Br. J. Pharmacol. (1992), 106, 774-782

'."

Macmillan Press

Ltd, 1992

Adenosine receptors on human airway epithelia and their relationship to chloride secretion 'E.R. Lazarowski, S.J. Mason, L. Clarke, *T.K. Harden & R.C. Boucher Departments of Medicine and *Pharmacology, University of North Carolina, Chapel Hill, North Carolina 27599, U.S.A. 1 We have characterized an adenosine receptor subtype present in human airway epithelial cells by measuring the changes in the intracellular levels of adenosine 3':5'-cyclic monophosphate (cyclic AMP) and the rate of transepithelial Cl- secretion. 2 Primary cultures of human nasal epithelium obtained from excised surgical airway epithelial tissues and the cell lines BEAS39 and CF/T43 derived from human airway epithelium were grown on plastic dishes and labelled with [3H]-adenine for measurement of intracellular cyclic AMP accumulation. Primary cultures were loaded with the calcium indicator fura-2 to measure [Ca2+]i and studied as polarized, ion transporting epithelia on collagen matrix supports for measurement of C1- secretion. 3 Adenosine analogues stimulated cyclic AMP accumulation with a rank order of potency characteristic of an A2-receptor: 5-N-ethyl-carboxamidoadenosine (NECA) > adenosine > R-phenylisopropyladenosine (R-PIA), 6-N-cyclopentyladenosine (CPA)> S-PIA. NECA increased cyclic AMP accumulation in normal and cystic fibrosis (CF) primary cells as well as in the CF/T43 and BEAS39 cell lines with K0.5 values ranging from 0.3 to 3 LM. Preincubation with NECA resulted in the homologous desensitization of airway epithelial cells. The effect of NECA was specifically inhibited by the adenosine receptor antagonist, aminophylline, in a competitive manner. 4 The Al-adenosine receptor agonists CPA and R-PIA did not inhibit isoprenaline-stimulated cyclic AMP accumulation in CF/T43 cells, and potentiating effects of the adenosine analogues were observed on forskolin-stimulated cyclic AMP accumulation. Adenosine analogues did not cause significant changes in intracellular Ca2+ ([Ca2+],) in airway epithelium. 5 Adenosine analogues, applied to either the serosal or mucosal side of the polarized amiloride pretreated primary cultures, induced changes in I,, with a rank order of potency of agonists similar to that observed for stimulation of cyclic AMP accumulation. Intracellular microelectrode studies indicated that the locus of action was the apical membrane Cl- conductance. Adenosine failed to stimulate C1secretion in CF airway epithelium. 6 These results provide evidence for the existence of an A2-adenosine receptor that modulates intracellular levels of cyclic AMP in human airway epithelium. Activation of this receptor might lead to stimulation of Cl- secretion in amiloride pretreated normal but not CF cells. Keywords: Epithelium; adenosine receptors; cyclic AMP; aminophylline; chloride secretion; cystic fibrosis; purinoceptors

Introduction Extracellular adenosine and adenine nucleotides modulate a variety of physiological processes by interacting with specific membrane bound receptors, termed PI-purinoceptors for those activated by adenosine and P2-purinoceptors for receptors activated by adenosine 5'-triphosphate (ATP) (Burnstock, 1978). The adenosine receptors have been subclassified into two major subtypes, Al and A2, according to their ability to increase (A2) or decrease (Al) adenylate cyclase activity and the relative potencies of a series of adenosine analogues (Van Calker et al., 1979; Londos et al., 1980; Daly, 1982; Stone, 1991). The A2-receptor exhibits a potency order of 5-N-ethyl-carboxamidoadenosine (NECA)>adenosine> R-phenylisopropyladenosine (R-PIA), N-6-cyclopentyladenosine (CPA) > S-PIA, while the order of potency at the Al-receptor is R-PIA, CPA> adenosine> NECA> S-PIA. Activation of both classes of adenosine receptors is competitively antagonized by methylxanthines (Sattin & Rall, 1970; Fredholm, 1980; Daly et al., 1983; Parsons & Stiles, 1985). Recent findings suggest that adenosine receptors may regulate second messenger systems in addition to adenylate cyclase. For example, an adenosine-mediated increase of intracellular Ca2+ ([Ca2+],) in rat mast cells has been reported to be secondary to activation of phosphoinositide turnover

Author for correspondence.

(Ali et al., 1990). Adenosine receptors may also regulate cardiac K + conductance by a mechanism not involving adenosine 3':5'-cyclic monophosphate (cyclic AMP) (Bohm et al., 1986). Ribeiro & Sebastiao (1986) have proposed that a third type of adenosine receptor, termed A3, is coupled to calcium channels and exhibits an agonist potency order not consistent with that of the Ar- or A2-receptor subtypes. In epithelia, adenosine receptor activation is functionally coupled to regulation of ion transport rates. Adenosine regulates the rate of chloride ion secretion in a variety of epithelial tissues including frog cornea (Spinowitz & Zadunaisky, 1979), rabbit ileum (Dobbins et al., 1984), canine trachea (Pratt et al., 1986), rabbit kidney cortical collecting duct (CCD) (Schwiebert et al., 1990) and T84 human colonic epithelial cells (Barrett et al., 1989). In a number of these target tissues the effect of adenosine on Cl- transport is accompanied by elevation of cyclic AMP suggesting involvement of an A2-adenosine receptor. Cl- transport by airway epithelia is likely to play an important role in modulating the volume and composition of pulmonary secretions. A defective regulation of Cl- transport by airway epithelia may be involved in lung diseases, e.g. cystic fibrosis (CF) (Frizzell et al., 1986; Welsh & Liedtke, 1986; Boucher et al., 1988; Riordan et al., 1989; Rommens et al., 1989). We previously observed that extracellular nucleotides such as ATP and UTP stimulate Cl- secretion in both normal and CF airway epithelial cells pretreated with amiloride (Mason et al., 1991). This effect was associated with an

AIRWAY EPITHELIAL ADENOSINE RECEPTORS

elevation of intracellular Ca2+ (Mason et al., 1991) and inositol phosphates (Brown et al., 1991). The objective of the present study was to determine the presence of adenosine receptors on human airway epithelia and to establish whether adenosine receptors modulate Cl- secretion in amiloride pretreated human airway epithelia.

