Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The adenosine A₁ receptor is a member of the seven-transmembrane G-protein-coupled receptor superfamily (Libert et al., 1992; Olah and Stiles, 1995; Shryock and Belardinelli, 1997). Adenosine A₁ receptors couple to pertussis toxin (PTX)-sensitive G-proteins (G_i1, G_i2, G_i3, and G_o) and stimulate numerous intracellular signaling events, such as inhibition of adenylyl cyclase, the closure of voltage-sensitive Ca²⁺ channels on nerve terminals, and the opening of potassium channels (Olah and Stiles, 1995; Figler et al., 1996; Srinivas et al., 1997). Stimulation of A₁ receptors also activates inositol phospholipid hydrolysis and calcium mobilization via PTX-sensitive G-proteins in many cell systems. These latter effects have been observed in both cells that express endogenous A₁ receptors (Gerwins and Fredholm, 1992; Dickenson and Hill, 1993; Rugolo et al., 1993) and those that have been transfected with the A₁ receptor cDNA (Freund et al., 1994; Megson et al., 1995). In addition to these direct effects, adenosine A₁ receptor activation can augment inositol phosphate and calcium responses stimulated by G_Q-coupled receptors (Gerwins and Fredholm, 1992; Dickenson and Hill, 1993; Megson et al., 1995; Okajima et al., 1995; Peakman and Hill, 1995).

It seems likely that G_i/o -subunits are involved in both the direct coupling of A₁ receptors to phospholipase C and the augmentation of G_Q/11-coupled receptor responses (Gerwins and Fredholm, 1992; Dickenson and Hill, 1998). Expression of G-scavenging proteins, such as the carboxy terminus of  adrenoceptor kinase 1 (residues 495-689; Koch et al., 1994), can partially attenuate the direct stimulation of phospholipase C by A₁-agonists without affecting the inhibition of forskolin-stimulated cAMP accumulation (Dickenson and Hill, 1998). Overexpression of G-protein -subunits also leads to a larger stimulation by a G_Q/11-coupled receptor agonist of phospholipase C activity in COS cells (Tomura et al., 1997). A characteristic feature of the activation of phospholipase C-₂ by purified G-protein subunits is the observation that higher concentrations of -subunits are required than _Q/11-subunits (Camps et al., 1992; Gudermann et al., 1997). These data suggest that the potency and efficacy of A₁ receptor agonists for inhibition of adenylyl cyclase and for stimulation/augmentation of intracellular calcium signaling and protein kinase C activation (via diacyglycerol derived from agonist-stimulated inositol phospholipid hydrolysis; Nishizuka, 1992; Singer et al., 1997) may differ markedly depending on the level of receptor expression in a given cell or tissue (Kenakin, 1995a,b; MacEwan et al., 1996).

The present study was undertaken to investigate how the potency and relative intrinsic activity of three different A₁ receptor agonists for inhibition of adenylyl cyclase activity and stimulation of inositol phospholipid hydrolysis change when the expression level of the human A₁ receptor is increased. It has been proposed that, at high levels of receptor expression, the fidelity of receptor-effector coupling may be lost, enabling receptors to couple to alternative G-protein families (Gudermann et al., 1997). For example, the adenosine A₃ receptor, the 5-HT_2C receptor, and the thrombin receptor have been reported to couple to both G_i/o and G_Q/11 families of G-protein (Gudermann et al., 1997; Berg et al., 1998). Furthermore, recent evidence has suggested that the relative efficacies of agonists may differ depending on the effector pathway that is activated, raising the possibility of "agonist trafficking" (Kenakin, 1995a,b; Berg et al., 1998). In this study, we provide evidence to support the contention that agonist-selective active states of A₁ adenosine receptor exist, which leads to a different stimulus pattern at the level of G-proteins.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of Recombinant Human Adenosine A₁ Receptors in Chinese Hamster Ovary Cells. The pSVL plasmid containing the human adenosine A₁ receptor cDNA was obtained from the American Type Culture Collection. The adenosine A₁ receptor cDNA was extracted on BstZ1/ApaI and subcloned into the NotI/ApaI site of the eukaryotic expression vector pcDNA3 to create pcDNA3A₁R. CHO-K1 cells (European Collection of Animal Cell Cultures, Porton Down, Salisbury, UK) were transfected with pcDNA3A₁R using transfectam [according to the manufacturer's instructions (Promega Corp., Madison WI)]. Stably transfected CHO-K1 cells were selected using 500 µg/ml geneticin (G418; Life Technologies Inc., Gaithersburg, MD) for 2 weeks. CHO-K1 cells resistant to G418 were subsequently cloned by the dilution cloning method. Transfected CHO cells were cultured in 75-cm² flasks (Costar, Acton, MA) in Dulbecco's modified Eagle's medium/nutrient F-12 (1:1) supplemented with 2 mM L-glutamine, 10% (v/v) fetal calf serum, and 500 µg/ml G418. Cells were maintained at 37°C in a humidified 5% CO₂ atmosphere until confluency and were subcultured (1:5 split ratio) using trypsin (0.05% w/v)/EDTA (0.02% w/v) solution. Cells for [³H]inositol phosphate and [³H]cAMP determinations were grown in 24-well cluster dishes (Costar).

