INTRODUCTION

Multiple pathways mediate stimulation of cardiac cell contractility. The classical and best characterized pathway is that mediated by the -adrenergic receptor, giving rise to an increase in cAMP with subsequent calcium entry and increase in myocyte contractility (1, 2, 3, 4). The -adrenergic receptor-mediated augmentation of cardiac contractility involves stimulation by inositol phosphates and protein kinase C that is cyclic AMP-independent (5, 6, 7, 8). Adenosine exerts pronounced contractile effects in the heart. In the cardiac ventricular cell, adenosine attenuates the increase in the force of contraction elicited by -adrenergic agonist, known as the anti-adrenergic effect, while adenosine by itself has no direct effect on the basal level of contractility. The anti-adrenergic effect is mediated by the inhibitory A₁ subtype of the adenosine receptor (for review, see Refs. 9, 10, 11, 12, 13). Using fetal chick embryo ventricular cells as a model, prior studies demonstrated the presence of a high affinity adenosine A_2a receptor capable of mediating positive inotropic response when the A₁ receptor is blocked (14, 15). The marked positive inotropic response mediated by the A_2a receptor suggests a significant physiological or pathophysiological role of this receptor in the cardiac cell. Since the inhibitory A₁ receptor coexists with the stimulatory A_2a receptor and since the A₁ receptor mediates the anti-adrenergic effect of adenosine, the question arises regarding specifically the function of the A_2a receptor in modulating the basal as well as the -adrenergic stimulated contractile states. Furthermore, the mechanism underlying the A_2a receptor-mediated positive inotropic response is not known. Although activation of the A_2a receptor can be coupled to stimulation of cAMP accumulation, it is not known whether cAMP mediates the increase in myocyte contractility.

The objectives of this study, using cultured fetal ventricular cells as a model system, were to investigate the mechanism underlying the adenosine A_2a receptor-mediated positive inotropic response and to study the contractile function subserved by the A_2a receptor. Cardiac ventricular cells cultured from chick embryos retain many of the biological properties of the intact heart and represent a useful model for study of cardiac function and contractility (16, 17, 18, 19, 20, 21, 22, 23). Activation of adenosine receptors in these cultured heart cells produced physiological effects similar to those elicited by adenosine in the mammalian heart (14, 15, 23). The present data demonstrate that the A_2a receptor and the -adrenergic receptor do not share a common pathway leading to stimulation of myocyte contractility or calcium entry; that a cyclic AMP-independent, G_s-mediated mechanism is largely responsible for the A_2a receptor-mediated stimulatory response; and that a physiological action of the activated A_2a receptor is to oppose the inhibitory effect of A₁ receptor agonist on basal and -adrenergic stimulated contractile states.

EXPERIMENTAL PROCEDURES

Methods

Atrial and ventricular cells were cultured from chick embryos 14 days in ovo according to previously described procedures (15, 23). Cells grew to confluence on day 3 of culture and exhibited rhythmic spontaneous contraction. Unless otherwise indicated, all experiments were carried out on day 3 of culture. Cultures were treated with adenosine deaminase (2 units/ml) for 24 h to keep the endogenous adenosine at a minimal level; they were also treated with pertussis toxin (5 ng/ml for 24 h) to uncouple the inhibitory A₁ receptor from its effector(s) as described previously (14, 15). Blocking of the A₁ receptor facilitated quantitation of the A₂ receptor-mediated functional responses. In experiments in which cholera toxin was used to activate the stimulatory G protein (G_s), the cultures were treated with 2 µg/ml cholera toxin for 3 h. This dose and duration gave consistent stimulation of cyclic AMP and ⁴⁵Ca²⁺ influx.

