Opposing Actions of Inositol 1,4,5-Trisphosphate and Ryanodine Receptors on Nuclear Factor of Activated T-cells Regulation in Smooth Muscle*

The nuclear factor of activated T-cells (NFAT), originally identified in T-cells, has since been shown to play a role in mediating Ca2+-dependent gene transcription in diverse cell types outside of the immune system. We have previously shown that nuclear accumulation of NFATc3 is induced in ileal smooth muscle by platelet-derived growth factor in a manner that depends on Ca2+ influx through L-type, voltage-dependent Ca2+ channels. Here we show that NFATc3 is also the predominant NFAT isoform expressed in cerebral artery smooth muscle and is induced to accumulate in the nucleus by UTP and other Gq/11-coupled receptor agonists. This induction is mediated by calcineurin and is dependent on sarcoplasmic reticulum Ca2+ release through inositol 1,4,5-trisphosphate receptors and extracellular Ca2+ influx through L-type, voltage-dependent Ca2+ channels. Consistent with results obtained in ileal smooth muscle, depolarization-induced Ca2+ influx fails to induce NFAT nuclear accumulation in cerebral arteries. We also provide evidence that Ca2+release by ryanodine receptors in the form of Ca2+ sparks may exert an inhibitory influence on UTP-induced NFATc3 nuclear accumulation and further suggest that UTP may act, in part, by inhibiting Ca2+ sparks. These results are consistent with a multifactorial regulation of NFAT nuclear accumulation in smooth muscle that is likely to involve several intracellular signaling pathways, including local effects of sarcoplasmic reticulum Ca2+release and effects attributable to global elevations in intracellular Ca2+.


From the Department of Pharmacology, University of Vermont, Burlington, Vermont 05405
The nuclear factor of activated T-cells (NFAT), originally identified in T-cells, has since been shown to play a role in mediating Ca 2؉ -dependent gene transcription in diverse cell types outside of the immune system. We have previously shown that nuclear accumulation of NFATc3 is induced in ileal smooth muscle by plateletderived growth factor in a manner that depends on Ca 2؉ influx through L-type, voltage-dependent Ca 2؉ channels. Here we show that NFATc3 is also the predominant NFAT isoform expressed in cerebral artery smooth muscle and is induced to accumulate in the nucleus by UTP and other G q/11 -coupled receptor agonists. This induction is mediated by calcineurin and is dependent on sarcoplasmic reticulum Ca 2؉ release through inositol 1,4,5-trisphosphate receptors and extracellular Ca 2؉ influx through L-type, voltage-dependent Ca 2؉ channels. Consistent with results obtained in ileal smooth muscle, depolarization-induced Ca 2؉ influx fails to induce NFAT nuclear accumulation in cerebral arteries. We also provide evidence that Ca 2؉ release by ryanodine receptors in the form of Ca 2؉ sparks may exert an inhibitory influence on UTP-induced NFATc3 nuclear accumulation and further suggest that UTP may act, in part, by inhibiting Ca 2؉ sparks. These results are consistent with a multifactorial regulation of NFAT nuclear accumulation in smooth muscle that is likely to involve several intracellular signaling pathways, including local effects of sarcoplasmic reticulum Ca 2؉ release and effects attributable to global elevations in intracellular Ca 2؉ .
Nuclear factor of activated T-cells (NFAT) 1 was originally identified as the transcription factor responsible for mediating the Ca 2ϩ -dependent transcription of genes involved in T-cell activation (1,2) but has since been shown to play a role in mediating Ca 2ϩ -dependent gene transcription in diverse cell types outside of the immune system, including neurons (3), endothelial cells (4,5), cardiac muscle (6), skeletal muscle (7,8), and smooth muscle (9,10). The potential physiological roles for this transcription factor are also diverse and include the developmental regulation of slow twitch/fast twitch skeletal muscle fiber types (11) and smooth muscle cell precursor migration and vascular development during embryogenesis (12). NFAT has also been implicated in the pathogenesis of cardiac (6) and skeletal (7,8) muscle hypertrophy and might be predicted to play a similar role in smooth muscle hypertrophy associated with, for example, atherosclerosis and bladder dysfunction.