Measurement of [³H]cAMP Accumulation. Confluent cell monolayers were incubated for 2 h at 37°C with 500 µl of Hanks'/HEPES buffer (pH 7.4) containing [³H]adenine (37 kBq/well). The cells were washed once and then incubated in 1 ml/well Hanks'/HEPES buffer containing the cAMP phosphodiesterease inhibitor, rolipram (10 µM) for 15 min at 37°C. Agonists were added (in 10 µl of medium) 5 min before the incubation with 3 µM forskolin (10 min). Incubations were terminated by the addition of 50 µl of concentrated HCl. [³H]cAMP was isolated by sequential Dowex-alumina chromatography as previously described (Megson et al., 1995). After elution, the levels of [³H]cAMP were determined by liquid scintillation counting.

Measurement of [³H]Inositol Phosphate Accumulation. Confluent cell monolayers were loaded for 24 h with [³H]myo-inositol (37 kBq/well) in 24-well cluster dishes in inositol-free Dulbecco's modified Eagle's medium containing 1% fetal calf serum. Prelabeled cells were then washed once with 1 ml/well Hanks'/HEPES buffer, pH 7.4, and incubated at 37°C for 30 min in the presence of 20 mM LiCl (290 µl/well). Where appropriate antagonists were added at the beginning of this incubation period. Agonists were then added in 10 µl of medium, and the incubation was continued for 40 min (unless otherwise stated) at 37°C. Incubations were terminated by aspiration of the incubation medium and the addition of 900 µl of cold (20°C) methanol/0.12 M HCl (1:1, v/v). Cells were left a minimum of 2 h at 20°C before isolation of total [³H]inositol phosphates in the supernatant of the disrupted cell monolayers by anion exchange chromatography. Aliquots (800 µl) of the supernatant were neutralized by the addition of 135 µl of 0.5 M NaOH, 1 ml of 25 mM Tris-HCl (pH 7.0), and 3.1 ml of distilled water and added to columns of Dowex 1 anion exchange resin (X8, 100-200 mesh, chloride form). [³H]Inositol and [³H]glycerophosphoinositol were removed with 20 ml of distilled water and 10 ml of 25 mM ammonium formate, respectively. Total [³H]inositol phosphates were then eluted with 3 ml of 1 M HCl, and the columns were regenerated with 10 ml of 1 M HCl followed by 20 ml distilled water. Radioactivity was quantified by scintillation counting in the gel phase (scintillator plus, Packard).

[³H]DPCPX Binding. CHO cells from two confluent 162-cm² flasks (which provide sufficient membrane protein for 48 tubes) were detached using Dulbecco's phosphate-buffered saline solution (Sigma Chemical Co., St. Louis, MO) containing 5 mM EDTA at 37°C for 5 min. After centrifugation (150g for 5 min), membranes were prepared by resuspending the cells in 10 ml of ice-cold Tris-EDTA buffer (50:1 mM; pH 7.4), followed by homogenization using a glass homogenizer (approximately 20 strokes) and centrifugation at 20, 000g for 15 min. The resulting pellet was resuspended in 600 µl of Tris-EDTA buffer and kept on ice until required.

