The primordial intercellular signaling molecule ATP acts through two families of cell-surface P2 receptors – the P2Y family of G-protein-coupled receptors and the P2X family of ligand-gated cation channels. Multiple P2 receptors are expressed in a variety of cell types. However, the significance of these networks of receptors in any biological system remains unknown. Using osteoblasts as a model system, we found that a low concentration of ATP (10 µM, ATPlow) induced transient elevation of cytosolic Ca2+, whereas a high concentration of ATP (1 mM, ATPhigh) elicited more sustained elevation. Moreover, graded increases in the Ca2+ signal were achieved over a remarkable million-fold range of ATP concentrations (1 nM to 1 mM). Next, we demonstrated that ATPlow caused transient nuclear localization of the Ca2+-regulated transcription factor NFATc1; whereas, ATPhigh elicited more sustained localization. When stimulated with ATPhigh, osteoblasts from P2X7 loss-of-function mice showed only transient Ca2+-NFATc1 signaling; in contrast, sustained signaling was observed in wild-type cells. Additional experiments revealed a role for P2Y receptors in mediating transient signaling induced by low ATP concentrations. Thus, distinct P2 receptors with varying affinities for ATP account for this wide range of sensitivity to extracellular nucleotides. Finally, ATPhigh, but not ATPlow, was shown to elicit robust expression of the NFAT target gene Ptgs2 (encoding COX-2), consistent with a crucial role for the duration of Ca2+-NFAT signaling in regulating target gene expression. Taken together, ensembles of P2 receptors provide a mechanism by which cells sense ATP over a wide concentration range and transduce this input into distinct cellular signals.
Purine and pyrimidine nucleotides represent a primordial and pervasive class of intercellular chemical messengers (Burnstock and Verkhratsky, 2009). Many stimuli, including shear stress, mechanical stretch, osmotic swelling and hypoxia, trigger release of nucleotides from cells via mechanisms such as vesicular exocytosis and plasma membrane channels or transporters (Burnstock, 2007b). Once in the extracellular milieu, nucleotides act on target cells through two families of P2 nucleotide receptors – the P2Y family of G-protein-coupled receptors and the P2X family of ligand-gated cation channels (Abbracchio et al., 2006; Khakh and North, 2006).
Eight P2Y subtypes (P2Y1, P2Y2, P2Y4, P2Y6, P2Y11–14) and seven subtypes of P2X (P2X1–7) have been identified in mammals, and are involved in regulating processes such as secretion, cell proliferation, differentiation, motility and death (Khakh and North, 2006; Burnstock, 2007a; Burnstock and Verkhratsky, 2010). Many P2Y receptors, including P2Y1, P2Y2, P2Y4 and P2Y6, couple to activation of phospholipase C (PLC), resulting in formation of inositol (1,4,5)-trisphosphate and subsequent release of Ca2+ from intracellular stores (Burnstock, 2007a). In contrast, homo- or heteromultimers of three ionotropic P2X receptor subunits form functional channels permeable to Na+, K+ and Ca2+ (Browne et al., 2010). Consequently, activation of P2X receptors results in membrane depolarization and, in many cases, Ca2+ influx. P2Y receptors are activated by one or more of ATP, ADP, UTP, UDP or UDP-glucose, with P2Y1, P2Y2, P2Y4, and P2Y11–13 exhibiting some degree of sensitivity for ATP; in contrast, members of the P2X receptor family are activated solely by ATP (Burnstock, 2007a).
All mammalian cell types express multiple P2 receptor subtypes, each with varying affinities for purine and pyrimidine nucleotides (Volonté et al., 2006; Burnstock and Verkhratsky, 2009). Bone-forming osteoblasts can express P2Y1, P2Y2, P2Y6, P2Y12–14 and P2X1–7, and several have been shown to regulate distinct processes in these cells (Orriss et al., 2010). For instance, activation of P2Y1 or P2Y2 enhances osteoblast responses to systemic factors such as parathyroid hormone (Bowler et al., 1999; Bowler et al., 2001). ATP or UTP, acting through P2Y2, inhibit matrix mineralization in cultures of differentiating osteoblasts (Hoebertz et al., 2002; Orriss et al., 2007). In contrast, activated P2X7 receptors enhance osteoblast differentiation and increase bone formation (Ke et al., 2003; Panupinthu et al., 2008). More recently, P2Y12 and P2Y13 have also been shown to positively regulate osteoblast proliferation and differentiation in vitro and in vivo (Syberg et al., 2012; Wang et al., 2012). Though the functions of individual P2Y and P2X receptors have been examined in osteoblasts and other cell types, mechanisms by which multiple P2 receptor subtypes act in concert to regulate cellular differentiation and function remain unclear in any system (Volonté et al., 2006).
The nuclear factor of activated T-cells (NFAT) family of transcription factors includes four members that are regulated by cytosolic Ca2+ (NFATc1–4) (Crabtree and Olson, 2002). In resting cells, NFATc1–4 are phosphorylated and localized to the cytosol. Receptor-mediated elevation in the concentration of cytosolic free Ca2+ ([Ca2+]i) leads to activation of calcineurin, a serine/threonine phosphatase that dephosphorylates NFATc1–4, resulting in their nuclear translocation and transcriptional activation (Hogan et al., 2003). Though classically described as master regulators of T-cell development and function, NFATc1–4 also play essential roles in the differentiation and function of neuronal, muscle and bone cells, including osteoblasts and osteoclasts (Hogan et al., 2003; Sitara and Aliprantis, 2010; Moore and Goldberg, 2011). However, whether P2 receptors signal through NFAT in bone-forming osteoblasts remains unexplored.
