Introduction

Top Summary Introduction Procedures Results Discussion References

In recent years, it has become well established that many G protein-coupled receptors may exist in a spontaneously active form in the absence of agonist (Costa, 1992; Lefkowitz, 1993; Mewes, 1993; Chidiac, 1994; Milano, 1994; Kenakin, 1996). This agonist-independent spontaneous activity of receptors has been most readily observed in cell lines in which receptors are overexpressed or mutated (Adie and Milligan, 1994; Barker et al., 1994; Chidiac et al., 1994; Bond et al., 1995; Newman-Tancredi et al., 1997) and in reconstituted membrane systems with purified receptors and appropriate G proteins (Cerione et al., 1984; Freissmuth et al., 1991). In cells in which a constitutive response to agonist-independent receptor activity can be measured, antagonists that reduce the levels of activity and functional response are referred to as inverse agonists or antagonists with negative intrinsic activity, whereas antagonists that do not reduce activity are called neutral antagonists or antagonists with no intrinsic activity (Costa and Herz, 1989). For example, results of studies using NIH 3T3 fibroblast cells expressing constitutively active wild-type 5-hydroxytryptamine_2C receptors revealed a wide spectrum of negative intrinsic activity for 5-hydroxytryptamine _2C receptor antagonists, from full inverse agonists to neutral antagonists (Barker et al., 1994).

A₁AdoRs are tightly coupled to G proteins (Stiles, 1985; Leung and Green, 1989), and activation of these receptors leads to inhibition of adenylyl cyclase activity (Cooper et al., 1980). A constitutive functional response caused by activity of A₁AdoRs in the absence of an agonist has not yet been conclusively identified. However, evidence for constitutive inhibition by A₁AdoRs of adenylyl cyclase activity was reported by Ma and Green (1992). These investigators found that the A₁AdoR antagonists CPX and BW-A844U [3-(4-amino)phenethyl-1-propyl-8-cyclopentylxanthine] increased adenylyl cyclase activity in detergent-permeabilized embryonic chick ventricular myocytes in the absence of agonist. The action of CPX and BW-A844U was prevented by pertussis toxin treatment (Ma and Green, 1992). The authors postulated that "precoupled" A₁AdoRs in cardiac myocytes exert a tonic inhibition of adenylyl cyclase and that CPX stimulates adenylyl cyclase activity by uncoupling the A₁AdoR from its G protein.

The goals of this study were to demonstrate constitutive activity of A₁AdoRs in intact cells and to identify both neutral antagonists and antagonists with inverse agonist activity at the A₁AdoR. We used intact and broken cell preparations of CHO cells in which human A₁AdoRs were cloned and expressed in high density (Kollias-Baker et al., 1997) as our experimental model. Using these CHO:A₁AdoR cells, we demonstrate that activity of A₁AdoRs in the absence of agonist causes activation of inhibitory G proteins and a tonic reduction of cAMP content. Evidence is also presented that certain A₁AdoR antagonists (i.e., neutral antagonists) have no intrinsic activity in the absence of agonist, whereas others (i.e., inverse agonists or antagonists with inverse agonist activity) cause a response that is the inverse of that caused by agonist.