NFAT represents a family of transcription factors composed of four well characterized members, designated NFATc1 (NFAT2/c), NFATc2 (NFAT1/p), NFATc3 (NFAT4/x), and NFATc4 (NFAT3). A fifth putative member of the family (NFAT5) is a calcineurin-insensitive, constitutively nuclear phosphoprotein that has limited sequence similarity to other members of the NFAT family (13).
NFAT activation is regulated primarily through control of its subcellular localization (2). Elevation of global Ca 2ϩ produced by a variety of mechanisms activates the Ca 2ϩ /calmodulin-dependent protein phosphatase, calcineurin (14). Subsequent calcineurin-mediated dephosphorylation of specific NFAT serine residues induces a conformational change in the NFAT molecule that exposes nuclear localization signals, allowing import of NFAT into the nucleus (15)(16)(17)(18)(19)(20). The distribution of NFAT between nuclear and cytoplasmic compartments is dynamically regulated by the activity of nuclear kinases, which oppose the action of calcineurin (3, 28 -31). Additional mechanisms, which modulate the ability of calcineurin to associate with and/or dephosphorylate NFAT in the cytosol, provide regulation at the level of nuclear import (17,(32)(33)(34). In the nucleus, NFAT associates with a transcriptional co-activator, an interaction that is required for significant NFAT-mediated transcriptional activity. NFAT family members have been shown to cooperatively bind to DNA with variety of cofactors, including AP-1 (24 -26), GATA (6,8,27), and MEF2 (11), and in this way integrate Ca 2ϩ /calcineurin signaling with other signaling pathways, such as Ras, Rac, and protein kinase C.
A sustained, global increase in intracellular Ca 2ϩ has generally been considered a defining feature of NFAT-activating stimuli (35)(36)(37)(38). More recent evidence has shown that temporal modulation of Ca 2ϩ signals in the form of Ca 2ϩ oscillations or waves increases the coupling efficiency of Ca 2ϩ signals to NFAT activation in nonexcitable cells (39,40). Surprisingly, transient increases in intracellular Ca 2ϩ induced by a depolar-izing stimulus and mediated by flux through L-type voltagedependent Ca 2ϩ channels (VDCC), have also been shown to effectively stimulate a sustained increase in NFAT activity in hippocampal neurons (3). In contrast, and counter to expectations, sustained increases in intracellular Ca 2ϩ induced by depolarization fail to induce nuclear accumulation of the NFATc3 isoform in ileal smooth muscle, although PDGF, which activates multiple intracellular pathways, is a potent stimulus for NFATc3 nuclear translocation in this tissue (10).
Smooth muscle exhibits a diverse array of Ca 2ϩ signals, including Ca 2ϩ waves that traverse the length of the cell and display distinctive frequency and amplitude properties (41)(42)(43)(44), and localized transient releases of Ca 2ϩ through sarcoplasmic reticulum (SR) ryanodine receptors (RyRs) in the form of Ca 2ϩ sparks (45). The role that various Ca 2ϩ signals and Ca 2ϩdependent transcription factors may play in the physiological or pathological regulation of gene expression in this phenotypically plastic tissue is largely unknown. Here we show that, in native cerebral artery smooth muscle cells, UTP and other G q/11 -coupled receptor agonists induce a calcineurin-mediated nuclear accumulation of NFATc3. This induction is dependent on SR Ca 2ϩ release through IP 3 receptors (IP 3 R) and further depends on extracellular Ca 2ϩ influx through L-type VDCC. We also show that UTP may act to induce NFATc3 nuclear accumulation, at least in part, by suppressing Ca 2ϩ sparks, suggesting a novel inhibitory role for Ca 2ϩ sparks in the regulation of Ca 2ϩ -sensitive transcription factors.

EXPERIMENTAL PROCEDURES
Chemicals-Epidermal growth factor (EGF) and platelet-derived growth factor BB (PDGF-BB) were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY), pinacidil was from RBI (Research Biochemical Inc.), fluo-4 and pluronic acid were from Molecular Probes, Inc. (Eugene, OR), ryanodine was from LC Laboratories, xestospongin C was from Calbiochem, and 2-aminoethoxydiphenyl borate (2-APB) was from Tocris. FK506 was kindly provided by Fujisawa. All other drugs and chemical reagents were from Sigma.
