on 1.5% agarose gels. All RT-PCR products were sequenced in one [adenosine receptors and Ca²⁺-dependent intermediate-conductance K⁺ (I_K)-1 channels] or both (nucleoside transporters) directions by Taq dideoxyterminator cycle sequencing with an automated DNA sequencer (model 373A, Applied Biosystems, Foster City, CA).

Nucleoside transport. Experiments were carried out at 20°C in HEPES-buffered Ringer's solution (HPBR) containing (in mM) 135 NaCl, 5.0 KCl, 3.33 NaH₂PO₄, 1.0 CaCl₂, 1.0 MgCl₂, 10 glucose, and 5.0 HEPES (pH = 7.4 at 20°C) or in Na⁺-free HPBR containing (in mM) 140 N-methyl-D-glucamine, 5.0 KH₂PO₄, 1.0 CaCl₂, 1.0 MgCl₂, 10 glucose, and 5.0 HEPES (pH = 7.4 at 20°C). Confluent monolayers of A549 cells grown on permeable filters were washed six times with HPBR or Na⁺-free HPBR and then incubated in the same solution (±1 µM NBTI) for 30 min. Uptake was initiated by adding 10 µM ¹⁴C-labeled adenosine or uridine (0.5 µCi/ml, Amersham Pharmacia Biotech) in HPBR or Na⁺-free HPBR (±1 µM NBTI) to either the apical or the basolateral compartment. Incubations with adenosine also included 1 µM deoxycoformycin to inhibit adenosine deaminase activity. Uptake was terminated after 30 s to 3 min by 10 rapid washes of the cell culture inserts in an ice-cold "stop" solution containing (in mM) 100 MgCl₂ and 10 Tris · HCl (pH = 7.4 at 0°C) (32). The monolayers were dissolved in 0.2 ml 5% (wt/vol) SDS and counted for radioactivity using a Beckman LS 6000IC liquid scintillation counter (Irvine, CA). Nonmediated (passive) uptake was determined in the presence of 1 µM NBTI and excess (5 mM) unlabeled uridine. The protein content of representative monolayers was measured using the Bio-Rad protein standard assay procedure. The flux values shown are means ± SE of n = 5 inserts. Each experiment was repeated at least three times on different batches of cells.

Drugs. Adenosine and deoxycoformycin were prepared in H₂O as 10 mM and 1 mM stock solutions, respectively. NBTI was prepared as a 3 mM stock solution in methanol, clotrimazole as a 30 mM stock solution in ethanol, and amiloride as a 10 mM stock solution in H₂O. All the above drugs were obtained from Sigma. 9-Chloro-2-(2-furyl)[1,2,4]triazolo[1,5-c]quinazolin-5-amine (CGS-15943) was prepared as a 2 mM stock solution in DMSO, 1-3-dipropyl-8-cyclopentylxanthine (DPCPX) was prepared as a 10 µM stock solution in 0.1 N NaOH, and 3,7-dimethyl-1-propargylxanthine (DMPX) was prepared as a 5 mM stock solution in H₂O; all three were purchased from RBI (Natick, MA). Finally, 4,4'-diisothiocyanostilbene-2,2'-disulfonic acid (DIDS; Molecular Probes, Eugene, OR) was prepared as a 5 mM stock solution in H₂O, 1-ethyl-2-benzimidazalinone (1-EBIO; Aldrich, Milwaukee, WI) as a 600 mM stock solution in ethanol, and 10,10-bis(4-pyridinylmethyl)-9(10H)-anthracenone (XE-991, a generous gift from Dr. B. S. Brown, DuPont, Wilmington, DE) as a 10 mM stock solution in 0.1 N HCl.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effect of exogenous and autocrine adenosine on whole cell current in A549 cells. Figure 1A shows typical recordings of whole cell current in A549 cells obtained in nonsymmetrical cationic solutions (pipette 135 mM KCl, bath 135 mM NaCl). Addition of adenosine (100 µM) to the bath solution significantly activated whole cell current (n = 8, P < 0.01). The effect of autocrine adenosine on whole cell current was studied with NBTI, an inhibitor of ENTs. We reasoned that if the uptake of extracellularly produced adenosine were inhibited, its concentration would increase, leading to activation of adenosine receptors. Figure 1B shows that the addition of NBTI (10 µM) to the bath solution had an effect on the whole cell current that was similar to that caused by the addition of exogenous adenosine (n = 15, P < 0.01). Similar experiments performed with NBTI at a concentration that specifically inhibits hENT1 but not hENT2 (1 µM) showed that there was no significant difference (P > 0.05, n = 6) in whole cell current activation by these two concentrations of inhibitor (data not shown). These results suggested that hENT1 may mediate the majority of nucleoside transport in A549 cells.