	Experimental Procedures

Top Summary Introduction Procedures Results Discussion References

Materials. PD81,723 [(2-amino-4,5-dimethyl-3-thienyl)-[3-(trifluoromethyl)phenyl]methanone], the allosteric enhancer of agonist binding to the A₁AdoR (Bruns and Fergus, 1990), and the A₁AdoR antagonists N-0861, N-0840, WRC-0571 [C⁸-(N-methylisopropyl)-amino-N⁶(5'-endohydroxy)-endonorbornan-2-yl-9-methyladenine], and WRC-0342 [N⁶-(5'-endohydroxy)-endonorbornan-2-yl-9-methyladenine] (Martin et al., 1996) were gifts from Dr. Noel Cusack (Discovery Therapeutics, Richmond, VA). Adenosine deaminase, GTPS, adenosine, and HEPES were purchased from Sigma Chemical (St. Louis, MO). Forskolin, 5'-N-ethylcarboxamidoadenosine, CGS-15943 [9-chloro-2-(2-furyl)[1,2,4]triazolo[1,5-c] quinazolin-5-amine], CPA, CPT, XAC, and CPX were purchased from Research Biochemicals (Natick, MA). Ham's F-12 cell culture medium and fetal bovine serum were from GIBCO Life Technologies (Gaithersburg, MD). Antibiotic G-418, Falcon 150-mm culture plates, and Costar 12-well culture plates were from Fisher Scientific (Pittsburgh, PA). CHO cells were purchased from the American Type Culture Collection (Rockville, MD). [³H]CPX, carrier-free ¹²⁵I, and [³⁵S]GTPS were purchased from DuPont-New England Nuclear Research Products (Boston, MA). Penicillin/streptomycin antibiotic mixture was purchased from Mediatech (Washington, DC). The composition of HEPES-buffered Hanks' solution was: 130 mM NaCl, 5.0 mM KCl, 1.5 mM CaCl₂, 0.41 mM MgS0₄, 0.49 mM MgCl₂, 0.5 mM Na₂HPO₄, 0.44 mM KH₂PO₄, 5.6 mM dextrose, and 5 mM HEPES (pH 7.4). Succinyl cAMP tyrosyl methyl ester (Sigma) was radiolabeled with ¹²⁵I in the presence of chloramine T as described by Steiner et al. (1972).

Cell culture. CHO cells, referred to herein as CHO:Wild cells, were cultured as monolayers on plastic culture dishes in Ham's F-12 medium supplemented with 10% fetal bovine serum, 100 units of penicillin G, and 100 µg of streptomycin in a humidified atmosphere of 5% CO₂/95% air at 37°. The density of [³H]CPX binding sites in CHO:Wild cells was 26 ± 2 fmol/mg protein (four experiments). Cells were subcultured twice weekly after detachment using 1 mM EDTA in Ca²⁺/Mg²⁺-free HEPES-buffered Hanks' solution. CHO cells stably expressing the recombinant human A₁AdoR (CHO:A₁AdoR cells) were prepared as described previously (Kollias-Baker et al., 1997) and cultured as for CHO:Wild cells, but with 0.5 mg/ml G-418 in the culture medium. Three different clones of CHO:A₁AdoR cells were used for experiments, and all results were confirmed with cells from two or three clones. The density of A₁AdoRs in these cells was 4000-8000 fmol/mg protein, as determined by assay of [³H]CPX specific binding.

Radioligand binding. CHO cells grown onto 150-mm culture dishes were rinsed with HEPES-buffered Hanks' solution and then removed with a cell scraper and homogenized in ice-cold 50 mM Tris·HCl, pH 7.4. Cell membranes were pelleted by centrifugation of the cell homogenate at 48,000 × g for 15 min. The membrane pellet was washed twice by resuspension in fresh buffer and centrifugation. The final pellet was resuspended in a small volume of 50 mM Tris·HCl, pH 7.4, and stored in aliquots of 1 ml at 80° until used for assays.

cAMP assay. Culture medium was aspirated from nearly confluent cells grown in 12-well culture clusters as adherent monolayers, and warm (37°) HEPES-buffered Hanks' solution was added to each well in the cluster. After 6 min, the Hanks' solution was aspirated, and fresh warm Hanks' solution containing appropriate drugs and adenosine deaminase (Type VIII, 2 units/ml), 0.3 µM forskolin, and 20 µM rolipram was added. Adenosine deaminase was present to degrade endogenous adenosine. Forskolin and rolipram were present to increase adenylyl cyclase activity and inhibit phosphodiesterase activity, respectively, and thereby amplify drug-induced changes in cell cAMP content. This incubation solution was aspirated at the end of 6 min of incubation and replaced immediately with 1 ml of ice-cold 50 mM HCl. cAMP contents of the acid extracts of cells were measured by radioimmunoassay. Cellular extract (100 µl), antibody to cAMP (a generous gift from Dr. Gary Brooker, Georgetown University), and ¹²⁵I-labeled succinyl cAMP tyrosyl methyl ester (~20,000 dpm) were mixed and incubated in glass tubes overnight at 2°. Hydroxyapatite (75 µl; 50:50 v/v in water) was then added to each sample, and the mixtures were incubated for 10-30 min at 2°. An assay was terminated by vacuum filtration of samples and collection of bound radioactivity on filter paper, using a Brandel cell harvester. The radioactivity of antibody-bound ¹²⁵I-labeled cAMP ester adsorbed to hydroxyapatite and trapped on filter paper was quantified by a gamma counter. cAMP content of sample extracts was estimated by comparison of sample results with results of parallel assays of standards of known cAMP content (0.06-8 pmol).