[³⁵S]GTPS Binding. Cells from four confluent 162-cm² flasks (which provide sufficient membranes for 98 tubes) were initially washed using Dulbecco's phosphate-buffered saline solution and then detached in Tris-HCl buffer (50 mM, pH 7.4) using a cell scraper. After centrifugation (1000g for 5 min), cells were combined and resuspended in 20 ml of Tris-HCl buffer and homogenized using a glass homogenizer (approximately 20 strokes). Homogenates were centrifuged twice at 20,000g for 10 min, and the resulting membrane pellet was resuspended in 3 ml of Tris-HCl buffer and stored at 20°C.

6

7

Data Analysis. EC₅₀ and IC₅₀ (concentrations of drug producing 50% of the maximal stimulation or inhibition) values were obtained by computer-assisted curve fitting by use of the computer program InPlot (GraphPad Software Inc., San Diego, CA). Statistical significance was determined by Student's unpaired t test (P < .05 was considered statistically significant). All data are presented as means ± S.E.. The n in the text refers to the number of separate experiments. GraphPAD was also used to perform nonlinear regression analysis for fitting data from saturation experiments.

Chemicals. [2-³H]myo-inositol, [2,8-³H]adenine, [³⁵S]GTPS, and DPCPX were from NEN DuPont (Hertsfordshire, UK). Rolipram was purchased from Calbiochem (Nottingham, UK). Adenosine deaminase, ATP, forskolin, theophylline, Triton X-100, GTPS, 5'-(N-ethylcarboxamido)adenosine (NECA), Igepal CA630, SDS, (R)-N⁶-(2-phenylisopropyl)adenosine (R-PIA), and N⁶-cyclopentyladenosine (CPA) were purchased from Sigma. 8-Cyclopentyl-1,3-dipropylxanthine was from Research Biochemicals Inc. (Natick, MA) PTX was obtained from Calbiochem (Darmstadt, Germany). Dulbecco's modified Eagle's medium/nutrient mix F-12 (1:1) and fetal calf serum were from Sigma. All other chemicals were of analytical grade.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

A₁ Receptor Expression. Two cell lines (CHOA1H and CHOA1L) were used in the present study that had levels of human adenosine A₁ receptor expression that differed by more than 1 order of magnitude (16.5-fold). The expression level in CHOA1L was 203.1 ± 16.5 fmol/mg protein (log K_D 8.67 ± 0.08; n = 4) and that in CHOA1H was 3350.4 ± 315.8 fmol/mg protein (log K_D 8.14 ± 0.04; n = 4). Studies of the displacement of [³H]DPCPX (1 nM) binding by the A₁ receptor antagonist xanthine amine congener (XAC) in the two cell lines yielded similar apparent log K_I values for XAC (7.89 ± 0.18, n = 3 and 7.58 ± 0.13, n = 4, in the high and low expressing cells, respectively).

cAMP Accumulation. CPA, NECA, and R-PIA were able to attenuate forskolin-stimulated (3 µM) cAMP accumulation in both CHOA1H and CHOA1L cell lines (Fig. 1a and Table 1). The maximum level of inhibition of forskolin-stimulated cAMP accumulation produced by all three agonists was increased in the high expressing cells (CHOA1H; Table 1 and Fig. 1a). This increase in maximal response in CHOA1H cells was accompanied by a marked decrease in agonist IC₅₀ values (by 2 orders of magnitude; Table 1 and Fig. 1a). As we have observed previously (Megson et al., 1995), the inhibition of adenylyl cyclase activity by CPA in cells expressing moderate levels of human A₁ receptor (200-300 fmol/mg protein) can be completely prevented by 24-h treatment with PTX (100 ng/ml; Fig. 2b; CHOA1L cells). Interestingly, treatment of CHOA1L cells with PTX revealed a small stimulation of cAMP accumulation when NECA was used as agonist (in the presence of 3 µM forskolin; Fig. 2a). The log EC₅₀ value obtained for NECA for this response (in the presence of PTX), in three of the four experiments in which the effect was large enough for accurate determination, was 7.5 ± 0.2. In the same experiments, the log IC₅₀ for NECA inhibition of forskolin-stimulated cAMP accumulation was similar (6.8 ± 0.3; n = 4; Fig. 3a) to that obtained in other experiments (Table 1).