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Adenosine and NBTI, acting in combination, display no additive effects on whole cell current, suggesting a common pathway of activation. A: representative traces showing no further activation of whole cell current when NBTI (10 µM) is added to cells treated with adenosine (100 µM, n = 3). B: summary of experiments done with adenosine in combination with NBTI. All values are means ± SE of chord conductance density. *P < 0.05.

View larger version (30K): [in this window] [in a new window]   Fig. 3.   CGS-15943 inhibits NBTI-stimulated whole cell current. A: representative traces showing inhibition of 10 µM NBTI-stimulated whole cell current with 1 µM CGS-15943 (n = 6). B: summary of experiments done with NBTI followed by CGS-15943. All values are means ± SE of chord conductance density. *P < 0.05.

View larger version (40K): [in this window] [in a new window]   Fig. 4.   1-3-Dipropyl-8-cyclopentylxanthine (DPCPX), but not 3,7-dimethyl-1-propargylxanthine (DMPX), inhibits adenosine-stimulated current. A: representative traces showing inhibition of 100 µM adenosine-stimulated whole cell current with 10 nM DPCPX (n = 6). B: representative traces showing that 10 µM DMPX does not inhibit adenosine-stimulated whole cell current (n = 6). C: summary of experiments performed with adenosine followed by DPCPX. D: summary of experiments performed with adenosine followed by DMPX. All values in C and D are means ± SE of chord conductance density. *P < 0.05.

Identification of channel types activated by adenosine. To determine the channel types involved in the response to adenosine, ion replacement and pharmacological studies were performed. Replacement of K⁺ in the pipette and bath solutions by Cs⁺ reduced basal conductance by 60%, indicating that K⁺ channels contribute to basal whole cell current (n = 3, P < 0.05). Under these conditions, addition of adenosine to the bath solution had no effect on the whole cell current, indicating that K⁺ channels are the major targets for adenosine action (n = 3, P > 0.05).

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Clotrimazole inhibits whole cell current and the subsequent response to adenosine. A: representative traces showing clotrimazole (10 µM) inhibition of whole cell current (n = 8) and subsequent lack of response to 100 µM adenosine (n = 3). B: summary of experiments showing the effect of clotrimazole alone and in combination with adenosine. Values are means ± SE of chord conductance density. *P < 0.05.

View larger version (28K): [in this window] [in a new window]   Fig. 6.   Amiloride affects neither basal whole cell current nor the subsequent response to adenosine. A: representative traces showing no effect of amiloride (10 µM, n = 6) on whole cell current and no subsequent alteration of response to 100 µM adenosine (n = 4). B: summary of experiments showing the effect of amiloride alone and in combination with adenosine on cell membrane chord conductance density. Values are means ± SE. *P < 0.05.

View larger version (25K): [in this window] [in a new window]   Fig. 7.   4,4'-Diisothiocyanostilbene-2,2'- disulfonic acid (DIDS) inhibits basal whole cell current but does not prevent the subsequent response to adenosine. A: representative traces showing DIDS (50 µM) inhibition of whole cell current (n = 12) and subsequent response to 100 µM adenosine (n = 12). B: summary of experiments showing the effect of DIDS alone compared with baseline (*P < 0.05) and DIDS in combination with adenosine compared with DIDS alone (**P < 0.05). Values are means ± SE of chord conductance density.

Identification of adenosine receptors, nucleoside transporters, and I_K channels using RT-PCR. The regulatory actions of adenosine are mediated via four subtypes of G protein-coupled receptors distinguished as A₁, A_2A, A_2B, and A₃. Gene expression of these receptors was investigated with RT-PCR. As shown in Fig. 8A, A549 cells express mRNA for A₁, A_2A, and A_2B but not A₃ receptors. Interestingly, studies with selective adenosine receptor antagonists indicate that only the A₁ receptor is involved in the regulation of whole cell current.

