Journal o~Neurochemislry Raven Press, Ltd., New York 0 1992 International Society for Neurochcmistry

The Adenosine Analogue N6-L-Phenylisopropyladenosine Inhibits Catecholamine Secretion from Bovine Adrenal Medulla Cells by Inhibiting Calcium Influx *Yi-Juang Chern, Marga Bott, *Po-Ju Chu, *Yi-Jen Lin, *Lung-Sen Kao, and Edward W. Westhead Program in Molecular and Cellular Biology, University of Massachusetts, Amherst, Massachusetts, U.S.A.; and *Institute of Biomedical Sciences, Academia Sinica, Tuiwan, Republic of China

Abstract: We reported earlier that adenine nucleotides and adenosine inhibit acetylcholine-induced catecholamine secretion from bovine adrenal medulla chromaffin cells. In this article, we used an adenosine analogue, N6-L-phenylisopropyladenosine (PIA), to study the mechanism underlying inhibition of catecholamine secretion by adenosine. PIA inhibits secretion induced by a nicotinic agonist, 1,l-dimethyl-4-phenylpiperazinium, or by elevated external K+. The half-maximal effect on 1,l -dimethyl-4-phenylpiperazinium-induced secretion occurred at -5 X M. The inhibition is immediate and reversible. Fura-2 measurements of cytosolic free Ca2+indicate that PIA inhibits CaZ+ elevation caused by stimulation; measurements of 45Ca2+

influx show that PIA inhibits uptake of Ca2+.PIA does not inhibit calcium-evoked secretion from digitonin-permeabilized cells, nor does PIA cause any significant change in the dependence of catecholamine secretion on calcium concentration. These data suggest that inhibition by PIA occurs at the level of the voltage-sensitive calcium channel. Key Words: Adenosine analogue-Catecholamine secretion-Calcium uptake-Calcium channel-Chromaffin cell. Chern Y.-J. et al. The adenosine analogue N6-L-phenylisopropyladenosine inhibits catecholamine secretion from bovine adrenal medulla cells by inhibiting calcium influx. J. Neurochem. 59, 1399-1404 (1992).

One of the most important functions of extracellular adenosine is to modulate transmitter release and central neuronal activity (Fredholm and Dunwiddie, 1988; Barry, 1990; Nishimura et al., 1990). In chromaffin cells, adenine nucleotides in the secretory vesicles are at a concentration of 150 mM(e.g., Winkler and Westhead, 1980). On stimulation of perfused adrenal glands, ATP released simultaneously with hormones is rapidly degraded into a mixture of adenine nucleotides and adenosine (Douglas et al., 1965). Because adenine nucleotides and adenosine have been recognized as important modulators of various cellular functions (Lee and Fain, 1989; Eterovic et al., 1990; Kizaki et al., 1990), they may play a feedback role in modulating secretion from the adrenal medulla. It has been previously reported that these cells

have receptors for adenine nucleotides that can mediate inhibition of secretion (Chern et al., 1987; Sasakawa et al., 1989). In the present article, we demonstrate that an adenosine analogue, N6-L-phenylisopropyladenosine (PIA), inhibits catecholamine secretion by regulating calcium entry through the plasma membrane.

Isolated bovine adrenal chromaffin cells were prepared by collagenase digestion and further purified by centrifugation on a Percoll gradient as described by Kilpatrick et al. (1980). The cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM; GIBCO, Grand Island, NY,U.S.A.) s u p

Received October 23, 199 I ; revised manuscript received March 16, 1992; accepted April 3, 1992. Address correspondence and reprint requests to Dr. E. W. Westhead at Program in Molecular and Cellular Biology,Lederle Graduate Research Towers, 435 Mom11 Science Center, University of Massachusetts, Amherst, MA 01003, U.S.A.

This work is taken, in part, from the Ph.D. thesis of Y.-J. Chern (1988). available from University Microfilms, Ann Arbor, MI, U.S.A. Abbreviations used: [Ca*+],,cytosolic frce calcium ion; DMEM, Dulbecco’s modified Eagle’s medium; DMPP, 1,1 -dimethyl-4phenylpiperazinium:PIA, N6-~-2-phenylisopropyladenosine.