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Effect of NECA on forskolin-stimulated (3 µM) [3H]cAMP accumulation (a) and [3H]inositol phosphate accumulation (b) in high () (CHOA1H) and low () (CHOA1L) expressing cells transfected with the human adenosine A1 receptor. Values represent means ± S.E.M. obtained from three (a), four (CHOA1L, b), or five (CHOA1H, b) separate experiments. Triplicate determinations were made at each concentration in each individual experiment.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Effect of NECA (a) and CPA (b) on forskolin-stimulated (3 µM) [3H]cAMP accumulation in low expressing CHOA1L cells under control conditions and after treatment with PTX (24 h, 100 ng/ml). Values represent combined means ± S.E.M. obtained from four (a) or three (b) separate experiments. Data are presented as percentage conversions of [3H]adenine to [3H]cAMP. Triplicate determinations were made at each concentration in each individual experiment. , Control; , +PTX; , basal; , basal + PTX; , 3µM Forskolin+PTX; , 3µM Forskolin.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Concentration-response curves for the effect of NECA on forskolin-stimulated (3 µM) [3H]cAMP accumulation in control CHOA1H cells and CHOA1H cells that had been treated with PTX (24 h, 100 ng/ml). Values represent means ± S.E.M. of triplicate determinations in a single experiment. Similar data were obtained in eight further experiments (see Table 2). Note that the x-axis has been extended to higher concentrations compared to that presented in Fig. 1.

Stimulation of [³⁵S]GTPS Binding. The direct interaction between G-protein-coupled receptors and PTX-sensitive G_i/o-proteins can be followed by measurement of the binding [³⁵S]GTP in cell membranes (Weiland and Jakobs, 1994). This is made possible because of the greater intrinsic guanine nucleotide exchange and GTPase activity of the G_i/o family of proteins, together with the higher levels of expression of G_i/o-proteins compared with other G-proteins (Fong et al., 1998). The agonist potencies of CPA, NECA, and R-PIA for eliciting [³⁵S]GTPS binding were very similar in both cell lines (Table 3 and Fig. 4). Furthermore, there was little evidence for any difference in relative intrinsic activity (maximal stimulation over basal levels) between them, particularly in the CHOA1H cells (Table 3). What is clear, however, is that both the basal and agonist-stimulated specific binding of [³⁵S]GTPS is 10- to 20-fold greater in those cells expressing 16.5-fold higher numbers of A₁ receptors (Table 3). This is consistent with an ability of A₁ receptors to recruit more G_i/o-proteins in the CHOA1H cells and suggests that the availability of G_i/o-proteins is not rate limiting. The higher basal level of [³⁵S]GTPS binding in the CHOA1H cells suggests that the A₁ receptor may be constitutively active and producing agonist-independent receptor activation. Consistent with this hypothesis, the inverse agonists DPCPX, theophylline, and XAC (Shryock et al., 1998) were able to reduce basal [³⁵S]GTPS-specific binding (Fig. 5).

View this table: [in this window] [in a new window]   TABLE 3 A1 agonist concentration-response parameters for stimulation of [35S]GTPS binding in membranes derived from high and low expressing CHO-K1 cells transfected with the human adenosine A1 receptor Values represent means ± S.E.M. obtained from three to four separate experiments (actual number given in parentheses).

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Effect of CPA on [35S]GTPS binding in membranes derived from high (a) and low (b) expressing CHO-K1 cells transfected with the human adenosine A1 receptor. Values represent means ± S.E.M. from three separate experiments. The open bar shows the basal levels of [35S]GTPS binding. Please note the difference in y-axis scale between the CHOA1L (b) and CHOA1H (a) cells.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Inhibition of basal [35S]GTPS binding by DPCPX (a) and XAC (b) in high (CHOA1H, filled circles) and low (CHOA1L, open circles) expressing CHO cells transfected with the human adenosine A1 receptor. Values represent means ± S.E.M. from eight (a, filled circles) or four (a, open circles; b, filled circles) separate experiments. Histograms show the basal levels of [35S]GTPS binding in CHOA1H (filled bars) and CHOA1L (cross-hatched bars).