View larger version (55K): [in this window] [in a new window]   Fig. 8.   Expression of adenosine receptors, Ca2+-dependent intermediate-conductance K+ (IK) channels, and nucleoside transporters, characterized by RT-PCR. A: A549 cells express transcripts for A1, A2A, and A2B but not A3 receptors. B: A549 cells express IK-1 channel mRNA. A representative positive control, using primers specific for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA, is also included. C: A549 cells express transcripts for both human (h) equilibrative nucleoside transporter (ENT)1 and hENT2 but not for concentrative nucleoside transporters hCNT1, hCNT2, and hCNT3. Arrows indicate corresponding DNA marker band sizes for reference. bp, Base pairs.

Functional studies of nucleoside transport by A549 cells. Apical and basolateral uptake of adenosine was measured as a function of time at room temperature and, as shown in Fig. 9, was linear for 3 min (apical > basolateral). Subsequent initial rate measurements to determine the basolateral and apical pathways for adenosine and uridine transport were carried out using a 2-min incubation and are shown in Fig. 10. Apical transport of adenosine was not significantly reduced by removal of extracellular Na⁺ but was substantially inhibited by 1 µM NBTI, a concentration sufficient to block all hENT1-mediated transport activity (Fig. 10A). NBTI-insensitive adenosine uptake was reduced further in the presence of excess unlabeled uridine. Because adenosine and uridine are both transported by hENT1 and hENT2, this result identifies additional hENT2-mediated and passive components of adenosine uptake. At the concentration of adenosine tested (10 µM), the ratio of the contribution of hENT1 (NBTI sensitive) to that of hENT2 (NBTI insensitive), corrected for passive uptake, was 9:1. Basolateral adenosine transport (Fig. 10B) as well as apical and basolateral transport of uridine (Fig. 10, C and D), a universal ENT/CNT permeant, showed similar characteristics (hENT1-to-hENT2 flux ratios 12:1, 5:1, and 4:1, respectively). These results were consistent with the electrophysiological data, which indicated that the majority of nucleoside transport that effects whole cell current is mediated by hENT1.

View larger version (13K): [in this window] [in a new window]   Fig. 9.   Time course of adenosine uptake across the apical and basolateral membranes of A549 cells. Uptake of 10 µM adenosine (pmol/mg protein) is greater across the apical membrane than the basolateral membrane, and it is linear across both membranes for the first 3 min. Data are means ± SE.

View larger version (25K): [in this window] [in a new window]   Fig. 10.   Functional characterization of nucleoside uptake in A549 cells. Initial rates of 10 µM adenosine (A, B) and uridine (C, D) uptake (pmol · mg protein1 · min1) across the apical (A, C) and basolateral (B, D) membranes were measured over a 2-min time course. In all experiments, nucleoside uptake was not significantly effected by the removal of Na+, indicating an absence of concentrative Na+-nucleoside cotransport. Adenosine and uridine uptake across both apical and basolateral membranes was sensitive to inhibition by 1 µM NBTI. However, a small component of nucleoside uptake was not inhibited by NBTI but was sensitive to inhibition by excess unlabeled uridine (5 mM), indicating a minor contribution of NBTI-insensitive nucleoside transport to total nucleoside transport. The remaining nucleoside uptake was via passive (non-transporter-mediated) mechanisms. Similar patterns were observed for adenosine and uridine uptake, indicating broad permeant selectivity for both NBTI-sensitive and NBTI-insensitive mediated transport in A549 cells.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Extracellular adenosine was shown previously to modulate the function of both cation (3, 6, 29, 39) and anion channels (5, 6, 10, 26, 33, 39) in several epithelia. There are also reports suggesting that the presence of nucleoside transporters in epithelial cells may regulate these effects by controlling the effective concentration of adenosine in the vicinity of its receptors (28, 32). The results presented in this paper confirm and extend these observations by identifying and characterizing the ion channels, adenosine receptors, and nucleoside transporters involved in the regulation of whole cell current in A549 cells.