-

1399

MATERIALS AND METHODS Cell preparation

1400

Y.-J. CHERN ET AL.

plemented with 10% fetal calf serum (Hyclone, UT, U.S.A.), penicillin (10 U/ml), gentamicin (2 mdml), and nystatin (2 pg/ml). Immediately after isolation, cells were plated onto Petri dishes or onto dishes covered with a layer of polystyrene beads and maintained in an incubation chamber gassed with 5% COJ95% air at 37°C. Experiments were performed 3-9 days after cell preparation.

Measurement of catecholamine release Catecholamine secretion was determined by electrochemi d l y measuring the release of either endogenous catecholamine in a constant-pressure flowing stream as previously described (Herrera et al., 1985) or by measuring release of [3H]norepinephrine from prelabeled cells in culture dishes. To measure on-line release of endogenous catecholamine, Percoll-purified cells were added to 35-mm-diameter culture dishes containing a layer of polystyrene beads (37-74 pm) at a density of 8 x lo6 cells per dish. After the cells attached, beads were placed in an HPLC fitting as the cell chamber (1.8 X 2.5 mm) and this chamber connected to a source of buffer and to an electrochemical detector (BAS, East Lafayette, IN, U.S..4.). Six-second pulses of stimulus were injected into the flowing solution by a manual injection valve. Except as indicated, the cell chamber was constantly pcrfused with Locke’s solution (154 mMNaCI, 5.6 mM KCI, 2.2 mM CaCI,, 10 mM glucose, and 5 mM HEPES buffer adjusted to pH 7.3) at a flow rate of 0.5-1 ml/min. Because switching is very rapid, it does not lead to measurable change in flow rate or pressure. There was no change in electrochemical signal due to switching alone. In this article, low levels of stimulant have been used to ensure that each stimulation leads to release of 90% by 1 mM PIA. To determine the rate at which the inhibition by PIA takes effect, cells were stimulated with 6-s pulses of 5 pLMDMPP at 5-min intervals. Pulses of stimulant werc alternately in the absence of PIA (controls) and in the presence of PIA. Varied lengths of preincubation with PIA were produced by switching to buffer containing 100pM PIA at varied times before stimulation. From 30 s to 4 min of preincubation with PIA, the extent of inhibition increased from 42 to 56% (n = 4, in both cases). When PIA was introduced simultaneously with DMPP, inhibition was 40 f 2% (n = 4). Most of the inhibitory effect of PIA is therefore immediate, but there appears to be some increase in the effect with time. The inhibitory effect of PIA is dependent on the strength of stimulation, being more pronounced at lower levels of stimulation. For example, in experiments like those of Fig. 1, 100 pA4 PIA inhibited secretion induced by 5 pLM DMPP by 60% (n = 3), but inhibited secretion induced by 50 pA4 DMPP only 33% (n = 3) (see Chern et al., 1988, and Fig. 1 for calculation method). This “competitive” aspect of PIA inhibition was seen with Kf and with DMPP as stimulants, ruling out the nicotinic receptor as the site of inhibition and suggesting the possibility that PIA effects inhibition by reducing Ca2+entry. As shown with other cell types, adenosine and its analogues may modify cellular functions by reducing the calcium conductance (Dolphin et al., 1986; MacDonald et al., 1986). We measured the [Ca2+],in the absence or presence of PIA on stimulation with DMPP or elevated external K+ using a fluorescent calcium indicator, fura-2 (Fig. 3). At 100 p M , PIA inhibited the increase in [Ca2+Iiinduced by 5 pM DMPP 5 1 k 26% (n = 11;p < 0.00 1, paired Student’s t test), and that by 31 m M Kt 42 f 18% (n = 6; p J. Nmirochem.. Vol. 59, No. 4, 1992

Y.-J. CHERN ET AL.