[³H]Inositol Phosphate Accumulation. CPA, NECA, and R-PIA were able to stimulate [³H]inositol phosphate accumulation in both cell lines (Table 4 and Fig. 6). In CHOA1L cells, the level of stimulation was small (approximately 1.5-fold over basal; Table 4); whereas in CHOA1H cells, it was similar or greater than the response to ATP (Fig. 6). Interestingly, NECA was 8-fold less potent in producing this response in the CHOA1H than in the lower expressing cells (Table 4 and Fig. 1b). R-PIA had the same potency in both cell lines. We have previously shown that the direct effect of CPA on [³H]inositol phosphate accumulation, in cells expressing moderate levels of human A₁ receptors (approximately 300 fmol/mg protein), is completely sensitive to inhibition by PTX treatment (Megson et al., 1995). In the present study, after PTX treatment of CHOA1H cells, there were residual PTX-resistant responses to all three agonists (Table 5). However, NECA was by far the most efficacious agonist, producing a much greater maximal PTX-resistant response (290%; CPA = 100%) than either CPA (100%) or R-PIA (130%; Table 5). In contrast, NECA had the lowest potency (in terms of log EC₅₀ value) of the three agonists (Table 5).

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Augmentation by CPA of ATP-stimulated [3H]inositol phosphate accumulation in CHOA1H (a) and CHOA1L (b) cells. Values represent means ± S.E.M. of triplicate determinations in a single experiment. Similar data were obtained in eight (a) or two (b) further experiments. Data were obtained for CPA in the absence (filled circles) or presence (open circles) of 0.1 mM ATP. Histograms show the basal (C) and ATP-stimulated (0.1 mM) control responses. The filled bars represent the controls measured in the same 24-well plate as the filled circles, and the cross-hatched bars represent those measured in the same plate as the open circles.

Augmentation of ATP- or UTP-Stimulated [³H]Inositol Phosphate Accumulation. CPA was able to augment the inositol phosphate responses to P_2Y2 receptor stimulation in both cell lines, although to a much lower extent in CHOA1L cells (Fig. 6). This effect was concentration dependent in both cell lines (Table 5 and Fig. 6). In CHOA1H cells, the maximal augmentation was 8.57 ± 0.15-fold (additive response = 2.98 ± 0.20-fold, n = 3; response to ATP = 1). An augmentation of the ATP response of similar magnitude in CHOA1H cells could also be demonstrated with both NECA (Fig. 7 and Table 5) and R-PIA (Table 5). Interestingly, the potency of all three agonists was increased (lower EC₅₀ values) in the presence of 100 µM ATP (CPA 4.0-fold, NECA 8.3-fold, R-PIA 2-fold; Table 5). The influence of ATP on agonist EC₅₀ values was greatest in CHOA1H cells with NECA as agonist (Fig. 8 and Table 5), but this effect was much lower in CHOA1L cells (Fig. 8).

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Augmentation by NECA of ATP-stimulated [3H]inositol phosphate accumulation in CHOA1H (a) and CHOA1L (b) cells. Values represent means ± S.E.M. of triplicate determinations in a single experiment. Similar data were obtained in four (a) or three (b) further experiments. Data were obtained for NECA in the absence (open circles, open squares) or presence (filled circles, filled squares) of 0.1 mM ATP. Some cells (filled and open squares) were treated for 24 h with 100 ng/ml PTX before assay. Control responses, in the absence of NECA, are shown at C. , NECA; , NECA + ATP; , PTX NECA; , PTX NECA + ATP.

