Amiloride Antagonizes ,6-Adrenergic Stimulation of cAMP Synthesis and CI- Secretion in Human Tracheal Epithelial Cells Pamela B. Davis, Catherine L. Silski, and Carole M. Liedtke Department of Pediatrics, Case Western Reserve University at Rainbow Babies and Childrens Hospital, Cleveland, Ohio
Amiloride, a potent blocker of the sodium channel in airway epithelium, has been administered by aerosol as a therapeutic agent for cystic fibrosis. Because amiloride in high concentration has been reported to interfere with cell functions, including adrenergic responses, we tested the ability of amiloride to inhibit {3-adrenergic responses in human tracheal epithelial cells. Amiloride (10-4 M), applied from the basolateral surface of a cell monolayer, inhibited the changes in transepithelial potential and short circuit current to isoproterenol (10-6 M). The stimulation of cyclic adenosine monophosphate (cAMP) synthesis by isoproterenol was inhibited in dose-dependent fashion by amiloride (P = 0.007 by multivariate ANOVA with multiple samples correction). Amiloride did not affect baseline transepithelial potential, short circuit current, basal cAMP levels, cAMP response to prostaglandin E1 , or basal adenylate cyclase activity measured directly in membrane preparations. Therefore, it is unlikely that amiloride exerts a nonspecific toxic effect on adenylate cyclase, receptor-cyclase coupling, or substrate or cofactor supply. The binding of ['lSI]iodocyanopindolol (ICYP), a {3-adrenergic receptor antagonist, to membranes from human tracheal epithelial cells could be displaced by amiloride with IC so = 410 I-tM; displacement was 70% at 10-3 M amiloride. These data are most consistent with the hypothesis that amiloride inhibits {3-adrenergic responses in airway epithelial cells by occupying {3-adrenergic receptor sites. Therapeutic administration of amiloride should take into account its affinity for adrenergic receptors.
Amiloride, a blocker of the Na/H antiport and the Na channel (I) was first used clinically as a potent diuretic. However, interest in its impact on the lung has increased. Amiloride inhibits airway reactivity to antigen in a guinea pig model (2) and is in clinical trial in aerosol form for the treatment of the lung disease of cystic fibrosis (3) and as an adjunct to aminoglycosides for treatment of lung infection with Pseudomonas cepacia in cystic fibrosis patients (4). For therapeutic effect, amiloride concentrations > 10-4 M in airway surface liquid have been sought, for such a concentration of amiloride is required to achieve blockade of the luminal sodium conductance (5). However, in some tissues, amiloride at concentrations of 10-4 M also has substantial toxic effects, including inhibi(Received in original form April 17, 1990 and in revisedform May 17, 1991) Address correspondence to: Pamela B. Davis, M.D., Ph.D., Pediatric Pulmonary Division, Rainbow Babies and Childrens Hospital, 2101 Adelbert Road, Cleveland, OH 44106. Abbreviations: adenosine triphosphate, ATP; cyclic adenosine monophosphate, cAMP; Dulbecco's modified Eagle's medium, DMEM; transepithelial conductance, G,,; human tracheal epithelial cells, HIE cells; isobutylmethylxanthine, IBMX; iodocyanopindolol, ICYP; short-circuit current with asymmetrical perfusion medium, Isc; short-circuit current with symmetrical perfusion solution, Isc-,q; prostaglandin E1 , PGE 1 ; transepithelial resistance, R,,; transepithelial voltage, V". Am. J. Respir. Cell Mol. BioI. Vol. 6. pp. 140-145, 1992
tion of receptor tyrosine kinases and protein synthesis (I). In renal epithelia, amiloride also interacts directly with adrenergic receptors (6). In airway epithelia, we have shown that the {3-adrenergicsystem is a potent stimulator of intracellular cyclic adenosine monophosphate (cAMP) (7, 8) which, in turn, regulates chloride secretion, ciliary beat frequency, and possibly secretion of macromolecules. In this tissue, interference with these functions might substantially affect the ability to maintain homeostasis. Therefore, we tested the hypothesis that amiloride, at concentrations of 10-4 M and greater, interferes with {3-adrenergicresponses in human tracheal epithelial (HTE) cells by determining the effect of amiloride on isoproterenol-stimulated cAMP synthesis and chloride secretion. Amiloride might have nonspecific effects on cellular metabolism or pH which interfere with signal transduction to adenylate cyclase, G" or adenosine triphosphate (ATP) (substrate) supply for adenylate cyclase, which could blunt all receptor-mediated cAMP responses. To test this possibility, we also measured cAMP accumulation in response to a nonadrenergic agonist as well as adenylate cyclase activity directly.