In the lung, most of the extracellular adenosine is derived from cleavage of the nucleotide AMP by the enzyme 5'-nucleotidase, which is located on the outer surface of the cell plasma membrane (35). AMP, in turn, may be generated from epithelium-derived extracellular ATP and other adenosine nucleotides, including cAMP. Because ATP is present at millimolar concentrations in the cytoplasm, it is possible that release of ATP from injured airway cells contributes to the increased concentration of adenosine found in the airway surface liquid of asthmatic patients. In addition, it has been suggested that cells may secrete ATP, by Ca²⁺-dependent vesicular exocytosis or through ATP transporters whose identity is still controversial (22). Intracellular levels of adenosine are normally kept low mainly by its conversion to AMP by the enzyme adenosine kinase, which creates an inwardly directed gradient for adenosine entry into the cell. However, the possibility of direct adenosine release from the airway epithelium cannot be excluded, particularly under conditions of stress.

The results presented in this study show that the application of NBTI, a selective inhibitor of hENT1-mediated adenosine transport, had effects on whole cell current similar to those of the application of exogenous adenosine. Furthermore, the effect of NBTI was not additive with that of adenosine and was inhibitable by the adenosine receptor antagonist CGS-15943, indicating that the effect of NBTI is mediated through the activation of adenosine receptors. Therefore, adenosine transporters could regulate epithelial electrolyte secretion by controlling adenosine concentration in the vicinity of its receptors.

It is important to note that CGS-15943 alone had no effect on the baseline current, indicating that endogenous adenosine does not affect the baseline whole cell current. However, several observations suggest that the effect of autocrine adenosine in vivo may be different from that in vitro. First, epithelial cells are normally covered by a thin (~10 µm) layer of airway surface liquid in vivo, whereas cells in our experiments were covered by a ~1-cm-thick layer of bath solution. Second, in vivo, the whole epithelial monolayer contributes to adenosine generation, in contrast to a single cell in patch-clamp studies. Third, other cell types (e.g., mast cells) in the vicinity of the epithelium may contribute to extracellular adenosine concentration in vivo.

The regulatory actions of adenosine are mediated via four subtypes of G protein-coupled receptors, distinguished as A₁, A_2A, A_2B, and A₃ (17, 25). Activation of each of these receptors has been linked with the regulation of ion transport in epithelial tissues (3, 4, 26, 39). The results of RT-PCR experiments have shown that A549 cells express A₁, A_2A, and A_2B but not A₃ receptors. However, functional studies with specific adenosine receptor antagonists indicate that only the A₁ receptor is involved in the regulation of whole cell current by adenosine. Because A₁ receptors are linked to G_i1/2/3 proteins and their activation increases inositol 1,4,5-trisphosphate generation and intracellular Ca²⁺ concentration (17), it is likely that adenosine stimulates whole cell current by activation of Ca²⁺-dependent ion channels.

Adenosine receptor activation has been shown to activate CFTR Cl channels (11), non-CFTR Cl channels (5, 33), amiloride-sensitive Na⁺ channels (3, 29), and Ca²⁺-dependent K⁺ channels (39). Because A549 cells do not express CFTR (16), they constitute a convenient model for the study of the regulation of non-CFTR anion channels. DIDS reduced whole cell current in A549 cells, indicating significant contribution of non-CFTR anion channels to the basal whole cell current. Interestingly, subsequent application of adenosine activated whole cell current, indicating that DIDS-sensitive anion channels may not be targeted by adenosine.

A549 cells were recently shown to contain amiloride-sensitive Na⁺ channels with molecular and biophysical properties similar to those of alveolar type II cells (27). However, the data from the present study showed whole cell current activation in the presence of amiloride, suggesting that amiloride-sensitive Na⁺ channels are not affected by adenosine. Similar to other examples of tissue-specific regulation of Na⁺ channels (for review, see Ref. 31), this result is clearly different from the effect of adenosine on amiloride-sensitive Na⁺ channels in the kidney (29) and intestine (3), in which adenosine has been shown to be a potent regulator of their function.