Measurement of contractile amplitude in cultured cardiac cells was carried out according to previously described methods (15, 18). The contractile amplitude of the cultured cell was determined by an optico-video motion detection system with a video motion analyzer (Colorado Video Inc., Boulder, CO) as described previously. The perfusion medium contained the various adenosine analogs indicated as well as the following components: 4 mM HEPES (pH 7.4), 137 mM NaCl, 3.6 mM KCl, 0.5 mM MgCl₂, 0.6 mM CaCl₂, 5.5 mM glucose, and 6% horse serum. Measurement of contractile amplitude was carried out on only one cell/coverslip, and each culture dish contained five coverslips. After achieving a steady state of beating in medium without adenosine analogs, the medium was switched to that containing the indicated adenosine drugs. Both the basal contractile amplitude and the amplitude measured during adenosine analog exposure were determined. The stimulatory effect of the various adenosine analogs on the contractile state was predominantly on the amplitude of contraction (14, 15). There was no significant consistent effect of any of the analogs on the spontaneous rate of contraction.

Cultured ventricular cells were treated with pertussis toxin and adenosine deaminase as described above. On day 3 of culture, the media were replaced with culture medium lacking fetal bovine serum, and cells were incubated with the indicated agonist(s) and antagonist(s). Cyclic AMP was extracted and assayed according to a previously described radioimmunoassay method (Ref. 15; Amersham Corp.). The effect of agonist on cyclic AMP accumulation was linear for 10 min, at which time, cyclic AMP was extracted for assay.

Determination of ⁴⁵Ca uptake was made according to a modification of a previously described method (18). Cultures were incubated with L-[3,4,5-³H]leucine (152.2 Ci/mmol) for 24 h prior to the ⁴⁵Ca uptake measurement. [³H]Leucine incorporated into the cellular protein allowed normalization of ⁴⁵Ca content to milligrams of cellular protein. For measurement of ⁴⁵Ca uptake, growth media in which ventricular cells were grown and maintained on 12-well culture plates were replaced with HEPES-buffered solution (pH 7.35) containing 1.0 mM CaCl₂, 4 mM KCl, and 0.5 mM MgCl₂ (buffer A) at 37 °C for 10 min and then incubated in the same medium except that calcium was omitted and 1 mM EGTA was added (buffer B) and subsequently incubated in buffer A containing ⁴⁵Ca (5-10 µCi/ml) and adenosine analogs for the times indicated. The step of exposure of myocytes to calcium-free medium helped minimize mixing of ⁴⁵Ca with unlabeled Ca²⁺ near or at the cell surface, optimized ⁴⁵Ca at the cell surface for subsequent uptake, and resulted in reproducible quantitation of ⁴⁵Ca uptake. Cells were then washed free of the ⁴⁵Ca media by four rinses with ice-cold buffer A containing 1 mM lanthanum. This washing procedure removed >99% of the extracellular marker ⁵¹Cr-labeled EDTA and thus ensured complete removal of the extracellular ⁴⁵Ca. The presence of lanthanum in the wash buffer allowed displacement of any ⁴⁵Ca that may have attached to the extracellular surface of the cells. Influx of ⁴⁵Ca was quantitated for 90 s to allow determination of its uptake into the rapidly exchanging pool, which has been shown to be due either to calcium entry through the L-type calcium channel or to Na⁺/Ca²⁺ exchange (18). Determination of ⁴⁵Ca influx during this initial phase of its uptake in the presence of various agents permitted assessment of their effects on the activity of the calcium channel or Na⁺/Ca²⁺ exchange (24, 25). For all data comparing the effect of different agents on ⁴⁵Ca influx, one-way ANOVA¹ followed by group comparison with the t test was carried out at each time point of ⁴⁵Ca influx.

The effect of CGS21680 (2-[4-(2-carboxyethyl)phenylethylamino]-5-N-ethylcarboxamidoadenosine) on the inositol 1,4,5-trisphosphate (IP₃) level in the cultured ventricular cells was determined. Growth media in which ventricular cells were grown were replaced with HEPES-buffered solution (pH 7.35) containing 1.0 mM CaCl₂, 4 mM KCl, and 0.5 mM MgCl₂. Cells were then exposed to CGS21680, carbamylcholine, or ATP at the concentrations indicated. The reaction was terminated by the addition of 0.2 volume of ice-cold trichloroacetic acid. The trichloroacetic acid was removed by extraction with a solution of 1,1,2-trichloro-1,2,2-trifluoroethane/trioctylamine. The IP₃ in the aqueous phase was determined by competing with [³H]IP₃ for binding to the IP₃ receptor supplied (DuPont). The concentration of IP₃ in the sample was inversely proportional to the amount of [³H]IP₃ in the membrane pellet. The increase in IP₃ in response to carbamylcholine or ATP was detectable within 15 s, and the level remained elevated at 30 s. Data obtained at 30 s of exposure are presented as means ± S.E. of triplicates and are typical of four experiments.