Methods

Cell culture Cell lines derived from airway epithelium from a normal subject (BEAS39, Reddel et al., 1988) and from a CF patient (CF/T43, Jetten et al., 1989) by introduction of SV40T, were grown to confluence in hormone supplemented serum-free medium (KGM) on plastic tissue culture 12 well plates (Costar, Cambridge, MA, U.S.A.) Human nasal epithelium from normal or cystic fibrosis subjects was obtained from freshly excised nasal tissue by standard techniques (Wu et al., 1985). Epithelial cells were harvested from these specimens and cultured to confluence on plastic 12 well plates for cyclic AMP studies, on non-fluorescent vitrogen coated glass coverslips for intracellular Ca2+ measurements, or on permeable collagen matrix supports for bioelectric measurements. An F12 based, serum-free medium supplemented with hydrocortisone 5ngml[', insulin lOLgml', transferrin 5JAgml-', endothelial cell growth substance 3.7 pg ml-', and 30 nM T3 (F12, 4.lX) was the culture medium for primary cells. In previous studies using cholera toxin (CTX)-containing medium, negligible Cl- secretory response to adenosine was observed (Mason et al., 1991). For this reason, CTX was omitted from the culture medium for these studies.

Measurement of cyclic AMP accumulation Cyclic AMP accumulation was monitored essentially as described previously (Hughes & Harden, 1986). Briefly, cells were incubated for 3 h with 1 ml of HEPES-buffered (20 mM, pH 7.5) Eagle's modified essential medium (EMEM) containing 2-33iCi ml-' of [3H]-adenine. The medium was aspirated, the cells were washed twice with 1 ml of HEPES-EMEM and fresh aliquots of 1 ml HEPES-EMEM were added. Except where indicated, papaverine 200 JM, was added O min before incubation with agonists. The reactions were carried out at 37°C and they were terminated at the indicated times by aspiration of the medium followed by the immediate addition of 1 ml 5% trichloroacetic acid containing 0.5 mM cyclic AMP. [3H]-cyclic AMP was separated from [3H]-ATP by sequential passage of the trichloroacetic acid extract over Dowex and alumina columns as described previously (Harden et al., 1982). Aliquots of the cyclic AMP fractions were assayed spectrophotometrically at 259 nm to correct for recovery of cyclic AMP. Experiments were carried out with triplicate samples that differed less than 10% from the mean. The results are expressed as the precentage conversion of [3H]-ATP to [3H]-cyclic AMP. Except where indicated, the cell cultures were supplemented with adenosine deaminase 2 u ml-' for 12 h before measurement of cyclic AMP accumulation. Measurement of intracellular calcium Primary nasal epithelial cells grown to confluence on vitrogen-coated coverslips were loaded with a final concentration of 3pJM Fura-2 acetomethylester (Fura-2/AM) at 37°C for 30min. [Ca2+]i was measured as described by Mason et al. (1991) with a modular microspectrofluorimeter (SPEX Industries Inc., Edison, NJ, U.S.A.).

Bioelectric studies Confluent monolayers of primary human nasal epithelium were mounted in modified Ussing chambers (Knowles et al.,

775

1983; Boucher et al., 1988). The cultures were maintained at 37°C and continuously perfused with a standard Krebs Ringer bicarbonate solution with amiloride (100 pM) added to the apical bath to block transepithelial Na+ transport. The current required to clamp the spontaneous transepithelial potential difference (V,), the short circuit current (4), was measured under voltage clamp (short circuit) conditions. Changes in I4, (AI), transepithelial resistance (R.) and V, were measured as described previously (Mason et al., 1991). Intracellular microelectrode techniques for electrophysiological measurement have been described in detail elsewhere (Willumsen et al., 1989; Clarke et al., 1991). In brief, the culture cups were inserted in a horizontally-oriented Ussing chamber and perfused as indicated above. V, was continuously measured and constant current pulses (-40 AA cm-2, 1) were passed across the tissue for measurement of transepithelial resistance (R, = AV,/1). The equivalent short circuit current (Iq) was calculated from Ohm's law. Cells were impaled vertically across the apical membrane with conventional microelectrodes backfilled with 3 mM KCI (tip resistance -- 130 mQ). The apical membrane potential (Va) and Vt were referenced to the apical bath and the basolateral membrane potential (Vb = V, - Va) was referenced to the basolateral bath. The fractional resistance of the apical membrane (fRa) was calculated from the equation, fRa = Ra/(Ra + Rb) = AVa/AVt where Ra and Rb are the resistance of the apical and basolateral membranes, respectively, and AVa and AV, are the changes in the voltage signals in response to I.

Chemicals and solutions Forskolin, NECA, R-PIA, 2-methylthio-ATP, papaverine and aminophylline were from Research Biochemicals Inc. (Natick, MA, U.S.A.). Isoprenaline, cyclic AMP, CPA, 3isobutyl-l-methylxanthine (IBMX), adenosine deaminase, ADP, grade I neutral alumina and all nucleosides were purchased from Sigma (St. Louis, MO, U.S.A.). AG 50W-X4 Dowex resin was purchased from Bio-Rad Laboratories (Richmond, CA, U.S.A.). UTP, ATP and other adenine nucleotides were from Boehringer Mannheim Biochemicals (Indianapolis, IN, U.S.A.). Fura-2/AM and Fura-2 pentapotassium salt were purchased from Molecular Probes (Eugene, OR, U.S.A.). [2,8-3H]-adenine was purchased from Amersham Corp (Arlington Heights, IL, U.S.A.). Hormones for culture medium were from Collaborative Research Inc (Waltham, MA, U.S.A.). Keratinocyte growth medium (KGM) was purchased from Clonetics (San Diego, CA, U.S.A.). Eagle's modified essential medium and Ham's F-12 cell culture medium were purchased from Gibco (Grand Island, NY, U.S.A.).

Results Under resting conditions, there was no detectable accumulation of [3H]-cyclic AMP from [3H]-ATP in [3H]-adenine prelabelled CF/T43 cells without or with cyclic nucleotide phosphodiesterase (PDE) inhibitors (Table 1). Exposure of the cells to isoprenaline or the adenosine analogue, NECA, resulted in stimulation of [3H]-cyclic AMP accumulation (Table 1). The methylxanthine PDE inhibitors such as IBMX are known to antagonize competitively P,-purinoceptors (Parsons & Stiles, 1985). In contrast, papaverine does not interact with A,- or A2-adenosine receptors. As illustrated in Table 1, 200 JAM papaverine potentiated cyclic AMP elevation at concentrations of NECA that had no effect on cyclic AMP accumulation in the presence of 200 JM IBMX. The response to isoprenaline was similar, irrespective of whether IBMX or papaverine was used as a PDE inhibitor (Table 1). Based on these and other results, 200 JAM papaverine was routinely included as a PDE inhibitor in all assays. The nature of the effect of methylxanthines on the adenosine receptor response is examined further below.

776

E.R. LAZAROWSKI et al.