Ion replacement studies demonstrated that K⁺ ions make a major contribution to the basal whole cell current in A549 cells. Studies from several laboratories have shown that Ca²⁺- and cAMP-mediated agonists regulate I_K and KCNQ channels, respectively. I_K channels, which are apparently absent in excitable tissues, are predominantly expressed in peripheral tissues including endothelia, epithelia, and the hematopoietic system (23, 24). These channels are thought to play a crucial role in the regulation of Cl and HCO secretion in human airway epithelial cells (13). Activation of the basolateral cAMP-dependent K⁺ channel K_vLQT1 (KCNQ1), in parallel with the apically located CFTR, has been shown to play an important role in maintaining cAMP-dependent Cl secretion in human airways (30). In this study we found that XE-991, a specific inhibitor of cAMP-dependent K⁺ channels (40), had no effect on basal whole cell current or the subsequent response to 100 µM adenosine, suggesting that these channels do not contribute to the cell membrane conductance. In contrast, studies with an opener (1-EBIO) and a blocker (clotrimazole) of I_K channels showed that these channels are a major contributor to current. Similarly, the fact that clotrimazole abolished the current response to adenosine indicated that I_K channels were a target for adenosine action.

Epithelial nucleoside transport has been most extensively studied in intestine, kidney, liver, and choroid plexus (1, 7, 46). Enterocytes of the small intestine, for example, contain transcripts for all five of the CNT and ENT isoforms (21, 36, 37, 44, 45) and express CNT1/2 functional activity in their apical membrane and ENT1 and/or ENT2 functional activity at the basolateral membrane (reviewed in Ref. 46). Cultured T84 cells, a model of intestinal crypt cells, express basolaterally restricted ENT1/2 functional activity and nucleoside uptake across the apical membrane having the characteristics of passive diffusion (32, 42). Similar properties have been described for the colonic epithelial cell line Caco-2 (42), although an earlier study found these cells to express CNT3-type transport activity at the apical surface (2). Immunocytochemical analyses have demonstrated the presence of CNT1 protein in the apical membrane of rat small intestine but not at the basolateral membrane, and a similar apical localization was identified for kidney proximal tubule (20). In rat liver parenchymal cells, CNT1 was abundant in bile canalicular membranes but largely excluded from sinusoidal membranes (20), which, instead, are enriched in CNT2 immunoreactivity (15). Choroid plexus expresses CNT3-type functional activity (43). Much less is known about nucleoside transport in other epithelia, although pharmacological and RT-PCR studies suggest the presence of CNT2 but not CNT1 in rat epididymal epithelium (9).

In the present study, we have used complementary molecular and functional approaches to investigate the nucleoside transport capabilities of A549 cells. In the first series of experiments, RT-PCR was used in conjunction with isoform-specific oligonucleotide primers to test for the presence of hCNT1, hCNT2, hCNT3, hENT1, and hENT2 mRNA. In the second series, transport studies were used to confirm the identity of the expressed nucleoside transporters and to investigate their vectorial distribution (apical vs. basolateral membrane). Our results show that A549 cells lack transcripts for hCNT1, hCNT2, and hCNT3. Functionally, we also failed to detect any Na⁺-dependent adenosine or uridine transport activity. Thus A549 cells represent another example of an epithelial cell line lacking Na⁺-dependent mechanisms of nucleoside transport. Instead, RT-PCR analyses identified transcripts for hENT1 and hENT2. Both transport activities were detected in apical as well as basolateral membranes (hENT1 > hENT2). Although hENT1 and hENT2 both transport adenosine and uridine and are broadly selective for other purine and pyrimidine nucleosides, the two transporters are not functionally equivalent. For example, hENT1 has generally higher apparent substrate affinities, whereas hENT2 is also capable of interacting with nucleobases (12, 18, 19). The two transporters may therefore fulfill complementary but distinct physiological functions.

In summary, the results of this study show that both adenosine receptors and transporters control adenosine effects on K⁺ channel function in A549 cells. Extracellularly generated adenosine is either transported via ENT1 (or, to a lesser extent, ENT2) into the cell or can activate adenosine receptors expressed on the cell surface. Inhibition of adenosine transport leads to an increase in adenosine concentration in the extracellular space and activation of adenosine receptors. Nucleoside transport may therefore represent an endogenous regulatory mechanism for adenosine-dependent control of ion secretion in human airway epithelial cells. A better understanding of this system could lead to the development of a novel therapeutic strategy in asthma and other respiratory disorders characterized by altered composition and quantity of airway surface liquid.