Abstract
The agonist selectivity for adenosine di- and triphosphates was determined for the human P2Y1 receptor stably expressed in human 1321N1 astrocytoma cells and was studied under conditions in which nucleotide metabolism was both minimized and assessed. Cells were grown at low density on glass coverslips, encased in a flow-through chamber, and continuously superfused with medium, and Ca2+responses to nucleotides were quantified. Superfusion with high performance liquid chromatographically purified ADP, ATP, 2-methylthio-ADP, and 2-methylthio-ATP resulted in rapid Ca2+ responses, with EC50 values of 10 ± 5, 304 ± 51, 2 ± 1, and 116 ± 50 nm, respectively. Similar peak responses were observed with maximal concentrations of these four agonists and with the hydrolysis-resistant adenine nucleoside triphosphate adenosine-5′-O-(3-thiotriphosphate). No conversion of [3H]ATP to [3H]ADP occurred under these conditions. Similar full agonist activities of ATP, 2-methylthio-ATP, and ADP were observed in human embryonic kidney 293 cells, which natively express the P2Y1 receptor. In contrast to these results, Leon et al. [FEBS Lett 403:26–30 (1997)] and Hechler et al.[Mol Pharmacol 53:727–733 (1998)] recently reported that, whereas ADP and 2-methylthio-ADP were agonists, ATP and 2-methylthio-ATP were weak antagonists in studies of the human P2Y1 receptor expressed in human Jurkat cells. To assess whether differences in the degree of receptor reserve might explain this discrepancy of results, P2Y1 receptor-expressing 1321N1 cells were incubated for 24 hr with adenosine-5′-O-(2-thiodiphosphate), with the goal of down-regulating the level of functional receptors. Pretreatment with adenosine-5′-O-(2-thiodiphosphate) resulted in a 10-fold rightward shift in the concentration-effect curve for ADP; in contrast, the agonist activity of ATP was completely abolished. Taken together, our results indicate that adenosine di- and triphosphates are agonists at the human P2Y1 receptor. However, the intrinsic efficacy of ATP is less than that of ADP, and the capacity of ATP to activate second messenger responses through this receptor apparently depends on the degree of P2Y1 receptor reserve.
P2Y receptors are G protein-coupled receptors that are activated by extracellular nucleotides (Fredholm et al., 1994). Molecular cloning and heterologous protein expression have led to unambiguous identification of at least five mammalian P2Y receptor subtypes (Communi et al., 1997; Fredholm et al., 1997). Antagonists capable of discriminating among the subtypes of P2Y receptors are not generally available, and pharmacological characterization of these receptors has been limited to descriptions of the rank order of potencies for activation by nucleotide agonists.
Metabolism of extracellular nucleotides complicates delineation of the agonist selectivity of P2Y receptors (Harden et al., 1997). For example, 1321N1 human astrocytoma cells (a cell line commonly used for heterologous expression of P2Y receptors) express both ectonucleotidase and extrafacial nucleoside diphosphokinase activities, which modify exogenously applied nucleotides (Lazarowski et al., 1997a, 1997b). Endogenous nucleotides also are released from these cells (Lazarowski et al., 1995, 1997a) and other cells (Osipchuk and Cahalan, 1992; Grierson and Meldolesi, 1995; Schlosseret al., 1996; Grygorczyk and Hanrahan, 1997), constitutively or as a result of mechanical stimulation, and promote elevation of basal levels of second messengers and receptor desensitization (Hardenet al., 1997). The purity of nucleotides obtained from commercial sources also is important (Nicholas et al., 1996;Harden et al., 1997; Leon et al., 1997), and marked effects on the apparent agonist selectivity of P2Y2, P2Y4, and P2Y6 receptors were demonstrated when precautions were taken to ensure nucleotide purity and stability (Nicholas et al., 1996).
The P2Y1 receptor, which promotes phospholipase C-catalyzed generation of inositol phosphates and subsequent mobilization of intracellular calcium, has been characterized by an agonist potency order of 2MeSADP > 2MeSATP > ADP > ATP (Webb et al., 1993; Filtz et al., 1994;Henderson et al., 1995; Schachter et al., 1996;Boyer et al., 1996). However, the previously described agonist activity of ATP and ATP analogues at the P2Y1 receptor recently has been questioned (Leonet al., 1997; Hechler et al., 1998). Whereas ADP and 2MeSADP were potent full agonists, ATP and 2MeSATP purified from contaminating nucleoside diphosphates by HPLC did not exhibit agonist activity and acted as weak antagonists of the human P2Y1 receptor-promoted calcium response in suspensions of Jurkat T cells heterologously expressing this receptor. This observed adenine nucleoside diphosphate specificity of the P2Y1 receptor is strikingly similar to the pharmacological selectivity of a receptor on platelets that is referred to as the P2T receptor (Hourani and Hall, 1994), which led to the suggestion that the two receptors may be the same signaling protein (Leon et al., 1997).
