Pharmacological characterization of adenosine receptors in PGT-β mouse pineal gland tumour cells

Abstract

1. The adenosine receptor in mouse pinealocytes was identified and characterized using pharmacological and physiological approaches.

2. Expression of the two adenosine receptor subtypes A_2B and A₃ was detected in mouse pineal glands and PGT-β cells by polymerase chain reaction and nucleotide sequencing.

3. Adenosine and 5′-N-ethylcarboxamidoadenosine (NECA) evoked cyclic AMP generation but the A_2A-selective agonist 2-(4-(2-carboxyethyl)phenylethylamino)adenosine-5′-N-ethylcarboxamideadenosine (CGS 21680) and the A₁-specific agonists R-N⁶-(2-phenylisopropyl)adenosine (R-PIA) and N⁶-cyclopentyladenosine (CPA) had little effect on intracellular cyclic AMP levels. The A_2B receptor selective antagonists alloxazine and enprofylline completely blocked NECA-mediated cyclic AMP accumulation.

4. Treatment of cells with the A₃-selective agonist N⁶-(3-iodobenzyl)-5′-(N-methylcarbamoyl)adenosine (IB-MECA) inhibited the elevation of the cyclic AMP level induced by NECA or isoproterenol in a concentration-dependent manner with maximal inhibition of 40 – 50%. These responses were blocked by the specific A₃ adenosine receptor antagonist MRS 1191. Pretreatment of the cells with pertussis toxin attenuated the IB-MECA-induced responses, suggesting that this effect occurred via the pertussis toxin-sensitive inhibitory G proteins.

5. IB-MECA also caused a concentration-dependent elevation in [Ca²⁺]_i and IP₃ content. Both the responses induced by IB-MECA were attenuated by treatment with U73122 or phorbol 12-myristate 13-acetate.

6. These data suggest the presence of both A_2B and A₃ adenosine receptors in mouse pineal tumour cells and that the A_2B receptor is positively coupled to adenylyl cyclase whereas the A₃ receptor is negatively coupled to adenylyl cyclase and also coupled to phospholipase C.

Introduction

Adenosine receptors are found on many neuronal cells of the central and peripheral nervous system. They modulate the general synaptic transmission of neurotransmitters, including norepinephrine, via the regulation of synaptic ion channel activity (Haas & Selbach, 2000; Cunha, 2001). The adenosine receptors have been divided into four subtypes, A₁-, A_2A-, A_2B-, and A₃-receptors on the basis of differences in their affinities for selective ligands, their second-messenger responses, and in the amino acid sequences of the adenosine receptor proteins (Fredholm et al., 2000). The A₁- and A₃-adenosine receptors are negatively coupled to adenylyl cyclase and stimulate phospholipase C (PLC) activity, whereas A_2A- and A_2B-adenosine receptors are positively linked to adenylyl cyclase (Palmer & Stiles, 1995). However, whereas the A_2A-adenosine receptor is equally sensitive to the 5′-substituted compounds 2-(4-(2-carboxyethyl)phenylethylamino)adenosine-5′-N-ethylcarboxamide-adenosine (CGS 21680) and 5′-N-ethylcarboxamidoadenosine (NECA), the A_2B-adenosine receptor exhibits a much higher sensitivity to NECA than to CGS 21680 (Klotz, 2000). In contrast, the A₁-adenosine receptor exhibits a higher sensitivity to N⁶-substituted adenosine analogues, such as R-N⁶-(2-phenylisopropyl)adenosine (R-PIA), N⁶-cyclopentyladenosine (CPA), and 2-chloro-N⁶-cyclopentyladenosine (CCPA) (Klotz et al., 1998), whereas the A₃-adenosine receptor preferentially responds to N⁶-(3-iodobenzyl)-5′-(N-methylcarbamoyl)adenosine (IB-MECA), 2-chloro-N⁶-(3-iodobenzyl)-5′-(N-methylcarbamoyl)adenosine (CI-IB-MECA), and 4-aminobenzyl-5′-N-methylcarboxamidoadenosine (AB-MECA) (Jacobson et al., 1993; Olah et al., 1994; Gallo-Rodriguez et al., 1994). Recently, MRE 3008F20 and the isoquinoline analogue VUF5574 were indicated as new potent and selective A₃ adenosine receptor antagonists (Baraldi et al., 2000; van muijlwijk-Koezen et al., 2000). NECA has also been known to bind to A₃ receptors, but the affinity of NECA for the A₃ receptor is >100 fold lower than that for IB-MECA (Jacobson et al., 1995). It has been suggested that the A₁- and A₃-adenosine receptors modulate the spontaneous firing of ion channels and neurotransmitter release via a mechanism other than the inhibition of adenylyl cyclase (Ralevic & Burnstock, 1998; Haas & Selbach, 2000).