1402

I

IK+

1

K+

L5&mmmm

I K+

PIA

4 IC

2 10

J DMPP

I

I

k

4

DMPP

DMPP

Y FIG. 3. PIA inhibition of the rise in intracellular free Ca2' level caused by 5 pM DMPP or 31 mM K+. Fura-2-loaded chromaffin cells, cultured on a quartz plate, were stimulated with 5 pM DMPP or elevated external K+, as indicated, in the presence or absence of 100 pM PIA. The interval between successivestimulations was 5 min. Concentrations of cytosdii Ca2+ are approximate (see Materials and Methods).

< 0.001) (Fig. 3). This inhibition of the rise in intracellular free calcium is reversible as is the inhibition of secretion. Adenosine or PIA by themselves do not cause any significant change in the intracellular free calcium level, which shows that they neither activate phospholipase C nor cause Ca2+entry. To test directly the possibility that PIA modulates the cytosolic free calcium level by inhibiting calcium entry, we measured the effect of PIA on 45Ca2tuptake caused by stimulation. As shown in Fig. 4, the unstimulated value of Ca2+uptake is 180 pmol/ lo6 cells. Subtracting this value from the total uptake in the presence of DMPP or K+ gives net stimulated uptake. For DMPP without and with PIA, these values are 210 and 120 pmol/ 1O6 cells, respectively, for an inhibition by PIA of 43 -t 10%(n = 6; p < 0.01, two-tailed Student's t test). For K' stimulations, the corresponding values are 25 f 9% inhibition (n = 5;p < 0.02). This is representative of four experiments using different batches of cells. Apparently, the PIA effect on the Ca2+transient (Fig. 3) is modulated at the levcl of Ca" entry. Because adenosine receptors often modulate the activity of adenylate cyclase, we have measured cyclic AMP after treatment with adenosine or with PIA. Control cells showed a cyclic AMP level of 7.8 -t 0.7 pmol/ lo6cells; cells treated with 100pMadenosine or PIA for 15 min showed levels of 7.4 k 0.5 pmol/106 cells and 8.3 k 0.5 pmol/106 cells, respectively (n = 6, in each case). Neither value was significantly different from the control value, although we have shown that adenosine can potentiate the elevation of cyclic AMP synergistically with low levels of forskolin (Chern et al., 1988). Thc above cyclic AMP data are representative of three different experiments. J. Neurochem., Vol. 59, No. 4, 1992

We have shown earlier that inhibition of secretion and inhibition of Ca2+entry by ATP appears to be mediated by a pertussis toxin-sensitive G protein (Dived-Pierluissi et al., 1989, 199I). Preincubation of chromaffin cells with 200 ng/ml of pertussis toxin for 3 h at 37"C, sufficient to prevent inhibition of secretion and Ca2+entry by ATP (Divers&-Pierluissiet al., 199I), had no effect on inhibition by PIA. Incubation of cells with 100 Mcholera toxin for 3 h, 37"C, conditions that we find cause maximum elevation of cyclic AMP in these cells, likewise did not alter the inhibitory response to PIA. Many G proteins are inactivated by treatment with the sulfhydryl reagent, N-ethylmaleimide (Fredholm and Lindgren, 1986). Treatment of the cells with this reagent at 30 pM for 40 min also left the inhibitory response to PIA intact. Although the PIA effect may be mediated by a G protein insensitive to these rcagents, PIA inhibition clearly does not use the same G protein that mediates inhibition by ATP. To test whether PIA inhibits catecholamine secretion from chromaffin cells at a step distal to secondmessenger generation, we examined the effect of PIA on calcium-dependcnt secretion from digitonin-permeabilized cells. Figure 5 shows that 100 p M PIA did not affect the maximal secretion, nor did PIA change the calcium dependence of the secretory response. Similar experiments were performed with adenosine with the same results (data not shown). These results support the hypothesis that PIA inhibits secretion by regulating only the calcium influx on stimulation, not by modulating the affinity of the secretory process for calcium. DISCUSSION We earlier reported that adenosine can inhibit secretion (Chern et al., 1987) and that, in synergy with 600

I

NONE

DMPP

56mM K*

FIG. 4. PIA inhibition of calcium uptake induced by DMPP OT K'. Calcium uptake was started by adding 45Ca2+ (1.6 mCi/mol) with 12 f l DMPP or 56 mM K' in the presenceor absence of PIA, and stopped after 3 min by washing the cells with a buffer containing N-methyl-o-glutamineand LaCI,, as described in Materials and Methods. Error bars show SEM.