Table 1 Effect of phosphodiesterase (PDE) inhibitors on cyclic AMP accumulation in CF/T43 cells % Conversion of [31]-ATP to [3H]-cyclic AMP

Control IBMX Papaverine Isoprenaline (10 JM) Isoprenaline (10 JM) + IBMX Isoprenaline (10 JM) + papaverine NECA (1 gM) NECA (1 JM) + IBMX NECA (1 JM) + papaverine NECA (1 00 AM) NECA (100 JM) + IBMX NECA (100 JM) + papaverine

0.00 0.00 0.00 0.38 ± 0.03 4.46 ± 0.32 4.36 ± 0.42 0.00 0.00 2.78 ± 0.08 0.86 ± 0.04 2.04 ± 0.06 2.86 ± 0.05

Cells were labelled and washed as described in Methods and preincubated for 10 min with vehicle, 200 JAM isobutylmethylxanthine (IBMX), or 200 JM papaverine. The cells were challenged for 5 min with the indicated concentration of isoprenaline or 5-N-ethyl-carboxamidoadenosine (NECA). The data are means ± s.d. from one experiment performed in triplicate and are representative of 3 different experi-

this value is brought into question by the fact that the slope of the line of the transformed data (- 1.7) differs significantly from the value (- 1) expected for a competitive antagonist. Lack of achievement of equilibrium conditions under the necessary restricted conditions of this experiment seems likely. Forskolin, which directly activates adenylate cyclase (Seamon & Daly, 1981), is also known to potentiate the actions of receptors activating this enzyme through G,, e.g. P-adrenoceptors and adenosine receptors (Seamon et al., 1981; Daly et al., 1982). We examined the effect of forskolin

2.50

oE

C.2 2.0-

1.5-

ments.

. CL 1.0I -

The pharmacological characteristics of the adenosine receptor involved in the elevation of cyclic AMP in CF/T43 cells was examined. Accumulation of cyclic AMP was stimulated with an order of potency of NECA > adenosine > CPA, R-PIA > S-PIA (Figure 1). Thymidine, cytidine, uridine, guanidine or inosine did not stimulate cyclic AMP accumulation at millimolar concentrations (data not shown). These results suggest that an A2-adenosine receptor activates adenylyl cyclase in CF/T43 cells. The time course of accumulation of cyclic AMP in CF/T43 cells was compared in the presence of maximal concentrations of NECA or isoprenaline. The initial rate of cyclic AMP accumulation in the presence of NECA was approximately one half of that observed for isoprenaline (Figure 2). After a 3-4 min exposure of cells to NECA no further increase in cyclic AMP occurred and addition of fresh NECA did not induce further cyclic AMP formation. In contrast, addition of isoprenaline to NECA-treated cells resulted in a further increase in cyclic AMP formation with an initial rate similar to that observed in cells treated with isoprenaline alone (Figure 2). These results suggest that CF/T43 cells underwent a rapid and receptor-specific desensitization to NECA, as previously described for A2-receptors in other cells (Kenimer & Nirenberg, 1981; Newman & Levitzki, 1983; Ramkumar et al., 1991). The theophylline analogue, aminophylline, has been shown to block the effect of adenosine on A2-receptors (Hartzell, 1979). Preincubation of CF/T43 cells with aminophylline resulted in a concentration-dependent reduction in NECAstimulated cyclic AMP accumulation (Figure 3). In contrast, aminophylline had no effect on the action of isoprenaline (Figure 3). The apparent potency of aminophylline for inhibition of the NECA response varied with the time of preincubation with antagonist and with the length of drug challenge (data not shown). Thus, concentration-effect curves for NECA were generated in the presence of varying concentrations of aminophylline, under conditions in which aminophylline and NECA were added concurrently. Incubations were for 2 min in order to minimize the contribution of agonist-induced desensitization. The concentration-effect curve for NECA was shifted to the right in a parallel fashion by increasing concentrations of aminophylline (Figure 4). These data suggest competitive inhibition of the NECA response by aminophylline. The apparent inhibition constant for aminophylline was estimated by Schild analysis (Arunlakshana & Schild, 1959). The calculated pA2 value was 4.8 which corresponds to a Ki value of 5.3JAM. The validity of

//C

a) 0.5 c 0 C

Ls-i~~

o-80 0-8

10-7

lo-6 0Al-5

10-3

10-4

[Agonist] M Figure 1 Effect of adenosine analogues on the accumulation of cyclic AMP in CF/T43 cells. Cells were labelled and preincubated as indicated in Methods and challenged for 2min with the indicated concentrations of NECA (0), adenosine (@), CPA (0), R-PIA (A), S-PIA (A), ADP (O), ATP (A) or UTP (*). Adenosine deaminase was omitted from the incubations. The data are representative of results from four experiments. For abbreviations, see text.

6.0 .2 5.0I 4.0a- -

Hg

3.0-

I 0-

-Z30)2.0 CX

0 c

01.0o0

0

3

6

9

12

15

Time (min)

Figure 2 Time-dependent effect of isoprenaline and NECA on the accumulation of cyclic AMP in CF/T43 cells. Cells were labelled and preincubated with 200 J1M papaverine as indicated in Methods. The cells were challenged with 1O JM isoprenaline (0) or 100 JAM NECA (0) for the indicated times or they were preincubated with 100AM NECA for 5 min and further incubated for 3 and 5 min with 10 JM isoprenaline (0) or 100IJM NECA (U). Results are the mean of triplicate determinations. The data are representative of results from three experiments. For abbreviations, see text.

AIRWAY EPITHELIAL ADENOSINE RECEPTORS 125 -

c

o

100-

E o _

0

._

wM

5

75-~

a

(0 u

0 50-

.)_

I

2-

25-

O-

0

50

150

100

200

250

[Aminophyllinel .LM Figure 3 Effect of aminophylline on NECA and i,soprenaline stimulated accumulation of cyclic AMP in CF/T43 cel[Is. The cells were preincubated for 10 min with the indicated concenltrations of aminophylline and challenged for 5 min with 10 gM isC)prenaline (0) or NECA (0). The results are plotted as the percentaLge of the response in the absence of aminophylline and are the nnean of triplicate determinations. For abbreviations, see text.

4.0 a-

< 3.5C.2 o

0

>-3.00~~~~~~~~

2-2.0

T CL 011.0 C

o0.