Materials

The adenosine analogs (R)-PIA and CGS21680 were from Research Biochemicals Inc. (Natick, MA). Pertussis and cholera toxins were from List Biological Laboratories Inc. (Campbell, CA). Adenosine deaminase was obtained from Boehringer Mannheim. 8-(3-Chlorostyryl)caffeine (CSC) was a generous gift from Dr. Kenneth Jacobson (NIDDK, National Institutes of Health). Embryonic chick eggs were from Spafas Inc. (Storrs, CT). The cyclic AMP radioimmunoassay kit was obtained from Amersham Corp. [³H]Leucine, ⁴⁵Ca, and the inositol 1,4,5-trisphosphate radioreceptor assay kit were obtained from DuPont. (R_p)-cAMP-S was from Biolog (La Jolla, CA).

RESULTS

Effects of Adenosine A_2a and -Adrenergic Receptor Activation on Cardiac Cell Contractility and Calcium Influx

The A₁ receptor coexists with the A_2a receptor and mediates inhibition of the stimulatory responses caused by A_2a agonist in cultured fetal chick embryo ventricular cells (14, 15). In cells where the A₁ receptor was first inactivated by prior treatment with pertussis toxin, the A_2a receptor-selective agonist CGS21680 caused a marked increase in myocyte contractile amplitude that was additive to the increase elicited by maximally effective concentration of the -adrenergic agonist isoproterenol (Fig. 1A), indicating that the two receptor pathways do not share a common mechanism of positive inotropic response. The positive inotropic response to isoproterenol is mediated by an increase in the trans-sarcolemmal influx of extracellular calcium. CGS21680 was also able to stimulate a significant increase in ⁴⁵Ca influx (Fig. 1B), which was again additive to the calcium influx effect elicited by isoproterenol (Fig. 1C), providing further evidence that the two receptors do not share a common pathway. In comparing the ability of CGS21680 and isoproterenol to stimulate myocyte contractile amplitude, the CGS21680 (10 µM)-induced increase in contractile amplitude (20 ± 2%, mean ± S.E., n = 14) was less than that caused by isoproterenol (31 ± 2.5%, n = 15; p < 0.01). The extent of the CGS21680 (1 µM)-stimulated increase in ⁴⁵Ca influx, measured at the 60-s time point (45 ± 6%, n = 5), was also less than the isoproterenol-induced stimulation (79 ± 8%, n = 6; p < 0.01).

and -adrenergic agonists on cardiac contractility and calcium influx.

Activation of the -adrenergic receptor in the cardiac cell causes a large increase in cAMP accumulation, which, in turn, serves as a primary mediator of its positive inotropic effect (1, 2, 3, 4). In contrast to isoproterenol, CGS21680 (1 µM) was able to induce only a modest increase in cAMP, but elicited a marked increase in ⁴⁵Ca influx. The percent increases above basal levels in the presence of CGS21680 and isoproterenol were 40 ± 5% (n = 5) and 317 ± 30% (n = 16), respectively (the basal cAMP level was 13.2 ± 1.6 pmol/mg (n = 12)). Isoproterenol (3 nM) stimulated an increase in cAMP similar to that elicited by 1 µM CGS21680, but was not able to cause any increase in either ⁴⁵Ca influx or myocyte contractility (Fig. 2). (R_p)-cAMP-S, which can antagonize the action of intracellular cAMP (27, 28, 29, 30, 31), blocked all of the isoproterenol-stimulated increase in calcium influx and myocyte contractility (Fig. 2B), but had only a modest inhibitory effect on the CGS21680-induced calcium response (Fig. 2C) ((R_p)-cAMP-S caused a 24.8 ± 2.4% inhibition of the CGS21680-stimulated calcium response at the 90-s time point (n = 5).) (R_p)-cAMP-S alone had no effect on the basal contractility or the basal uptake of ⁴⁵Ca (data not shown). Similar percent inhibition by (R_p)-cAMP-S was also evident at the 15-, 30-, and 60-s time points of ⁴⁵Ca influx (Fig. 2C). Increasing (R_p)-cAMP-S to 100 µM did not result in further inhibition of the CGS21680-induced increase in ⁴⁵Ca influx (inhibition = 26.4 ± 1% at the 90-s time point (n = 5)). The ability of CGS21680 to stimulate myocyte contractility appeared to be minimally affected by (R_p)-cAMP-S (Fig. 2C). However, a small inhibition of contractile amplitude by (R_p)-cAMP-S may be difficult to quantitate and cannot be ruled out completely. Although the data are consistent with a role of cyclic AMP in mediating the calcium influx response to the A_2a agonist, they demonstrate the existence of a cAMP-independent mechanism in causing a stimulation of calcium influx and provide further evidence that the A_2a and -adrenergic receptors use different pathways to cause stimulatory responses.