Tissue Samples-Adult female CD-1 mice (20 -25 g; Charles River Laboratories) were euthanized by peritoneal injection of pentobarbital solution (200 mg/kg). Cerebral arteries were dissected from the brain in ice-cold physiological saline solution (containing 135 mmol/liter NaCl, 5.9 mmol/liter KCl, 1.2 mmol/liter MgCl 2 , 11.6 mmol/liter Hepes, 11.5 mmol/liter glucose, pH 7.4) and cleaned of connective tissue. Endothelium-denuded vessels were prepared by passing an air bubble through the lumen of the artery.
Western Blot Analysis-Cerebral arteries (midcerebral, posterior, cerebellar, and basilar) from 2-4 mice were pooled and homogenized in sample preparation buffer (50 mM Tris-Cl, pH 6.8, 2% SDS, 100 mM dithiothreitol, 10% glycerol, and 0.1% bromphenol blue). Aliquots of cerebral artery and control tissue extracts, prepared in a similar manner, were separated by SDS-PAGE on 8% gels using the Laemmli buffering system. Proteins were transferred to immunoblot polyvinylidene difluoride membranes (Bio-Rad) and blocked by rocking for 1 h at room temperature in blocking buffer (Tris-buffered saline with 0.1% Tween 20 (TBST) and 5% nonfat dry milk). Blots were exposed to primary antibodies for 1 h, multiply washed with TBST, and treated with horseradish peroxidase-conjugated secondary antibody for 45 min, followed by a final series of washes with TBST. Primary (rabbit polyclonal anti-NFATc3; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and secondary (donkey anti-rabbit IgG; Santa Cruz Biotechnology) antibodies were prepared in blocking buffer, and all steps were performed at room temperature. Blots were developed using an enhanced chemiluminescence substrate (SuperSignal West Dura; Pierce) according to the manufacturer's instructions.
Immunofluorescence-Arteries were treated at room temperature as specified and then mounted onto glass slides. After air-drying for 5 min, the arteries were fixed with 4% formaldehyde in phosphate-buffered saline (pH 7.4) for 15 min, permeabilized with 0.2% Triton-X-100 in phosphate-buffered saline for 10 min, and blocked for 2 h with 2% bovine serum albumin in phosphate-buffered saline. Primary antibody (rabbit anti-NFATc3 (Santa Cruz Biotechnology) diluted 1:250 in 2% bovine serum albumin/phosphate-buffered saline) was applied overnight at 4°C. Secondary antibody (Cy5 anti-rabbit IgG (Jackson Immuno Research Laboratories), 1:500 dilution) was applied for 1 h at 25°C. Nuclei were identified using the fluorescent nucleic acid dye YOYO-1 (1:30,000 dilution). After washing, the vessels were mounted (Aqua Polymount mounting medium; Polysciences) and examined at ϫ40 magnification using a Bio-Rad 1000 laser-scanning confocal microscope. NFATc3 was detected by monitoring Cy5 fluorescence using an excitation wavelength of 650 nm and an emission wavelength of 670 nm. Specificity of immune staining was confirmed by the absence of fluorescence in arteries incubated with primary or secondary antibodies alone. For scoring of NFATc3-positive nuclei, multiple fields for each vessel were imaged and counted by two independent observers under double-blind conditions. For quantification, a cell was considered positive if co-localization (yellow) was observed in the nucleus, whereas a cell was considered negative if no co-localization (green only) was visualized.