Several previous studies have addressed that the regulation of pineal functions depends on a variety of first messengers including the nucleoside adenosine (Ebadi & Govitrapong, 1986). Studies of rat pinealocytes have shown that NECA stimulates cyclic AMP accumulation and melatonin synthesis by acting on A_2B-adenosine receptors which have been linked to the activation of adenylyl cyclase (Nikodijevic & Klein, 1989; Babey et al., 1994). In cultured chicken pinealocytes, nonmetabolizable adenosine analogues exerted an inhibitory action on pineal melatonin production elicited by forskolin probably via the A₁ adenosine receptors (Falcon et al., 1988a). In addition, pharmacological profiles and molecular studies indicated that A₁- and A₃-adenosine receptors are found most prominently in sheep pineal membranes (Falcon et al., 1997; Linden et al., 1993). These studies indicate that the distribution of the adenosine receptor subtypes and the effect of the receptors on pineal function are species-specific. At present, the small size and the inefficiency of obtaining pure pinealocytes in primary culture make it difficult to study the mouse pineal gland. Recently, therefore, a transgenic mouse was established by targeted expression of the SV40 T-antigen directed to the pineal gland, and a clonal neuroendocrine pineal cell line, PGT-β, was developed from a pineal tumour (Son et al., 1996). Utilizing the pineal tumour cell line, we now attempted to characterize the adenosine receptors of the mouse pineal gland. In the present study, we also provide first evidence that an interaction between A_2B and A₃ adenosine receptors does occur during signal transduction.

Methods

Cell culture methods

The cells were cultured in Dulbecco's modified Eagle's medium (DMEM) (GIBCO BRL, Gaithersburg, MD, U.S.A.) supplemented with 10% (v v⁻¹) heat-inactivated bovine calf serum (Hyclone, Logan, UT, U.S.A.) and 1% (v v⁻¹) antibiotics containing 5000 units ml⁻¹ penicillin G (sodium) and 5000 μg ml⁻¹ streptomycin sulphate in 0.85% saline buffer (GIBCO BRL), pH 7.4. Cell cultures were maintained in a humidified atmosphere of 5% CO₂ at 37°C. Cells grown to confluence were removed from the dishes after a 5 min incubation with 0.25% (w v⁻¹) trypsin containing 1 mM EDTA (GIBCO BRL). They were subcultured about twice weekly.

Reverse transcriptase-polymerase chain reaction (RT – PCR)

Total RNA was extracted from PGT-β cells and mouse pineal glands (10 mice, CBA/J) by the acid guanidinium thiocyanate-phenol-chloroform (Chomczynski, 1993). Total RNA concentrations were approximately quantitated by spectrophotometry. One microgram of total RNA was added to 0.5 μg oligo(dT) in DEPC-treated water and incubated at 70°C for 5 min and 4°C for 5 min. A total of 1 mM of all four deoxynucleotides (dNTPs), 5 μl 5X reverse transcriptase (RT) buffer, and 200 units of superscript II reverse transcriptase (GIBCO BRL) were added and the reactions incubated at 42°C for 1 h followed by 10 min at 75°C and stored at 4°C. For PCR amplification, an aliquot of the cDNA synthesis reaction was added to a reaction buffer containing 1 mM dNTPs, 1 mM of oligonucleotide primers, and 2 units of Taq DNA polymerase. Forty amplification cycles were conducted as follows: denaturation at 95°C for 1 min, annealing at a temperature specific for each set of primers for 1 min, and extension at 72°C for 1 min in a Minicycler (MJ Research, Watertown, MA, U.S.A.). The resultant amplification products were analysed by gel electrophoresis in a 1.0% agarose gel stained with ethidium bromide.

The following sequence of oligonucleotide primers were used for the amplification of each mouse adenosine receptor: the sense primer 5′-aaatgtactggtgatttggg-3′ (bases 385 – 404) and the antisense primer 5′-tgatgcagttcaagatgtgt-3′ (bases 1057 – 1076) for A₁; the sense primer 5′-tattgccatcgacagataca-3′ (bases 521 – 540) and the antisense primer 5′-aaggggaaaactctgaagac-3′ (bases 1442 – 1461) for A_2A; the sense primer 5′-agctccatctttagcctctt-3′ (bases 304 – 323) and the antisense primer 5′-gtcataagcccagactgaga-3′ (bases 1014 – 1033) for A_2B; the sense primer 5′-ctgtttgcctggggaagtaag-3′ (bases 13 – 33) and the antisense primer 5′-gagtttgtttcggatgatgt-3′ (bases 923 – 942) for A₃. Amplified PCR products were sequenced according to the enzymatic method of Sanger et al. (1992).

Measurement of cyclic AMP generation

Intracellular cyclic AMP generation was determined by [³H]-cyclic AMP competition assay for binding to cyclic AMP binding protein as previously described (Park et al., 1997) with some modification. To determine the cyclic AMP production induced by adenosine or its analogues, the pineal gland tumour cells were detached by trypsin treatment and aliquoted by 5×10⁵ cells per Eppendorf tube. The cells were stimulated with agonists for 20 min in the presence or absence of the phosphodiesterase inhibitor Ro 20 – 1724 (50 μM). The reaction was then quickly terminated by three repeated cycles of freezing and thawing. The samples were centrifuged at 2500×g for 5 min at 4°C. The cyclic AMP assay is based on the competition between [³H]-labelled cyclic AMP and unlabelled cyclic AMP present in the sample for binding to a crude cyclic AMP-binding protein prepared from bovine adrenal cortex following the method of Brown et al. (1971). Each sample was incubated with 50 μl [³H]-labelled cyclic AMP (5 μCi) and 100 μl binding protein for 2 h at 4°C. Separation of the protein-bound cyclic AMP from the unbound cyclic AMP was achieved by adsorption of the free cyclic AMP onto charcoal (100 μl) followed by centrifugation at 12,000×g at 4°C. The 200 μl of supernatant was then placed into an Eppendorf tube containing 1.2 ml scintillation cocktail to measure the radioactivity. The cyclic AMP concentration in the sample was determined based on a standard curve and expressed as pmol number of cells⁻¹.