777

potentiated cyclic AMP elevation at concentrations of NECA (10-8 M) that did not cause apparent cyclic AMP accumulation if added alone (Figure 5a). Preincubation of CF/T43 cells with increasing concentrations of aminophylline or IBMX resulted in a concentration-dependent inhibition of the formation of cyclic AMP in response to 30 fM NECA plus 10 1M forskolin. The IC50 values for aminophylline and IBMX were 35 ^4M and 72 JM, respectively (Figure 5b). The inhibitory effect of the methylxanthines was not observed in CF/T43 cells stimulated with forskolin alone (not shown). To test whether the CF/T43 cells selected for study accurately reflected adenosine receptor activation in human airway epithelium, the ability of NECA to stimulate the formation of cyclic AMP in primary culture of human airway epithelial cells from normal and CF patients, as well as in a cell line (BEAS39) derived from a normal subject, was examined. Incubation of the cells for 2 min with NECA caused concentration-dependent increases in cyclic AMP accumulation in each cell type studied (Figure 6). NECA, at a maximal stimulatory concentration, was similarly efficacious in stimulating cyclic AMP accumulation in normal and CF cells, being 3-4 times more effective in the cell lines than in primary cells (Figure 6). The K0.5 values for NECA (in jaM) were as follows: normal primary cells, 0.6 (n = 3); CF primary cells, 3.0 (n = 2); BEAS39 cells, 3.0 (n = 2), and CF/ T43 cells, 1.1 (n = 5). We also examined whether an A,-adenosine receptor capable of inhibiting cyclic AMP accumulation was present on CF/T43 cells. Combined addition of NECA and submaximal concentrations of isoprenaline resulted in slightly (24% ± 11) enhanced cyclic AMP accumulation over that observed with isoprenaline alone. No significant inhibitory effect of CPA or R-PIA on isoprenaline-stimulated cyclic AMP levels occurred (not shown). These data suggest that Al-adenosine receptors are not expressed on airway epithelial cells. The effect of P2-purinoceptor agonists on the levels of cyclic AMP in CF/T43 cells was also studied. Incubation of the cells with either ADP, ATP or UTP (Figure 1), or with ADPPS, 2-methylthio-ATP or ocp-methylene ATP (not shown) did not result in increased cyclic AMP accumulation. Neither ATP, adenosine-5'-0-3-thiotriphosphate (ATPyS) or UTP (each at 100 gM) caused changes in a 1 to 10 min time course of isoprenaline-stimulated cyclic AMP accumulation

C) 0 .0 HIM III,,,,,

,,,,,M.

,I,,, ,

,,,,,

,

10-8 10-7 10-6 10-5 10-4 10- 3

[NECA] M

Figure 4 Effect of aminophylline on the stimulati,on of cyclic AMP accumulation by NECA. CF/T43 cells were lalbelled and preincubated as indicated in the legend of Figure 22. The cells were incubated for 2 min with the indicated concentrrations of NECA alone (0) or with the concurrent addition of 10 gr1 4 (0), 20 gM (0), 50 jaM (U), or 100 JiM (A) aminophylline. The resiults are plotted as percentage conversion of [3H]-ATP to [3H]-cyclic )AMP and are the mean of triplicate determinations. For abbreviati ons, see text.

.)

>. 10-

I

2-8 F C

Ir 20

,, > &

on cyclic AMP accumulation in isoprenalin e or adenosinestimulated CF/T43 cells. Addition of isopre naline with forskolin to CF/T43 cells, resulted in cyclic AMIP accumulation at levels 2-3 times larger than that expelcted for simple additivity between the two stimuli (not ;shown). Similar results were obtained in CF/T43 cells stimiulated with forskolin combined with NECA (Figure 5a) or with adenosine, R-PIA or CPA (not shown). Incubation of cells for 5 min with 1 gAM NECA or higher resulted in maxKimal accumulation of cyclic AMP, and a similar level of cyclic AMP accumulation was produced by 10tiM forslkolin. However,

the addition of forskolin in combination with NECA resulted in marked potentiation of their respective elffects. Forskolin

b

a < 12-

-0

-

6 6-

4

2

-

C 10-

lo-8 lo-0 6 10-4 10-9 10-7 10-5

10--10

[NECA]

0

50

100

150

200

[Antagonist] p.M

M

Figure 5 Potentiation of NECA-stimulated cyclic AMP accumulation by forskolin. CF/T43 cells were labelled and preincubated as indicated in the legend of Figure 2. (a) The cells were challenged for 5 min with the indicated concentrations of NECA with (0) or without (0) the concurrent addition of 10 1M of forskolin. (b) The cells were preincubated for 10 min with the indicated concentrations of aminophylline (0) or IBMX (0) and challenged for 5 min with 30 jAM NECA plus 10pM of forskolin. Results are plotted as the percentage conversion of [3H]-ATP to [3H]-cyclic AMP and they are the mean of triplicate determinations. For abbreviations, see text.

778

E.R. LAZAROWSKI et al. a 0.5

-

0.4 -

0.3

-

0.2

-

0.1

0.0

b

III I 111_ 111_ 111_11

10-9

10-8 10-710-610-510-410-3

b 0.5 a- 0.4

0.3.2 I

0.20.1

a- 0.0 I

10-910-810-710-610-510-410 -3 H a,) 2.5

C -

2.00 0

1.5-

Q) -0

1.0-

recently demonstrated that extracellular adenine nucleotides stimulate Cl- secretion, as measured by changes in I4, in amiloride-pretreated cells (Mason et al., 1991). Pretreatment of monolayers with amiloride (100f1M) effectively removes the Na' absorptive component of 4,, so that the residual I4, reflects a basal Cl- secretory activity (Boucher et al., 1986; Willumsen et al., 1989). The effect of adenosine receptor agonists on IA, in amiloride pretreated polarized human nasal epithelium is illustrated in Figure 7. The observed rank order of potency of Cl- secretory stimulation for the adenosine analogues was NECA > adenosine > R-PIA > CPA when applied to either the apical (Figure 7a) or basolateral (Figure 7b) surface. Representative tracings of I4, in normal and CF primary nasal epithelium are shown in Figure 8. Exposure of amiloride pretreated normal nasal epithelium to a NECA concentration causing maximal cyclic AMP accumulation resulted in the complete lack of response to a second addition of NECA and in a markedly diminished response to isoprenaline (Figure 8a). In contrast, the 4,, response to ATP following NECA and isoprenaline was in the range reported for untreated tissues (Mason et al., 1991). Amiloride pretreated CF tissue showed no C1- secretory response to NECA or isoprenaline, but a large response was observed to ATP (Figure 8b). Additivity between PI-purinoceptor and P-adrenoceptor agonists and between PI and P2-purinoceptor agonists for stimulation of Cl- secretion was tested in normal tissues (Figure 9). When amiloride pretreated cells were stimulated

0.5 0.0

a

10 910-8io0-710-610-510-410-3

d 2.5

-

2.0

-

1.5

-

0

-

120

-

Q 90-

1.0

0.5

150

60 -

0.0 10

30

910-810-710-610-510-410-3

0U

[NECA] M Figure 6 Effect of NECA on [3H]-cyclic AMP accumulation in human airway epithelial cells. Primary tissue from human airway epithelia excised from control subjects (a) or cystic fibrosis (CF) patients (b) and BEAS39 (c) and CF/T43 (d) cells were labelled with [3H]-adenine as indicated in Methods. Cells were preincubated for 10 min with 200 AM papaverine and then challenged with the indicated concentration of NECA for 2 min. The results are plotted as the percent conversion of [3H]-ATP to [3H]-cyclic AMP and are the mean of triplicate determinations. For abbreviations, see text.