Protein determination. Protein content of cell extracts and cell membrane preparations was determined according to the Bradford method using a kit purchased from BioRad (Richmond, CA). Albumin was used as a standard. Protein content of a nearly confluent monolayer of CHO cells was ~125 µg/well of cells in a 12-well culture cluster.

Data analysis. Values of experimental measurements are expressed either as mean ± standard error or as the negative logarithm of the mean ± standard error (for values of EC₅₀ and K_i, which presumably are log normally distributed). Statistical analysis of differences among values in experiments with multiple-comparison groups was based on analysis of variance followed by Dunnett's test (SigmaStat; Jandel, San Rafael, CA) unless otherwise indicated. Differences between group mean values were considered significant at p < 0.05. Results of radioligand binding assays were analyzed using the computer program KELL for Windows (Elsevier-Biosoft, Cambridge, UK).

	Results

Top Summary Introduction Procedures Results Discussion References

Effects of A₁AdoR agonists and antagonists on G protein activation. The activation of G proteins by G protein-coupled A₁ adenosine receptors was assessed by quantification of binding of [³⁵S]GTPS to membranes prepared from CHO cells expressing these receptors (i.e., CHO:A₁AdoR cells) and associated G proteins. Adenosine deaminase (5 units/ml) was added to the incubation medium to degrade endogenous adenosine. Results of assays of binding of [³⁵S]GTPS to membranes from CHO:A₁AdoR cells are shown in Fig. 1. Both the A₁AdoR agonist CPA and the allosteric enhancer PD81,723 increased the binding of [³⁵S]GTPS, by 241 ± 7 (nine experiments) and 25 ± 3% (four experiments), respectively (p < 0.05 compared with control). Neither of the A₁AdoR antagonists N-0861 (nine experiments), N-0840 (seven experiments), and WRC-0342 (four experiments) caused a significant change of [³⁵S]GTPS binding from control. In contrast, the selective and nonselective A₁AdoR antagonists WRC-0571, CPX, CGS-15943, and XAC reduced the binding of [³⁵S]GTPS in a concentration-dependent manner (Fig. 1). For instance, a 1 µM concentration of WRC-0571, CPX, CGS-15943, or XAC significantly reduced the binding of [³⁵S]GTPS by 28 ± 3% (three experiments), 42 ± 1% (eight experiments), 53 ± 7% (seven experiments), or 47 ± 7% (four experiments), respectively (Fig. 1). Thus, in the absence of an agonist of the A₁AdoR, N-0861, N-0840, and WRC-0342 acted as neutral antagonists, whereas WRC-0571, CPX, CGS-15943, and XAC acted as inverse agonists in this assay. A neutral antagonist is expected to attenuate both the action of an agonist to increase [³⁵S]GTPS binding and the action of an antagonist with inverse agonist activity to decrease [³⁵S]GTPS binding. These expectations were confirmed. As shown in Fig. 2A, N-0861 (1 µM) caused a rightward shift of the CPA concentration-response relationship and a significant change in the value of pEC₅₀ for CPA from 8.54 ± 0.23 (EC₅₀ = 2.9 nM) to 7.74 ± 0.41 (EC₅₀ = 18 nM) (five experiments). The action of CPX to reduce binding of [³⁵S]GTPS was also antagonized by 1 µM N-0861 (Fig. 2B). Thus, the data shown in Figs. 1 and 2 suggest that neutral antagonists (N-0861, N-0840, and WRC-0342) of the A₁AdoR and antagonists with inverse agonist activity (CPX, WRC-0571, XAC, and CGS-15943) were distinguished by analysis of their actions in [³⁵S]GTPS binding assays.