Measurement of IP₃

IP₃ concentration in the cells was determined by [³H]-IP₃ competition assay in binding to IP₃ binding protein (Suh et al., 1995). The PGT-β cells were stimulated with agonists and the reaction was terminated by aspirating the medium off the cells followed by addition of 0.3 ml ice-cold 15% (w v⁻¹) trichloroacetic acid containing 10 mM EGTA. The samples were centrifuged at 5000×g for 10 min at 4°C. The trichloroacetic acid in the extract was removed by four extractions with diethyl ether. Finally the extract was neutralized with 200 mM Trizma base and its pH adjusted to about 7.4. 20 μl of the cell extract was added to 20 μl of assay buffer (0.1 M tris(hydroxymethyl)aminomethane buffer containing 4 mM EDTA and 4 mg ml⁻¹ bovine serum albumin) and 20 μl of [³H]-IP₃ (0.1 μCi ml⁻¹). Then 20 μl of solution containing the binding protein was added and the mixture incubated for 15 min on ice and centrifuged at 2000×g for 5 min. The pellet was resuspended in 100 μl of water, and 1 ml of scintillation cocktail was added to measure the radioactivity. IP₃ concentration in the sample was determined based on a standard curve and expressed as pmol μg protein⁻¹. The IP₃ binding protein was prepared from bovine adrenal cortex according to the method of Challiss et al. (1990).

Determination of intracellular Ca²⁺ level

The level of intracellular Ca²⁺ was measured using the fluorescent Ca²⁺ indicator fura-2/AM as previously described (Suh et al., 1997). Briefly, PGT-β cells were grown to confluency in 75-cm² polystyrene dishes and then loaded with fura-2/AM to a final concentration of 3 μM in complete medium at 37°C for 50 min. After the loading, the cells were detached by trypsin treatment and washed twice with Locke's solution (mM: NaCl 154, KCl 5.6, MgCl₂ 1.2, CaCl₂ 2.2, HEPES 5.0, and glucose 10, pH 7.4) to remove extracellular dye. Sulfinpyrazone which is known to inhibit the dye leakage by blocking the organic-anion transport systems (di virgilio et al., 1988) was added to all solutions to the final concentration of 250 μM. About 5×10⁵ cells of the cell suspension were then transferred to a quartz cuvette and placed in a thermostatically controlled cell holder at 37°C, and the cell suspension was continuously stirred. Changes in fluorescence ratios were measured at the dual excitation wavelengths of 340 and 380 nm and the emission wavelength of 500 nm by an alternative wavelength time scanning method. Calibration of the fluorescence signal in term of [Ca²⁺]_i was performed according to Grynkiewicz et al. (1985).

Analysis of data

All quantitative data are expressed as mean±s.e.mean. Comparison between two groups was analysed using Student's unpaired t-test. Differences were considered to be significant when the degree of confidence in the significance was 95% or better (P<0.05).

Materials

ATP, ADPβS, forskolin, enprofylline, ethidium bromide, pertussis toxin, phorbol ester, isoproterenol, and EGTA were obtained from Sigma Chemical Co. (St. Louis, MO, U.S.A.). NECA, 2-chloroadenosine, adenosine, AB-MECA, CGS 21680, R-PIA, CPA, IB-MECA, alloxazine, U73122, 3-ethyl-5-benzyl-2-methyl-4-phenylethynyl-6-phenyl-1,4-(±)-dihydropyridine-3,5-dicarboxylate (MRS 1191), and 4-[(3-butoxy-4-methoxyphenyl)methyl]-2-imidazolidinone (Ro 20-1724) were purchased from Research Biochemicals Inc. (Natick, MA, U.S.A.). [³H]-IP₃ and [³H]-adenosine from DuPont NEN Research Products (Boston, MA, U.S.A.). Fura-2 pentaacetoxymethyl ester (fura-2/AM) was obtained from Molecular Probes (Eugene, OR, U.S.A). Taq DNA polymerase and other restriction enzymes were obtained from Promega (Madison, WI, U.S.A.).

Results

Expression of A_2B and A₃ adenosine receptors in the mouse pineal gland

In order to investigate which subtype of adenosine receptor was expressed on mouse pinealocytes, we used RT – PCR analysis of total RNA prepared from mouse pineal glands (10 mice, CBA/J) and PGT-β cells. We used primers designed to specifically amplify a fragment of the mouse adenosine receptor cDNAs. As shown in Figure 1, the amplified products were of the expected sizes for A_2B (730 bp) and A₃ (930 bp). Also, no cDNA of A₁ and A_2A receptors was amplified from the above RNAs (data not shown), while the signal for α-tubulin is commonly detected. RT – PCR analysis of total RNA prepared from PGT-β cells further demonstrated the expression of mRNA for A_2B and A₃ (Figure 1). However, cDNA of A₁ and A_2A receptors was not detected in the PGT-β cells (data not shown). In contrast, A_2B receptors, but not A₁, A_2A and A₃, were selectively detected in total RNA prepared from mouse NIH-3T3 fibroblasts, which is consistent with the previous reports (Brackett & Daly, 1994). Nucleotide sequence analysis confirmed that the amplified DNA products of pinealocytes were authentic mouse A_2B and A₃ adenosine receptors. The results, therefore, suggest that two subtypes of adenosine receptor, A_2B and A₃, are specifically co-expressed on mouse pinealocytes and PGT-β cells.