10-9 10-8 10-7 10-6

10-5

o-4

b

150 1 120 _R 0

90

-

9

a)

'1> 600

30 -

(isoprenaline = 10 AM, data not shown). These data suggest that cyclic AMP levels are not regulated by P2-purinoceptors in CF/T43 cells. We have previously shown that the P2-purinoceptor agonist, ATP, induces a rapid increase in [Ca2+], (net increase: 739 ± 171 nM) in human nasal epithelium (Mason et al., 1991). To determine whether airway epithelial cells possess a PI-receptor subtype linked to elevation of [Ca2+]i, we studied the effect of the adenosine analogues on fura-2-loaded human nasal epithelial cells. In contrast to the effect of ATP, addition of 100 gM adenosine or NECA caused delayed (1-2 min after addition) and negligible (10-20 nM) changes in [Ca2+]i. Addition of CPA and R-PIA initiated no change in [Ca2]j (data not shown). In bioelectric studies of human nasal epithelium, we have

01111111

1111111

1111111

1111111

10-9 10-8 1o-7 10-6 10-5 Agonist [Ml

1111111

1o-4

Figure 7 Effect of adenosine analogues on IS, changes in human airway epithelium. Confluent monolayers of primary human nasal epithelium were pretreated with amiloride as indicated in Methods followed by apical (a) or basolateral (b) addition of NECA (U), adenosine (0), R-PIA (0) or CPA (A). The results are based on the agonist-stimulated peak change in ISC and are plotted as percentage change in IS. The data represent the mean of 3-12 determinations from tissues obtained from 3-10 different normal donors. ISC values of amiloride pretreated unstimulated cells were 15.5 ± 4.6 pA cm-2 in (a) and 15.7 ± 6.1 LA cm-2 in (b). For abbreviations, see text.

AIRWAY EPITHELIAL ADENOSINE RECEPTORS

1

NECA

A

I 17

b

r

1 p1AL

NECA

Iso

Figure 8 Representative bioelectric tracings of 4,, in human airway epithelium. Primary cultures of human nasal epithelium from a normal (a) or a cystic fibrosis (b) donor were pretreated with amiloride as indicated in Methods and stimulated with consecutive (apical) addition of 100 pM NECA, 10 AM isoprenaline (Iso) and 100 9AM ATP. Control post-amiloride I, were 24 ftA cm-2 in (a) and 38 gA cm2 in (b). The data are representative of experiments performed with tissue from at least four different donors.

779

individual membrane potentials and resistances were studied during cellular impalement with conventional microelectrodes. As shown in Figure 10, upper trace, luminal application of NECA (100 AM) to amiloride pretreated human nasal epithelium cultures induced an increase in transepithelial potential difference (Vt) and a decrease in transepithelial resistance (R.), indicated by a decrease in the amplitude of the current-induced voltage deflection. These changes indicate an increase in the equivalent short circuit current. As shown in the lower trace, NECA application induced an abrupt depolarization of the apical membrane potential (Va). Note also that the apical membrane voltage response to the current pulses decreased to a relatively greater degree than in the Vt tracing, indicating a decrease in the fractional resistance of the apical membrane. Like the previous Ussing chamber data (see Figure 8), NECA stimulated a significant increase in Clsecretion (4q) (Table 2). During the response, the cellular changes were (1) depolarization of Va and (2) a decrease in fRa, which coupled with a decrease in Rt, indicated a decrease in the resistance (i.e., increased conductance) of the apical membrane. Thus, the Cl- secretory response resulting from NECA application was associated with the induction of a depolarizing conductance in the apical membrane, which is consistent with activation of an apical membrane Cl- conductance.

Discussion a 25 -

b 25 l

20 -

20 -

1;-

Eo

E

15-

0

15-

In this study we have expanded our previous investigations of the effect of adenosine and adenine nucleotides in human airway epithelial cells. Previously, we have shown that extracellular ATP stimulates the formation of inositol phosphates (Brown et al., 1991), raises [Ca2+]j, and increases the rate of Cl- transport (Mason et al., 1991) in human nasal epithelium. Here, we demonstrate that extracellular adenosine modulates the intracellular levels of cyclic AMP in human

i

al 10 c 0)

a1 105

-20 -

5-1

NECA

5

--150

o I.

Figure 9 I4, changes in human airway epithelium stimulated by combined addition of NECA, isoprenaline and ATP. Human nasal epithelium from normal donors were treated with amiloride as indicated in Methods. (a) The cells was challenged with 100 9M NECA (hatched columns) either before (left) or after (right) the addition of 109M isoprenaline (open columns). (b) The cells were stimulated by 100 AM ATP alone (left) or after the addition of 100 AM NECA (right). In each set of experiments, the addition of the second agonist followed stabilization of the response to the first one. All additions were on the apical surface of the monolayer. For abbreviations, see text.

with sequential additions of NECA and isoprenaline, the I,, response to the second agonist was substantially reduced, irrespective of the order of the additions (Figure 9a). These results suggest that these agonists stimulated Cl- transport at least in part through a common mechanism. In contrast, the ATP response following NECA was not significantly different from the response to ATP alone suggesting additivity of these agents (Figure 9b). Cyclic AMP-dependent activation of Cl- secretion in human airway epithelium, e.g. in response to isoprenaline, results from a primary activation of an apical membrane Clconductance (Widdicombe et al., 1985; Boucher et al., 1988). To test whether activation of A2-receptors with NECA induces Cl- secretion by a similar mode, changes in the

E -10 >7

-50

-505 -40E > -30-

-20-

I

I

3 2 Time (min)

4

.

0

1

Figure 10 Representative microelectrode tracings of transepithelial and apical membrane responses to NECA. Amiloride pretreated human nasal epithelium was challenged with 1009AM NECA added to the apical bath. Upper trace, transepithelial potential difference (Vt). Lower trace, apical membrane potential (Va). Voltage deflections are in response to constant current pulses (-40 JLA cm-2).

780

E.R. LAZAROWSKI et al.

Table 2 Effect of 5-N-ethyl-carboxamidoadenosine (NECA) on Cl- secretion in amiloride pretreated nasal epithelial cell cultures Transepithelial parameters

vt

(mV) -2.9 ± 0.9 A Basal B NECA -7.2 ± 1.0 D (B-A) -4.3±1.4

R.