Materials and Methods Cell Culture HTE cells were obtained from necropsy specimens and cultured as previously described (7) by a modification of the
Davis, Silski, and Liedtke: Amiloride Antagonizes {3-Adrenergic Responses
method ofWu (9). Cells were plated on uncoated plastic tissue culture ware at approximately 2 x 10-5 viable cells/ cm-, Cultures were incubated at 3r C in an atmosphere of 95% air/5% CO 2, Media was changed at 48-h intervals and 16 h before beginning experiments. Cells prepared in this way demonstrate the typical cobblestone appearance, specific staining with fluorescent anti-keratin antisera (10), negative transmonolayer potential, and the presence of microvilli, tonofilamen ts, and tight junctions on electron microscopic examination. The cells became confluent by 7 to 10 days in culture and were studied at this time. For electrophysiologic studies, 0.5 x 106 cells were grown on 0.6-cm 2 Millicell HA filter cups coated with vitrogen. Culture medium was changed as described above. Daily measurement of resistance was made starting on day 3 with a Millicell (Millipore Corp., Bedford, MA) electrical resistance system. Filters were tested for use in the Ussing chamber when resistance exceeded 350 Wcm 2 • Electrophysiology of Monolayers When cultures attained a conductance (Gs) of < 3 mS/cm 2, the cultures were inserted into a modified Ussing chamber that was built at Case Western Reserve University to accommodate 0.6-cm 2 Millipore HA filter cups. The temperature of the perfusing media and chamber were maintained at 37° C in a 5% CO 2/95% air atmosphere. Apical and basolateral monolayer surfaces were initially perfused with Dulbecco's modified Eagle's medium (DMEM):F12 (1:1) (GIBCO, Grand Island, NY). In most experiments, the apical perfusion solution was changed to nominally Cl-free medium of the following composition (in mM): Na gluconate, 136.9; K gluconate, 5.4; NaHC0 3 , 25.0; Ca gluconate, 1.3; MgS04·7H 20, 0.9; NaHP04 • 7H 20, 0.3; and glucose, 6 (pH 7.4). The serosal solution was perfused with DMEM:FI2 in all subsequent experiments. Electrical measurements were made with conventional four electrode circuits. Electrodes were built at Case Western Reserve University. The transepithelial voltage (V,,) was measured between a pair of flowing junction electrodes. The electrodes consisted of Ag/Ag Cl wires bathed in 4 M NaCl solution, which was encased in a plastic jacket with a ceramic tip. The electrodes were placed 2 mm away from the monolayer. Voltage-sensing electrodes were connected to a high-impedance voltmeter (71OC-l dual voltage clamp; University ofIowa, Iowa City, IA). Current from an external DC source was passed by a pair of platinum wires that had been treated with chloroplatinic acid to increase their surface area. During experiments, short-circuit current (Isc_,q with symmetrical perfusion solution of L with asymmetrical perfusion medium) was recorded continuously, except at 20s intervals, at which time V" was clamped to ±1 mV to calculate transepithelial resistance (R,). The series resistance of solution and vitrogen-coated filters without cells was 595.6 ± 18 (n = 10) Wcm 2. Rio of filters with cells was corrected. All measurements of Vte were made in reference to the basolateral solution using the resistance compensation circuit. cAMP Stimulations Cultures were washed for 15 min at 37° C with DMEM, pre incubated 20 min at 3r C with DMEM containing 0.25
141
mM isobutylmethylxanthine (IBMX), and then treated with agonists and antagonists as indicated at 37° C. Medium was aspirated and cAMP extracted from the cells with 0.1 N HCl for at least 30 min and frozen at -20° C for later radioimmunoassay (11). Protein was extracted with 0.1 N NaOH for at least 72 h, then measured by the methods of Lowry and associates (12). To determine the effect of amiloride on cellular cAMP levels, experiments were performed in both the presence and absence of IBMX, with added amiloride at concentrations of 10-6 to 10-3 M. Because of the variability from trachea to trachea (7, 13), each experiment testing the effect of amiloride on