and -adrenergic agonist effects on cardiac contractility and calcium influx.

Activation of the A_2a receptor could cause an increase in the trans-sarcolemmal calcium influx by stimulating the voltage-sensitive L-type calcium channel or by augmenting Na⁺/Ca²⁺ exchange. Verapamil (1 µM) blocked the spontaneous contraction of cultured ventricular cells and completely attenuated the CGS21680-elicited increase in ⁴⁵Ca influx in the absence (Fig. 3A) or presence (data not shown) of (R_p)-cAMP-S, indicating that the effect on calcium influx is due to stimulation of the L-type calcium channel. Similar results were obtained using nifedipine (1 µM) as the calcium channel antagonist (data not shown). To examine whether Na⁺/Ca²⁺ exchange is involved in this calcium response, the effect of CGS21680 was determined in media in which choline chloride substituted for NaCl in the presence of verapamil. Decreasing the extracellular Na⁺ induced a marked stimulation of ⁴⁵Ca influx in the presence of verapamil (Fig. 3B), indicating a Na⁺/Ca²⁺ exchange-mediated increase in calcium influx. The addition of CGS21680 did not cause any further stimulation of ⁴⁵Ca influx. In the absence of verapamil, CGS21680 increased the extent of ⁴⁵Ca influx further, above that caused by lowering the extracellular sodium concentration (Fig. 3C).

Stimulation of cAMP accumulation by adenosine A_2a agonist is likely mediated by the G_s-induced activation of adenylyl cyclase (9, 10, 11, 12, 13). Although the present data do not indicate a major role for the A_2a receptor-G_s-cAMP pathway in the calcium influx response, activation of G_s mediated by the A_2a receptor may be involved in causing a cAMP-independent stimulation of calcium influx. Direct activation of G_s by cholera toxin, which stimulated cAMP accumulation (67 ± 4%, mean ± S.E., n = 5), induced a large increase in ⁴⁵Ca influx that was only partially blocked by 100 µM (R_p)-cAMP-S (Fig. 3D). The percent inhibition of cholera toxin-stimulated ⁴⁵Ca influx by (R_p)-cAMP-S was 40 ± 8, 54 ± 15, 43 ± 8, and 48 ± 6% at the 15-, 30-, 60-, and 90-s time points, respectively (p < 0.05, paired t test). The (R_p)-cAMP-S-insensitive portion of the increase in ⁴⁵Ca influx was abolished by 1 µM verapamil (Fig. 3D), which was also able to block all of the ⁴⁵Ca influx caused by isoproterenol or by cholera toxin (data not shown). Thus, direct activation of G_s by cholera toxin can stimulate, through the L-type calcium channel, both cAMP-dependent and -independent increases in ⁴⁵Ca influx.