Unambiguous description of the pharmacological selectivity of P2Y receptors requires the use of purified nucleotides in a rapid assay of agonist action that is performed under conditions that minimize the confounding influences of ectoenzyme activities and endogenous nucleotide release. To this end, we have adapted a digital imaging system to record intracellular calcium responses in human P2Y1 receptor-expressing 1321N1 cells grown at low density and continuously superfused with buffer. This method provides a test system that both minimizes and allows accurate quantification of the metabolism of superfused molecules. Both ATP and 2MeSATP were active as P2Y1 receptor agonists in this assay system, using cells that either heterologously or natively expressed the P2Y1 receptor. Our results suggest that the extent of the agonist activity of adenosine triphosphates at the human P2Y1 receptor is dependent on the extent of receptor reserve.
Experimental Procedures
Materials.
2MeSADP and 2MeSATP were obtained from Research Biochemicals International (Natick, MA), ADP, ADPβS, and ATPγS were from Boehringer Manheim Biochemicals (Indianapolis, IN), ATP and UTP were from Pharmacia (Piscataway, NJ), [3H]ATP (15–30 Ci/mmol) was from Amersham (Arlington Heights, IL), grade I apyrase was from Sigma Chemical Co. (St. Louis, MO), and fura-2/acetoxymethyl ester was from Molecular Probes (Eugene, OR). The sources of all other reagents have been reported (Schachter et al., 1996).
HPLC.
Separation of nucleotides was accomplished by HPLC with a Zorbax strong anion exchange column (4.6 mm × 25 cm). Samples were eluted with a linear gradient of NaCl (0–1 m) in a buffer of 10 mmNH4HPO4, pH 6.5. Nucleotides were detected by absorbance at 260 nm and were identified from the elution profiles of authentic standards. Assays were routinely carried out with freshly purified nucleotides.
Intracellular calcium measurement.
1321N1 human astrocytoma cells stably expressing the human P2Y1 receptor (hP2Y1–1321N1 cells) (Schachter et al., 1996) and human HEK293 cells (Schachter et al., 1997) were grown on glass coverslips for 2–3 days to a cell density of approximately 20% of confluence. Intracellular calcium was quantified essentially as described (Palmer et al., 1994). Coverslips containing fura-2 (0.5–1.0 μm)-loaded cells were encased in a flow-through chamber (0.2-ml volume) and superfused continuously (1.4 ml/min) with Hanks’ buffered saline solution, with or without agonist. Solution changes were accomplished by means of a valve attached to a gravity-driven six-well reservoir. The flow-through chamber was secured to the stage of a Nikon Diaphot inverted fluorescence microscope. The cells were exposed to alternating excitation wavelengths of 340 and 380 nm, and fluorescence emission at 510 nm was monitored by a silicone-intensified tube camera. The 340/380-nm fluorescence emission ratio was determined and converted to the intracellular Ca2+ concentration using the equation ofGrynkiewicz et al. (1985). Data were recorded and processed using an Image 1/FL digital imaging system (Universal Imaging Corp., West Chester, PA).
Results
We previously observed in assays of inositol lipid hydrolysis that ATP and 2MeSATP, although somewhat less potent than ADP and 2MeSADP, respectively, were full or nearly full agonists at the P2Y1 receptor (Filtz et al., 1994;Schachter et al., 1996). Considering the contrasting results of Gachet and co-workers (Leon et al., 1997; Hechleret al., 1998), we compared the activities of adenine nucleoside di- and triphosphates at the human P2Y1 receptor under conditions designed to circumvent the confounding influences of secreted endogenous nucleotides and nucleotide metabolism (see Experimental Procedures). Rapid calcium responses to nucleotide agonists were recorded from small numbers of continuously superfused cells. The commercially obtained preparations of ATP and 2MeSATP used in this study were contaminated by approximately 1% and 5%, respectively, with the corresponding nucleoside diphosphate. Therefore, all nucleotides were purified by strong anion exchange HPLC just before testing of the hP2Y1–1321N1 cells.