In the present study, we show that endogenous co-expression of multiple P2Y and P2X receptors provides a novel mechanism for dose-to-duration encoding of Ca2+-NFAT signaling. Using osteoblasts as a model system, we found that increases in Ca2+ signaling could be achieved over a million-fold range of ATP concentrations. Low concentrations of ATP acting through P2Y receptors caused transient elevation of Ca2+ and brief nuclear localization of NFATc1, but failed to induce expression of NFAT target genes. In contrast, high ATP concentrations acting through P2X7 elicited sustained Ca2+-NFATc1 signaling and robust NFAT transcriptional activity. Taken together, these data show that P2 receptor networks provide a mechanism by which cells sense ATP over a wide range of concentrations and transduce this input into distinct cellular signals.
Effect of ATP concentration on the duration of cytosolic Ca2+ signals
Cells of the osteoblast lineage express multiple subtypes of P2Y and P2X nucleotide receptors (Orriss et al., 2012). Given the central role for Ca2+ in P2 receptor signaling, we investigated the functional significance of endogenous P2 receptor networks by assessing Ca2+ signaling elicited by extracellular ATP. MC3T3-E1 osteoblast-like cells were loaded with the Ca2+-sensitive dye fluo-4, and changes in [Ca2+]i were monitored using real-time imaging of live-cells by confocal microscopy (Fig. 1A). A low concentration of ATP (10 µM, which activates some P2Y and all P2X receptor subtypes except P2X7; ATPlow) induced large, transient elevations in [Ca2+]i that recovered to baseline within ∼240 s after stimulation (data are whole-field Ca2+ responses, Fig. 1A,B; supplementary material Movie 1). In contrast, a higher ATP concentration (1 mM, which activates some P2Y and all P2X receptor subtypes including P2X7; ATPhigh) elicited larger, more sustained elevations in [Ca2+]i exceeding 20 min in duration (Fig. 1A,B; supplementary material Movie 2). Vehicle elicited a small elevation in [Ca2+]i (Fig. 1A,B), consistent with fluid shear-induced release of ATP and slight P2 receptor activation (supplementary material Figs S1, S2), as described previously (Li et al., 2005).
To further characterize ATP-induced Ca2+ signaling, changes in [Ca2+]i were assessed in response to ATP concentrations from 1 nM to 10 mM. We first plotted the peak amplitude of the Ca2+ signal against ATP concentration (Fig. 1C). Remarkably, graded increases in the peak Ca2+ response were observed over a million-fold range of ATP concentrations (1 nM to 1 mM). The amplitude at 10 mM ATP was slightly reduced, likely reflecting chelation of extracellular Ca2+ by this high concentration of ATP. In contrast, only ATP concentrations >100 µM elicited a sustained Ca2+ signal (measured as amplitude 10 min post-treatment, Fig. 1D). Similar patterns of concentration dependence were observed when percentages of cells exhibiting elevation of [Ca2+]i were examined (supplementary material Fig. S3).
Finally, we quantified the magnitude of responses as the area under the fluorescence–time curve, providing a combined measure of both amplitude and duration. In this case, concentration dependence was clearly biphasic (Fig. 1E). The first phase began at ∼1 nM and plateaued at ∼1 µM ATP, whereas the second phase began at ∼100 µM and plateaued at ∼1 mM ATP.
P2X7 is essential for sustained cytosolic Ca2+ signaling elicited by high concentrations of ATP
Extracellular ATP activates a number of P2Y and P2X receptor subtypes at concentrations within the nM to µM range; in contrast, P2X7 is stimulated solely at ATP concentrations exceeding 100 µM (Khakh and North, 2006; Burnstock, 2007a). Therefore, we investigated whether P2X7 receptors mediate the effects of high ATP concentrations. Changes in [Ca2+]i were examined following treatment of osteoblast-like cells with vehicle, ATPlow or BzATP (300 µM) in the absence or presence of A 438079 (10 µM), a specific P2X7 antagonist (Nelson et al., 2006; Donnelly-Roberts and Jarvis, 2007) (Fig. 2A–C). Although not specific for P2X7, BzATP is a more potent agonist than ATP at the P2X7 receptor (North, 2002). In the absence of antagonist, BzATP elicited a large, sustained elevation in cytosolic Ca2+, whereas ATPlow induced a more transient increase (Fig. 2A–C). A 438079 abolished the sustained phase of the BzATP-induced response, converting it to a transient increase comparable to that elicited by ATPlow (Fig. 2A–C). A 438079 had no effect on Ca2+ elevation induced by ATPlow, consistent with the specificity of this P2X7 receptor antagonist.
To confirm the role of P2X7 in mediating effects of high concentrations of ATP or BzATP on cytosolic Ca2+ signaling, changes in [Ca2+]i were examined in primary calvarial osteoblasts isolated from wild-type mice and mice with loss of P2X7 function (knockout) (Fig. 2D–F). Stimulation of wild-type calvarial osteoblasts with ATPlow induced transient elevations in [Ca2+]i that returned to baseline within ∼240 s, whereas ATPhigh and BzATP (300 µM) both elicited more sustained elevations of [Ca2+]i exceeding 20 min in duration. In contrast, responses to ATPhigh in calvarial osteoblasts from knockout mice were comparable to those induced by ATPlow, and treatment with BzATP had no significant effect on [Ca2+]i (Fig. 2D–F). Taken together, pharmacological and genetic evidence establish that the sustained elevations in [Ca2+]i induced by BzATP or high concentrations of ATP are mediated by activation of P2X7 receptors.
