Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The endogenous nucleoside adenosine binds to and activates four different subtypes of G protein-coupled receptors (GPCR): adenosine A₁, A_2A, A_2B, and A₃ receptors. The different subtypes can be distinguished pharmacologically and differ also in their coupling to second messenger systems (Fredholm et al., 1994, 1996). Adenosine A₁ and A₃ receptors inhibit adenylyl cyclase and stimulate phospholipase C via activation of the pertussis toxin-sensitive G proteins G_i and/or G_o. Adenosine A_2A and A_2B receptors are positively coupled to adenylyl cyclase, but also may activate alternative signaling pathways (Feoktistov et al., 1994; Feoktistov and Biaggioni, 1995; Mirabet et al., 1997).

This diversity in intracellular signaling downstream of adenosine receptors and the fact that adenosine has been reported, depending on the receptor subtype activated, to both enhance and inhibit mitogenesis (Jonzon et al., 1985; Neary et al., 1996), postischemic cell death (Fredholm, 1997), and apoptosis (Shneyvays et al., 1998; Vitolo et al., 1998) led us to examine the effects of adenosine receptor stimulation on extracellular regulated kinases 1/2 (ERK1/2). ERK1/2 belong to the family of mitogen-activated protein kinases (MAPKs), which can be activated by receptor tyrosine kinases via the classical MAPK cascade, and mediate proliferation, differentiation, cell survival, and cell death (Seger and Krebs, 1995). G protein-coupled receptors, G_i/G_o as well as G_s coupled, also have been shown to activate the MAPK cascade (Sugden and Clerk, 1997; Gutkind, 1998; Luttrell et al., 1999).

The activation of ERK1/2 via adenosine A₁ (Dickenson et al., 1998), A_2A (Sexl et al., 1997; Seidel et al., 1999), and A_2B (Feoktistov et al., 1999; Gao et al., 1999) receptors has recently been detected in a variety of systems. However, activation of A_2A receptors constitutively expressed in the rat pheochromocytoma cell line PC12 (Arslan et al., 1997) or stably transfected into Chinese hamster ovary (CHO) cells (Hirano et al., 1996) inhibited ERK activation by other stimuli. Little or nothing is known about A₃ receptors. It seemed important to compare activation of MAPKs mediated by all the human adenosine receptor subtypes against an identical cellular background. Transfection of CHO cells with the different adenosine receptors presents a useful model to compare the intracellular events after receptor stimulation in a cellular environment with an identical signaling machinery. Such receptor-expressing cells have been previously characterized biochemically as well as pharmacologically (Klotz et al., 1998). Our results show a transient time- and dose-dependent phosphorylation of ERK1/2 after 5'-N-ethylcarboxamidoadenosine (NECA) and adenosine stimulation of CHO cells expressing the human A₁, A_2A, A_2B, or A₃ receptor.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. All cell culture media, fetal calf serum, and supplies were from Life Technologies (Täby, Sweden). NECA, 9-chloro-2-(2-furyl)[1,2,4]triazolo[1,5-c]quinazolin-5-amine (CGS 15943), 2-[p-(2-carboxyethyl)phenylethylamino]-5'-N-ethylcarboxamidoadenosine (CGS 21680), and (R)-N⁶-phenylisopropyladenosine (R-PIA) were from Research Biochemicals International (Wayland, MA). Polyvinylidene difluoride-Immobilon P membrane was from Millipore Corp. (Bedford, MA). Phosphospecific rabbit anti-phospho-Thr-202/Tyr-204-ERK1/2 was from New England Biolabs, Inc. (Beverly, MA) and rabbit polyclonal anti-human cyclophilin A was from Upstate Biotechnology (Lake Placid, NY). Goat anti-rabbit horseradish peroxidase antibody was from Pierce (Rockford, IL). Enhanced chemiluminescence detection (ECL) kit was from Amersham International (Buckinghamshire, England).

Cell Culture. CHO cells transfected with the four human adenosine receptors (Klotz et al., 1998) were grown adherent at 37°C and 5% CO₂, 95% air in Dulbecco's modified Eagle's medium/F-12 (1:1), 0.2 mg/ml geneticin (G-418), 100 U/ml penicillin, 100 U/ml streptomycin, 2 mM L-glutamine, and 10% fetal calf serum. Untransfected control cells were maintained under the same conditions, but without G-418. All cells were split three times a week at a ratio of 1:20. For detection of ERK1/2 phosphorylation 1.5 × 10⁶ cells were plated in 10-cm Petri dishes and grown overnight. After serum deprivation (growth medium with 0.5% fetal calf serum without G-418) overnight, the cells were washed two times with Dulbecco's modified Eagle's medium, 20 mM HEPES, pH 7.4, at 37°C and then stimulated with drugs as described in the figure legends. The reaction was stopped at 0, 5, 10, 15, or 30 min with two quick washes with 2 ml of ice-cold PBS and lysis in 500 µl of 70 mM -glycerophosphate, 0.5% Triton X-100, 2 mM MgCl₂, 1 mM dithiothreitol, 1 mM NaF, 1 mM Na₃VO₄, 20 µg/ml aprotinin, and 5 µg/ml leupeptin on ice. Time points earlier than 5 min have not been studied because the method used for cell lysis did not allow us to stop the reaction precisely. The cell lysate was centrifuged for 5 min in a microfuge and the supernatant was denatured with Laemmli buffer. Equal amounts of each sample were analyzed by polyacrylamide gel electrophoresis (4% stacking gel; 12% resolving gel) with the Hoefer Mini VE system.