PIA INHIBITS CATECHOLAMINE VIA CALCIUM CHANNELS

I 0

0.1

1

10

100

FIG. 5. Effect of PIA on calciuminduced catechohmine secretion from cells permeabilizedwith digitonin. Chromaffincells were incubated with (closed circles) or without (open circles) 100 f l PIA for 10 min; secretion was then induced by adding 20 f l digitonin and various concentrations of calcium as described in Materials and Methods. Error bars show SEM from four separate experiments.

low levels of forskolin, adenosine enhances secretion. PIA has proved useful in the study of inhibition because it manifests the enhancing effect less and hence inhibits more completely and more consistently than adenosine. Although the concentrations needed to produce inhibition are in the high micromolar region, it is worth noting that the secretory vesicles release adenosine nucleotides at 150 mM, and that in the intact gland, hydrolysis of the nucleotides to adenosine is very rapid (Douglas et al., 1965). In this article, we show that inhibition is mediated by diminishing the rise in intracellular free calcium that occurs on stimulation. Because PIA inhibits calcium uptake induced by the nicotinic agonist, DMPP, and by elevated external K+, but has no effect on secretion from digitonin-permeabilized cells, the target site for PIA inhibition is apparently the voltage-dependent calcium channel. Others have reached the same conclusion in previous studies on muscle cells and bovine coronary artery (Fenton et al., 1982; Mustafa and Askar, 1986). The concentration of PIA that is required for half-maximal inhibition is higher than that needed for typical A, receptors, but given the potentially high adenosine concentration to which these cells may be exposed on secretion, the effectsseen here are likely to be of physiological significance. We have previously shown that ATP and ADP inhibit secretion from chromaffin cells (Chern et al., 1987) and, more recently, that both nucleotides inhibit the L-type calcium current (Divers&-Pierluissiet al., 1991). We have consistently found that inhibition by ATP of both secretion and Ca2+ current is prevented by preincubation of cells with pertussis toxin (Divers&-Pierluissiet al., 1989, 1991). In contrast, the inhibitory effect of PIA or adenosine is not affected by pretreating cells with the toxin. PIA may inhibit voltage-dependent calcium channels directly, or indi-

-

1403

rectly by activating a potassium channel to induce an increase in K+ outward current, leading to a secondary decrease in Ca2+inward current (Fredholm and Dunwiddie, 1988). The evidence that ATP and PIA inhibit Ca2+entry through different pathways allows us to consider the possibility that ATP works directly on the Ca2+ channel whereas PIA modulates the channel indirectly by altering membrane potential. Modulation of calcium entry may not be the sole explanation for the action of adenosine in all cases. Silinsky (1986) has proposed that adenosine may inhibit secretion of several neurotransmitters by reducing the intracellular affinity of the secretory apparatus for Ca2+.Adenosine and its analogues have also been reported to potentiate the concanavalin A-induced histamine release by increasing the sensitivity of the release process to Ca2+ in mast cells (Lohse et al., 1988). In this study, we have shown that PIA does not inhibit calcium-dependent secretion from digitoninpermeabilized chromaffin cells, nor change the dependence of secretion on calcium concentration. Thus, PIA inhibition appears to operate by reducing the calcium fluxes on stimulation, not by reducing the affinity of some component of the secretory process for calcium. There are at least two classes of adenosine receptors, A, and A,, known to be involved in the action of adenosine; both affect cyclic AMP production. A, receptors mediate inhibition of adenylate cyclase, whereas A2 receptors mediate stimulation of the cyclase. However, the action of adenosine is not always cyclic AMP dependent (Bruckner et al., 1985; Dunwiddie and Proctor, 1987), and that appears to be the case with the inhibition we have measured. In other systems, the action of adenosine, dependent or independent of cyclic AMP, has been reported to be mediatcd by a GTP-binding protein (Scott and Dolphin, 1987; Bohm ct al., 1989; Green et al., 1990). The lack of effect of pertussis toxin or of N-ethylmaleimide in chromaffin cells does not exclude the possibility that a toxin-insensitive GTP-binding protein is involved in thc PIA inhibition. The present study shows that the adenosine analogue, PIA, inhibits secretion by reducing stimulation-induced calcium fluxes through the plasma membrane. The inhibition is not mediated by cyclic AMP, nor by a GTP-binding protein sensitive to pertussis toxin or cholera toxin. Electrophysiological techniques will be used in the future to determine the signaling pathway through which the inhibition of calcium influx by adenosine and its analogue is achieved. Acknowledgment We express appreciationto Ms. ShangZhe Xu and Mr. Ming-Hsien Tsai for their excellent technique in preparing and maintaining the cell cultures. This work was supported in part by a grant from the NIH (U.S.A.) (NS 26606), by BRSG Grant RR07048, and by National Science Council (R.O.C.)Grant NSC 80-0412-B00 1-05. J. Neurochem.. Vol. 59, No. 4, 1992