CGS21680 (at 1 or 10 µM) had no effect on the level of inositol 1,4,5-trisphosphate (41 ± 5 pmol/mg (control) versus 45 ± 4 pmol/mg (1 µM CGS21680) and 38 ± 6.5 pmol/g (10 µM CGS21680)) (data were means ± S.E. of quadruplicates and were representative of three other experiments). As a positive control, both the muscarinic cholinergic agonist carbamylcholine (300 µM) and the purinergic agonist ATP (300 µM) caused a large increase in the level of inositol trisphosphate (carbamylcholine, 70 ± 5 pmol/mg; and ATP, 92.5 ± 6 pmol/mg) (data were typical of three other experiments). Thus, at the concentrations of CGS21680 that caused a marked increase in calcium influx or myocyte contractility, there was no stimulation of inositol 1,4,5-trisphosphate accumulation. These data indicate that the CGS21680-mediated stimulatory contractile or ⁴⁵Ca influx response is not mediated by inositol 1,4,5-trisphosphate.

The present data predict that concomitant activation of both the A_2a and A₁ receptors by adenosine agonist should alter the basal as well as the -adrenergic stimulated levels of contractility. (R)-PIA is an A₁ receptor-selective agonist that can also activate the A_2a receptor at the higher concentrations (14, 26). Prior studies indicated that (R)-PIA was an A₁ receptor agonist capable of inhibiting the isoproterenol-stimulated increase in myocyte contractility in chick embryo cardiac cells (14, 15, 23). The present data demonstrate that (R)-PIA (10 µM) can stimulate both ⁴⁵Ca influx and contractility in cells in which the A₁ receptor pathway is blocked (Fig. 4A). The A_2a receptor-selective antagonist CSC blocked the (R)-PIA-induced stimulation of calcium influx and contractility, indicating that the stimulatory effects of (R)-PIA are mediated by the A_2a receptor. (R)-PIA further increased the level of ⁴⁵Ca influx and myocyte contractility stimulated by isoproterenol (Fig. 4, B and C), an additive effect that was consistent with agonist activity of (R)-PIA at the A_2a receptor. Thus, (R)-PIA can activate both the A₁ and A_2a receptors in these ventricular cells. Whether activation of the A_2a receptor can modulate the basal and anti-adrenergic contractile responses elicited by A₁ receptor agonist was examined next.

Blocking of the A_2a receptor caused not only a depression of the basal contractility (Fig. 5A), but also a further inhibition of the isoproterenol-stimulated positive inotropic response by (R)-PIA (Fig. 5B). A previous study demonstrated that atrial myocytes cultured from the same chick embryos exhibited no positive inotropic response to CGS21680 or other adenosine receptor agonists (14), indicating the absence of a functional A_2a receptor in the atrial myocyte. If atrial myocytes do not express A_2a receptors, the presence of CSC should not influence the ability of (R)-PIA to inhibit the basal or isoproterenol-stimulated increase in contractility. In fact, Fig. 5C demonstrates that CSC had no effect on the ability of (R)-PIA to inhibit basal contractile amplitude (basal contractile amplitude = 77 ± 2%, mean ± S.E., n = 10 ((R)-PIA) versus 78 ± 1.6%, n = 10 ((R)-PIA plus CSC); p > 0.1, paired t test). Furthermore, CSC did not affect the (R)-PIA-mediated inhibition of the isoproterenol-stimulated increase in contractility (maximal isoproterenol stimulation by (R)-PIA alone of 76 ± 3% versus percent maximum of 75 ± 4% by (R)-PIA plus CSC; p > 0.1, paired t test). These data serve as a control for the experiments carried out in the ventricular cells and indicate that CSC has no intrinsic nonspecific contractile effect.