Figure 1.

Expression of A_2B and A₃ mRNAs in mouse pineal glands and PGT-β cells. Total RNA was extracted for RT – PCR analysis as described under Methods. Both of the transcripts, A_2B (730 bp) and A₃ (930 bp), were detected in mouse pineal glands and PGT-β cells, whereas A_2B was selectively detected in NIH-3T3 mouse fibroblasts. α-tubulin (Tu) was used as the loading control. Molecular weight markers are in lane M. Typical results obtained in more than three separate experiments are shown.

A_2B adenosine receptor-mediated adenylyl cyclase activation

Stimulation of the PGT-β cells with adenosine evoked concentration-dependent cyclic AMP generation, the half-maximal effective concentration (EC₅₀) of adenosine being ∼87 μM (Figure 2A). Figure 2B shows the time course of cyclic AMP generation induced by 300 μM adenosine, a submaximal effective concentration of the agonist. The peak level of cyclic AMP generation was obtained about 20 min after stimulation, which was then followed by a slow decline to basal levels. In the presence of the phosphodiesterase inhibitor Ro 20-1724 (50 μM), the accumulation of cyclic AMP induced by adenosine was clearly detectable within 1 min and was saturated after about 10 min. Figure 2C shows that treatment of PGT-β cells with the general A₂-adenosine receptor agonist NECA produced a larger maximal cyclic AMP generation, whereas the specific A_2A-adenosine receptor agonist CGS 21680 had little effect. The non-hydrolyzable adenosine analogue 2-chloroadenosine (2-CADO) also elevated the cyclic AMP level in a concentration-dependent manner. The EC₅₀ for NECA was 3.2±0.5 μM, which was about 10 fold the potency of 2-CADO (29.6±3.1 μM). The selective A₃-adenosine receptor agonists IB-MECA and AB-MECA also continuously increased the cyclic AMP level at concentrations of up to 300 μM, which may be due to their ability to bind to multiple types of adenosine receptors at higher concentrations (Shearman & Weaver, 1997). However, addition of the selective A₁-adenosine receptor agonists R-PIA and CPA had little effect on the cyclic AMP level. Figure 2D shows that addition of the specific A_2B-adenosine receptor antagonists alloxazine (Brackett & Daly, 1994) and enprofylline (Auchampach et al., 1997) inhibited the adenosine-induced cyclic AMP generation in a concentration-dependent manner. The half effective inhibitory concentrations (IC₅₀) for alloxazine and enprofylline were 2.9±0.4 and 33.2±4.3 μM, respectively. Treatment of the cells with maximal effective concentrations of alloxazine (30 μM) and enprofylline (300 μM) almost completely inhibited the adenosine-mediated cyclic AMP accumulation (data not shown). These data, therefore, indicate that the effect of adenosine on cyclic AMP generation in these cells occurs primarily through A_2B-adenosine receptors.

Figure 2.

Adenosine-induced cyclic AMP production in PGT-β cells. (A) Cells pretreated with Ro 20-1724 (50 μM) for 15 min were stimulated with various concentrations of adenosine for 20 min. (B) Cells were stimulated with 300 μM adenosine for the indicated lengths of time (0, 1, 3, 5, 10, 20, 30, and 60 min) in the absence or presence of Ro 20-1724. (C) Concentration-dependence curves for cyclic AMP generation evoked by adenosine analogues. Cells were stimulated with various concentrations of the analogues for 20 min in the presence of Ro 20-1724. (D) Interference of adenosine receptor antagonists with adenosine-stimulated cyclic AMP production. Cells were stimulated with 300 μM adenosine in the presence or absence of various concentrations of antagonists for 20 min. Net increase in cyclic AMP production is expressed as percentage of the level obtained upon treatment with adenosine alone. The cyclic AMP levels were measured as described in the Methods, and each point is the mean (±s.e.mean) of three independent experiments.

To determine the mode of activation of adenylyl cyclase coupled to the A_2B-adenosine receptor in comparison to other activating signals, the effects of co-treatments with NECA, isoproterenol, and forskolin were examined. Since we had demonstrated the presence of the β₂-adrenergic receptors on PGT-β cells (Suh et al., 1999), we looked for interaction between the signallings of the β₂-receptor and the adenosine receptor. Figure 3A shows that the addition of 5 μM forskolin led to a huge synergistic augmentation of the effect of NECA and isoproterenol on cyclic AMP production, suggesting that the PGT-β cells contained adenylyl cyclase II or IV, which are enzymes that are synergistically activated when they interact with G_sα and forskolin (Feinstein et al., 1991; Tang & Gilman, 1991). In contrast, co-treatment with the maximal effective concentration of NECA (100 μM) and various concentrations of isoproterenol in the presence of 50 μM MRS 1191, which prevents the involvement of inhibitory A₃ receptor activation, resulted in additively increased cyclic AMP production (Figure 3B). The results thus indicated that the cyclic AMP accumulation stimulated by a G_sα-coupled receptor is additively stimulated by other G_s protein-mediated signalling.