View larger version (27K): [in this window] [in a new window]   Fig. 1.   Effects of A1AdoR ligands on specific binding of [35S]GTPS to membranes prepared from CHO cells expressing the recombinant human A1AdoR. Membranes were incubated for 45 min with indicated drug, 0.4 nM [35S]GTPS, and adenosine deaminase (5 units/ml) as indicated in the text. Control specific binding of [35S]GTPS was 3398 ± 466 dpm. Values of normalized responses to drugs are mean ± standard error of triplicate determinations in each of three to nine experiments. Specific binding of [35S]GTPS in the presence of 1 µM of CPA, PD81,723, WRC-0571, CPX, XAC, or CGS-15943 was significantly different from control (p < 0.05, by one-way analysis of variance and Student-Newman-Keuls multiple-comparison procedure).

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Attenuation by the neutral antagonist N-0861 (1 µM) of both the CPA-induced increase of [35S]GTPS binding (A) and the CPX-induced decrease of [35S]GTPS binding (B) to membranes prepared from CHO cells expressing the recombinant human A1AdoR. Control (absence of drug) specific binding of [35S]GTPS was 2492 ± 275 dpm in A and 3939 ± 495 dpm in B. Values of normalized responses to drugs are mean ± standard error of triplicate determinations in each of five (A) or three (B) experiments. Experiments were done as described in legend to Fig. 1 and the text. Specific binding of [35S]GTPS in the absence and presence of 1 µM N-0861 was significantly different for experiments presented in both A and B (p < 0.05, by two-way repeated measures analysis of variance for two factors: drug concentration and presence/absence of N-0861).