Superfusion of hP2Y1–1321N1 cells with HPLC-purified ATP, ADP, 2MeSATP, or 2MeSADP resulted in rapid increases in intracellular Ca2+ concentrations (Fig.1). The superfusion time required to elicit a response and the rate of elevation of intracellular Ca2+ concentrations were identical for all four agonists. Moreover, maximally effective concentrations of 2MeSATP and ATP elicited responses that were comparable in magnitude to those produced by ADP, although in some experiments the maximal effect of ATP was somewhat less than that of ADP. Assuming that ADP exhibited a relative efficacy of 1.00 for eliciting a maximal response, the averaged relative efficacies of HPLC-purified ATP and 2MeSATP were 0.80 ± 0.08 (mean ± standard error of 18 experiments) and 1.00 ± 0.05 (mean ± standard error of 12 experiments), respectively. Although the response tracings of ADP versus ATP were often superimposable (Fig. 1), in some experiments differences in the descending portion of response tracings were noted (data not shown).
Complete concentration-effect curves were generated for the adenine nucleoside di- and triphosphates (Fig.2). The nucleoside diphosphates were more potent than their triphosphate counterparts, and a rank order of potency for stimulation of Ca2+ mobilization of 2MeSADP > ADP > 2MeSATP > ATP was observed (Fig. 2, Table 1) The position of ADP in this order of potency differs from that we previously reported for the human P2Y1 receptor in studies of agonist-promoted inositol phosphate formation (Schachter et al., 1996).
Although nucleotides were exposed only briefly to a few cells under continuous flow in these experiments, sufficient conversion of nucleoside triphosphate to diphosphate might occur to account for the agonist activity of ATP and 2MeSATP. To examine this possibility, the conversion of [3H]ATP to [3H]ADP was directly measured under the same conditions in which intracellular calcium responses were recorded. A 10 μm solution of ATP containing HPLC-purified [3H]ATP was superfused over coverslips, with or without hP2Y1–1321N1 cells, for 30 sec. The effluent was collected and3H-labeled nucleotides were resolved by HPLC. [3H]ADP was not present in amounts that were statistically significantly above the background levels in control superfusate. Moreover, no increase in [3H]ADP levels was detected in the superfusate from chambers containing cells (Table 2). Thus, conversion of the nucleoside triphosphates to their diphosphate forms cannot account for the observed responses to ATP and 2MeSATP. This conclusion is supported by the observation that the hydrolysis-resistant analogue ATPγS (made diphosphate-free by the treatment described in the legend to Fig.3) elicited Ca2+responses in hP2Y1–1321N1 cells (Fig. 3).
Disparate conclusions regarding the relative activities of diphosphates and triphosphates at the human P2Y1 receptor could be explained by differences in the levels of functional receptors in various P2Y1 receptor test systems. Agonists with low intrinsic efficacy might be equally effective, compared with full agonists, for stimulation of a maximal response in cells exhibiting considerable receptor reserve. In contrast, responses to a low-efficacy agonist would be submaximal or absent in cells with no receptor reserve. Thus, the possibility that ATP is a low-efficacy agonist at the human P2Y1 receptor was examined by subjecting hP2Y1–1321N1 cells to agonist-induced desensitization. Cells were incubated for 24 hr with 100 μm ADPβS (ADPβS is a full agonist with an EC50 of 66 nm) and then rechallenged with a range of concentrations of ADP or ATP. The resulting concentration-effect curve for ADP was shifted approximately 10-fold to the right and slightly downward in ADPβS-pretreated cells. In contrast, prolonged preincubation with ADPβS essentially eliminated the capacity of ATP to elicit a calcium response in hP2Y1–1321N1 cells (Fig. 4). Although 30 μm ATP exhibited no agonist activity in ADPβS-desensitized cells, this concentration of ATP partially blocked thecalcium response to 300 nm ADP in these pretreated cells (Fig. 5). Thus, the agonist action of ATP, but not ADP, was largely lost under conditions that likely substantially reduced the number of functional P2Y1 receptors. These results suggest that ATP is a low-efficacy agonist at the human P2Y1 receptor and that the capacity of ATP to elicit a calcium response is dependent on the levels of functional P2Y1 receptors.