(fl.cm2) 298 251 -47 *

'eq

(iuA cm-2)

29 -9.5 ± 3.1 21 -28.6 ± 4.0 ± 13 -19.1 ± 5.9 ±

±

*

Va (mV)

Cellular parameters Vb

(mV)

fR.

± 31.9 -33.1 ±4.4 aD.56 ± 0.06 ± 4 .3 -30.6±4.4 0).43 ± 0.05 6.8 ± I .6 2.5 ± 1.7 -0D.14 ± 0.01

- 30.2 - 23.4

*

*

Basal and NECA (100 gM, luminal application) data values were measured during a continuous cellular impalement of each primary culture with microelectrodes as indicated in Methods. The values represent the means ( ± s.e.mean) from four experiments. All NECA values were recorded simultaneously with the maximal change in Iq. Amiloride (100 tM) was added to the luminal bath prior to experiment. *P < 0.05 (paired t test). Abbreviations: Vo, Va and Vb, transepithelial, apical and basolateral membrane potential differences, respectively; R., transepithelial resistance; Ieq equivalent short-circuit current; and fRa, fractional resistance of the apical membrane.

airway epithelium, and that this effect appears to be mediated by a cell surface receptor that responds to adenosine but not to adenine nucleotides. The effect of adenosine on cyclic AMP suggests the existence of a stimulatory A2-adenosine receptor on airway epithelium. Most notably, the A2-adenosine receptor agonist, NECA, is the most potent and effective agonist at this receptor. The order of potency of the adenosine analogues is consistent with that observed in other tissues for stimulatory A2-adenosine receptors (Van Calker, 1979; Londos et al., 1980; Daly, 1982; Stone, 1990). The potency of NECA for stimulation of cyclic AMP accumulation in epithelial cells was lower than that reported for other tissues, e.g. rat liver (Londos, 1980), and rat brain (Daly et al., 1983). However, Daly (1982) has pointed out that the apparent potencies of adenosine analogues show considerable range in different tissues and the K0.5 values of NECA observed with primary tissue and cell lines from human airway epithelium are within this range. The phenomenon of homologous desensitization is known to occur following the activation of receptors coupled in a stimulatory fashion to the enzyme adenylyl cyclase, e.g. Padrenoceptors (Harden, 1983; Sibley & Lefkowitz, 1985). Previous studies in NG108-15 neuroblastoma-glioma hybrid cells (Kenimer & Nirenberg, 1981), rat normal kidney (RNK) fibroblasts (Newman & Levitzki, 1983) and DDT, MF-2 smooth muscle cells (Ramkumar et al., 1991) have shown that the A2-receptor undergoes homologous desensitization during exposure of cells to adenosine analogues. Consistent with these reports, we have shown (Figure 2) that NECAtreated CF/T43 cells became refractory to further stimulation with the adenosine analogue but not with the P-receptor agonist isoprenaline. It is widely recognized that membrane-bound adenosine receptors are competitively antagonized by methylxanthines (Sattin & Rall, 1970; Fredholm, 1980; Daly et al., 1983; Parsons & Stiles, 1985). As shown in Figures 3-5, the effect of NECA on cyclic AMP accumulation in CF/T43 cells was inhibited by IBMX and aminophylline and this effect was specific for adenosine and its analogues. Recently, it has been reported that adenosine stimulatory receptors can co-exist in the same cell type with those coupled to inhibition of adenylate cyclase (Ramkumar et al., 1990). However, our results show that A,-adenosine receptor agonists R-PIA or CPA did not inhibit the effect of isoprenaline on cyclic AMP accumulation and potentiating interactions rather than inhibitory ones were observed by the adenosine analogues on forskolin-stimulated CF/T43 cells (Figure 5). The later observation supports the premise that the action of adenosine is mediated through G,. A recent report showed that adenosine stimulates phospholipase C and Ca2" mobilization in RBL-2H3 cells with a rank order of potency of agonists characteristic of an A2adenosine receptor (Ali et al., 1990). However, there was only a negligible increase in [Ca2+], in NECA or adenosine-stimu-

lated human airway epithelial cells. Furthermore, we have shown recently that adenosine does not stimulate inositol phosphate accumulation in CF/T43 cells (Brown et al., 1991). While adenylate cyclase is likely to be the earliest effector of A2-adenosine receptor activation, the spectrum of distal targets of this pathway in human airway epithelium has not been fully identified. There have been previous reports that adenosine-stimulated cyclic AMP accumulation and cyclic AMP-dependent protein kinase in epithelial cells is associated with increased Cl- secretion (Spinowitz & Sadunaitzki, 1979; Dobbins et al., 1984; Pratt et al., 1986; Barrett et al., 1990). Similarly, cyclic AMP plays an important role in regulating Cl- transport in human airway epithelium. Evidence for this notion comes from studies showing that isoprenaline and other agents that increase cyclic AMP levels as well as cyclic AMP-dependent protein kinase activate Cl- transport or channels in normal human airway epithelia but not in CF cells (Frizzel et al., 1986; Welch et al., 1986; Schoumacher et al., 1987; Li et al., 1988; Boucher et al., 1988; 1989). From these studies, we might expect adenosine and its analogues to activate Cl- transport in normal epithelium but not in cells from CF patients. Importantly, the direction of stimulated Cl- transport, absorption versus secretion, is only known for amiloride pretreated human nasal epithelia. In this condition adenosine would be expected to induce an increase in secretion. The data presented here confirm these predictions. Indeed, activation of Cl- secretion by adenosine (or analogues) primarily reflects its actions on raising cyclic AMP levels as indicated by the following data: (1) adenosine analogues showed the same order of potency for stimulating Cl- secretion and cyclic AMP accumulation; (2) NECA-stimulated Cl- secretion and cyclic AMP accumulation occurred with similar concentration-dependence; (3) analysis of the microelectrode data indicate that NECA, like isoprenaline, initiates Cl- secretion via an action on the apical membrane Cl1 conductance, a conductance previously shown to be regulated by cyclic AMP-dependent mechanisms; and (4) NECA failed to induce Cl- secretion in cells obatined from patients with a genetic inability to activate a Cl- conductance in response to cyclic AMP, i.e., CF patients. Data from studies of other cell systems indicate that regulatory mechanisms in addition to cyclic AMP might activate Cl- secretion subsequent to adenosine stimulation of airway epithelial cells. Recently, it has been shown that transepithelial Cl- secretion can be stimulated by Ca2` ionophores and by agonists that increase [Ca2+]i and the formation of inositol phosphates (Boucher et al., 1989; Mason et al., 1991; Brown et al., 1991). Furthermore, adenosine-stimulated Cl- secretion in CDD cells was associated with increased inositol phosphate formation (Schwiebert et al., 1990). However, as discussed above with reference to the absence of changes in [Ca2+] in response to adenosine, this is not the case in human airway epithelium. In conclusion, BEAS39, CF/T43 and primary cultures of