Effects of A₁AdoR agonists/antagonists on cAMP content of intact cells. To further demonstrate the distinctions among agonists, neutral antagonists, and antagonists with inverse agonist activity, experiments were done to study the effects of these classes of compounds on cAMP contents of intact CHO:A₁AdoR and CHO:Wild cells. Two neutral antagonists (N-0861, WRC-0342) and two antagonists with inverse agonist activity (WRC-0571, CPX) were used in these experiments. The contents of cAMP of both CHO:Wild and CHO:A₁AdoR cells in the absence of drugs were 7 ± 2 (six experiments) and 5 ± 2 pmol/mg protein, respectively. Additions of adenosine deaminase (2 units/ml) and CPX (0.1 µM) to the incubating medium of CHO:A₁AdoR cells caused small (9-38%) increases of cAMP content. To increase cAMP content of cells and thereby amplify the effects of A₁AdoR agonists and antagonists, forskolin, an activator of adenylyl cyclase, and rolipram, an inhibitor of cAMP phosphodiesterase, were used. In the presence of 0.3 µM forskolin and 20 µM rolipram, cAMP contents of CHO:Wild and CHO:A₁AdoR cells increased to 462 ± 7 and 147 ± 13 pmol/mg protein (30 and 36 determinations, respectively, in three experiments). This result suggests that forskolin-stimulated adenylyl cyclase activity was lower in CHO:A₁AdoR cells than in CHO:Wild cells. The addition of adenosine deaminase (2 units/ml) to the incubation medium caused a further increase of cAMP content of CHO:A₁AdoR cells but not of CHO:Wild cells (Fig. 3). In two separate experiments, it was determined that an increase of adenosine deaminase activity from 1 to 2, 5, and 10 units/ml did not significantly increase cAMP content of CHO:A₁AdoR cells above that caused by 1 unit/ml adenosine deaminase (not shown). Results of assays of the action of adenosine (0.1 nM to 1 µM) to reduce cAMP content in the presence of 1 µM forskolin and 20 µM rolipram indicated that the concentration of endogenous adenosine in the incubation medium (i.e., adenosine that diffuses from CHO cells into the extracellular medium) in the absence of adenosine deaminase was ~1-2 nM. This concentration of adenosine was thus apparently sufficient to cause a reduction of cAMP content of CHO:A₁AdoR cells by 40% compared with cAMP content of CHO:A₁AdoR cells incubated with adenosine deaminase (Fig. 3). The A₁AdoR antagonist CPX (50 nM) caused an additional increase of cAMP content of CHO:A₁AdoR cells by ~2-fold in the presence of adenosine deaminase (2 units/ml), whereas it did not change cAMP content of CHO:Wild cells (Fig. 3). The data indicate that both endogenous adenosine and spontaneous activity of A₁AdoRs in the absence of agonist act to inhibit adenylyl cyclase activity and reduce cAMP content of CHO:A₁AdoR cells treated with forskolin.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Endogenous adenosine-mediated and agonist-independent A1AdoR-mediated inhibition of forskolin-stimulated cAMP accumulation in CHO cells stably expressing the recombinant human A1AdoR (CHO:A1AdoR). CHO:Wild and CHO:A1AdoR cells were incubated with 0.3 µM forskolin (F), 20 µM rolipram (R), 2 units/ml adenosine deaminase (ADA), and CPX (50 nM for CHO:A1AdoR cells, 1 µM for CHO:Wild) in the combinations indicated. Bars, mean ± standard error for 6-36 independent determinations. *, Significantly different from F+R, p < 0.05. , Significantly different from F+R+ADA, p < 0.05.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Antagonism by inverse agonists (WRC-0571, CPX) and neutral antagonists (N-0861, WRC-0342) of the action of CPA on cAMP content of CHO:A1AdoR cells in the presence of forskolin. CHO:A1AdoR cells were incubated with 0.3 µM forskolin, 20 µM rolipram, and 2 units/ml of adenosine deaminase in the absence (control) and presence of 3 nM CPA alone or together with either 1 µM WRC-0571, 1 µM CPX, 50 µM N-0861, or 50 µM WRC-0342. Control cAMP content of cells incubated with forskolin, rolipram, and adenosine deaminase was 22.1 ± 0.6 pmol/well. Bars, mean ± standard error percent of control for 29-32 determinations in four experiments. *, Significantly different from control, p < 0.05.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Concentration-response relationships for stimulation by four A1AdoR antagonists of cAMP accumulation in CHO:A1AdoR cells. CHO:A1AdoR cells were incubated with 0.3 µM forskolin, 20 µM rolipram, and 2 units/ml of adenosine deaminase in the absence (control) or presence of either CPX (1 nM to 1 µM), WRC-0571 (1 nM to 1 µM), N-0861 (0.1-50 µM), or WRC-0342 (0.1-50 µM). Each point indicates mean ± standard error of 11-16 determinations in four experiments. The EC50 values for CPX- and WRC-0571-mediated increases of cAMP content were 10 and 19 nM, respectively.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   Antagonism by WRC-0342 and N-0861 of actions of CPX and WRC-0571 to increase cAMP content of CHO:A1AdoR cells. Top, CHO:A1AdoR cells were incubated with 0.3 µM forskolin, 20 µM rolipram, and 2 units/ml of adenosine deaminase in the absence (control) and presence of 50 nM CPX alone or together with either 25 µM WRC-0342, 50 µM WRC-0342, or 25 µM N-0861. Bottom, cells were incubated as described for the top but in the absence (control) and presence of 50 nM WRC-0571 alone or together with either 25 µM WRC-0342, 50 µM WRC-0342, or 25 µM N-0861. Bars, mean ± standard error of 10-73 determinations in three to seven experiments. *, Significantly different from control, p < 0.05. **, Significantly different from either CPX (top) or WRC-0571 (bottom), p < 0.05.

	Discussion

Top Summary Introduction Procedures Results Discussion References

The two major findings of this study were that antagonists of A₁AdoRs could cause responses (increased cAMP content of intact cells and decreased binding of [³⁵S]GTPS to membranes) in CHO cells expressing the transfected human A₁AdoR and that two classes of A₁AdoR antagonists could be distinguished by their effects on binding of [³⁵S]GTPS to CHO:A₁AdoR membrane preparations and on cAMP content of intact CHO:A₁AdoR cells. Based on these findings and other supportive evidence, we conclude that in the absence of agonist, the spontaneous activity of human A₁AdoRs expressed at high density (4000-8000 fmol/mg of protein) in CHO cells causes constitutive inhibition of adenylyl cyclase and that N-0861, N-0842, and WRC-0342 are neutral antagonists of the A₁AdoR, whereas CPX, WRC-0571, XAC, and CGS-15943 are antagonists with inverse agonist activity (inverse agonists).