Source of Ca2+ underlying the transient and sustained elevations of [Ca2+]i elicited by ATP
Several P2Y receptors couple to activation of PLC, resulting in release of Ca2+ from intracellular stores (Burnstock, 2007a). In contrast, P2X receptors form channels that in many cases permit Ca2+ influx (Browne et al., 2010). We next determined the source of Ca2+ required for the transient and sustained elevations of [Ca2+]i. To assess the contribution of intracellular stores, changes in [Ca2+]i were examined following treatment with vehicle, ATPlow or BzATP (300 µM) in the absence or presence of the PLC inhibitor U 73122 (1 µM) (Fig. 3A–C). U 73122 abolished the response induced by ATPlow and the transient component of the Ca2+ signal elicited by BzATP. In contrast, the inhibitor had no significant effect on the sustained phase of the BzATP-induced response (Fig. 3A–C).
To determine the role of extracellular Ca2+, changes in [Ca2+]i were examined following treatment of cells with vehicle, ATPlow or BzATP (300 µM) in the presence or absence of extracellular Ca2+ (Fig. 3D–F). Removal of extracellular Ca2+ abolished the sustained phase of the BzATP-induced response. In contrast, there was no significant effect on the transient component of the responses to ATPlow and BzATP (Fig. 3D–F). Taken together, these data are consistent with ATPlow activating P2Y receptors, leading to release of Ca2+ from intracellular stores. On the other hand, BzATP activates both P2Y receptors and P2X7, leading to transient release of Ca2+ from intracellular stores and sustained influx, respectively.
We further investigated the nature of the P2Y receptors mediating transient Ca2+ signaling. UTP (10 µM) elicited responses comparable to those elicited by ATPlow, providing further evidence for involvement of P2Y receptors. Next, we examined responses to sequential addition of ATPlow and UTP and observed marked cross-desensitization (supplementary material Fig. S4), consistent with the agonist specificities of P2Y2 and/or P2Y4 (Burnstock, 2007a).
Effect of ATP concentration on the duration of NFATc1 nuclear localization
Elevations in [Ca2+]i can activate calcineurin, resulting in NFATc1 translocation from the cytoplasm to the nucleus (Crabtree and Olson, 2002; Hogan et al., 2003). To characterize the dynamics of NFATc1 activation in response to ATP, osteoblast-like cells were transfected with EGFP-tagged NFATc1, and subcellular localization was monitored by confocal imaging of live cells. In the absence of agonists, NFATc1–EGFP was uniformly distributed throughout the cytoplasm, with little if any fluorescence in the nucleus (supplementary material Fig. S5A). ATP (0.01–1 mM) and BzATP (300 µM) all induced prompt translocation of NFATc1–EGFP to the nucleus. Low concentrations of ATP (10–100 µM) elicited transient nuclear localization of NFATc1–EGFP, with recovery of cytosolic fluorescence less than 90 min after stimulation (supplementary material Fig. S5 and Movie 3). In contrast, ATPhigh and BzATP both induced more sustained nuclear localization of NFATc1–EGFP that persisted for at least 2 h (supplementary material Fig. S5 and Movie 4).
To quantify NFATc1 activation, osteoblast-like cells expressing NFATc1–EGFP were treated and fixed at various time points. The percentage of cells exhibiting nuclear localization of NFATc1–EGFP was determined and expressed as a function of nucleotide concentration or time. When assessed 15 min after addition of nucleotide, nuclear localization of NFATc1–EGFP was induced in a concentration-dependent manner over a relatively narrow range for ATP and UTP (from ∼0.1 to 1 µM) and BzATP (from ∼1 to 100 µM) (Fig. 4A). Responses to low concentrations of ATP and UTP and higher concentrations of BzATP indicate involvement of multiple P2 receptor subtypes. Moreover, at this early time point, NFATc1 responds to ATP over a relatively narrow dynamic range of concentrations.
We next examined the time course of NFATc1 nuclear translocation. Consistent with the live-cell data (supplementary material Fig. S5), nuclear localization of NFATc1–EGFP elicited by ATPlow was transient, with recovery of cytosolic localization in all cells by 90 min (Fig. 4B,C). In contrast, ATPhigh and BzATP both induced sustained localization of NFATc1–EGFP to the nucleus that persisted for up to 3 h (Fig. 4B,C). Thus, elevations in [Ca2+]i elicited by increasing ATP concentrations are transduced into graded increases in the duration of NFATc1 nuclear localization. Moreover, effects of BzATP and ATPhigh implicate the P2X7 receptor in mediating prolonged NFATc1 nuclear localization.
P2X7 is essential for sustained NFATc1 nuclear localization elicited by high concentrations of ATP
To investigate whether P2X7 receptors mediate sustained NFATc1 nuclear localization, MC3T3-E1 cells were transfected with EGFP-tagged NFATc1 and treated with vehicle, ATPlow or BzATP (300 µM) in the absence or presence of A 438079 (10 µM). A 438079 abolished the sustained phase of the BzATP-induced response, converting it to a transient response comparable to that elicited by ATPlow (Fig. 5A,B). In contrast, this antagonist had no effect on the transient nuclear localization of NFATc1–EGFP stimulated by ATPlow.