cAMP production was performed on cells from the same trachea, some treated with amiloride and some untreated. Receptor agonists tested were i-isoproterenol and prostaglandin E2 (PGE2). For each of these, three doseresponse curves (10-8 to 10-5 M) for the agonist were performed, one in the absence of amiloride, one in the presence of 10-3 M amiloride, and one in the presence of 10-4 M amiloride, in cells from seven separate tracheas. {3-Adrenergic Receptor Binding Studies These studies were conducted as previously described (7). Cells on the culture plate were scraped into DMEM and washed with DMEM, washed with hypotonic buffer, pelleted, and the pellet resuspended in distilled water. After 15 min on ice, the cells were homogenized with a Polytron homogenizer for 15 s on setting 8 and centrifuged for 10 min at 4° C at 40,000 x g. The membrane pellet was resuspended in receptor assay buffer (50 mM Tris HCl [pH 7.4], 10 mM MgCI 2). The radioactive ligand was [125I]iodocyanopindolol (lCYP) , 2,200 Ci/mmol. Displacement studies were performed at concentration ofICYP = 30 pM, a concentration close to the Kd , at 37° C for 90 min, at which time equilibrium binding was achieved. Bound ligand was separated from free by dilution with cold buffer and rapid filtration under vacuum through Whatman GFC filters (Whatman, Hillsboro, OR). The filters were dried and counted in an LKB gamma counter with efficiency 24 % for 1251. Nonspecific binding was defined as that not displaced by propranolol (10-6 M). ICYP displacement curves were analyzed by nonlinear curve fitting (14) or by log-logit analysis. Adenylate Cyclase Assays Membrane preparations were made as described above for receptor-binding studies, except that the final membrane pellet was suspended in buffer containing 25 mM Tris (pH 7.4), MgCb (5 mM), ATP (1 mM), creatine phosphate (5 mM), creatine phosphokinase (20 U), IBMX (0.25 mM), dithiothreitol (1 mM), guanyl5'ylimidodiphosphate (GMPPNP) (1 mM). Assay was begun by addition of ATP and incubation at 37° C, and terminated by boiling. Samples were processed as previously described (15), and cAMP was assayed by radioimmunoassay (11). Chemical Sources ICI 118,551 was kindly provided by Imperial Chemical Industries (Cheshire, UK). Reagents for the competitive protein binding assay for cAMP were purchased from Amersham Corp. (Arlington Heights, IL). 125[ICYP] was purchased from New England Nuclear (Boston, MA). Culture media
142
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992
TABLE I
Effect of amiloride on l-isoproterenol-induced Cl: secretion* Net Change in Bioelectric Property
(mil)
I-Isoproterenol I-Isoproterenol + arniloride Amiloride Amiloride + I-isoproterenol A23187
3.2 -2.6 0.1 0.8 11.8
± ± ± ± ±
0.8 0.8 0.3 0.2 3.3
(11) (7)t (8) (7)+
(5)§
1.3 ± 0.3 (II) -0.5 ± 0.4 (8)t -0.6 ± 0.6 (9) 0.7 ± 0.2 (10) 2.9 ± 1.2 (5)
0.1 ± 0.01 (11) -0.2 ± 0.1 (8)
o
o 0.3 ± 0.2 (5)
Definition of abbreviations: lsc = short-circuit current; V" = transepithelial voltage; G" = transepithelial conductance. • Values are reported as mean ± SE for the number of filters in parentheses. Human tracheal epithelial cells from seven tissue samples were cultured on Millicell HA filters as described in MATERIALS AND METHODS. Filter cups were mounted in an Ussing chamber, and I" was continuously recorded. Monolayers perfused on the apical surface of Cl-free medium displayed a current of 23.2 ± 1.9 p.A/cm' (n = 10), G" of 1.5 ± 0.1 mS/cm' (n = 8), and V" of 16.8 ± 1.8 mV (n = 8). I-Isoproterenol and amiloride were perfused on the basolateral surface only. Because basolateral perfusion with A23187 produced only a small, insignificant change in current, A23187 was added by apical perfusion. Net change in bioelectric parameter was calculated as values obtained after perfusion with indicated drug less values before exposure to drug. Concentration of drugs: I-isoproterenol, 10- 6 M; amiloride, 10-4 M; A23187, 2 x 10- 5 M. t P < 0.00 I, compared with i-isoproterenol alone. +P < 0.01, compared with I-isoproterenol alone. § P < 0.01, compared with current before addition of A23187.