Immunoblotting. The proteins were transferred onto a polyvinylidene difluoride membrane by using Towbin's buffer for transfer. After blocking for 1 h in 3% dry milk powder dissolved in 50 mM Tris, 150 mM NaCl, 0.05% Tween 20, the membrane was incubated with the phosphospecific rabbit anti-P-ERK1/2 1:1000 in 50 mM Tris, 150 mM NaCl, 0.05% Tween 20, 0.01% BSA, 0.01% NaN₃ either for 2 h at room temperature or overnight at 4°C. As an internal standard for normalization, the membrane was simultaneously incubated with a rabbit anti-human cyclophilin A antibody, which detects a protein of about 18 kDa. P-ERK1/2 and cyclophilin A were visualized with a goat anti-rabbit horseradish peroxidase antibody (1:10,000) in 10 mM Tris, 500 mM NaCl, pH 7.6, 0.5% Tween 20 for 1 h at room temperature and the enhanced chemiluminescence method according to the manufacturer's protocol.

cAMP Assay. Intracellular cAMP levels were determined by using a competitive protein-binding assay and [³H]cAMP. Briefly, cells were plated into 24-well microplates, grown overnight, and stimulated at 37°C. The reaction was stopped by addition of perchloric acid to a final concentration of 0.4 M and incubation on ice. After lysis and neutralization with KOH in 50 mM Tris, the supernatant was examined for cAMP content as described elsewhere (Kull et al., 1999a).

Data Analysis. We used the NIH Scion Image software for quantitative analysis of the immunoblots. For the graphical presentation and the statistical analysis of data (one-way ANOVA) we used GraphPad Prism2 software (GraphPad, San Diego, CA). Sigmoidal dose-response curves were calculated by using nonlinear regression.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Stimulation of Human Adenosine A₁, A_2A, A_2B, and A₃ Receptors Expressed in CHO Cells Led to Phosphorylation of the MAPK ERK1/2. The immunoblot signal seen using the dual phosphorylation site-selective antibody increased in a time- (Fig. 1) and dose-dependent (Fig. 2) manner. In cells not transfected with any adenosine receptor, NECA was virtually without effect on CAMP production and ERK1/2 phosphorylation (Figs. 3 and 4). After stimulation with the nonselective adenosine receptor agonist NECA, ERK1/2 phosphorylation reached a maximal value at 5 min in cells transfected with each type of the receptor. After 30 min NECA stimulation the levels of ERK1/2 phosphorylation had returned to control (184 ± 47, 147 ± 39, and 145 ± 10%, respectively) in cells transfected with A₁, A_2A, or A_2B receptor, whereas in cells transfected with the A₃ receptors phospho-ERK1/2 remained almost 4-fold (394 ± 137%) above control levels at this time point.

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Time course of ERK1/2 phosphorylation in quiescent CHO cells stably transfected with the human adenosine A1, A2A, A2B, or A3 receptor after NECA stimulation (A1, A2A, and A3 stimulated with 100 nM NECA; A2B stimulated with 10 µM NECA). For further details see Experimental Procedures. The immunoblots (A) are representative of at least three independent experiments per transfected cell line as summarized in the bar graphs (B). Error bars show S.D. P values for the statistical analysis (one-way ANOVA) with *P < .05, **P < .01, ***P < .001.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   ERK1/2 phosphorylation in serum-deprived CHO cells stably transfected with the human adenosine A1, A2A, A2B, or A3 receptor is dependent on NECA concentration. Cells were stimulated for 5 min. The immunoblots (A) show representative results of at least three experiments. These data are summarized in B that shows the normalized concentration-response curves. In each experiment the maximal level of ERK1/2 phosphorylation was set to 100% and that seen in unstimulated cells to 0% (). The graph for adenosine A2B receptor-transfected CHO cells also contains a concentration-response curve for the production of cAMP after NECA stimulation (). The error bars give S.D.