To confirm the role of P2X7 receptors in mediating sustained nuclear localization of NFATc1, primary calvarial osteoblasts were isolated from P2X7 knockout and wild-type mice. Cultures were treated with vehicle, ATPlow, ATPhigh or BzATP (300 µM), and fixed at 15 or 120 min (Fig. 5C,D). Endogenous NFATc1 was labeled using a monoclonal antibody and localized by immunofluorescence (Fig. 5C). The percentage of cells exhibiting nuclear localization was determined for each treatment at 15 and 120 min (Fig. 5D). Stimulation of wild-type calvarial osteoblasts with ATPlow induced transient NFATc1 nuclear localization, whereas ATPhigh and BzATP both elicited sustained localization (Fig. 5C,D). In contrast, responses of osteoblasts from knockout mice to ATPhigh and BzATP were transient and comparable to those elicited by ATPlow (Fig. 5C,D). Approximately 35% of wild-type osteoblasts responded to ATPhigh and BzATP at 120 min (Fig. 5D, right panel), in keeping with the percentage of primary calvarial osteoblasts reported previously to express functional P2X7 receptors (Panupinthu et al., 2007).
Taken together, these experiments establish that the sustained nuclear localization of NFATc1 elicited by BzATP or high concentrations of ATP is mediated by activation of P2X7 receptors; whereas, transient localization induced by low concentrations of ATP is due to activation of higher affinity P2 receptors.
Effects of nucleotides on NFAT transcriptional activity
To determine if differences in duration of P2 receptor-induced Ca2+-NFATc1 signaling give rise to corresponding alterations in expression of NFAT target genes, MC3T3-E1 cells were treated with vehicle, ATPlow, ATPhigh or BzATP (300 µM). Total RNA was isolated at various times, and expression of the NFATc1 target gene Ptgs2 (encoding COX-2) was assessed by real-time RT-PCR (Fig. 6A). BzATP and ATPhigh elicited dramatic increases in expression of Ptgs2 that peaked at 3 h before returning to baseline by 6 h after treatment. In contrast, ATPlow did not induce any increase in Ptgs2 expression.
The Ptgs2 promoter contains binding sites for many transcription factors in addition to NFAT (Kang et al., 2007). To more specifically assess changes in NFAT transcriptional activity, osteoblast-like cells were transfected with an NFAT luciferase reporter and treated with vehicle, ATPlow, ATPhigh or BzATP (300 µM). After 24 h, cell lysates were collected and luminescence was assessed as a measure of NFAT transcriptional activity (Fig. 6B). Similar to the pattern observed for Ptgs2, BzATP and ATPhigh both elicited significant increases in NFAT transcriptional activity. In contrast, ATPlow did not induce any significant change compared to vehicle (Fig. 6B). Taken together, sustained Ca2+-NFATc1 signaling elicited by high concentrations of ATP or BzATP is associated with robust expression of NFAT target genes.
To confirm a role for the Ca2+-calcineurin pathway in regulation of NFAT transcriptional activity, changes in Ptgs2 expression and NFAT reporter activity were examined (supplementary material Fig. S6). Osteoblast-like cells were treated with vehicle, ATPlow, ATPhigh or BzATP (300 µM) in the presence or absence of the calcineurin inhibitors cyclosporin A (1 µM) or FK506 (1 µM). Both cyclosporin A and FK506 significantly suppressed expression of Ptgs2 induced by ATPhigh and BzATP (supplementary material Fig. S6A). Moreover, in osteoblast-like cells transfected with the NFAT luciferase reporter, the effects of BzATP and ATPhigh were abolished by inhibition of calcineurin (supplementary material Fig. S6B). These observations are consistent with nucleotide-induced changes in Ptgs2 expression and NFAT reporter activity being mediated by activation of the Ca2+-calcineurin pathway.
P2X7 is essential for mediating effects of high concentrations of ATP on NFAT transcriptional activity
Changes in Ptgs2 expression and NFAT reporter activity were examined in osteoblast-like cells following treatment with vehicle, ATPlow, ATPhigh or BzATP (300 µM) in the absence or presence of A 438079 (10 µM). A 438079 completely abolished the effects of ATPhigh and BzATP both on expression of Ptgs2 (Fig. 7A) and on luminescence in osteoblast-like cells transfected with the NFAT luciferase reporter (Fig. 7B).
To confirm the involvement of P2X7 receptors, changes in Ptgs2 expression and NFAT reporter activity were examined in primary calvarial osteoblasts isolated from P2X7 knockout and wild-type mice. Stimulation of wild-type calvarial osteoblasts with ATPhigh or BzATP (300 µM) induced significant expression of Ptgs2 at 3 h, whereas ATPlow had no effect (Fig. 7C). In contrast, ATPlow, ATPhigh and BzATP all failed to elicit expression of Ptgs2 in calvarial osteoblasts from knockout mice (Fig. 7C). Moreover, in primary calvarial osteoblasts transfected with the NFAT luciferase reporter, ATPhigh induced an increase in luminescence that was absent in cells from knockout mice (Fig. 7D). These observations establish that the increase in NFAT transcriptional activity and target gene expression elicited by BzATP or high concentrations of ATP is mediated by activation of the P2X7 receptor.
In this study, we examined ATP-induced Ca2+-NFAT signaling through a network of endogenously expressed P2 nucleotide receptors. We show that ensembles of P2Y and P2X receptor subtypes impart sensitivity over a wide range of ATP concentrations, and provide a mechanism by which cells transduce ATP levels into distinct cellular signals (Fig. 8). Specifically, low concentrations of ATP act through P2Y receptors to elicit transient Ca2+-NFAT signaling; whereas, high ATP concentrations act through P2X7 to induce more sustained Ca2+-NFAT signaling that culminates in robust target gene expression. These findings demonstrate that the Ca2+-NFAT pathway functions in dose-to-duration encoding of P2 receptor stimuli.