Because most studies examining the relative agonist activities of ADP and ATP at the human P2Y1 receptor were carried out with heterologously expressed receptors, it also was important to test these activities at natively expressed P2Y1receptors. Thus, agonist activities of 2MeSATP and ATP were examined by measuring Ca2+ responses from a small number of superfused HEK293 cells, which natively express the P2Y1 receptor (Schachter et al., 1997). These cells also natively express a phospholipase C-coupled P2Y2 receptor that we have shown mediates a minor component of the inositol phosphate response to ATP. Consistent with our earlier findings, a maximal concentration of UTP stimulated calcium responses from approximately one third of the cells tested (Fig.6, A and B). In contrast, HPLC-purified 2MeSATP, which is inactive at the P2Y2 receptor (Lazarowski et al., 1995), and ATP elicited rapid calcium responses from essentially all HEK293 cells (Fig. 6, A and C). These results are consistent with the idea that the response of HEK293 cells to ATP is predominantly attributable to the P2Y1 receptor and that ATP and 2MeSATP are both agonists at the human P2Y1 receptor present at natively expressed levels.
Discussion
A P2Y1 receptor-expressing cell preparation was used under conditions that circumvent potential contributions of nucleotide metabolism or release to the rapid responses observed with superfused agonists. HPLC purification of test molecules immediately before assay also ensured that the effects measured were not a combination of responses to both nucleoside di- and triphosphates. Under these conditions, adenosine diphosphates were clearly the most potent agonists at the human P2Y1 receptor, as was recently emphasized by Gachet and co-workers (Leon et al., 1997; Hechler et al., 1998). However, in contrast to those results, we observed that ATP, 2MeSATP, and ATPγS also were P2Y1 receptor agonists.
The absence of a reliable binding assay for P2Y1receptors has relegated this (and other P2Y receptors) to a group of signaling proteins for which ligand binding affinities cannot be directly determined. Therefore, assays of the activation of phospholipase C or of downstream signaling responses have evolved as the primary means to assess ligand interactions at P2Y receptors. These assays provide accurate monitoring of second messenger responses but only indirectly reflect ligand-receptor interactions, as well as introducing limitations associated with most steady-state biochemical analyses. Occurrence of agonist-induced receptor desensitization is a potential problem. The existence of a complex set of ectoenzymes further complicates study of extracellular nucleotides, because triphosphates can be converted to diphosphates by nucleotidases (Zimmermann, 1996) and triphosphates can be formed from diphosphates by nucleoside diphosphokinases (Lazarowski et al., 1997a). Thus, simple elimination of contaminating nucleotides in stock solutions of a given nucleotide by HPLC purification does not preclude the contribution of rapid formation of similar molecules as metabolites during measurements of signaling responses.
We have emphasized the differences in apparent agonist selectivity that the aforementioned factors can produce in studies of the uridine nucleotide-activated P2Y2, P2Y4, and P2Y6 receptors (Nicholas et al., 1996). The results of Leon et al. (1997) with P2Y1 receptors stably expressed in Jurkat cells established that similar considerations apply in delineating agonist selectivities at the P2Y1receptor. Whereas previous studies from our laboratory and other laboratories showed that both 2MeSATP and ATP were potent full agonists at the P2Y1 receptor (Webb et al., 1993; Filtz et al., 1994; Henderson et al., 1995;Schachter et al., 1996), this was not the case with the Jurkat cell-expressed P2Y1 receptor when nucleotides were purified by HPLC before agonist activity testing and rapid measurements of Ca2+ accumulation were carried out to minimize metabolism or interconversion of agonists during drug testing (Leon et al., 1997). Although the results of Leon et al. (1997) prompted us to adopt a P2Y1 receptor test system that alleviated problems of nucleotide metabolism, our initial results were not entirely consistent with those reported in their well-controlled studies. That is, as we previously observed in the less definitive studies of inositol phosphate accumulation, ATP activated the P2Y1 receptor to promote Ca2+ accumulation, and the maximal effects of ATP were very similar to those of ADP. Thus, in our hands ATP is not simply an antagonist at the human P2Y1 receptor. The same conclusions can be drawn from our data with 2MeSATP and, perhaps more importantly, ATPγS, which is a nonhydrolyzable triphosphate analogue that is a full or nearly full agonist at the P2Y1 receptor.