Y.-J. CHERN ET AL.

I404

REFERENCES Artalejo C . R., Garcia A. G., and Aunis D. (1987) Chromaffin cell calcium channel kinetics measured isotopically through fast calcium, strontium, and barium fluxes. J. Biol. Chem. 262, 9 15-926. Barry S. R. (1990) Adenosine depresses spontancous transmitter release from frog motor nerve terminals by acting as an A 1-like receptor. Life Sci. 46, 1389-1 397. Bohm M., Gierschik P., Ungcrer M., and Erdmann E. (1989) Coupling of adenosine receptors to a pertussis toxin-sensitive G protein in the human heart. Eur. J. Pharmacol. 172,407-41 1 . Bruckner R., Fenner A., Meyer W., Nobis T. M., Schmitz W., and Scholz H. (1 985) Cardiac effects of adenosinc and adenosine analoguesin guinea-pigatrial and ventricular preparations: evidence against a role of cyclic AMP and cyclic GMP. J. Pharmacoi. Exp. Ther. 234,766-114. Chern Y.-J., Herrera M., Kao L. S., and Westhead E. W. (1987) Inhibition of catecholamine secretion from bovine chromaffin cells by adenine nucleotides and adenosine. J. Neurochem. 48, 1573-1 576. Chern Y.-J., Kim K.-T., Slakey L. L., and Westhead E. W. (1988) Adenosine receptors activate adenylate cyclase and enhance secretion from bovine adrenal chromaffin cells in the presence of forskolin. J. Neurochem. 50, 1484-1493. Divers&-PierluissiM., Kopell W. N., Kim K.-T., and Westhead E. W. (1989) Modulation of catecholamine secretion from adrenal chromaffin cells by adenosine and adenine nucleotides. Soc. Neurosci. Abstr. 15, 683. Diverk-Picrluissi M.. Dunlap K., and Westhead E. W. (1991) Multiple actions of extracellular ATP on calcium currents in cultured bovine chromaffin cells. Proc. Natl. Acad. Sci. USA 88, I26 I- 1265. Dolphin A. C., Forda S. R., and Scott R. H. (1986) Calcium-dependent currents in cultured rat dorsal root ganglion neurons are inhibited by an adenosine analogue. J. Physiol. (Lond.) 373, 47-6 I . Douglas W. W., Poisner A. M., and Rubin R. P. (1965) Efflux of adenine nucleotides from perfused adrenal glands exposed to nicotine and other chromaffin cell stimulants. J. Physiol. (Lond.) 183, 130-137. Dunwiddie T. V. and Proctor W. R. ( I 987) Mechanisms underlying physiologicalresponses to adenosine in the central nervous system, in Tbpics and Perspectives in Adenosine Research (Gerlach E. and Recker B. F., eds), pp. 499-508. Springer-Verlag, Berlin. Eterovic V. A., Li L., Palma A., and McNamee M. G. (1990) Regulation of nicotinic acetylcholine receptor function by adenine nucleotides. Cell. Mol. Neurobiol. 10, 423-433. Fabiato A. ( 1 988) Computer programs for calculating total from specified free or free from specified total ionic concentration in aqueous solutions containing multiple metals and ligands. Methods Enzymol. 157, 378-417. Fenton R. A., Bruttig S. P., Rubio R., and Berne R. M. (1982) Effect of adenosine on calcium uptake by intact and cultured vascular smooth muscle. Am. . I Physiol. 242, H797-H804.