P2 receptor networks impart sensitivity over a wide dynamic range of ATP concentrations
Previous studies examining the characteristics of individual P2Y and P2X receptor subtypes have demonstrated that each exhibits distinct affinities for purine and pyrimidine nucleotides. Of the P2 receptor subtypes sensitive to ATP, P2Y2, P2X1, P2X3 and P2X5 have the highest affinities, with responses seen at relatively low ATP concentrations (<1 µM); in contrast, P2X7 has the lowest affinity with activation occurring at much higher concentrations of ATP (Ralevic and Burnstock, 1998). Though informative, many of these earlier studies used heterologous expression to characterize individual P2 receptor subtypes. However, expression of a single P2Y or P2X subtype does not recapitulate responses mediated by a network of P2 receptors. Moreover, overexpression of a single receptor could perturb interactions among endogenous P2 receptors, such as P2Y dimerization and the formation of heteromeric P2X receptors. Thus, results of these earlier studies may not represent the physiological behavior of P2 receptor networks. For these reasons, we examined concentration-dependent responses to ATP mediated by a network of P2 receptors endogenously expressed in osteoblasts.
Characterization of the concentration dependence of ATP-induced Ca2+ signaling in osteoblasts revealed, for the first time in any system, that graded increases in the Ca2+ response could be achieved over a remarkably broad range of ATP concentrations (1 nM to 1 mM). The physiological effects of many other intercellular messengers are also mediated by multiple receptor subtypes. However, few if any of these receptor families exhibit as wide a range of affinities as those observed for ATP in the purinergic system. For instance, prostaglandin E2 (PGE2) acts through four G-protein-coupled receptors (EP1–4), which exhibit an ∼100-fold range in EC50 for PGE2 (Abramovitz et al., 2000; Hata and Breyer, 2004). LPA, another lipid mediator, acts through at least five G-protein-coupled receptors, LPA1–5, with a similar 100-fold range in EC50 (Bandoh et al., 2000; Anliker and Chun, 2004; Lee et al., 2006). On the other hand, the excitatory neurotransmitter L-glutamate, which like ATP signals through families of metabotropic and ionotropic receptors, elicits responses at low and high concentrations of L-glutamate (Conn and Pin, 1997; Traynelis et al., 2010). However, it has not been reported that L-glutamate can induce graded, dose-dependent increases in the amplitude or duration of a common intracellular signal. Nevertheless, given its similarities to the purinergic system, the principles described in the present study for ATP signaling may also be applicable to the glutaminergic system.
P2 receptor networks enable dose-to-duration encoding of Ca2+-NFAT signaling
In addition to relaying qualitative information about the presence or absence of a stimulus, receptors and their associated signaling pathways must also transmit quantitative information about stimulus intensity. ‘Dose-to-duration’ encoding refers to the process by which information about the concentration of a stimulus is transduced as duration of the signal, permitting distinct cellular responses to different agonist concentrations (Behar et al., 2008). In yeast, the switch from proliferation to differentiation and/or mating is dependent both on pheromone concentration and the duration of downstream mitogen-activated protein kinase (MAPK) activity. Behar et al. proposed a model in which negative feedback regulation of the yeast pheromone receptor causes dose-dependent decreases in its affinity for ligand to allow for signaling beyond initial saturation (Behar et al., 2008).
As opposed to regulating the affinity of a single receptor, we describe a novel mechanism for dose-to-duration encoding in which the presence of multiple P2Y and P2X receptors, with different affinities for ATP, determine the duration of Ca2+-NFATc1 signaling. We demonstrated that low concentrations of ATP (1 nM to 100 µM) induce transient Ca2+ elevation and NFATc1 activation, whereas higher ATP concentrations (300 µM to 10 mM) elicit more sustained Ca2+-NFATc1 signaling.
Ca2+ signaling is a central mechanism by which P2 receptors elicit changes in cell behavior in response to extracellular nucleotides. A number of groups have shown that ATP and other nucleotides can elicit transient or sustained Ca2+ signals in a variety of other cell types (Oshimi et al., 1999; Möller et al., 2000; Nobile et al., 2003; Korcok et al., 2004). The role of the Ca2+-regulated NFAT transcription factors in P2 receptor signaling has also been examined previously (Ferrari et al., 1999; Abbott et al., 2000; Kataoka et al., 2009; Shiratori et al., 2010). However, no previous studies have employed a concentration range capable of revealing the relationship between ATP dose and Ca2+-NFAT signal duration described in the present study. Moreover, we provide the first evidence demonstrating that P2 receptors utilize the duration of Ca2+-NFAT signaling to elicit distinct responses to different concentrations of extracellular ATP.
In general, signal duration can control distinct cell functions in a variety of biological systems. In the neuronal PC12 cell line, transient activation of extracellular-signal-regulated kinase (ERK) by epidermal growth factor stimulates proliferation; whereas prolonged ERK activation, elicited by nerve growth factor, promotes differentiation (Vaudry et al., 2002). Similarly, in yeast, vegetative growth patterns are mediated by transient MAPK signaling at low pheromone levels; whereas, growth arrest and mating are initiated by sustained MAPK signaling in response to high levels of pheromone (Dohlman and Thorner, 2001). In the present study, we found that transient Ca2+-NFAT signaling elicited by low concentrations of ATP resulted in no measurable increase in NFAT transcriptional activity. Nevertheless, these Ca2+ transients regulate other important aspects of osteoblast function. For example, such Ca2+ events activate ion channels, regulate cytoskeletal remodeling and stimulate secretion in other cell types (Berridge, 2012). In this regard, low concentrations of ATP have been reported to induce functional changes in cells of the osteoblast lineage. For instance, stimulation of P2 receptors with micromolar concentrations of ATP increases the proliferation of osteoblasts and osteoblast-like cells (Nakamura et al., 2000; Katz et al., 2011). In addition, micromolar concentrations of ATP and UTP activate the osteoblast master transcription factor runt-related transcription factor 2 in the osteoblast-like HOBIT cell line (Costessi et al., 2005). In addition to these previously reported effects of low ATP concentrations, we found that high concentrations of ATP elicited sustained Ca2+-NFATc1 signaling to stimulate NFAT transcriptional activity. Thus, we demonstrate that different ATP concentrations are transduced into distinct cellular signals.