There appeared to be differences in the maximal stimulation of ERK1/2 phosphorylation between cells transfected with different adenosine receptors. Thus, adenosine A₁ and A₃ receptor-transfected cells showed a much more pronounced increase in ERK1/2 phosphorylation at 5 min compared with the A_2A- or A_2B-transfected cells (Fig. 1). The maximal stimulation at 5 min in cells transfected with A₁, A_2A, A_2B, and A₃ receptor was (mean ± S.E.) 1305 ± 545, 546 ± 165, 158 ± 9, and 988 ± 99% compared with the unstimulated control (100%) (100 nM NECA for A₁, A_2A, and A₃ receptor; 10 µM NECA for A_2B receptor).

The potency of NECA on the adenosine receptors was similar and in the nanomolar range for all of the receptor subtypes (Fig. 2). The EC₅₀ values (mean and 95% CI) for NECA-stimulated ERK1/2 phosphorylation were A₁, 115 nM (56 to 236 nM); A_2A, 26 nM (12 to 56 nM); A_2B, 19 nM (6 to 62 nM); and A₃, 65 nM (36 to 119 nM).

The effect of NECA in A_2B-transfected cells was modest. It was therefore important to show that it is indeed receptor mediated. The results in Fig. 4 show that ERK1/2 phosphorylation in untransfected CHO cells was not increased and that the NECA effects in A_2B-transfected cells was abolished by a receptor antagonist (CGS 15943, 10 µM). In addition, we performed experiments with CGS 21680 and R-PIA as agonists on adenosine A_2B receptor-transfected CHO cells (data not shown). The dose-response curves for ERK1/2 phosphorylation of these compounds, compared with those reported for cAMP production (Klotz et al., 1998), were shifted to the left as well. The order of potency for the agonists tested at the human adenosine A_2B receptor, however, remained unchanged (NECA > R-PIA > CGS 21680). Furthermore, this high potency of agonists on ERK1/2 phosphorylation in these cells prompted the examination of NECA's ability to stimulate cAMP production. As seen in Fig. 2 there is a close to 100-fold difference in EC₅₀ = 1.4 µM (0.7 to 2.7 µM). A similar discrepancy does not exist in the other transfected cells (Table 1).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   NECA-dependent changes in cAMP levels detected in CHO cells transfected with the human adenosine A2B receptor and in untransfected CHO cells. The experiment was performed in triplicate. Error bars show S.D.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Receptor-mediated ERK1/2 phosphorylation induced by NECA in A2B receptor-transfected CHO cells. The left part shows that the adenosine receptor antagonist CGS 15943 (10 µM) abolishes the NECA-induced changes in ERK1/2 phosphorylation in CHO A2B cells, thus confirming the adenosine receptor dependence of the effect. The right part shows that neither NECA nor adenosine stimulates ERK1/2 phosphorylation in untransfected CHO cells.

The rather high potency of the adenosine analogs NECA, CGS 21680, and R-PIA suggested that the endogenous agonist adenosine also might exert effects in a physiological range of concentrations. We therefore tested the effect of adenosine on ERK1/2 phosphorylation in CHO cells at 100 nM, a concentration that occurs under basal physiological conditions in many body fluids, and at 3 µM, a concentration that is reached under hypoxic conditions. As seen in Fig. 5, in all the adenosine receptor-transfected cells, adenosine mimicked the response to NECA in time as well as dose dependence. Thus, at the high physiological concentration a clear-cut increase in ERK1/2 phosphorylation was seen in all adenosine receptor transfected cells after 5 min. This effect was smaller after 30 min. The lower adenosine concentration produced small, but clear effects in cells transfected with any of the four adenosine receptors. As shown for NECA, adenosine stimulation showed differing efficacy on the different receptor subtype-transfected cells, with A_2B mediating the smallest changes in ERK1/2 phosphorylation (Fig. 5).

View larger version (57K): [in this window] [in a new window]   Fig. 5.   Adenosine induces ERK1/2 phosphorylation in a time- and dose-dependent manner similar to NECA in quiescent CHO cells transfected with the four human adenosine receptor subtypes. A1, A2A, A2B, and A3 receptor expressing cells were stimulated with 100 nM and 3 µM adenosine. The immunoblots show one representative experiment of 2 to 3 per cell type. The increase in ERK1/2 phosphorylation after 5 min 100 nM adenosine compared with the unstimulated control was 2.92 ± 0.10-fold (n = 2), 1.34 ± 0.06-fold (n = 2), 2.14 ± 0.83-fold (n = 3), and 11.35 ± 0.21-fold (n = 2) (error in S.D.) for the A1, A2A, A2B, and A3 receptor, respectively.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Our results show that adenosine or the adenosine analog NECA can increase phosphorylation of the MAPK ERK1/2 in CHO cells expressing each of the human adenosine receptors A₁, A_2A, A_2B, or A₃, but not in untransfected CHO cells. Although the two A₂ receptors have been the subject of most previous studies (see Introduction) the maximal increase of ERK1/2 phosphorylation was lower in cells expressing these receptors than in cells with the A₁ or A₃ receptor. In fact, the adenosine analog produced the highest level of phosphorylation in cells expressing A₃ receptors, a receptor that has not previously been studied.