and additives were purchased from GIBCO and Biofluids (Rockville, MD). All other reagents were purchased from Sigma Chemical Co. (St. Louis, MO). EC50 was determined by log-logit analysis of dose response. Data Analysis Dose-response curves for cAMP agonists under the three conditions (no amiloride, 10-4 M amiloride, 10-3 M amiloride) were compared by multivariate ANOVA with multiple samples correction (to adjust for the fact that the responses at different concentrations of the same agonist are not truly independent) using the SPSS/PC+ statistical package (16). Adenylate cyclase values were compared by onetailed paired t test.
Results Functional Effects of Amiloride In intact airway epithelium and cultured monolayers of HTE cells, I-isoproterenol induces Cl secretion through the second messenger cAMP and protein kinase A (17-20). The kinetics of Cl- secretion can be monitored electrophysiologically by continuous recording of I,e in monolayers mounted in a Ussing chamber. We used this approach to determine whether amiloride could block I-isoproterenol-stimulated Cl: secretion. Monolayers displayed a G, of 1.5 ± 0.07 (n = 8) ms/cm' , V" of 4.1 ± 0.4 (n = 8) mY, and Isc-eq of 6.3 ± 0.7 (n = 8) Il-A/cm2 (Table 1). Monolayers treated with 0.25 mM IBMX and I-isoproterenol at concentrations ranging from 10-7 to 10-5 M failed to respond with an increase in I,e unless the apical perfusion medium was changed to Ct-free medium, as others have reported (17). Because of this, we perfused the apical surface with Cl-free medium in all subsequent experiments. Switching to apical Cl-free medium increased baseline current by 16.9 ± 1.7 Il-A/cm2 (n = 8). Cl secretion was investigated by the basolateral application of I-isoproterenol, as shown in Figure 1, left panel. After recording baseline Vte , perfusion with I-isoproterenol was started and a response observed after a 2-min delay. This time was required to deliver the basolateral perfusate to the
filter. Basolateral addition of Ill-M I-isoproterenol increased current by 3.2 ± 0.8 Il-A/cm2 (n = 11) (Table 1). Addition of 10-4 M amiloride to the basolateral perfusion solution significantly reduced the functional response of I-isoproterenol by 2.6 ± 0.8 (n = 7) Il-A/cm2 (Table 1). When amiloride was applied to the basolateral perfusion solution before I-isoproterenol addition (Figure 1, right panel), it alone did not significantly alter bioelectric properties but did block I-isoproterenol-stimulated current (Table 1). Apical application of 20 Il-M A23187 significantly increased current (Figure 1, right panel; Table 1), indicating that the monolayers retained the ability to respond functionally to increased intracellular Ca2+. cAMP Studies We studied this system to ascertain at which step amiloride altered l3-adrenergic-mediated responses. First, we assessed the ability of amiloride to interfere with cAMP production. The finding that arniloride blocks I-isoproterenol-stimulated current but does not affect basal current implied an interference with the l3-adrenergic stimulation system. Amiloride at concentrations ranging from 10-6 to 10-3 M did not affect basal HTE cell cAMP levels either in the presence or absence of the phosphodiesterase inhibitor IBMX (0.25 mM) (Table 2). We have shown previously that I-isoproterenol is a potent agonist for cAMP production in HTE cells (7, 8). The effect of amiloride on I-isoproterenol-stimulated cAMP production in the presence of IBMX is shown in Figure 2. ANOVA with multiple samples correction showed that there was a significant dose-related effect of amiloride on the cAMP response to I-isoproterenol (F = 16.46, P = 0.007). There was a significant interaction effect of amiloride with the concentration of I-isoproterenol being tested. Amiloride had its greatest effect at the lower concentrations of I-isoproterenol (F = 4.7, P = 0.03). In order to distinguish an effect of amiloride on adenylate cyclase or signal transduction from an effect on the l3-adrenergic receptor, we measured the effect of amiloride on the cAMP response to a nonadrenergic agonist, PGE 2 , which
Davis, Silski, and Liedtke: Amiloride Antagonizes i3-Adrenergic Responses
143
30..---------------------------,
15-.----------------,
10
"
E U
-, -< "-
5 AMIL
ISO + AMIL
u
H
•
o
1
IlltIIIIIlIIIIIW"11I 111H+t/; i
A23187
00:1------:r-----:-
E