Potential physiological roles of P2 receptor networks in osteoblasts
Mechanotransduction, the process by which mechanical stimuli are translated into cellular responses, has been suggested to be mediated by nucleotide release and subsequent P2 receptor signaling in bone (Dixon and Sims, 2000; Robling et al., 2006). Phenotypic examination of knockout mouse strains for various P2Y and P2X receptor subtypes has indeed revealed a number of skeletal phenotypes, helping to solidify the importance of purinergic signaling in bone. Whole-body deletion of P2Y6, P2Y12 or P2Y13 results in decreased bone resorption, and loss of P2Y13 leads to decreased bone formation (Orriss et al., 2011a; Orriss et al., 2011b; Su et al., 2012; Wang et al., 2012). Furthermore, genetically modified mice carrying a non-functional P2X7 receptor exhibit diminished periosteal bone formation, excessive trabecular bone resorption, and impaired skeletal responses to mechanical loading (Ke et al., 2003; Li et al., 2005). The findings of the present study provide the first evidence that expression of multiple P2 receptor subtypes increases the range over which differences in ATP concentration can be sensed by a cell. This phenomenon provides a novel mechanism by which osteoblasts may transduce differences in ATP concentration and, therefore, intensity of mechanical stimuli over a remarkably wide dynamic range.
We also demonstrate for the first time that P2X7 receptors couple to the Ca2+-NFATc1 pathway in osteoblasts. Our lab has previously demonstrated that activation of P2X7 receptors by exogenous nucleotides leads to production of LPA and PGE2 by osteoblasts, culminating in enhanced differentiation and matrix mineralization (Panupinthu et al., 2008). The Ca2+-NFATc1 pathway also plays an important role in the regulation of osteoblast differentiation (Koga et al., 2005; Winslow et al., 2006), but until now the pathways underlying NFATc1 activation in osteoblasts have remained obscure. In the present study, we found that Ca2+-NFATc1 signaling stimulates expression of COX-2 downstream of the P2X7 receptor in osteoblasts. Given the important role of COX-2 and PGE2 in osteoblast differentiation and responses to mechanical stimuli (Blackwell et al., 2010), we propose that the Ca2+-NFATc1 axis may play an important role in mediating P2X7 receptor signaling during skeletal development and mechanotransduction.
The ability of ATP to elicit distinct responses over a wide dynamic range of concentrations, as described for osteoblasts in the present study, may also occur in other cell types, many of which express multiple subtypes of P2 receptors. In this regard, greatly varying amounts of ATP are released into the extracellular fluid under different circumstances – smaller amounts during neurotransmission and paracrine signaling, and massive amounts in response to trauma and cell lysis (Burnstock, 2007b). The presence of an ensemble of P2 receptors would allow osteoblasts and other cell types to respond appropriately in each of these different situations.
Materials and Methods
Materials and solutions
α-Minimum essential medium (α-MEM), heat-inactivated fetal bovine serum (FBS), antibiotic solution (10,000 U/ml penicillin, 10,000 µg/ml streptomycin, and 25 µg/ml amphotericin B), trypsin solution, Dulbecco's phosphate buffered saline (DPBS), Medium 199 buffered with HEPES (25 mM) and HCO3− (4 mM) (M199), HCO3−-free MEM, Dulbecco's modified Eagle medium (high glucose) (DMEM) and HEPES were obtained from GIBCO (Life Technologies Inc., Burlington, ON, Canada). Fluo-4 acetoxymethyl ester (fluo-4-AM), indo-1-AM and Pluronic F-127 were obtained from Molecular Probes (Life Technologies). FuGENE 6 and X-tremeGENE 9 were from Roche Diagnostics (Laval, QC, Canada). TRIzol reagent and UltraPure distilled water (DNase/RNase-free) were obtained from Invitrogen (Life Technologies). RNeasy Mini Kit was from QIAGEN (Toronto, ON, Canada). TaqMan One-Step RT-PCR Master Mix Reagents kit, COX-2 (Ptgs2) primers and probe (Mm00478374_m1) and 18S rRNA primers and probe were obtained from Applied Biosystems (Life Technologies). Passive Lysis Buffer, 5× and Bright-Glo™ Luciferase Assay System were purchased from Promega (Madison, WI, USA). Biotinylated goat anti-mouse antibody, fluorescein-conjugated streptavidin (FITC) and Vectashield mounting medium with 4,6-diamidino-2-phenylindole (DAPI) were obtained from Vector Laboratories (Burlingame, CA, USA). NFATc1 mouse monoclonal antibody was obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Normal goat serum, collagenase type II, ethylene glycol-bis(β-aminoethylether)-N,N,N′,N′-tetraacetic acid tetrasodium salt (EGTA), adenosine 5′-triphosphate disodium salt (ATP), uridine 5′-triphosphate trisodium salt hydrate (UTP), 2′-3′-O-(4-benzoylbenzoyl)adenosine 5′-triphosphate triethylammonium salt (BzATP), 1-oleoyl-sn-glycero-3-phosphate (LPA) and suramin sodium salt were obtained from Sigma-Aldrich (St. Louis, MO, USA). U 73122 was from Enzo Life Sciences (Plymouth Meeting, PA, USA). 3-[[5-(2,3-Dichlorophenyl)-1H-tetrazol-1-yl]methyl]pyridine hydrochloride (A 438079 HCl) was from Tocris Bioscience (Ellisville, MO, USA). HEPES buffer consisted of (in mM): 135 NaCl, 5 KCl, 1 MgCl2, 1 CaCl2, 20 HEPES and 10 glucose, adjusted to pH 7.30±0.02, 290±5 mosmol/l. CaCl2 was omitted where indicated.