Differences in efficacy may be caused by differences in receptor expression and it has been pointed out that overexpression of receptors can lead to detection of signaling mechanisms that do not occur in cells where the receptors are expressed at a normal level (Kenakin, 1997). However, in the experimental model used in our study the transfected receptors are expressed at levels that resemble normal expression (Klotz et al., 1998) as shown by ligand binding for A_1, A_2A, and A₃ receptors. The number of A₁ receptors per A₁-transfected CHO cell is on the order of 40,000 (L. V. Lopes, B. Kull, and B. B. Fredholm, unpublished data), which is lower than the number reported, e.g., on the hamster smooth muscle cell line DDT₁ MF-2 [110,000 receptors/cell (Gerwins et al., 1990)]. Previously, it was found that the number of A₁ and A₃ receptors in CHO cells is similar (Klotz et al., 1998). We also have shown that the number of A_2A receptors in the CHO cells transfected with the human A_2A receptor is lower than the number of receptors expressed in striatopallidal neurons in the intact rat striatum (Kull et al., 1999b). It is difficult to estimate the number of A_2B receptors expressed in CHO cells because specific ligands for ligand-binding studies are not available. In unpublished experiments we have used [³H]ZM241385, which does bind to adenosine A_2B receptors with appreciable affinity (Ji and Jacobson, 1999). The results suggest that there are fewer than 200,000 A_2B receptors per transfected CHO cell. For all these reasons we conclude that there is no major risk that we have studied signal pathways that only operate when receptors are expressed at levels dramatically higher than those occurring naturally.

The potency of NECA on the adenosine A₁ and A_2A receptor resembles the values reported previously in the same cells for radioligand binding or cyclic AMP production (Klotz et al., 1998; Kull et al., 1999a). The situation is somewhat similar for the A₃ receptor: half-maximal ERK1/2 phosphorylation was observed at 65 nM NECA, which lies in the same range as described for the rat (Zhou et al., 1992) and the human (Salvatore et al., 1993) A₃ receptor by using cAMP generation as the assay. However, in A_2B receptor-transfected CHO cells, half-maximal stimulation of ERK1/2 phosphorylation occurred at concentrations two orders of magnitude lower than those required to half maximally activate cAMP production (Fig. 2). Indeed, NECA was more potent in causing ERK1/2 phosphorylation via A_2B receptors than via any of the other subtypes. The adenosine receptor antagonist CGS 15943 abolished the A_2B-mediated effect, and the order of potency of different agonists, such as NECA, CGS 21680, and R-PIA, was similar to the one reported earlier, thus confirming that the increase in ERK1/2 phosphorylation in A_2B-transfected CHO cells is really mediated by this receptor subtype.

Previous studies of the potency of agonists to activate/inhibit adenylyl cyclase have suggested that the A₁, A_2A, and A₃ receptors may be activated by concentrations of the endogenous agonist adenosine that occur under basal physiological conditions, whereas A_2B receptors may be activated by levels that occur only pathophysiologically (Fredholm et al., 1994). The present data suggest that this conclusion may not be as valid if other signaling pathways than the cAMP/cAMP-dependent protein kinase cascade, phospholipase C/inositoltrisphosphate formation or ion channel activation are considered. Consistent with this hypothesis is the finding that adenosine mediates the phosphorylation of ERK1/2 via any one of the human adenosine receptor subtypes already at physiologically normal concentrations of 100 nM.

Adenosine A_2B receptors have been suggested to play an important role during conditions such as hypoxia and ischemia when the adenosine concentration rises dramatically and can reach levels where A_2B receptors are activated. Our results show an activation of ERK1/2 at NECA concentrations in the lower nanomolar range in A_2B receptor-transfected CHO cells. This may have fundamental consequences for understanding the role of adenosine A_2B receptors. It may be that not only cAMP and calcium signals initiated by adenosine levels above 500 nM but also MAPK signaling at much lower, physiologically normal concentrations (30-300 nM) turn out to be important. In vitro studies have suggested that adenosine receptor-activated MAPK is important in the regulation of endothelial cell (Sexl et al., 1997) and astrocyte (Neary et al., 1998) proliferation, and in mast cell activation (Feoktistov et al., 1999), but the physiological role of adenosine receptor-mediated ERK1/2 activation remains to be studied.