Amiloride, a potent blocker of the sodium channel in airway epithelium, has been administered by aerosol as a therapeutic agent for cystic fibrosis. Because amiloride in high concentration has been reported to interfere with cell functions, including adrenergic responses, we tested the ability of amiloride to inhibit {3-adrenergic responses in human tracheal epithelial cells. Amiloride (10-4 M), applied from the basolateral surface of a cell monolayer, inhibited the changes in transepithelial potential and short circuit current to isoproterenol (10-6 M). The stimulation of cyclic adenosine monophosphate (cAMP) synthesis by isoproterenol was inhibited in dose-dependent fashion by amiloride (P = 0.007 by multivariate ANOVA with multiple samples correction). Amiloride did not affect baseline transepithelial potential, short circuit current, basal cAMP levels, cAMP response to prostaglandin E1 , or basal adenylate cyclase activity measured directly in membrane preparations. Therefore, it is unlikely that amiloride exerts a nonspecific toxic effect on adenylate cyclase, receptor-cyclase coupling, or substrate or cofactor supply. The binding of ['lSI]iodocyanopindolol (ICYP), a {3-adrenergic receptor antagonist, to membranes from human tracheal epithelial cells could be displaced by amiloride with IC so = 410 I-tM; displacement was 70% at 10-3 M amiloride. These data are most consistent with the hypothesis that amiloride inhibits {3-adrenergic responses in airway epithelial cells by occupying {3-adrenergic receptor sites. Therapeutic administration of amiloride should take into account its affinity for adrenergic receptors.
Amiloride, a blocker of the Na/H antiport and the Na channel (I) was first used clinically as a potent diuretic. However, interest in its impact on the lung has increased. Amiloride inhibits airway reactivity to antigen in a guinea pig model (2) and is in clinical trial in aerosol form for the treatment of the lung disease of cystic fibrosis (3) and as an adjunct to aminoglycosides for treatment of lung infection with Pseudomonas cepacia in cystic fibrosis patients (4). For therapeutic effect, amiloride concentrations > 10-4 M in airway surface liquid have been sought, for such a concentration of amiloride is required to achieve blockade of the luminal sodium conductance (5). However, in some tissues, amiloride at concentrations of 10-4 M also has substantial toxic effects, including inhibi(Received in original form April 17, 1990 and in revisedform May 17, 1991) Address correspondence to: Pamela B. Davis, M.D., Ph.D., Pediatric Pulmonary Division, Rainbow Babies and Childrens Hospital, 2101 Adelbert Road, Cleveland, OH 44106. Abbreviations: adenosine triphosphate, ATP; cyclic adenosine monophosphate, cAMP; Dulbecco's modified Eagle's medium, DMEM; transepithelial conductance, G,,; human tracheal epithelial cells, HIE cells; isobutylmethylxanthine, IBMX; iodocyanopindolol, ICYP; short-circuit current with asymmetrical perfusion medium, Isc; short-circuit current with symmetrical perfusion solution, Isc-,q; prostaglandin E1 , PGE 1 ; transepithelial resistance, R,,; transepithelial voltage, V". Am. J. Respir. Cell Mol. BioI. Vol. 6. pp. 140-145, 1992
tion of receptor tyrosine kinases and protein synthesis (I). In renal epithelia, amiloride also interacts directly with adrenergic receptors (6). In airway epithelia, we have shown that the {3-adrenergicsystem is a potent stimulator of intracellular cyclic adenosine monophosphate (cAMP) (7, 8) which, in turn, regulates chloride secretion, ciliary beat frequency, and possibly secretion of macromolecules. In this tissue, interference with these functions might substantially affect the ability to maintain homeostasis. Therefore, we tested the hypothesis that amiloride, at concentrations of 10-4 M and greater, interferes with {3-adrenergicresponses in human tracheal epithelial (HTE) cells by determining the effect of amiloride on isoproterenol-stimulated cAMP synthesis and chloride secretion. Amiloride might have nonspecific effects on cellular metabolism or pH which interfere with signal transduction to adenylate cyclase, G" or adenosine triphosphate (ATP) (substrate) supply for adenylate cyclase, which could blunt all receptor-mediated cAMP responses. To test this possibility, we also measured cAMP accumulation in response to a nonadrenergic agonist as well as adenylate cyclase activity directly.