DISCUSSION

Adenosine plays an important role in modulating the contractile state of the cardiac cell (for review, see Refs. 9, 10, 11, 12, 13). Previous studies carried out in this laboratory have demonstrated the existence of a high affinity adenosine A_2a receptor capable of mediating a marked positive inotropic response in cultured fetal chick embryo ventricular cells (14, 15). However, the mechanism underlying the A_2a receptor-mediated positive inotropic effect is not known. Although the A_2a receptor is coupled to stimulation of cyclic AMP accumulation, it is not clear whether cyclic AMP mediates all of the positive inotropic effect of A_2a receptor agonist. It is not known whether the A_2a receptor and the -adrenergic receptor share a similar mechanism of positive inotropic effect. Furthermore, the role of the A_2a receptor in regulating the contractile state of the cardiac cell is not known. Since the inhibitory A₁ receptor coexists with the stimulatory A_2a receptor in these cultured ventricular cells and since the A₁ receptor mediates the anti-adrenergic effect of adenosine, the question arises regarding the specific role of the A_2a receptor in modulating the basal as well as the -adrenergic stimulated contractile states. The main findings of this study are as follows. The mechanism underlying the A_2a receptor-mediated positive inotropic effect involves a cyclic AMP-independent, A_2a receptor/G_s-mediated stimulation of the L-type calcium channel and differs from the mechanism of the -adrenergic stimulated inotropic response. A normal function of the activated A_2a receptor is to attenuate the A₁ receptor-mediated anti-adrenergic response.

Four lines of evidence support the existence of a new mechanism responsible for the A_2a receptor-mediated positive inotropic response. First, the positive inotropic effects elicited by maximally effective concentrations of isoproterenol and CGS21680 were additive, indicating that the -adrenergic and adenosine A_2a receptors do not share a common positive inotropic mechanism. Similarly, the stimulation of calcium influx caused by maximally effective concentrations of isoproterenol and CGS21680 was also additive, providing further evidence for the distinct pathways used by the two receptors. Second, while the -adrenergic receptor is coupled to pronounced stimulation of cyclic AMP accumulation, A_2a receptor activation caused only a modest increase in cyclic AMP, with nearly 10-fold less stimulation of the cyclic AMP level. Third, at a concentration of isoproterenol that caused an increase in the cyclic AMP level similar to the increase elicited by the maximally effective concentration of CGS21680, there was no increase in either myocyte contractile amplitude or calcium influx. Fourth, while (R_p)-cAMP-S was able to block all of the positive inotropic and calcium influx responses elicited by isoproterenol, (R_p)-cAMP-S could inhibit only a small portion of the calcium influx response stimulated by CGS21680. These data indicate that a distinct mechanism, different from that used by the -adrenergic receptor, mediates the stimulatory contractile and calcium influx responses to adenosine A_2a receptor agonist.

Since verapamil or nifedipine (1 µM) was able to block all of the CGS21680-stimulated increase in calcium influx, an A_2a receptor-mediated stimulation of the L-type calcium channel appeared to be responsible for the increase in calcium influx. Lowering the extracellular sodium in the presence of verapamil caused a marked stimulation of calcium influx, secondary to a Na⁺/Ca²⁺ exchange-mediated increase in calcium entry. The inability of CGS21680 to cause a further increase in calcium influx in this type of medium suggests that CGS21680 does not stimulate Na⁺/Ca²⁺ exchange. In the absence of verapamil, CGS21680 was able to cause a further increase in the level of calcium influx that was stimulated by lowering the extracellular sodium. These data provide further evidence for the notion that the CGS21680-induced calcium effect is mediated by the L-type calcium channel and does not involve Na⁺/Ca²⁺ exchange. Since CGS21680 could not stimulate inositol 1,4,5-trisphosphate accumulation, it is unlikely that the inositol trisphosphate is involved in mediating the CGS21680-induced stimulatory responses.