Figure 3.

Enhancement of NECA-stimulated cyclic AMP production by forskolin. (A) Effect of forskolin on the NECA- and the isoproterenol-stimulated cyclic AMP production. PGT-β cells were stimulated with NECA (100 μM) or isoproterenol (1 μM) in the presence or absence of forskolin (5 μM) for 20 min after treatment with 50 μM Ro 20-1724, and the cyclic AMP generation was measured as described in the Methods. (B) Additive generation of cyclic AMP by co-treatment of the cells with NECA and isoproterenol. The cells were stimulated with various concentrations of isoproterenol with or without NECA (100 μM) in the presence of 50 μM MRS 1191 for 20 min. The experiments were performed three times and each point is the mean±s.e.mean.

Inhibition of A_2B receptor-mediated cyclic AMP generation by A₃ receptor activation via pertussis toxin-sensitive G proteins

Since the A₃-adenosine receptors are negatively coupled to adenylyl cyclase, we investigated the effect of the selective A₃ receptor agonist IB-MECA on the cyclic AMP generation mediated by G_s protein-coupled receptors. Figure 4A shows that treatment of the cells with IB-MECA decreased the NECA-stimulated cyclic AMP production with maximal inhibition of 40 – 50% obtained at 1 μM IB-MECA. However, this inhibitory effect of IB-MECA was antagonized by addition of the A₃ selective antagonist MRS 1191, indicating involvement of an adenosine A₃ receptor. IB-MECA also inhibited isoproterenol-stimulated cyclic AMP generation in a concentration-dependent manner, which could be prevented by addition of MRS 1191 (Figure 4B).

Figure 4.

Effect of IB-MECA on NECA- and isoproterenol-stimulated cyclic AMP generation. Cells preincubated with or without 50 μM MRS 1191 for 10 min were stimulated with various concentrations of IB-MECA and 10 μM NECA (A) or 300 nM isoproterenol (B) for 20 min. The cyclic AMP levels were measured as described in the Methods. The net increase in cyclic AMP production is expressed as percentage of the level obtained upon treatment with NECA or isoproterenol alone. Data are the mean±s.e.mean (bars) of triple experiments.

In order to elucidate the role of G protein in the IB-MECA-induced inhibition of cyclic AMP generation, we performed experiments with pertussis toxin which selectively uncouples G_i/G_o proteins from receptors by catalyzing the ADP-ribosylation of the α_i/α_o subunits (Simon et al., 1991). As shown in Table 1, the IB-MECA-mediated inhibitory effect was almost completely reversed by a 12-h pretreatment of the cells with pertussis toxin (300 ng ml⁻¹). These data indicate that the A₃ receptors are coupled to adenylyl cyclase through pertussis toxin-sensitive inhibitory G proteins and thus inhibits the G_s-coupled receptor-mediated adenylyl cyclase activation.

Table 1.

Effects of pertussis toxin on IB-MECA-induced inhibition of cyclic AMP production in mouse pineal gland tumour cells

A₃ adenosine receptor-mediated PLC activation

The effect of IB-MECA on PLC activation was studied in PGT-β cells. Figure 5A shows that IB-MECA in the presence of 2.2 mM CaCl₂ produced a rise in the [Ca²⁺]_i. The [Ca²⁺]_i peaked within 30 s after stimulation and was followed by a sustained increase above the basal level for more than 3 min. In the absence of extracellular Ca²⁺ (dotted trace), Ca²⁺ mobilization from the intracellular Ca²⁺ stores occurred. The results indicate that the increase in the [Ca²⁺]_i was caused not only by mobilization of Ca²⁺ from the intracellular stores but also by influx of Ca²⁺ from the extracellular medium. IB-MECA increased the [Ca²⁺]_i in a concentration-dependent manner (Figure 5B). Although adenosine and NECA also tend to induce an increase in [Ca²⁺]_i at high concentrations (⩾100 μM), the intensity of the elevation of the [Ca²⁺]_i was much below that of IB-MECA (Figure 5A). R-PIA and CPA at concentration of 100 μM had no effect on the [Ca²⁺]_i rise (data not shown).

Figure 5.

Effects of IB-MECA on the increase in [Ca²⁺]_i in PGT-β cells. (A) Typical pattern of [Ca²⁺]_i rise after treatment with IB-MECA (50 μM) in the presence and absence of 2.2 mM extracellular Ca²⁺. In Ca²⁺-free experiments, the Locke's solution did not contain Ca²⁺ but rather 200 μM EGTA. Cells were washed with Ca²⁺-free buffer twice and incubated in Ca²⁺-free solution for 3 min before stimulation with the agonist. (B) Concentration-dependent effect of adenosine analogues on [Ca²⁺]_i rise. Fura-2-loaded cells were treated with various concentrations of each nucleotide, and net increases in [Ca²⁺]_i were measured. Each concentration of IB-MECA was tested four times independently, and the data are mean values±s.e.mean (bars).