Confocal Ca 2ϩ Measurements-All imaging experiments were performed at room temperature. Arteries were loaded with 10 M fluo-4-AM in physiological saline solution and 0.05% pluronic acid for 60 min and then kept in physiological saline solution for 30 min to allow fluo-4 de-esterification. The vessel ends were anchored beneath two stainless steel hooks to maintain the artery at the bottom of the chamber and provide a fixed imaging area. Arteries were illuminated with a krypton/ argon laser at 488 nm and imaged using a Noran Oz laser-scanning confocal microscope and a ϫ60 water immersion lens (numerical aperture ϭ 1.2). For Ca 2ϩ sparks detection, images of 58.1 ϫ 54.5-m (256 ϫ 240 pixels) sections of the vessel wall were recorded every 16.7 ms (60 images/s). Under each condition, at least two different representative areas of the same artery were scanned for 10 s. Ca 2ϩ sparks were analyzed using custom software written in our laboratory by Dr. Adrian Bonev (using IDL 5.2; Research Systems Inc., Boulder, CO), which allows for off-line quantification of fluorescence changes in selected regions of a sample corresponding to boxes of defined dimensions positioned by eye within the sample. Ca 2ϩ spark amplitude (F/F 0 ) was obtained by determining the fluorescence intensity within a 2.37-m 2 (1.54 m (7 pixels) ϫ 1.54 m (7 pixels)) area corresponding to a detected spark event (F), and dividing by a base line (F 0 ) that was determined by averaging 30 images without Ca 2ϩ spark activity. Ca 2ϩ spark frequency under a given condition was calculated by measuring the number of sparks that occurred in a 58.1 ϫ 54.5-m area (ϳ20 cells) scanned for 10 s. For detection of Ca 2ϩ waves, images of the vessel wall (116.2 ϫ 108.0 m, or 512 ϫ 480 pixels) were recorded every 531.9 ms (1.88 images/s). Under each condition, at least two different representative areas of the same artery were scanned for 213 s. Ca 2ϩ waves were determined by analyzing recurrent changes in fluorescence intensity occurring in 2.2 ϫ 2.2-m (10 ϫ 10 pixels) regions of individual myocytes and defined as a change in F/F 0 Ͼ 1.3 that remained elevated for Ͼ200 ms. Changes in global F/F 0 were calculated by measuring the mean pixel value of images acquired at 1.88 images/s, before and after application of each drug. The same area was not scanned more than once to avoid introducing Ca 2ϩ signaling artifacts due to laser-induced cell damage.
Statistical Analysis-Results are expressed as means Ϯ S.E., where applicable. All statistical analysis was performed using GraphPad software (Prism 3.0). Statistical significance was determined using one-way analysis of variance analysis followed by Bonferroni or Tukey-Kramer tests (for comparisons between up to five groups or at least six groups, respectively).

NFAT Isoform Expression in Vascular Smooth Muscle-We
have used RT-PCR analysis and immunoblotting to identify NFAT isoforms expressed in native cerebral artery smooth muscle. Previous results from experiments employing the rat A7r5 aortic smooth muscle cell line suggested that the NFATc2 and NFATc1 isoforms are expressed in smooth muscle (9), although expression of additional isoforms could not be ruled out in this study. In an RT-PCR analysis of total RNA isolated from cerebral arteries, we find no evidence for NFATc2 expression using primer pairs that efficiently amplify NFATc2 from spleen (Fig. 1A). Instead, we find that NFATc3 and, to a lesser extent, NFATc4, are expressed in cerebral arteries (Fig. 1B). NFATc1 expression is generally very low to undetectable in unstimulated mouse cerebral arteries (Fig. 1B). The presence of NFATc3 protein in cerebral arteries was confirmed by Western analysis, which showed the presence of major bands corresponding to those identified in thymus (Fig. 1C), a tissue that expresses predominantly the NFATc3 isoform (46). The predominant expression of NFATc3 in smooth muscle has led us to focus on this isoform, although it is likely that other NFAT isoforms may play important roles in smooth muscle, as suggested by others (9,12,47).
Induction of NFATc3 Nuclear Accumulation by G q/11 -coupled Vasoconstrictor Agonists-UTP is an important vasoactive substance in the cerebral vasculature and has been previously shown to activate the NFATc1 isoform in cultured smooth muscle cells (47). In intact cerebral arteries, treatment with UTP induces nuclear translocation of NFATc3 as evidenced by colocalization of NFATc3 with the fluorescent nucleic acid dye, YOYO-1 ( Fig. 2A). These results are summarized in Fig. 2B, which shows that the number of NFATc3-positive nuclei in intact cerebral arteries is increased from 7.9% in untreated vessels to 43.2% in UTP-treated arteries. Similar results for control (5.8%) and UTP-treated conditions (50.7%) were obtained for endothelium-denuded arteries, indicating that this action of UTP on NFATc3 subcellular distribution is a direct effect on smooth muscle.
Other G q/11 -coupled vasoconstrictors and peptide ligands for certain tyrosine kinase-linked growth factor receptors are ca-pable of inducing NFATc3 translocation in cerebral artery smooth muscle, although the robustness of the response varies with the agonist used. Endothelin-1 is as effective as UTP in inducing NFATc3 nuclear accumulation (Fig. 2C), whereas angiotensin II and the peptide ligand EGF are much less effective. Prostaglandin F2␣ induces a small, but significant, increase in NFAT nuclear accumulation that is comparable in magnitude with that induced by angiotensin II and EGF. PDGF, which is a smooth muscle mitogen and potent activator of NFATc3 nuclear accumulation in ileal smooth muscle (10), is ineffective in cerebral artery smooth muscle.