One possible explanation for differences between our results and those of Leon et al. (1997) and Hechler et al.(1998) is that P2Y1 receptors might be expressed at much higher levels in the 1321N1 cells used in the current study, compared with the Jurkat cell-expressed P2Y1receptors. Because the studies of Leon et al. (1997) andHechler et al. (1998) were carried out with suspended cells at relatively high cell densities, release of endogenous ATP and basal activation of the expressed P2Y1 receptors also were more likely to occur than in the current study, which was carried out with relatively few cells in a monolayer under constant superfusion. Thus, lowering of functional receptor levels by agonist-induced desensitization before drug treatment is a possibility with the Jurkat cell-expressed receptors. Conversely, relatively high levels of P2Y1 receptor expression in 1321N1 cells may generate a large receptor reserve that, in turn, permits detection of agonist activity of nucleotides that exhibit much lower intrinsic efficacy than does ADP. The 20-fold greater potencies of ADP and 2MeSADP determined at P2Y1 receptors expressed in 1321N1 cells, compared with the values reported for the receptors expressed in Jurkat cells, suggests greater functional P2Y1 receptor expression in our studies, compared with those of Leon et al. (1997) and Hechler et al. (1998). Moreover, experiments with agonist-preincubated 1321N1 cells directly confirm this possibility. That is, preincubation of P2Y1 receptor-expressing 1321N1 cells with ADPβS resulted in a complete loss of response to ATP but only a shift to the right of the concentration-effect curve for ADP, with no substantial decrease in the maximal effect. Indeed, some antagonist activity of ATP was apparent in experiments with the desensitized cells. Thus, we conclude from these studies that both ADP and ATP are agonists at the human P2Y1 receptor and that the intrinsic efficacy of ADP is greater than that of ATP.
The physiological significance of our observations is not completely clear. The agonist activity of ATP is not restricted to the unnatural situation of heterologous overexpression of P2Y1receptors in 1321N1 cells, because ATP was also an agonist at human P2Y1 receptors natively expressed in HEK293 cells and is likely to be a physiologically important activator of P2Y1 receptors in various target tissues. Our results emphasize that it will be important in future studies to assess the accuracy with which in vitro tests of receptor activity reflect the extent of functional reserve of P2Y1receptors in vivo. ADP clearly is more potent than ATP and, under conditions of little P2Y1 receptor reserve, would be a more efficacious agonist. Therefore, the importance of ATP as an agonist at the P2Y1 receptor would depend on the receptor level in target tissues and on the extracellular concentration of ATP, relative to that of ADP. Clearly, the 30-fold greater potency of ADP, relative to ATP, at the P2Y1 receptor ensures that lower concentrations of ADP are detected. The studies of both Leon et al. (1997)and Kunapuli and co-workers (Jin et al., 1998; Danielet al., 1998) strongly suggest that the P2Y1 receptor is prominently involved in the platelet aggregation response to ADP and that, at least in in vitro studies, the agonist action of ADP is antagonized by ATP.
We doubt that the antagonist action of ATP that may exist under conditions of low P2Y1 receptor reserve has physiological importance, because this antagonist activity is observed only at concentrations of ATP much higher than those necessary to observe full agonist actions of ADP. However, the data of Leonet al. (1997) and Hechler et al. (1998) and those presented here indicate that the rate of metabolism of extracellular ATP to ADP may prove to be an important temporal determinant of the agonist action of released adenine nucleotides at the P2Y1 receptor. Clearly, multiple factors exist that can serve to regulate the extracellular action of adenine nucleotides.
Footnotes
- Received June 22, 1998.
- Accepted September 15, 1998.
-
Send reprint requests to: Dr. T. Kendall Harden, CB 7365, Department of Pharmacology, School of Medicine, University of North Carolina, Chapel Hill, NC 27599-7365. E-mail:tkh{at}med.unc.edu
-
This work was supported by United States Public Health Service Grants GM38213 and HL54889 and by a National Research Service Award (GM18464) to R.K.P.
Abbreviations
- 2MeSADP
- 2-methylthio-ADP
- 2MeSATP
- 2-methylthio-ATP
- ADPβS
- adenosine-5′-O-(2-thiodiphosphate)
- ATPγS
- adenosine-5′-O-(3-thiotriphosphate)
- HEK
- human embryonic kidney
- HPLC
- high performance liquid chromatography
- The American Society for Pharmacology and Experimental Therapeutics