Animals and cell culture
The P2X7 loss-of-function (knockout) mouse, generated as previously described (Solle et al., 2001), was obtained from Pfizer. Though P2X7 is present in this genetically modified mouse model, the protein is truncated at its C-terminus resulting in greatly diminished receptor function (Masin et al., 2012). Colonies of both wild-type and knockout mice were maintained in a mixed genetic background (129/Ola×C57BL/6×DBA/2) by crossbreeding of heterozygous mice. All procedures were approved by the Council on Animal Care at the University of Western Ontario and were in accordance with the guidelines of the Canadian Council on Animal Care.
Calvarial osteoblasts were isolated from 5- to 7-day-old mice using sequential collagenase digestion, as previously described (Panupinthu et al., 2008). Freshly isolated calvarial osteoblasts were plated at a density of 1.0–1.5×104 cells/cm2 on Nunc six-well plates (Thermo Fisher Scientific, Rochester, NY, USA) and maintained in α-MEM supplemented with 10% FBS and 1% antibiotic solution (culture medium) at 37°C and 5% CO2. After confluence was reached (∼3–5 days), cells were trypsinized and plated for experiments.
The MC3T3-E1 osteoblast-like cell line (subclone 4) was obtained from the American Type Culture Collection (Rockville, MD, USA). MC3T3-E1 cells were subcultured twice weekly and maintained in culture medium at 37°C and 5% CO2.
Fluorescence measurement of [Ca2+]i
For confocal microscopy, MC3T3-E1 cells or calvarial osteoblasts were plated at a density of 1.5×104 cells/cm2 on 35-mm glass-bottomed dishes (MatTek Corporation, Ashland, MA, USA) in culture medium. After 2 days, cells were placed in serum-free medium and incubated overnight. For experiments, cells were loaded with fluo-4 by incubation with fluo-4-AM (2 µg/ml) and 0.1% Pluronic F-127 for 30–45 min at 37°C and 5% CO2. Medium was then replaced with M199 supplemented with 1% antibiotic solution, and cells were observed by live-cell confocal microscopy (model LSM 510; Carl Zeiss Inc., Jena, Germany) at ∼28°C using a Plan-Apochromat 40×objective (1.2 NA) with 488-nm Ar+ ion laser excitation. The emission wavelength was filtered at 500–550 band pass and images were captured every 500 ms in time-lapse mode.
For spectrofluorimetry, MC3T3-E1 cells were loaded with indo-1 as previously described (Grol et al., 2012). For measurement of [Ca2+]i, 1 ml aliquots of indo-1-loaded cell suspensions (∼1.0×106 cells) were sedimented and resuspended in 2 ml Ca2+-containing or Ca2+-free Na+-HEPES buffer in a fluorometric cuvette at room temperature. Changes in [Ca2+]i were then monitored using a dual-wavelength spectrofluorimeter (Model RF-M2004; Photon Technology International, South Brunswick, NJ, USA) at 355 nm excitation and emission wavelengths of 405 and 485 nm.
Expression and localization of NFATc1–EGFP
The EGFP-tagged NFATc1 (NFATc1–EGFP) fusion protein expression vector was purchased from GE Healthcare (Amersham Place, UK). For live-cell studies, MC3T3-E1 cells were plated at a density of 1.5×104 cells/cm2 on 35-mm glass-bottomed dishes in culture medium. After 1 day, cells were transfected with the NFATc1–EGFP expression vector using FuGENE 6 according to manufacturer's instructions. At 1 day post-transfection, cells were placed in serum-free medium and incubated overnight. On the day of the experiment, medium was replaced with M199 supplemented with 1% antibiotic solution, and cells were observed by confocal microscopy at ∼28°C using a Zeiss Plan-Apochromat 40× objective (1.2 NA) with 488-nm Ar+ ion laser excitation. The emission wavelength was filtered at 500–550 nm band pass, and image stacks of 2-µm slices were captured every 5 min in time-lapse mode. Changes in subcellular localization were quantified by comparing the average fluorescence intensity in the nucleus (FN) to the average fluorescence intensity of an area of equal size in the cytosol (FC). Values of the ratio FN/FC were plotted as a function of time, and values for FN/FC exceeding 1 were taken to indicate nuclear localization.