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Concentration-response curves for NECA-stimulated [3H]inositol phosphate accumulation in CHOA1H (a) and CHOA1L (b) cells. Values represent means ± S.E.M. of five (a) or four (b) separate experiments. Data have been normalized to the response to 0.1 mM NECA in each experimental condition. Data represent curves obtained with NECA alone or in combination with 0.1 mM ATP. Some cells (PTX) were treated for 24 h with 100 ng/ml PTX before assay.

Immunoprecipitation of Individual G-Protein Subunits after Agonist-Stimulated [³⁵S]GTPS Binding in CHOA1H Cells. To investigate directly whether NECA and CPA have different relative intrinsic efficacies for the activation of individual G-protein subunits in CHOA1H cells, we have investigated whether these two agonist can stimulate [³⁵S]GTPS binding to G_s, G_i(1-3), and G_Q/11 (Fig. 9). CPA and NECA stimulated [³⁵S]GTPS binding to G_i(1-3) with similar relative intrinsic efficacies (Fig. 9). These data are consistent with the data obtained for inhibition of adenylyl cyclase (Table 1) and [³⁵S]GTPS binding in intact membranes (Table 3). The log EC₅₀ values obtained for activation of G_i(1-3) proteins by NECA (7.66 ± 0.32, n = 3) and CPA (8.09 ± 0.15, n = 3) were also similar to each other and to those values obtained from studies of intact membranes (Table 3).

View larger version (17K): [in this window] [in a new window]   Fig. 9.   Effect of NECA and CPA on [35S]GTPS binding to Gi(1-3) (a), GQ/11 (b), and GS (c). Binding of [35S]GTPS was measured after immunoprecipitation of individual G-subunits as described under Materials and Methods. Data have been expressed as percentages of the response to 10 µM CPA (b and c) or 1 µM (a) measured in each individual experiment, after subtraction of basal. Values represent means ± S.E.M. of three separate experiments.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The present study was undertaken to investigate how the potency and relative intrinsic activity of three different A₁ receptor agonists for stimulation of different intracellular pathways change when the expression level of the human A₁ receptor is increased.

Adenosine A₁ agonists produced a larger maximal inhibition of forskolin-stimulated cAMP accumulation in cells (CHOA1H) that had a much higher expression of human adenosine A₁ receptors than CHOA1L cells. The IC₅₀ values, deduced from concentration-response curves, for CPA, NECA, and R-PIA were all shifted 2 orders of magnitude to the left of those values obtained in the lower expressing cells. These data are consistent with the expected increase in signal amplification resulting from an increased A₁ receptor density, as predicted by traditional receptor theory (Clarke and Bond, 1998; McDonnell et al., 1998). This suggests that amplification occurs between the binding of A₁ agonists and the inhibition of adenylyl cyclase (as a consequence of the saturation of some of the intracellular signaling processes involved), leading to the generation of a receptor reserve.

In both cell lines, the agonists NECA, R-PIA, and CPA inhibited adenylyl cyclase activity to a similar extent (i.e., they appear to have the same relative intrinsic efficacy in each cell line). This is particularly pertinent in the lower expressing cell line, in which the maximum response to A₁ agonists represents an inhibition of only approximately 70% of the response to forskolin. These findings are consistent with the data obtained from measurement of [³⁵S]GTPS binding, in membranes from both cell lines, which primarily give an indication of the activation of G_i/o proteins (Fong et al., 1998; the responses obtained were completely sensitive to inhibition by PTX). It was notable that the EC₅₀ values for these [³⁵S]GTPS responses were similar in both high and low A₁ receptor-expressing cells. This observation, coupled with the approximately 30-fold increase in magnitude of the agonist-stimulated [³⁵S]GTPS response in the CHOA1H cells, confirms that any gain of function (in terms of agonist potency for adenylyl cyclase inhibition) resulting from signal saturation is downstream of the receptor-G_i/o-protein interaction.

It was noticeable in the low expressing CHOA1L cells that a small enhancement of forskolin-stimulated cAMP accumulation was observed in response to NECA, but not CPA, after PTX treatment (Fig. 2). This observation raises the possibility that NECA can stimulate alternative PTX-resistant G-protein pathways with higher relative intrinsic efficacy than CPA. The most likely target is G_S, although a role for G_Q/11 cannot be eliminated because activation of G_Q/11-coupled receptors has been shown to augment forskolin-stimulated cAMP formation in CHO cells (Burford et al., 1995). Consistent with a role for G_S, it was notable that NECA was more effective (in terms of maximal response) in stimulating binding of [³⁵S]GTPS to G_S in the high expressing CHOA1H cells than CPA (Fig. 9).