To test for phospholipase C activation upon A₃ adenosine receptor stimulation, we treated the cells with IB-MECA and examined phosphoinositide turnover. The time course of the IP₃ formation in the cells in response to IB-MECA is shown in Figure 6A. Addition of IB-MECA to the cells evoked a rapid increase in the level of IP₃ which reached a peak ∼30 s after administration of the agonist. The level of IP₃ then slowly declined over time returning to the basal level 10 min after the IB-MECA addition. Figure 6B illustrates the IB-MECA concentration-response curve for IP₃ in terms of peak IP₃ generation time (30 s). In these cells, IB-MECA induced concentration-related (10⁻⁶ to 10⁻⁴ M) increases in IP₃ formation. This is consistent with the concentration-dependent Ca²⁺ response. However, NECA treatment had little effect on the IP₃ generation. Preincubation of the cells with the selective A₃ adenosine receptor antagonist MRS 1191 abolished the formation of IP₃ induced by IB-MECA. Figure 6C shows the concentration-dependent inhibition by MRS 1191 of the IB-MECA-induced IP₃ generation, indicating that the IB-MECA-induced responses were mediated through A₃ adenosine receptors.

Figure 6.

Effect of IB-MECA on IP₃ production. (A) Time course of IP₃ generation stimulated by 50 μM IB-MECA. The cells were treated for the designated time (0, 0.25, 0.5, 1, 3, 5, or 10 min), and the reactions were stopped by addition of 15% (w v⁻¹) TCA containing 10 mM EGTA. (B) Concentration-dependent stimulation of IP₃ formation. The cells were treated with various concentrations of IB-MECA or NECA for 30 s, and the IP₃ production was measured by competition assay as described in the Methods. (C) Concentration-dependent inhibition of IB-MECA-stimulated IP₃ generation by the A₃ selective antagonist MRS 1191. The net increase in IP₃ generation is expressed as percentage of the level obtained after treatment with 50 μM IB-MECA alone. Data are mean±s.e.mean (bars) values from three experiments.

A 5 min pretreatment of the cells with the novel aminosteroid U73122, reportedly a selective inhibitor of phosphatidylinositol-specific PLC, abolished the IB-MECA-mediated [Ca²⁺]_i rise in a concentration-dependent manner with an IC₅₀ of 2.7±0.4 μM (Figure 7A). In addition, the pretreatment with U73122 (5 μM) also decreased the IP₃ formation elicited by IB-MECA by ∼60% (Figure 7B). However, U73122 did not have a significant effect on the cyclic AMP production resulting from treatment with IB-MECA or NECA (data not shown). These results indicate that the IB-MECA-mediated [Ca²⁺]_i rise and IP₃ generation resulted from A₃ receptor-coupled PLC activation.

Figure 7.

Effect of U73122 on IB-MECA-stimulated [Ca²⁺]_i rise and IP₃ generation. (A) Cells pretreated with various concentrations of U73122 for 5 min were stimulated with 50 μM IB-MECA. The [Ca²⁺]_i increase is expressed as percentage of the IB-MECA response in the absence of U73122. (B) The cells pretreated with vehicle or U73122 for 5 min were stimulated with IB-MECA or ATP for 30 s, and the reactions were stopped by addition of 15% (w v⁻¹) TCA containing 10 mM EGTA. The IP₃ generation was measured by competition assay as described in the Methods. Data are mean±s.e.mean (bars) values from three experiments.

Pertussis toxin inhibits the IB-MECA-mediated PLC activation

In order to elucidate the mechanism of the IB-MECA-induced PLC activation, we performed experiments with pertussis toxin. Table 2 shows that the IB-MECA-mediated [Ca²⁺]_i rise was partially inhibited by a 12-h pretreatment of the cells with pertussis toxin (300 ng ml⁻¹). In parallel, pertussis toxin also reduced the IB-MECA-induced IP₃ generation 35 – 45%. However, the ADPβS-induced [Ca²⁺]_i rise and IP₃ generation remained unaffected (data not shown), as previously described (Suh et al., 1997). The pertussis toxin effect on the IB-MECA-mediated responses indicates that the A₃ receptors are coupled to PLC, at least in part, through pertussis toxin-sensitive G proteins.

Table 2.

Effects of pertussis toxin on IB-MECA-stimulated [Ca²⁺]_i rise and IP₃ production in mouse pineal gland tumour cells

Effect of phorbol 12-myristate 13-acetate on IB-MECA-mediated [Ca²⁺]_i rise and IP₃ generation

The regulation of the IB-MECA-mediated signal transduction by protein kinase C (PKC) was investigated by treating the cells with the phorbol ester PMA. Figure 8A shows that pretreatment of the cells with PMA inhibited the subsequent IB-MECA-elicited [Ca²⁺]_i elevation. The inhibitory effect of PMA on IB-MECA-induced [Ca²⁺]_i rise was concentration-dependent and 300 nM PMA inhibited the IB-MECA response maximally by 85 – 90%, whereas 300 nM PMA inhibited the ATP-induced response by 25 – 30% (Figure 8B). The results suggest that the signal transduction pathways between the receptors and PLC are negatively regulated by PKC and that the A₃ receptor-mediated signalling is more sensitive to PKC regulation than that of the P2 purinergic receptor. In contrast, PMA treatment slightly increased the forskolin- and NECA-stimulated cyclic AMP accumulation (Figure 8C), indicating that the A_2B receptor-mediated signalling is not sensitive to PKC activation.