For fixed-cell studies, MC3T3-E1 cells were plated at a density of 1.5×104 cells/cm2 on 12-mm glass coverslips in Falcon 24-well plates (BD Biosciences, Mississauga, ON, Canada) in culture medium. After 1 day, cells were transfected with the NFATc1–EGFP expression vector as described above. At 1 day post-transfection, cells were placed in serum-free medium and incubated overnight. On the day of the experiment, cells were incubated with test substances for the indicated times. Cells were then fixed with paraformaldehyde (4%) in sucrose solution (2%), sealed using Vectashield mounting medium with DAPI, and visualized by fluorescence microscopy. Cells were categorized as positive for nuclear localization of NFATc1–EGFP if fluorescence intensity of the nucleus exceeded that of the cytoplasm. The proportion of cells exhibiting nuclear localization was then calculated. Representative images were acquired using a Zeiss Plan-Apochromat 40×objective (1.2 NA) at a slice thickness of 2 µm with one of the following configurations: (1) 405-nm diode or 488-nm Ar+ ion laser excitation with emission wavelengths filtered at 420–480 or 505–550 nm band pass, respectively; or (2) 730-nm Chameleon multiphoton or 488-nm Ar+ ion laser excitation with emission wavelengths filtered at 390–465 or 500–550 nm band pass, respectively.
Immunofluorescence localization of native NFATc1
Calvarial osteoblasts were plated at a density of 1.5×104 cells/cm2 on 12-mm glass coverslips in Falcon 24-well plates in culture medium. After 2 days, cells were placed in serum-free medium and incubated overnight. On the day of the experiment, cells were incubated with test substances for the indicated times. Cells were then fixed with paraformaldehyde (4%) in sucrose solution (2%), permeabilized with 0.1% Triton X-100 in DPBS for 10 min, and blocked for 1 h with 1% normal goat serum in DPBS (blocking solution). To detect subcellular localization of native NFATc1, cells were incubated overnight at 4°C with a mouse monoclonal antibody (1∶100 in blocking solution). The next day, cells were incubated with a biotinylated goat anti-mouse antibody (1∶200 in blocking solution) for 2 h followed by 15 min incubation with fluorescein-conjugated streptavidin (FITC; 1∶100 in DPBS). Stained samples were then sealed using Vectashield mounting medium with DAPI, and visualized by fluorescence microscopy. Cells were categorized as positive for nuclear localization of NFATc1 if fluorescence intensity of the nucleus exceeded that of the cytoplasm. Representative images were acquired using a Zeiss Plan-Apochromat 40×objective (1.2 NA) at a slice thickness of 2 µm with the appropriate excitation wavelengths and emission filters as described above.
Real-time RT-PCR analyses
MC3T3-E1 or calvarial osteoblasts were plated at a density of 1.5×104 cells/cm2 on Falcon 6-well plates in culture medium. After 2 days, cells were placed in serum-free medium and incubated overnight. On the day of the experiment, cells were incubated with test substances for the indicated times. Total RNA was isolated using TRIzol reagent and the RNeasy Mini Kit according to manufacturer's instructions. Real-time PCR was performed using the ABI Prism 7900 HT Sequence Detector (PerkinElmer) with 15 µl final reaction volumes containing 50 ng RNA sample, TaqMan One-Step RT-PCR Master Mix Reagents, and one of Ptgs2 or 18S rRNA primers and probes. Reverse transcription was performed at 48°C for 30 min followed by 40 cycles of amplification at an annealing temperature of 60°C. Reactions for each sample were performed in triplicate. All samples were normalized to 18S rRNA, and time zero or vehicle-treated controls using the delta-delta cycle threshold (ΔΔCt) method.
Luciferase reporter assay for NFATc1
The NFAT luciferase reporter plasmid (pGL3-NFAT luciferase, plasmid 17870) was obtained from Addgene (Cambridge, MA, USA). This plasmid possesses three copies of the NFAT site cloned upstream of the minimal IL-2 promoter (Clipstone and Crabtree, 1992). MC3T3-E1 and calvarial osteoblasts were transfected in suspension with the NFAT luciferase reporter vector using FuGENE 6 or X-tremeGENE 9 according to manufacturer's instructions. Cells were subsequently plated at a density of 3.0×104 cells/cm2 on Falcon 48-well plates in culture medium. At 1 day post-transfection, cells were placed in serum-free medium and incubated overnight. On the day of the experiment, cells were treated with test substances and subsequently incubated for 24 h. Cell lysates were then prepared by incubation with 65 µl of passive lysis buffer, 1× per well at room temperature for a minimum of 30 min with agitation. To assess luminescence, 15 µl of lysate was combined with 15 µl of Bright-Glo Luciferase Reagent in a 96-well white plate (Greiner Bio-One, Monroe, NC, USA). Reactions for each sample were performed in triplicate. Luminescence was measured using a 2-second integration per well on a LMAX II384 microplate reader (Molecular Devices, Downingtown, PA, USA).
Data are shown as means±s.e.m. Differences between two groups were assessed using Student's t tests. Differences among three or more groups were evaluated by one-way analysis of variance followed by a Tukey multiple comparisons test, or two-way analysis of variance followed by a Bonferroni multiple comparisons test. Differences were accepted as statistically significant at P<0.05.
We thank Dr Souzan Armstrong, Tom Chrones, Jamie Simek and Karen Nygard for assistance with confocal imaging; Andrea Lee for help with data analysis; and Kim Beaucage for constructive comments on the manuscript.
M.W.G., S.M.S. and S.J.D. conceived and designed the experiments; M.W.G. conducted the experiments; A.P. prepared the plasmids; M.W.G. and S.J.D. interpreted the results of the experiments; M.W.G. prepared the figures; M.W.G. and S.J.D. wrote the manuscript; M.W.G., A.P., S.M.S. and S.J.D. approved the final version of the manuscript.
This work was funded by from the Canadian Institutes of Health Research (CIHR) [grant number 102542]. M.W.G. was supported by a CIHR Frederick Banting and Charles Best Canada Graduate Scholarship Doctoral Award.