In the higher expressing CHOA1H cells, all three agonists were able to enhance forskolin-stimulated cAMP accumulation after ablation of G_i/o-signaling with PTX. In these cells, NECA was nearly 1 order of magnitude less potent (EC₅₀) than CPA or R-PIA in producing this response but was significantly (p < .05) more efficacious than the other two agonists (in terms of E_MAX; Table 2). Very similar data for both relative potency and intrinsic activity were obtained when the effects of NECA and CPA on G_S were directly measured (Fig. 9). These latter observations, however, are difficult to reconcile with traditional receptor theory. It is clear that the higher strength of signal produced at the A₁ receptor in the high expressing CHOA1H cells can channel into activation of a number of different signaling pathways (some of which interact with each other). Furthermore, traditional theory would predict that agonists would retain the same relative potency and efficacy orders on each pathway (even if synergistic interactions occur between them). However, one would not expect to observe a decrease in potency (i.e., an increase in EC₅₀) for NECA (relative to CPA), coupled with an increase in relative intrinsic activity (again compared to CPA), observed on one pathway (G_S-mediated activation of adenylyl cyclase) when compared to another (G_i-mediated inhibition of adenylyl cyclase). These data strongly suggest that different A₁ agonists can initiate specific stimulus profiles at the level of the G-protein. Interestingly, the potential for an agonist to direct signaling to a particular G-protein-mediated response has been raised previously (Kenakin, 1995a,b), and this concept has received support recently from studies of 5-HT_2C-mediated arachidonic acid release and inositol phosphate accumulation (Berg et al., 1998; Clarke and Bond, 1998).

As we have previously described (Megson et al., 1995), A₁ agonists can produce a small PTX-sensitive stimulation of [³H]inositol phosphate accumulation in transfected CHO cells, which appears to be mediated by G_i/o--subunits (Dickenson and Hill, 1998). These observations were confirmed in the present study in the low expressing CHOA1L cells (e.g., see Fig. 7b). In the high expressing CHOA1H cells, this inositol phosphate response was considerably larger. In the case of CPA and R-PIA, the EC₅₀ values obtained in the two cell lines were quite similar, and these observations, coupled with the large increase in response magnitude, point to the lack of a receptor reserve for these inositol phosphate responses. However, in the case of NECA, the EC₅₀ value in CHOA1H cells is nearly 1 order of magnitude greater than in the lower expressing cell line. Thus (as illustrated in Fig. 1), whereas the inhibition of adenylyl cyclase activity by NECA is shifted by 2 orders of magnitude to the left on increased A₁ receptor expression, the inositol phosphate response is shifted by 1 order of magnitude in the opposite direction.

As noted with the stimulation of cAMP accumulation in CHOA1H cells above, the relative intrinsic activity of NECA for the inositol phosphate response was significantly greater than for CPA in these cells (P < .02), suggesting higher efficacy. In contrast to the low expressing cells, however, there was a small but significant residual inositol phosphate response in CHOA1H cells (with all three agonists) after PTX treatment. In the case of NECA, this agonist had the lowest potency (i.e., highest EC₅₀ value) but produced a maximal response that was nearly 3-fold greater than that obtained with CPA. Again, these observations are difficult to reconcile with traditional receptor theory. It is tempting to speculate that the rightward shift in the concentration-response curve for NECA-induced inositol phosphate accumulation in CHOA1H cells (compared to CHOA1L) is related to the greater ability of this agonist to produce a PTX-resistant (presumably G_Q/11-mediated) inositol phosphate response. In keeping with this hypothesis, we have been able to show that NECA can stimulate directly GTPS binding to G_Q/11 in these cells with a greater relative intrinsic activity (again suggesting higher efficacy), but lower potency (higher EC₅₀), than CPA (Fig. 9).