Figure 8.

Inhibition of IB-MECA-induced [Ca²⁺]_i rise by PMA. (A) Cells pretreated for 5 min with vehicle or 300 nM PMA were stimulated with 50 μM IB-MECA or 300 μM ATP. These experiments were repeated more than five times and typical Ca²⁺ transients are presented. (B) Concentration-dependent effects of PMA on IB-MECA-and ATP-induced [Ca²⁺]_i rise. Cells pretreated for 5 min with vehicle or various concentrations of PMA were stimulated with 50 μM IB-MECA or 300 μM ATP. Net increase in cyclic AMP production is expressed as percentage of the level obtained upon treatment with IB-MECA or ATP alone. (C) Cells were pretreated for 5 min with vehicle or 300 nM PMA then stimulated with 100 μM NECA or 5 μM forskolin for 20 min. The cyclic AMP generation was measured as described in the Methods. Data are mean±s.e.mean (bars) values from three separate experiments.

Discussion

In the present study, we illustrated the presence of both A_2B and A₃ adenosine receptors on mouse pineal gland tumour cells and that A₃ receptors have dual coupling to PLC and adenylyl cyclase which is inhibited through pertussis toxin-sensitive G proteins. Our conclusion is based on several lines of evidence. First, RT – PCR and nucleotide sequencing analysis showed that A_2B and A₃ receptors were co-expressed on the cells. Second, the general A₂-adenosine receptor agonist NECA significantly raised cyclic AMP production, while the specific A_2A-adenosine receptor agonist CGS 21680 had little effect on the cyclic AMP level. In addition, NECA was more effective in stimulating cyclic AMP production than any of the other adenosine analogues, the difference being more than 2-orders of magnitude large. The A_2B-selective antagonists alloxazine and enprofylline profoundly inhibited cyclic AMP production induced by adenosine or NECA. The results are consistent with the properties of the A_2B-adenosine receptors (Gharib et al., 1992), indicating that the A_2B receptors are involved in adenosine-mediated cyclic AMP accumulation in the cells. Third, the A₃ selective agonist IB-MECA caused a 40 – 50% inhibition of isoproterenol- and NECA-stimulated cyclic AMP accumulation. However, pertussis toxin treatment reversed the inhibitory effects elicited by IB-MECA, suggesting that IB-MECA-responsive A₃ receptors are coupled to adenylyl cyclase via the pertussis toxin-sensitive G_i/G_o proteins. Fourth, IB-MECA induced Ca²⁺ release from the intracellular Ca²⁺ pools followed by sustained Ca²⁺ influx, which is typical for PLC-coupled receptor-mediated responses. The agonist also induced a concentration-dependent elevation of IP₃ contents in the cells. The IB-MECA-induced IP₃ production and [Ca²⁺]_i rise was, however, selectively inhibited by the commonly used A₃ receptor antagonist MRS 1191 (Jacobson et al., 1997) and the PLC inhibitor U-73122. The result suggests that both of these responses elicited by IB-MECA are mediated via PLC-linked A₃ receptors. Overall, the results suggest the involvement of A₃ receptors in the negative regulation of the A_2B receptor-mediated adenylyl cyclase activation through G_i proteins in PGT-β cells.

Multiple adenosine receptor coexistence has been identified in many cell types (Ralevic & Burnstock, 1998). These include A_2A-, A_2B-, and A₃ receptors on mast cells, A_2A and A_2B receptors on endothelial cells and PC12 cells. The functional significance of this is not entirely clear, but our results in terms of cross-talk between A_2B and A₃ receptors suggest an agonistic interplay between two separate signalling pathways. The results of our experiments revealed that stimulation of the adenosine A₃ receptors resulted in a concentration-dependent inhibition of cyclic AMP production induced by NECA and isoproterenol with maximal inhibition of 40 – 50% at ∼1 μM IB-MECA. However, the EC₅₀ (>1 μM) and the maximal effective value (>50 μM) of IB-MECA for IP₃ generation and [Ca²⁺]_i rise were 10 fold higher than those causing inhibition of cyclic AMP production. In addition, pertussis toxin treatment blocked the inhibitory effect of IB-MECA on cyclic AMP production and the stimulatory effect of IB-MECA on PLC activation (Tables 1 and 2). These observations strongly suggest that the A₃ receptor activation resulted in dual effects: a G_iα-mediated process inhibits adenylyl cyclase and a G_βγ-mediated process activates PLC. Indeed, the affinity of βγ subunits for their target enzyme is much lower (10 – 100 fold) in comparison to α subunits (Birnbaumer, 1992; Park et al., 1993; Sternweiss, 1994). Thus, the coupling of the adenosine A₃ receptor to adenylyl cyclase through G_i protein is more efficient than its coupling to PLCβ and more βγ subunits are needed to activate PLCβ than α subunits to inhibit adenylyl cyclase. This implies that sufficient G_i/o protein must be available to release βγ subunits to obtain a substantial PLCβ response.