Calcineurin Dependence of UTP-induced NFATc3 Nuclear Accumulation-Calcineurin activity is sensitive to inhibition by the chemically unrelated immunosuppressive agents FK506 and cyclosporin A, which inhibit calcineurin by distinct mechanisms (48). To determine whether UTP-induced NFATc3 nuclear accumulation is calcineurin-dependent, we treated intact cerebral arteries with each of these agents prior to and/or concurrent with UTP treatment. Inhibition of calcineurin activity with either of these compounds completely abrogates UTP-induced NFATc3 nuclear accumulation (Fig. 3), indicating that this process is calcineurin-dependent.
Role of Intracellular Ca 2ϩ Stores in UTP-induced NFATc3 Nuclear Accumulation-Calcineurin activity is strictly dependent on Ca 2ϩ /calmodulin (49). In intact cerebral arteries, UTP induces an increase in global intracellular Ca 2ϩ characterized by an initial Ca 2ϩ spike followed by a sustained elevated plateau phase ( Fig. 4A; see also Ref. 41). Although the magnitude of each phase is somewhat variable between and within vessels, this biphasic response is a consistent property of UTPinduced Ca 2ϩ elevation. To determine whether SR-mediated Ca 2ϩ release is involved in UTP-induced NFATc3 nuclear accumulation, we pretreated cerebral arteries with the SR Ca 2ϩ -ATPase inhibitor, thapsigargin, to deplete SR luminal Ca 2ϩ . This treatment prevents the increase in global Ca 2ϩ induced by UTP (Fig. 4B), indicating that intracellular calcium stores are required for this effect. To determine whether this calcium release from the SR induced by UTP contributes to UTP-induced NFATc3 nuclear accumulation, NFATc3 subcellular distribution was monitored immunohistochemically in cerebral arteries pretreated with thapsigargin. In these experiments, we employed cerebral arteries that had first been denuded of endothelium, as described under "Experimental Procedures." In endothelium-denuded cerebral arteries, prior depletion of Ca 2ϩ stores completely prevents UTP-induced NFATc3 nuclear accumulation (Fig. 4C). Under these conditions, thapsigargin alone has no effect on NFATc3 subcellular distribution. These data indicate that UTP mediates its effects on NFATc3 subcellular distribution, at least in part, through release of SR Ca 2ϩ .
Role of Extracellular Ca 2ϩ Influx through L-type VDCC in UTP-induced NFATc3 Nuclear Accumulation-In nonexcitable cells, the sustained increase in intracellular Ca 2ϩ required to maintain NFAT in the nucleus is provided by a capacitative mechanism by which depletion of intracellular Ca 2ϩ stores is coupled to extracellular Ca 2ϩ influx (50,51). In smooth muscle, where a potential role for capacitative Ca 2ϩ entry pathways remains speculative (55), the principle mediator of extracellular Ca 2ϩ influx is the L-type VDCC (56).
To determine whether influx of Ca 2ϩ through VDCC is required for UTP-induced NFATc3 nuclear accumulation, we employed two complementary approaches. First, we inhibited VDCC directly using the dihydropyridine, nisoldipine, which potently and specifically blocks these channels (57). We also indirectly decreased VDCC activity using pinacidil, an agent that activates ATP-sensitive K ϩ channels, resulting in membrane potential hyperpolarization and diminished VDCC activ- ity (58). Treatment with either nisoldipine or pinacidil completely abrogates UTP-induced NFATc3 accumulation in intact cerebral arteries (Fig. 5, A and B), indicating that Ca 2ϩ influx through VDCC is required for UTP-induced NFATc3 nuclear accumulation in native arterial smooth muscle.