A Two-State Equilibrium Model of Receptors Can Explain Our Results

The observation that a number of G protein-coupled receptors can be spontaneously active has led to a simple model in which receptors exist in an equilibrium between inactive (R) and active (R*) conformational states in the absence of agonist (Wreggett and DeLéan, 1984; Lefkowitz et al., 1993; Bond et al., 1995). Agonists bind preferentially to and stabilize the active state (R*) of the receptor and shift the equilibrium toward increased R*. In contrast, antagonists with inverse agonist activity bind preferentially to and stabilize the inactive state (R) of the receptor and shift the equilibrium toward increased R. On the other hand, neutral antagonists bind with equal affinity to both states of the receptor and hence do not alter the equilibrium between R and R*. The results of our experiments can be interpreted by use of this model. The agonist CPA and the allosteric enhancer PD81,723 both increased [³⁵S]GTPS binding to membranes prepared from CHO:A₁AdoR cells (Fig. 1) and decreased cAMP content of intact CHO:A₁AdoR cells (Fig. 4 and text), actions mediated by stabilization of the active R* state of the A₁AdoR. It is noteworthy that PD81,723 was not found to have agonist-independent actions on either guinea pig isolated heart (Amoah-Apraku et al., 1993) or rat hippocampal brain slices (Janusz et al., 1991). We speculate that an agonist-independent action of PD81,723 may be indicative of the presence of constitutively active A₁AdoRs. A₁AdoR antagonists (Fig. 4), and pretreatment of cells with pertussis toxin blocked the action of CPA, the former by a competitive effect to reduce agonist binding and the latter by blocking G protein-mediated signal transduction. In addition, in the absence of agonist, CPX and WRC-0571 reduced the number of constitutively active (R*) receptors and thereby decreased the binding of [³⁵S]GTPS to G proteins in CHO:A₁AdoR membranes and increased the cAMP content of intact CHO:A₁AdoR cells. The neutral antagonists N-0861 and WRC-0342 bound equally well to both R and R* forms of the A₁AdoR and consequently attenuated both the responses mediated by agonists and allosteric enhancer (Figs. 2A and 4 and text) and those mediated by antagonists with inverse agonist activity (Figs. 2B and 6).

Consideration of Alternative Explanations of Our Data

Several alternative explanations of our data were considered, including (1) the postulated constitutive activity of A₁AdoRs is caused by endogenous adenosine in the preparations and therefore all actions of antagonists are due to attenuation of the effects of this agonist; (2) the actions of A₁AdoR antagonists to increase cAMP content of intact cells are caused by their inhibition of cAMP phosphodiesterase activity; (3) the putative neutral antagonists N-0861 and WRC-0342 are really inverse agonists but were used at concentrations too low to see their "true" effect. Each of these explanations of our data is discussed below and dismissed.