AIRWAY EPITHELIAL ADENOSINE RECEPTORS

both normal and CF human airway epithelial cells express an A2-type adenosine receptor that stimulates cyclic AMP accumulation. This receptor is located functionally on both the apical and basolateral surface of polarized airway epithelia. The A2-receptor appears coupled via adenylate cyclase activation to regulating the rate of Cl- secretion in normal airway epithelium. Like the P-adrenoceptor, the stimulatory adenosine receptor is functionally uncoupled

781

from stimulating C1- secretion in CF airway epithelial cells due to a defect in a distal effector element, probably the CF protein itself, and not due to abnormality in A2-receptor or its coupling to adenylate cyclase. This work was supported by the National Institute of Health grant HL34322 and the Cystic Fibrosis Foundation grant R025.

References ALI, H., CUNHA-MELO, J.R., SAUL, W.F. & BEAVEN, M.A. (1990).

Activation of phospholipase C via adenosine receptors provides synergistic signals for secretion in antigen-stimulated RBL-2H3 cells. J. Biol. Chem., 265, 745-753. ARUNLAKSHANA, 0. & SCHILD, H.O. (1959). Some quantitative uses of drug antagonists. Br. J. Pharmacol. Chemother., 14, 48-58. BARRETT, K.E., COHN, J.A., HUOTT, P.A., WASSERMAN, S.I. &

DHARMSATHAPHORN, K. (1990). Immune-related intestinal chloride secretion II. Effect of adenosine on T84 cell line. Am. J. Physiol., 258 (Cell. Physiol. 27), C902-C912. BOHM, M., BRUCKNER, R., NEUMMAN, J., SCHMITZ, W., SCHOLZ,

H. & STARBATTY, J. (1986). Role of guanine nucleotide bindingprotein in the regulation by adenosine of cardiac potassium conductance and force of contraction. Evaluation with pertussis toxin. Naunyn-Schmiedebergs Arch. Pharmacol., 332, 403-405. BOUCHER, R.C., STUTTS, M.J., KNOWLES, M.R., CANTLEY, L. &

GATZY, J.T. (1986). Na' transport in cystic fibrosis respiratory epithelia. Abnormal basal rate and response to adenylate cyclase activation. J. Clin. Invest., 78, 1245-1252. BOUCHER, R.C., COTTON, C.U., GATZY, J.T., KNOWLES, M.R. &

YANKASKAS, J.R. (1988). Evidence for reduced Cl- and increased Na' permeability in cystic fibrosis human primary cell cultures. J. Physiol., 405, 77-103. BOUCHER, R.C., CHENG, H.C., PARADISO, A.M., STUTTS, M.J.,

KNOWLES, M.R. & EARP, H.S. (1989). Chloride secretory response of cystic fibrosis human airway epithelia. Preservation of calcium but not protein kinase C- and A mechanisms. J. Clin. Invest., 84, 1424-1431. BROWN, H.A., LAZAROWSKI, E.R., BOUCHER, R.C. & HARDEN, T.K.

(1991). Evidence that UTP and ATP regulate phospholipase C through a common extracellular 5'-nucleotide receptor in human airway epithelial cells. Mol. Pharmacol., 40, 648-655. BURNSTOCK, G. (1978). A basis for distinguishing two types of purinergic receptor. In Cell Membrane Receptors for Drug and Hormones: A Multidisciplinary Approach. ed. Straub, L.B. pp. 107-118. New York: Raven Press. CLARKE, L.L., PARADISO, A.M., MASON, S.J. & BOUCHER, R.C. (1992). Effect of bradykinin on Na' and Cl- transport in human nasal epithelium. Am. J. Physiol., (Cell Physiol.), 31 C644-C655. DALY, J.W. (1982). Adenosine receptors: target for future drugs. J. Med. Chem., 25, 197-207. DALY, J.W., PADGETT, W. & SEAMON, K.B. (1982). Activation of cyclic AMP-generating systems in brain membranes and slices by the diterpene forskolin. Augmentation of receptor-mediated responses. J. Neurochem., 38, 532-544. DALY, J.W., BUTTS-LAMB, P. & PADGETT, W. (1983). Subclasses of adenosine receptors in the central nervous system: Interaction with caffeine and related methylxanthines. Cell. Mol. Neurobiol., 3, 69-80. DOBBINS, J.W., LAURENSON, J.P. & FORREST, J.N. Jr (1984). Adenosine and adenosine analogues stimulate adenosine cyclic 3',5'monophosphate-dependent chloride secretion in the mammalian ileum. J. Clin. Invest., 74, 929-935. FREDHOLM, B.B. (1980). Are methylxanthine effects due to antagonism of endogenous adenine? Trends Pharmacol. Sci., 1, 129132. FRIZZELL, R.A., RECHKEMMER, G. & SHOEMAKER, R.L. (1986). Altered regulation of airway epithelial cell chloride channels in cystic fibrosis. Science, 233, 558-560. HARDEN, T.K. (1983). Agonist-induced desensitization of the fr adrenergic receptor-linked adenylate cyclase. Pharmacol. Rev., 35, 5-32. HARDEN, T.K., SCHEER, A.G. & SMITH, M.M. (1982). Differential modification of the interaction of cardiac muscarinic and betaadrenergic receptors with a guanine nucleotide binding component(s). Mol. Pharmacol., 21, 570-580.

HARTZELL, H.C. (1979). Adenosine receptors in frog sinus venus: slow inhibitory potentials produced by adenine compounds and acetylcholine. J. Physiol., 293, 23-49. HUGHES, A. & HARDEN, T.K. (1986). Adenosine and muscarine receptors attenuate cyclic AMP accumulation by different mechanisms in 1321 Ni astrocytome cells. J. Pharmacol. Exp. Ther., 237, 173-178. JETTEN, A.M., YANKASKAS, J.R., STUTTS, M.J., WILLUMSEN, N.J. &

BOUCHER, R.C. (1989). Persistence of abnormal chloride conductance regulation in transformed Cystic Fibrosis epithelia. Science, 244, 1472-1475. KENIMER, J.G. & NIREMBERG, M. (1981). Densensitization of adenylate cyclase to prostaglandin El or 2-chloroadenosine. Mol. Pharmacol., 20, 585-591. KNOWLES, M.R., STUTTS, M.J., SPOCK, A., FISHER, N., GATZY, J.T.