Materials and Methods Cell Culture HTE cells were obtained from necropsy specimens and cultured as previously described (7) by a modification of the
Davis, Silski, and Liedtke: Amiloride Antagonizes {3-Adrenergic Responses
method ofWu (9). Cells were plated on uncoated plastic tissue culture ware at approximately 2 x 10-5 viable cells/ cm-, Cultures were incubated at 3r C in an atmosphere of 95% air/5% CO 2, Media was changed at 48-h intervals and 16 h before beginning experiments. Cells prepared in this way demonstrate the typical cobblestone appearance, specific staining with fluorescent anti-keratin antisera (10), negative transmonolayer potential, and the presence of microvilli, tonofilamen ts, and tight junctions on electron microscopic examination. The cells became confluent by 7 to 10 days in culture and were studied at this time. For electrophysiologic studies, 0.5 x 106 cells were grown on 0.6-cm 2 Millicell HA filter cups coated with vitrogen. Culture medium was changed as described above. Daily measurement of resistance was made starting on day 3 with a Millicell (Millipore Corp., Bedford, MA) electrical resistance system. Filters were tested for use in the Ussing chamber when resistance exceeded 350 Wcm 2 • Electrophysiology of Monolayers When cultures attained a conductance (Gs) of < 3 mS/cm 2, the cultures were inserted into a modified Ussing chamber that was built at Case Western Reserve University to accommodate 0.6-cm 2 Millipore HA filter cups. The temperature of the perfusing media and chamber were maintained at 37° C in a 5% CO 2/95% air atmosphere. Apical and basolateral monolayer surfaces were initially perfused with Dulbecco's modified Eagle's medium (DMEM):F12 (1:1) (GIBCO, Grand Island, NY). In most experiments, the apical perfusion solution was changed to nominally Cl-free medium of the following composition (in mM): Na gluconate, 136.9; K gluconate, 5.4; NaHC0 3 , 25.0; Ca gluconate, 1.3; MgS04·7H 20, 0.9; NaHP04 • 7H 20, 0.3; and glucose, 6 (pH 7.4). The serosal solution was perfused with DMEM:FI2 in all subsequent experiments. Electrical measurements were made with conventional four electrode circuits. Electrodes were built at Case Western Reserve University. The transepithelial voltage (V,,) was measured between a pair of flowing junction electrodes. The electrodes consisted of Ag/Ag Cl wires bathed in 4 M NaCl solution, which was encased in a plastic jacket with a ceramic tip. The electrodes were placed 2 mm away from the monolayer. Voltage-sensing electrodes were connected to a high-impedance voltmeter (71OC-l dual voltage clamp; University ofIowa, Iowa City, IA). Current from an external DC source was passed by a pair of platinum wires that had been treated with chloroplatinic acid to increase their surface area. During experiments, short-circuit current (Isc_,q with symmetrical perfusion solution of L with asymmetrical perfusion medium) was recorded continuously, except at 20s intervals, at which time V" was clamped to ±1 mV to calculate transepithelial resistance (R,). The series resistance of solution and vitrogen-coated filters without cells was 595.6 ± 18 (n = 10) Wcm 2. Rio of filters with cells was corrected. All measurements of Vte were made in reference to the basolateral solution using the resistance compensation circuit. cAMP Stimulations Cultures were washed for 15 min at 37° C with DMEM, pre incubated 20 min at 3r C with DMEM containing 0.25
141
mM isobutylmethylxanthine (IBMX), and then treated with agonists and antagonists as indicated at 37° C. Medium was aspirated and cAMP extracted from the cells with 0.1 N HCl for at least 30 min and frozen at -20° C for later radioimmunoassay (11). Protein was extracted with 0.1 N NaOH for at least 72 h, then measured by the methods of Lowry and associates (12). To determine the effect of amiloride on cellular cAMP levels, experiments were performed in both the presence and absence of IBMX, with added amiloride at concentrations of 10-6 to 10-3 M. Because of the variability from trachea to trachea (7, 13), each experiment testing the effect of amiloride on cAMP production was performed on cells from the same trachea, some treated with amiloride and some untreated. Receptor agonists tested were i-isoproterenol and prostaglandin E2 (PGE2). For each of these, three doseresponse curves (10-8 to 10-5 M) for the agonist were performed, one in the absence of amiloride, one in the