If activation of G_Q/11 by NECA is an important determinant of the effect of this agonist on inositol phospholipid hydrolysis, then one could propose that the G_i/o--mediated phospholipase C response (the PTX-sensitive component) is partly dependent on amplification of the G_Q/11-mediated (PTX-resistant) component of the NECA response. Thus, the concentration at which NECA activates phospholipase C via G_Q/11 is largely responsible for the EC₅₀ of the final response, i.e., the G_i/o-derived -subunits only produce a substantial amplification of phospholipase C activity when concentrations of NECA reach those required to provide a direct stimulation of the enzyme via G_Q/11, although this might also involve an exchange of G-subunits between G_i/o-proteins and G_Q/11-proteins, as has been recently suggested by Quitterer and Lohse (1999). In the lower expressing CHOA1L cells, the small inositol phosphate response to NECA, obtained at lower concentrations of the agonist, is likely to be primarily due to G_i/o--subunits enhancing any basal G_Q/11-mediated phospholipase C activity.

It has been shown previously that adenosine A₁ receptor activation can augment inositol phosphate and calcium responses stimulated by G_Q/11-coupled receptors (Gerwins and Fredholm, 1992; Dickenson and Hill, 1993; Megson et al., 1995; Okajima et al., 1995; Peakman and Hill, 1995). In the present study, this effect is most marked in the high expressing cells, in which a large amplification of ATP-stimulated inositol phospholipid hydroylsis can be demonstrated. If the argument described above, for the rightward shift in the NECA concentration-response curve, is correct, then it should be possible to produce a leftward shift in the concentration-response curve for NECA by coactivation with a G_Q/11-coupled receptor (such as the P_2y2 receptor for ATP). This was achieved with NECA (and to a lesser extent with CPA and R-PIA) in the high expressing CHOA1H cells but not as expected in the lower expressing CHOA1L cells (Fig. 8). Interestingly, under these conditions the relative intrinsic activity of all three A₁ agonists in CHOA1H were effectively identical and similar to the relative values obtained from G_i-mediated responses such as GTPS binding (Table 3) or inhibition of adenylyl cyclase (Table 1). Thus, these data support the contention that the effect of NECA on G_Q/11-stimulated phospholipase C plays a major role in setting the potency (EC₅₀) and overall relative intrinsic efficacy of the final response in the absence of ATP.

It is noticeable that the relative intrinsic activity of NECA is much lower for direct activation of G_Q/11 than that obtained for PTX-resistant [³H]inositol phosphate accumulation. The EC₅₀ values for agonist-stimulated GTPS binding to G_Q/11 in CHOA1H cells are also lower than those obtained from measurement of the PTX-resistant inositol phosphate response. The simplest explanation for this is that the experimental conditions for the isolated membrane-based GTPS assays and the intact cell-based inositol phospholipid hydrolysis assays are very different. However, it should be noted that the GTPS binding to G_Q/11 has been undertaken in an intact system (i.e., where the equilibria between A₁ receptors and G_i/o-proteins in cell membranes has not been disrupted with PTX). Thus, the relative intrinsic activities should better match those obtained under control conditions (Table 5), which in fact they do. The difference in EC₅₀ values may also reflect the concentration-response relationships for activation phospholipase C by G_Q/11.

In summary, it is clear that increased A₁ receptor expression leads to the predicted amplification of G_i/o-protein-mediated inhibition of adenylyl cyclase, by three different A₁ agonists with similar efficacy and potency, and the creation of a substantial receptor reserve. In contrast, higher expression of A₁ receptors reveals a differential ability of A₁ agonists to stimulate responses mediated by PTX-resistant G-proteins. NECA has the highest relative intrinsic activity (suggesting higher efficacy) for these latter responses but surprisingly has the lowest potency. These data are difficult to reconcile with traditional receptor theory and suggest that agonists differ in the extent to which they can recruit other G-proteins. Cross-talk between different G-protein-mediated-signaling cascades, under these conditions, can have a substantial effect on both the agonist's potency and the relative efficacies of different A₁ agonists. These observations have important implications for the design of agonists that may produce differential responses via the same receptor.