We have previously shown that extracellular Ca 2ϩ influx through VDCC induced by depolarizing concentrations of potassium is, alone, an insufficient stimulus for induction of NFATc3 nuclear accumulation in ileal smooth muscle (10). In intact cerebral arteries, high K ϩ also fails to induce a significant change in NFATc3 subcellular distribution (Fig. 5C). Treatment with the Ca 2ϩ ionophore, ionomycin, is also an ineffective stimulus, as is concurrent treatment with the phorbol ester phorbol 12-myristate 13-acetate (Fig. 5C) and ionomycin or high K ϩ . These latter results suggest that the effectiveness of UTP and other G q/11 -coupled receptor agonists is not simply the result of simultaneous elevation of intracellular Ca 2ϩ and activation of protein kinase C.
Differential Effects of IP 3 R-and RyR-mediated Intracellular Ca 2ϩ Release-Stimulation of phospholipase C-mediated hydrolysis of phosphoinositol bisphosphate to yield IP 3 and diacylglycerol constitutes a common mechanism of action of G q/11coupled vasoconstrictor agonists. The subsequent release of SR Ca 2ϩ through IP 3 Rs increases global [Ca 2ϩ ] i and contributes to the Ca 2ϩ required for contraction. To confirm a role for IP 3 Rmediated Ca 2ϩ release in UTP-induced NFATc3 nuclear accu-mulation, we employed the structurally dissimilar, cell-permeable IP 3 R inhibitors, 2-APB and xestospongin C (59,60). Treatment with either of these agents completely prevents the nuclear accumulation of NFATc3 induced by treatment with UTP (Fig. 6A). Taken together with the results from SR Ca 2ϩ depletion, these data indicate that induction of NFATc3 nuclear accumulation by UTP is dependent on IP 3 R-mediated release of Ca 2ϩ from smooth muscle SR.
In addition to IP 3 Rs, the SR membrane also contains a ryanodine-sensitive class of Ca 2ϩ -activated Ca 2ϩ -release channels, known as RyRs. These channels play an important role in regulating vascular tone through transient release of Ca 2ϩ in the form of Ca 2ϩ sparks, which create high local concentrations of Ca 2ϩ capable of activating closely juxtaposed, large conductance, Ca 2ϩ -activated K ϩ (BK) channels (45). The resulting increase in K ϩ conductance promotes membrane hyperpolarization, thus reducing the activity of voltage-gated Ca 2ϩ channels and opposing vasoconstriction. In contrast to the inhibitory effect that blockers of IP 3 Rs exert on NFATc3 nuclear accumulation, we found that inhibition of RyR with ryanodine potentiates UTP-induced NFATc3 nuclear translocation (Fig.  6B). Since blocking RyR would be expected to lead to membrane depolarization due to diminished BK channel activity (45), it is conceivable that the activity of Ca 2ϩ /calmodulin-dependent calcineurin might be enhanced by virtue of an increased Ca 2ϩ flux through VDCC. In experiments in which the membrane potential was ionically clamped with 60 mM potassium, ryanodine exerted a similar potentiating effect on UTPinduced NFATc3 nuclear accumulation, strongly suggesting that membrane potential depolarization attributable to RyR inhibition is not the primary mechanism by which ryanodine exerts its enhancing effect on NFAT3 nuclear accumulation. To directly confirm that increased Ca 2ϩ influx associated with membrane depolarization does not augment UTP-induced NFATc3 nuclear accumulation, we treated cerebral arteries with UTP and high K ϩ . Membrane depolarization with high K ϩ had no significant effect on UTP-induced NFATc3 translocation (Fig. 6C). Ryanodine alone and ryanodine plus high K ϩ were also without effect (Fig. 6C).
The absence of a membrane potential component to the potentiating action of RyR blockade suggests that Ca 2ϩ flux through RyRs in the form of Ca 2ϩ sparks, acting independently of BK channel activation, may have an inhibitory effect on UTP-induced NFATc3 nuclear accumulation. Accordingly, increasing spark frequency in the presence of UTP would be predicted to exert an inhibitory effect on UTP-induced NFATc3 accumulation. To test this hypothesis, we pretreated cerebral arteries with 300 M caffeine, a concentration that induces a reproducible increase in Ca 2ϩ spark frequency ( Fig. 7B; see also Refs. 61 and 62). This caffeine-induced increase in spark frequency is accompanied by a complete abrogation of UTPinduced NFATc3 nuclear translocation (Fig. 7C). This postulated inhibitory effect of the spark pathway may also help to explain the failure of high K ϩ alone to stimulate NFATc3 nuclear accumulation (see Fig. 5C), since, in addition to increasing global intracellular Ca 2ϩ , high K ϩ increases SR Ca 2ϩ load, resulting in an increase in Ca 2ϩ release through RyRs that is reflected in a substantial increase in spark frequency (Fig. 7B). Finally, treatment with UTP alone induces a profound decrease in spark frequency with a time course that closely parallels the nuclear accumulation of NFATc3 (Fig. 7, A  and D). These data suggest that inhibition of Ca 2ϩ sparks may constitute an important component of the mechanism of action of UTP with respect to induction of NFATc3 nuclear accumulation.