Endogenous adenosine. Antagonism by A₁AdoR antagonists of the action of endogenous adenosine in both membrane and whole-cell preparations may cause a decrease of [³⁵S]GTPS binding and an increase of cAMP content that could be falsely interpreted to be evidence of inverse agonism. Therefore, adenosine deaminase (5 units/ml in membrane preparations for [³⁵S]GTPS binding assay and 2 units/ml in whole-cell preparations) was present in all assays of A₁AdoR agonist and antagonist activity. The addition of adenosine deaminase (0.1-10 units/ml) was found to increase cAMP content of CHO:A₁AdoR cells (Results and Fig. 3), which suggests that adenosine in the bathing medium was acting on A₁AdoRs to inhibit cellular adenylyl cyclase activity. The effect of adenosine deaminase was maximal at 1 unit/ml. In the presence of 2 units/ml of adenosine deaminase, both CPX (50 nM) and WRC-0571 further increased cAMP content of CHO:A₁AdoRs (Fig. 3). These actions of CPX and WRC-0571 therefore are unlikely to be due to the antagonism of endogenous adenosine. Also, WRC-0571 and CPX increased the cAMP content of CHO:A₁AdoR cells by 74% and 64% in the presence of adenosine deaminase, whereas N-0861 and WRC-0342 increased cAMP content of CHO:A₁AdoR cells by only 29% and 0%, respectively (Fig. 5). In contrast, each antagonist caused a similar maximal reduction of [³H]CPX binding to A₁AdoRs in CHO:A₁AdoR membranes. Furthermore, N-0861 and WRC-0342 antagonized the increases of cAMP content of CHO:A₁AdoR cells caused by either CPX or WRC-0571 (Fig. 6). Because all four A₁AdoR antagonists are expected to fully and equally attenuate a response mediated by endogenous adenosine, our findings that the actions of CPX and WRC-0571 are clearly different from those of N-0861 and WRC-0342 are not consistent with a mechanism of action involving inhibition of the effect of endogenous adenosine.

Phosphodiesterase activity. Adenosine receptor antagonists may act to inhibit phosphodiesterase activity and thereby increase cAMP content of cells. This action is a potential explanation of results of our assays of cAMP content of intact cells but not of results of assays of [³⁵S]GTPS binding to membrane preparations. For example, the phosphodiesterase inhibitor rolipram increased cAMP contents of both CHO:A₁AdoR and CHO:Wild cells (Fig. 3) but did not alter binding of [³⁵S]GTPS. For assay of nonspecific actions (e.g., inhibition of phosphodiesterase activity) of A₁AdoR antagonists on cAMP content, CHO:Wild cells were used in this study. Only N-0861 (50 µM) of the four A₁AdoR antagonists tested caused an elevation of cAMP content of CHO:Wild cells in the presence of 0.3 µM forskolin. Assuming that N-0861 (50 µM) caused similar nonspecific increases of cAMP content in CHO:A₁AdoR and CHO:Wild cells, the small increase in cAMP content of CHO:A₁AdoR cells in the presence of 50 µM N-0861 (Fig. 5) also is unrelated to antagonism of A₁AdoRs and is likely to be caused by inhibition of phosphodiesterase activity. The A₁AdoR antagonists CPX, WRC-0571, and WRC-0342 did not appear to increase cAMP content of CHO:Wild cells at the concentrations tested and therefore are not likely to cause inhibition of phosphodiesterase activity in our experiments. Furthermore, the findings that treatment of CHO:A₁AdoR cells with pertussis toxin abolished effects of both A₁AdoR agonist and antagonists on cAMP content and that N-0861 and WRC-0342 attenuated increases of cAMP content caused by WRC-0571 and CPX strongly indicate that actions of A₁AdoR antagonists in our experiments are not due to inhibition of phosphodiesterase activity but rather to antagonism of A₁AdoRs.

Antagonist affinity for A₁AdoRs. The neutral A₁AdoR antagonists N-0861, N-0840, and WRC-0342 at concentrations up to 0.1 mM did not reduce binding of [³⁵S]GTPS to membranes prepared from CHO:A₁AdoR cells. In contrast, both antagonists competitively reduced specific binding of [³H]CPX to CHO:A₁AdoR cell membranes with K_i values of ~1 µM. These findings suggest that although the affinities of N-0861, N-0840, and WRC-0342 for the A₁AdoR are low in comparison to those of CPX, WRC-0571, XAC, and CGS-15943, the concentrations of each antagonist used in functional assays of [³⁵S]GTPS binding and cAMP content were sufficiently high to cause occupancy of nearly all A₁AdoRs. It therefore is not likely that N-0861, N-0840, or WRC-0342 at higher concentrations than those used in our experiments will act as inverse agonists. This conclusion is supported by the finding that actions of N-0861 and WRC-0342 were not additive with, but were antagonistic of, those caused by CPX and WRC-0571 (Figs. 2B and 6).