Although (R_p)-cAMP-S inhibited partially the A_2a receptor-mediated calcium influx response, the percent inhibition was modest (25%). These data indicate that a cAMP-independent pathway is capable of, and likely plays a primary role in, mediating the A_2a agonist-induced increase in calcium influx. Activation of G_s by A_2a receptor agonist likely mediates the cAMP-independent stimulation of the L-type calcium channel, based on the following evidence. First, direct activation of G_s by cholera toxin resulted in stimulation of both cAMP and calcium influx; cholera toxin-stimulated calcium influx exhibited both (R_p)-cAMP-S-sensitive and -insensitive components. Both components can be blocked by the L-type channel antagonists verapamil and nifedipine. Since activation of G_s can stimulate the L-type calcium channel independent of cAMP (32, 33), these data are consistent with the notion that cholera toxin-activated G_s can cause a cAMP-independent stimulation of the L-type calcium channel. Second, A_2a agonist-stimulated calcium influx also exhibited (R_p)-cAMP-S-sensitive and -insensitive components, which were blocked by verapamil or nifedipine. These data indicate that the A_2a receptor is also capable of mediating a cAMP-independent activation of the L-type channel. Similarities in the action and effect of CGS21680 compared with those of cholera toxin are consistent with the notion that the A_2a agonist-induced augmentation of nifedipine-sensitive calcium influx involves a cAMP-independent, A_2a receptor/G_s-mediated stimulation of the L-type calcium channel. Preliminary data demonstrate that transfection of the myocytes with rat G_s causes an increased A_2a agonist-mediated, nifedipine-sensitive ⁴⁵Ca influx in the presence of (R_p)-cAMP-S, providing evidence that exogenous G_s can couple the A_2a receptor to a cAMP-independent stimulation of the calcium channel. Although -adrenergic receptor activation, via G_s, can stimulate the L-type calcium channel directly independent of cAMP in the cardiac cell membrane (32, 33), the present data suggest that a cyclic AMP-independent, G_s-mediated pathway, which functionally coupled the cell-surface A_2a receptor to stimulation of the L-type calcium channel, exists in the intact cardiac cell.

These results indicate that cAMP mediates most if not all of the -adrenergic stimulated increases in myocyte contractility and calcium influx in the cultured chick embryo ventricular cell, similar to findings in other cardiac cells (1, 2, 3, 4). The reason for the ability of the G_s-linked adenosine A_2a receptor, but not the G_s-linked -adrenergic receptor, to couple to stimulation of the calcium channel independent of cAMP is not clear. Whether this differential ability of the two G_s-linked receptors to couple to activation of the calcium channel can be explained by a predetermined coupling of the A_2a receptor to a pool of G_s capable of mediating a cAMP-independent stimulation of the calcium channel, which are separate from those coupled to the -adrenergic receptor, remains unknown.

The existence of an A_2a receptor-mediated signaling mechanism distinct from that of the -adrenergic receptor raised the possibility that activation of the A_2a receptor can modulate not only the basal but also the anti-adrenergic contractile responses elicited by adenosine A₁ receptor agonist. Prior studies indicated that (R)-PIA is an A₁ receptor agonist capable of inhibiting basal (atrial cells) and isoproterenol-stimulated (both atrial and ventricular cells) increases in contractile amplitude (9, 10, 11, 12, 13, 14, 15, 23). The present data demonstrate that (R)-PIA at the higher concentration can also activate the A_2a receptor in the cultured ventricular cell. Blocking the A_2a receptor with the selective antagonist CSC caused not only a decrease in basal contractile amplitude, but also a further inhibition of isoproterenol-stimulated contractility in response to (R)-PIA. These data indicate that activation of the A₁ receptor, if unopposed by the A_2a subtype, could cause a direct negative inotropic effect as well as an enhanced anti-adrenergic response in the ventricular cell. The lack of effect of CSC on the A₁ receptor-mediated anti-adrenergic or direct negative inotropic effect in atrial cells, which do not express an A₂ receptor-mediated functional response (14), provides an important control and suggests that the effect of CSC in ventricular cells is not due to a nonspecific contractile response to CSC. These data are consistent with a prior study indicating that CSC acts as a selective antagonist at the adenosine A_2a receptor in the cultured chick embryo ventricular cell (15). The data also indicate that potential agonist effect on the A_2a receptor should be considered when studying the cardiac action of adenosine agonist.

Overall, this study demonstrates a new stimulatory signaling mechanism mediated by the adenosine A_2a receptor and elucidates a novel contractile function of the A_2a receptor in the intact cardiac ventricular cell. Whether other cell-surface cardiac receptor(s) are also coupled to such G_s-mediated stimulatory pathway remains to be determined. Demonstration of this new cAMP-independent, G_s-mediated stimulatory mechanism in the cardiac cell, however, should have significant general implications for the regulation of basic cardiac function.