& BOUCHER, R.C. (1983). Abnormal ion permeation through cystic fibrosis respiratory epithelium. Science, 221, 1067-1070. LI, M., MCCANN, J.D., LIEDTKE, C.M., NAIRN, A.C., GREENGARD,

P. & WELSH, M.J. (1988). Cyclic AMP-dependent protein kinase open chloride channels in normal but not cystic fibrosis airway epithelium. Nature, 331, 358-360. LONDOS, C., COOPER, D.M.F. & WOLFF, J. (1980). Subclasses of external adenosine receptors Proc. Natl. Acad. Sci. U.S.A., 77, 2551 -2554. MASON, S.J., PARADISO, A.M. & BOUCHER, R.C. (1991). Regulation of ion transport and intracellular calcium by adenosine triphosphate in human normal and cystic fibrosis airway epithelium. Br. J. Pharmacol., 103, 1649-1656. NEWMAN, M.E. & LEVITZKI, A. (1983). Desensitization of normal rat kidney cells to adenosine. Biochem. Pharmacol., 32, 137-140. PARSONS, W.J. & STILES, G.L. (1985). Methylxanthines as adenosine receptor antagonists. Ann. Intern. Med., 103, 643. PRATT, A.D., CLANCY, G. & WELSH, M.J. (1986). Mucosal adenosine stimulates chloride secretion in canine tracheal epithelium. Am. J. Physiol., 251 (Cell Physiol., 20), C167-C174. RAMKUMAR, V., BARRINGTON, W.W., JACOBSON, K.A. & STILES,

G.L. (1990). Demonstration of both A1 and A2 adenosine receptors in DTT1 MF-2 smooth muscle cells. Mol. Pharmacol., 37, 149-156. RAMKUMAR, V., OLAH, M.E., JACOBSON, K.A. & STILES, G. (1991). Distinct pathways of desensitization of Al- and A2- adenosine receptors in DDT, MF-2 cells. Mol. Pharmacol., 40, 639-647. REDDEL, R.R. YANG KE, GERWIN, B.I., MCMENAMIN, M.G., LECHNER, J.F., SU, R.T., BRASH, D.E., PARK, J.B., RHIM, J.S. & HAR-

RIS, C.C. (1988). Transformation of human bronchial epithelial cells by infection with SV40 or adenovirus-12 SV40 hybrid virus, or transfection via strontium phosphate coprecipitation with a plasmid containing SV40 early region genes. Cancer Res., 48, 1904-1909. RIBEIRO, J.A. & SEBASTIAO, A.M. (1986). Adenosine receptors and calcium: Basis for proposing a third (A3) adenosine receptor. Prog. Neurobiol., 26, 179-207. RIORDAN, J.R., ROMMENS, J.M., KEREM, B.-T., ALON, N., ROZMAHEL, R., GRZELCZAK, Z., ZIELENSKI, J., LOK, S., PLAVSIC, N., CHOU, J.-L., DRUM, M.L., IANNUZZI, M.C., COLINS, F.S. & TSUI,

L.C. (1989). Identification of the cystic fibrosis gene: Cloning and characterization of complementary DNA. Science, 245, 10661073. ROMMENS, J.M., IANNUZZI, M.C., KEREM, B.-T., DRUM, M.L., MELMER, G., DEAN, M., ROZMAHEL, R., COLE, J.L., KENNEDY, D., HIDAKA, N., ZSIGA, M., BUCHWALD, M., RIORDAN, J.R., TSUI, L.-C. & COLLINS, F.S. (1989). Identification of the cystic fibrosis gene: chromosome walking and jumping. Science, 245, 10591065.

782

E.R. LAZAROWSKI et al.

SATTIN, A. & RALL, T.W. (1970). The effect of adenosine and adenine nucleotides on cyclic adenosine 3',5'-phosphate content of guinea pig cerebral cortex slices. Mol. Pharmacol., 6, 13-23. SCHWIEBERT, E., LIGHT, D., DIETEL, P., FEJES-TOTH, G., NARAYFEJES-TOTH, A. & STANTON, B. (1990). Protein kinase C and a G protein, a-3, regulate a C1- channel in the cortical collecting duct (CCD). Kidney Int., 37, 216. SEAMON, K.B. & DALY, J.W. (1981). Activation of adenylate cyclase by the diterpene forskolin does not require the guanine nucleotide regulatory protein. J. Biol. Chem., 256, 9799-9801. SEAMON, K.B., PADGETT, W. & DALY, J. (1981). Forskolin: unique diterpene activator of adenylate cyclase in membranes and in intact cells. Proc. Natl. Acad. Sci. U.S.A., 78, 3363-3367. SCHOUMACHER, R.A., SHOEMAKER, R.L., HALM, D.R., TALLANT, E.A., WALLACE, R.W. & FRIZZELL, R.A. (1987). Phosphorylation fails to activate chloride channels from cystic fibrosis airway cells. Nature, 330, 752-754. SIBLEY, D.R. & LEFKOWITZ, R.J. (1985). Molecular mechanisms of receptor desensitization using the P-adrenergic receptor-coupled adenylate cyclase system as a model. Nature, 317, 124-129. SPINOWITZ, B.S. & ZADUNAISKY, J.A. (1979). Action of adenosine on chloride active transport of isolated frog cornea. Am. J. Physiol., 237, (Renal Fluid Electrolyte Physiol., 6), F121-F127.

STONE, T.W. (1991). Receptor for adenosine and adenine nucleotides. Gen. Pharmacol., 22, 25-31. VAN CALKER, D.M., MULLER, M. & HAMPRECHT, B. (1979). Adenosine regulates via two different types of receptors, the accumulation of cyclic AMP in culture brain cells. J. Neurochem., 33, 99-1005. WELSH, M.J. & LIEDTKE, C.M. (1986). Chloride and potassium channels in cystic fibrosis airway epithelia. Nature, 322, 467-470. WIDDICOMBE, J.H., WELSH, M.J. & FINKBEINER, W.E. (1985). Cystic fibrosis decreases the apical membrane chloride permeability of monolayers cultured from cells of tracheal epithelium. Proc. Nati. Acad. Sci. U.S.A., 82, 6167-6171. WILLUMSEN, N.J., DAVIS, C.W. & BOUCHER, R.C. (1989). Intracellular Cl- activity and cellular pathways in cultured human airway epithelium. Am. J. Physiol., 256, C1033-C1044. WU, R., YANKASKAS, J., CHENG, E., KNOWLES, M.R. & BOUCHER,

R.C. (1985). Growth and differentiation of human nasal epithelial cells in culture. Serum free, hormone-supplemented medium and proteoglycan synthesis. Am. Rev. Respir. Dis., 132, 311-320. (Received January 9, 1992 Accepted February 17, 1992)