Ca 2ϩ Signal Modulation and UTP-induced NFATc3 Nuclear Accumulation-In addition to mediating global increases in intracellular Ca 2ϩ , vasoconstrictor agonists, including UTP, are capable of inducing Ca 2ϩ waves in smooth muscle that are mediated by IP 3 Rs and/or RyRs (41,42,63,64). These recurrent Ca 2ϩ waves give rise to oscillatory Ca 2ϩ signals similar to those that have been shown to increase the efficiency of NFAT activation in nonexcitable cells (39,40).
To address the possible contribution of Ca 2ϩ waves to UTPinduced NFATc3 nuclear accumulation, we sought initially to define the Ca 2ϩ signaling properties associated with treatment of mouse cerebral arteries with UTP. Spatial and temporal changes in intracellular Ca 2ϩ were determined in intact cere- bral arteries loaded with the Ca 2ϩ -binding indicator dye, fluo-4, using confocal microscopy as described under "Experimental Procedures." After application of UTP (10 M), mouse cerebral arteries exhibited a complex pattern of repetitive Ca 2ϩ signals (Fig. 8A). The frequency and amplitude of these events ranged from 0.3 to 2.2 waves/min and 0.3-5.5 F/F 0 , respectively, and ranged in duration from 4 to 32 s (129 waves analyzed from six arteries). UTP treatment increased the percentage of cells exhibiting Ca 2ϩ waves and wavelike signals ϳ8-fold, from 5.8% under control conditions to 47.8% following UTP exposure (Fig. 8B). This effect of UTP on wave probability is similar, in general, to that observed in rat cerebral arteries (41), although the wave properties are less uniform in mouse arteries and include propagating Ca 2ϩ features with low frequencies and very large amplitudes that have not been previously described (Fig. 8C). As expected, pretreatment with the IP 3 R inhibitors, xestospongin C and 2-APB, prevented the induction of Ca 2ϩ waves by UTP (no waves detected in n ϭ 3-5 arteries for each condition). Similarly, pretreatment with the RyR inhibitor, ryanodine, prevented UTP-induced waves and also abolished on-going Ca 2ϩ waves when added after UTP (no waves detected in n ϭ 3 arteries).
Whether Ca 2ϩ wave and/or wave-like features contribute to the efficacy of UTP-induced NFATc3 translocation is uncertain, since IP 3 R and RyR inhibitors, which abolish UTP-induced Ca 2ϩ waves, have opposite effects on NFATc3 nuclear accumulation (see Fig. 6, A and B). Although UTP-induced Ca 2ϩ waves are abolished by ryanodine, UTP-induced [Ca 2ϩ ] i elevation is retained in the presence of ryanodine (Fig. 8D). It is, therefore, likely that these Ca 2ϩ features that remain in the presence of ryanodine encode information that is sufficient to induce NFATc3 nuclear accumulation in cerebral artery smooth muscle. DISCUSSION We have shown that UTP and a number of other G q/11coupled vasoconstrictor agonists effectively induce the nuclear accumulation of NFATc3 in cerebral artery smooth muscle. UTP-induced NFATc3 nuclear accumulation is blocked by the chemically unrelated compounds cyclosporin A and FK506, which inhibit calcineurin activity by distinct mechanisms, indicating that this action of UTP is dependent on calcineurin. We have also found that release of Ca 2ϩ from intracellular stores and influx of extracellular Ca 2ϩ are both required for the induction of NFATc3 nuclear accumulation by UTP, since inhibition of either pathway completely abrogates UTP-induced NFATc3 nuclear accumulation. A specific role for Ca 2ϩ release through IP 3 Rs is strongly suggested, based on the inhibition of UTP action by the cell-permeable IP 3 R inhibitors, xestospongin C and 2-APB.