presence of 10-3 M amiloride, and one in the presence of 10-4 M amiloride, in cells from seven separate tracheas. {3-Adrenergic Receptor Binding Studies These studies were conducted as previously described (7). Cells on the culture plate were scraped into DMEM and washed with DMEM, washed with hypotonic buffer, pelleted, and the pellet resuspended in distilled water. After 15 min on ice, the cells were homogenized with a Polytron homogenizer for 15 s on setting 8 and centrifuged for 10 min at 4° C at 40,000 x g. The membrane pellet was resuspended in receptor assay buffer (50 mM Tris HCl [pH 7.4], 10 mM MgCI 2). The radioactive ligand was [125I]iodocyanopindolol (lCYP) , 2,200 Ci/mmol. Displacement studies were performed at concentration ofICYP = 30 pM, a concentration close to the Kd , at 37° C for 90 min, at which time equilibrium binding was achieved. Bound ligand was separated from free by dilution with cold buffer and rapid filtration under vacuum through Whatman GFC filters (Whatman, Hillsboro, OR). The filters were dried and counted in an LKB gamma counter with efficiency 24 % for 1251. Nonspecific binding was defined as that not displaced by propranolol (10-6 M). ICYP displacement curves were analyzed by nonlinear curve fitting (14) or by log-logit analysis. Adenylate Cyclase Assays Membrane preparations were made as described above for receptor-binding studies, except that the final membrane pellet was suspended in buffer containing 25 mM Tris (pH 7.4), MgCb (5 mM), ATP (1 mM), creatine phosphate (5 mM), creatine phosphokinase (20 U), IBMX (0.25 mM), dithiothreitol (1 mM), guanyl5'ylimidodiphosphate (GMPPNP) (1 mM). Assay was begun by addition of ATP and incubation at 37° C, and terminated by boiling. Samples were processed as previously described (15), and cAMP was assayed by radioimmunoassay (11). Chemical Sources ICI 118,551 was kindly provided by Imperial Chemical Industries (Cheshire, UK). Reagents for the competitive protein binding assay for cAMP were purchased from Amersham Corp. (Arlington Heights, IL). 125[ICYP] was purchased from New England Nuclear (Boston, MA). Culture media
142
AMERICAN JOURNAL OF RESPIRATORY CELL AND MOLECULAR BIOLOGY VOL. 6 1992
TABLE I
Effect of amiloride on l-isoproterenol-induced Cl: secretion* Net Change in Bioelectric Property
(mil)
I-Isoproterenol I-Isoproterenol + arniloride Amiloride Amiloride + I-isoproterenol A23187
3.2 -2.6 0.1 0.8 11.8
± ± ± ± ±
0.8 0.8 0.3 0.2 3.3
(11) (7)t (8) (7)+
(5)§
1.3 ± 0.3 (II) -0.5 ± 0.4 (8)t -0.6 ± 0.6 (9) 0.7 ± 0.2 (10) 2.9 ± 1.2 (5)
0.1 ± 0.01 (11) -0.2 ± 0.1 (8)
o
o 0.3 ± 0.2 (5)
Definition of abbreviations: lsc = short-circuit current; V" = transepithelial voltage; G" = transepithelial conductance. • Values are reported as mean ± SE for the number of filters in parentheses. Human tracheal epithelial cells from seven tissue samples were cultured on Millicell HA filters as described in MATERIALS AND METHODS. Filter cups were mounted in an Ussing chamber, and I" was continuously recorded. Monolayers perfused on the apical surface of Cl-free medium displayed a current of 23.2 ± 1.9 p.A/cm' (n = 10), G" of 1.5 ± 0.1 mS/cm' (n = 8), and V" of 16.8 ± 1.8 mV (n = 8). I-Isoproterenol and amiloride were perfused on the basolateral surface only. Because basolateral perfusion with A23187 produced only a small, insignificant change in current, A23187 was added by apical perfusion. Net change in bioelectric parameter was calculated as values obtained after perfusion with indicated drug less values before exposure to drug. Concentration of drugs: I-isoproterenol, 10- 6 M; amiloride, 10-4 M; A23187, 2 x 10- 5 M. t P < 0.00 I, compared with i-isoproterenol alone. +P < 0.01, compared with I-isoproterenol alone. § P < 0.01, compared with current before addition of A23187.
and additives were purchased from GIBCO and Biofluids (Rockville, MD). All other reagents were purchased from Sigma Chemical Co. (St. Louis, MO). EC50 was determined by log-logit analysis of dose response. Data Analysis Dose-response curves for cAMP agonists under the three conditions (no amiloride, 10-4 M amiloride, 10-3 M amiloride) were compared by multivariate ANOVA with multiple samples correction (to adjust for the fact that the responses at different concentrations of the same agonist are not truly independent) using the SPSS/PC+ statistical package (16). Adenylate cyclase values were compared by onetailed paired t test.