The inhibitory adenosine receptors have been shown to desensitize with different time courses according to a subtype-specific manner. Desensitization of the A₁ adenosine receptor typically occurs over the period of several hours, or even days, and is temporally associated with receptor down-regulation (Ramkumar et al., 1991). In contrast, the native A₃ adenosine receptor undergoes a rapid functional desensitization detectable within a few minutes of agonist exposure (Ramkumar et al., 1993; Trincavelli et al., 2000). This is associated with the uncoupling of the receptor-G protein complex, as has been indicated by the reduction in the number of high affinity binding sites (Palmer & Stiles, 1995; Palmer et al., 1997). It was also reported that the agonist-stimulated phosphorylation of the C-terminal domain of the A₃ adenosine receptor by one or more G protein-coupled receptor kinases is responsible for initiating the events that lead to a rapid desensitization of the A₃ receptor (Palmer & Stiles, 1995). The results of the present study show that A₃ receptor-mediated PLC activation was also highly sensitive to PKC activation, while the A_2B receptor-mediated adenylyl cyclase activation was not effected by short-term treatment with PMA in our experimental conditions. In addition, the A₃ receptor-mediated signalling was more sensitive to PMA than was the P2 purinergic receptor. Although the phosphorylation process of A₃ adenosine receptors by second messenger-activated kinase PKC has not yet elucidated, it seems likely that A₃ receptors might be primarily phosphorylated by PKC as evidenced by the involvement of G protein-coupled receptor kinase.

Many studies have suggested that adenosine plays an important role in the trans-synaptic regulation of the pineal function and the modulation of melatonin production through its effect on serotonin N-acetyltransferase (NAT) regulation. Recently, in vitro studies with isolated mammalian pineal glands have shown that adenosine elevates NAT gene expression and enzyme activity (Nicholls et al., 1997). PGT-β cells express functionally active forms of two characteristic marker enzymes of the pinealocyte, tryptophan hydroxylase and NAT, and the activities of these enzymes can be enhanced by pharmacological stimulation with forskolin and epinephrine through the mediation of β₂-receptors (Son et al., 1996; Suh et al., 1999). However, NAT activity was only slightly increased (1.2∼1.5 fold) upon stimulation of the cells with NECA (data not shown). This weak activation of NAT may be due to the origin of the cells, since these pineal tumour cells originated from a transgenic mouse that was a hybrid of a NAT-expressing mouse (CBA/J) and a NAT-deficient mouse (C57BL/6J) (Son et al., 1996). Goto et al. (1989) demonstrates that most laboratory mice do not have pineal melatonin because of a genetic defect in the activity of NAT. For example, C57BL/6J mice do not have NAT activity because of a mutation in an autosomal gene required for normal activity of NAT (Goto et al., 1994; Roseboom et al., 1998). Recently, experiments with a BALB/c mouse, which did not make pineal melatonin, showed that the night-to-day ratio of NAT transcripts in the retina was much lower than in other mammalian retinae (Sakamoto & Ishida, 1998). However, our finding of the existence of adenosine receptors in pineal gland provides new insights into the biology of adenosine in the pineal gland.

We (Suh et al., 1997) and another group (Ferreira et al., 1994) have shown that P₂ purinergic receptors are also present in mammalian pineal glands and involved in the modulation of adrenergic receptor-mediated signalling. The results are consistent with evidence that the high concentrations of ATP stored in synaptic vesicles and the release of catecholamine are accompanied by ATP release to rat pineal gland (Mortani Barbosa et al., 2000). Previous studies have shown that extracellular adenosine is produced from extracellular ATP, cyclic AMP, or S-adenosylmethionine (Nikodijevic & Klein, 1989; Falcon et al., 1988b), presumably by 5′-nucleotidase enzymes on the external surface of pinealocytes. The results suggest that neuronal stimulation of the pineal gland may involve the activity of nucleotides through P2 purinoceptors and adenosine receptors. Furthermore, the data presented here clearly indicate that the signal flow from the A₃-adenosine receptor negatively regulates the A_2B-adenosine and β-adrenergic receptor-mediated signal transduction leading to cyclic AMP production.

Abbreviations

AB-MECA

4-aminobenzyl-5′-N-methylcarboxamidoadenosine

[Ca²⁺]_i

intracellular free Ca²⁺ concentration

2-CADO

2-chloroadenosine

CGS 21680

2-(4-(2-carboxyethyl)phenylethylamino)adenosine-5′-N-ethylcarboxamideadenosine

CPA

N⁶-cyclopentyladenosine

IB-MECA

N⁶-(3-iodobenzyl)-5′-(N-methylcarbamoyl)adenosine

IP₃

inositol 1,4,5-trisphosphate

MRS 1191

3-ethyl-5-benzyl-2-methyl-4-phenylethynyl-6-phenyl-1,4-(±)-dihydropyridine-3,5-dicarboxylate

NAT

serotonin N-acetyltransferase

NECA

N-ethylcarboxamidoadenosine

PKC

protein kinase C

PLC

phospholipase C

R-PIA

R-N⁶-(2-phenylisopropyl)adenosine

Ro 20-1724

4-[(3-butoxy-4-methoxyphenyl)methyl]-2-imidazolidinone