Surprisingly, it appears that release of SR Ca 2ϩ through RyR exerts an inhibitory effect on NFATc3 nuclear accumulation, since blocking these receptors with ryanodine potentiates UTPinduced NFATc3 nuclear translocation. The potentiating effect of RyR inhibition, which does not appear to reflect ryanodineinduced membrane depolarization, is consistent with experi-  ). B, summary data, quantified as described under "Experimental Procedures," from multiple repeats of experiments described for A. C, summary data from multiple repeats of experiments employing depolarizing stimuli (60 mM K ϩ ) and the Ca 2ϩ ionophore, ionomycin (1 M) for 30 min (room temperature) in the presence or absence of the protein kinase C activator, phorbol 12-myristate 13-acetate (PMA) (100 nM), depicting the failure of these treatments to induce NFATc3 nuclear accumulation in cerebral arteries (m, number of mice; i, analyzed images; c, total number of cells counted).
ments showing that increasing Ca 2ϩ spark frequency with micromolar caffeine inhibits UTP-induced NFATc3 nuclear accumulation, as well as data indicating that UTP profoundly decreases spark frequency. Collectively, these data provide strong support for the idea that Ca 2ϩ sparks may exert a novel inhibitory effect on NFATc3 nuclear accumulation in addition to their well characterized role in providing negative feedback regulation of vascular tone through activation of BK channels (45).
The role of Ca 2ϩ waves in the action of UTP on NFATc3 subcellular localization is less clear. UTP induces a global elevation in [Ca 2ϩ ] i as well as a striking increase in the frequency of smooth muscle cells exhibiting Ca 2ϩ waves or wavelike features. Ca 2ϩ waves are though to be due to IP 3 R-mediated Ca 2ϩ release (42,63,64) and may further depend on Ca 2ϩ -dependent Ca 2ϩ release through ryanodine receptors (41). Therefore, it was not unexpected that inhibitors of IP 3 R or RyR prevented Ca 2ϩ waves. However, treatments that inhibit RyR-potentiated UTP-induced NFATc3 nuclear accumulation, unlike inhibitors of IP 3 R, which had a profound inhibitory effect on NFATc3 nuclear accumulation. Since inhibitors of IP 3 R and RyR block waves but have opposite effects on NFATc3 nuclear accumulation, it is unlikely that Ca 2ϩ waves are required, although it is conceivable that they may contribute to the efficacy of UTP. In contrast, the UTP-induced IP 3 R-mediated global Ca 2ϩ transient, which was not inhibited by ryanodine (Fig. 8D), is probably necessary and sufficient for NFATc3 nuclear accumulation in cerebral artery smooth muscle cells.
The differential effects of SR Ca 2ϩ release through IP 3 R and RyR suggests that local Ca 2ϩ action at release sites may be an important component of the observed effects. In this view, IP 3 R-mediated Ca 2ϩ release is coupled to cellular elements that promote calcineurin activation and NFAT dephosphorylation, whereas RyR-mediated Ca 2ϩ release may act through associated or closely apposed molecular components to negatively regulate NFAT import or promote its nuclear export. Alternatively, it is possible that inhibition of RyR-mediated Ca 2ϩ efflux may simply serve to increase Ca 2ϩ efflux through IP 3 Rs by increasing SR Ca 2ϩ load, thus amplifying IP 3 R-mediated effects, rather than acting directly to regulate NFAT import/export. Consistent with results obtained in ileal smooth muscle, depolarization-induced increases in global [Ca 2ϩ ] i fail to induce NFATc3 nuclear accumulation in cerebral arteries. This result is in striking contrast to the results obtained in neurons, where even brief depolarization with 90 mM K ϩ induces a sustained nuclear accumulation of NFAT (3). The observed differences between the responses of these two excitable cell types to this common stimulus may reflect significant tissue-specific differences in the role and regulation of NFAT. An alternative pos-sibility is that NFAT isoform-specific regulatory features are important, since neurons express predominantly NFATc4 rather than NFATc3.
In addition to highlighting potentially important tissue-specific differences in NFAT regulation, our results provide the first evidence that two distinct intracellular Ca 2ϩ release mechanisms, represented by IP 3 receptors and ryanodine receptors, can have opposing regulatory effects on a Ca 2ϩ -dependent transcription factor.