J. Neurochem., Val. 59. No. 4, 1992

Fredholm B. B. and Dunwiddie T. V. (1 988) How does adenosinc inhibit transmitter release? Trends Pharmacol. Sci. 9, 130134. Fredholm B. B. and Lindgren E. ( 1986) Possible involvement of the Ni-protein in the prejunctional inhibitory effect of a stable adenosine analogue (R-PIA) on noradrenaline release in the rat tuppocampus. Acta Physiol. Scand. 126, 307-309. Green A., Johnson J. L., and Milligan G. (1990) Down-regulation of Gi subtypes by prolonged incubation of adipocytes with an Ai adenosine meptor agonist. J. Biol.Chem. 265,5206-52 10. Herrera M., Kao L. S., Curran D. J., and Westhead E. W. (1985) flow-injection analysis of catecholamine secretion from bovine adrenal medulla cells on microbeads.Anal. Biochem. 144, 218-227. Kilpatrick D. L., Ledbetter F. H., Carson K. A., Kirshner A. G., Slepetis R., and Kirshner N. (1980) Stability of bovine adrenal medulla cells in culture. J. Neurochem. 3 5 , 6 7 9 4 9 2 . Kim K. T. and Westhead E. W. (1989) Cellular responses to Ca” from extracellular and intracellular sourccs are different as shown by simultaneous measurements of cytosolic Ca2+and secretion from bovine chromaffin cells. Proc. Natl. Acad. Sci. USA 86,9881-9885. Kizaki H., Suzuki K., Tadakuma K. T., and Ishimura Y. (1990) Adenosine receptor-mediated accumulation of cyclic AMPindued T-lymphocyte death through internuclcosomal DNA cleavage. J. Biol. Chem. 265, 5280-5284. Lee H. M. and Fain J. N. (1989) Regulation of oxytocin-induced phosphoinositide breakdown in adipocytes by adenosine, isoproterenol and insulin. Biochem. Biophys. Acta 1013, 73-79. Lohse M. J., Klotz K.-N., Salzer M. J., and Schwabe U. (1988) Adenosine regulates the Ca2+sensitivity of mast cell mediator release. Proc. Nud. Acad. Sci. USA 85, 8875-8879. MacDonald R. L., Skenitt J. H., and Wen M. A. (1986) Adenosine agonists reduce voltage-dependent calcium conductance of mouse sensory neurons in cell culture. J. Physiol. (Lond.) 370, 75-90. Mustafa S. J. and Askar A. 0. (1986) Effect of calcium entry blockers and adenosine on the relaxation of largc and small coronary arteries. Life Sci. 38, 877-885. Nishimura S., Mohri M., Okada Y., and Mori M. (1990) Excitatory and inhibitory effects of adenosine on the neurotransmission in thc hippocampal slices ofguinca pig. Brain Res. 525, 165169. Sasakawa N., Nakaki T., Yamarnoto S., and Kato R. (1 989) Stimulation by ATP of inositol trisphosphate accumulation and calcium mobilization in cultured adrenal chromaffin cells. J. Neurochem. 52,44 1-447. Scott R. H. and Dolphin A. C. (1987) Inhibition of calcium currents by an adenosine analogue 2-chloroadenosine, in Topics and Perspectives in Adenosine Research (Gerlach E. and Bccker B. F., eds), pp. 549-558. Springer-Verlag, Bcrlin. Silinsky E. M. (1 986) Inhibition of transmitter release by adcnosine: are CaZ+currcnts depressed or are the intracellular effects of Ca2+impaired? Trends Pharmacol. Sci. 7, 180- 185. Winkler H. and Westhead E. W. (1980) The molecular organization of adrenal chromaffin granules. Neuroscience 5, 18031823.