Results Functional Effects of Amiloride In intact airway epithelium and cultured monolayers of HTE cells, I-isoproterenol induces Cl secretion through the second messenger cAMP and protein kinase A (17-20). The kinetics of Cl- secretion can be monitored electrophysiologically by continuous recording of I,e in monolayers mounted in a Ussing chamber. We used this approach to determine whether amiloride could block I-isoproterenol-stimulated Cl: secretion. Monolayers displayed a G, of 1.5 ± 0.07 (n = 8) ms/cm' , V" of 4.1 ± 0.4 (n = 8) mY, and Isc-eq of 6.3 ± 0.7 (n = 8) Il-A/cm2 (Table 1). Monolayers treated with 0.25 mM IBMX and I-isoproterenol at concentrations ranging from 10-7 to 10-5 M failed to respond with an increase in I,e unless the apical perfusion medium was changed to Ct-free medium, as others have reported (17). Because of this, we perfused the apical surface with Cl-free medium in all subsequent experiments. Switching to apical Cl-free medium increased baseline current by 16.9 ± 1.7 Il-A/cm2 (n = 8). Cl secretion was investigated by the basolateral application of I-isoproterenol, as shown in Figure 1, left panel. After recording baseline Vte , perfusion with I-isoproterenol was started and a response observed after a 2-min delay. This time was required to deliver the basolateral perfusate to the
filter. Basolateral addition of Ill-M I-isoproterenol increased current by 3.2 ± 0.8 Il-A/cm2 (n = 11) (Table 1). Addition of 10-4 M amiloride to the basolateral perfusion solution significantly reduced the functional response of I-isoproterenol by 2.6 ± 0.8 (n = 7) Il-A/cm2 (Table 1). When amiloride was applied to the basolateral perfusion solution before I-isoproterenol addition (Figure 1, right panel), it alone did not significantly alter bioelectric properties but did block I-isoproterenol-stimulated current (Table 1). Apical application of 20 Il-M A23187 significantly increased current (Figure 1, right panel; Table 1), indicating that the monolayers retained the ability to respond functionally to increased intracellular Ca2+. cAMP Studies We studied this system to ascertain at which step amiloride altered l3-adrenergic-mediated responses. First, we assessed the ability of amiloride to interfere with cAMP production. The finding that arniloride blocks I-isoproterenol-stimulated current but does not affect basal current implied an interference with the l3-adrenergic stimulation system. Amiloride at concentrations ranging from 10-6 to 10-3 M did not affect basal HTE cell cAMP levels either in the presence or absence of the phosphodiesterase inhibitor IBMX (0.25 mM) (Table 2). We have shown previously that I-isoproterenol is a potent agonist for cAMP production in HTE cells (7, 8). The effect of amiloride on I-isoproterenol-stimulated cAMP production in the presence of IBMX is shown in Figure 2. ANOVA with multiple samples correction showed that there was a significant dose-related effect of amiloride on the cAMP response to I-isoproterenol (F = 16.46, P = 0.007). There was a significant interaction effect of amiloride with the concentration of I-isoproterenol being tested. Amiloride had its greatest effect at the lower concentrations of I-isoproterenol (F = 4.7, P = 0.03). In order to distinguish an effect of amiloride on adenylate cyclase or signal transduction from an effect on the l3-adrenergic receptor, we measured the effect of amiloride on the cAMP response to a nonadrenergic agonist, PGE 2 , which
Davis, Silski, and Liedtke: Amiloride Antagonizes i3-Adrenergic Responses
143
30..---------------------------,
15-.----------------,
10
"
E U
-, -< "-
5 AMIL
ISO + AMIL
u
H
•
o
1
IlltIIIIIlIIIIIW"11I 111H+t/; i
A23187
00:1------:r-----:-
E
