Sustained Calcium Entry through P2X Nucleotide Receptor Channels in Human Airway Epithelial Cells*

Purinergic receptor stimulation has potential therapeutic effects for cystic fibrosis (CF). Thus, we explored roles for P2Y and P2X receptors in stably increasing [Ca 2 (cid:1) ] i in human CF (IB3-1) and non-CF (16HBE14o (cid:2) ) airway epithelial cells. Cytosolic Ca 2 (cid:1) was measured by fluorospectrometry using the fluorescent dye Fura-2/ AM. Expression of P2X receptor (P2XR) subtypes was assessed by immunoblotting and biotinylation. In IB3-1 cells, ATP and other P2Y agonists caused only a transient increase in [Ca 2 (cid:1) ] i derived from intracellular stores in a Na (cid:1) -rich environment. In contrast, ATP induced an increase in [Ca 2 (cid:1) ] i that had transient and sustained components in a Na (cid:1) -free medium; the sustained plateau was potentiated by zinc or increasing extracellular pH. Benzoyl-benzoyl-ATP, a P2XR-selective agonist, increased [Ca 2 (cid:1) ] i

In cystic fibrosis (CF), 1 cyclic AMP-and protein kinase A-dependent transepithelial Cl Ϫ transport is impaired because of mutations in the CF gene that encodes for the protein, the "cystic fibrosis transmembrane conductance regulator" or CFTR (1). Originally, CFTR was thought to function exclusively as a low conductance Cl Ϫ channel (2,3). More recently, it has become clear that CFTR also regulates a series of other transporters and ion channels, such as the Cl Ϫ /HCO 3 Ϫ exchanger, the Na ϩ :HCO 3 Ϫ cotransporter, epithelial Na ϩ channels, K ϩ channels, and aquaporin water channels (4,5). Although the exact mechanisms of the regulation of these proteins by CFTR are not yet fully understood, it is clear that impaired Cl Ϫ transport is shared as a key disease phenotype by CF epithelia from all affected tissues and that this pathway is lost in CF. Therefore, activation of a cAMP-independent Cl Ϫ secretory pathway through exploitation of a naturally expressed epithelial protein could be of interest for CF therapy. In certain cases, stimulation of Ca 2ϩ -dependent Cl Ϫ channels can correct the impaired HCO 3 Ϫ secretion in CF cells (6,7). It is widely accepted that CFTR plays a crucial role in ATP release from cells (8 -10). The same is true for mdr ABC transporters in hepatocytes and heterologous cells (11,12). Once ATP is released into the extracellular space, it can bind to purinoceptors regulating a variety of functions in different epithelia (13)(14)(15). ATP and other agonists of purinoceptors are known to increase intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) potently in airway epithelial cells which, in turn, leads to stimulation of Cl Ϫ secretion (14 -17) and inhibition of Na ϩ absorption (18 -22). In fact, earlier studies have proposed the use of UTP and non-hydrolyzable UTP analogs as therapeutic agonists targeted to the P2Y 2 receptors in the treatment of CF lung disease (23,24).
Purinoceptors are divided into two classes: P1 or adenosine receptors, and P2, which recognize primarily extracellular ATP, ADP, UTP, and UDP. The P2 receptors are further subdivided into two subclasses. P2X receptors are extracellular ATP-gated calcium-permeable non-selective cation channels that are modulated by extracellular Ca 2ϩ , Mg 2ϩ , H ϩ , and metal ions such as Zn 2ϩ and/or Cu 2ϩ (25). P2Y receptors couple to heterotrimeric G proteins and phospholipases (primarily phospholipase C␤) to raise intracellular free calcium concentration (26). In CF epithelial cells from multiple tissues, expression of P2X and P2Y receptors appears unaffected, offering the possibility to increase [Ca 2ϩ ] i through targeting a naturally expressed receptor in the apical or basolateral membrane domains (27,28). Nonetheless, in different CF epithelial cell models, the desensitization of P2Y receptors and the transient nature of the Ca 2ϩ response upon chronic and repeated delivery of a P2Y-specific agonist have made it difficult to generate stable stimuli for ion secretion (7,29).
In this study, we used both CF (IB3-1) (30) and non-CF (16HBE14o Ϫ ) (31) human airway epithelial cell models, to dissect out P2X-specific and P2Y-specific mechanisms of trigger-ing an increase in [Ca 2ϩ ] i . We characterized a broad range of P2Y-selective, P2X-selective, and non-discriminant P2Y and P2X agonists under different chemical and ionic conditions to explore possible strategies to elicit an increase in [Ca 2ϩ ] i that is sustained and prolonged. Results described herein, using Fura-2/AM-based imaging, show that activation of P2Y and P2X receptors increases [Ca 2ϩ ] i by completely distinct mechanisms. P2Y receptors elicit a transient increase in [Ca 2ϩ ] i derived from intracellular endoplasmic reticulum (ER) stores, whereas P2X receptors trigger a sustained rise in [Ca 2ϩ ] i , allowing Ca 2ϩ influx from the extracellular space. In addition, biochemical evidence shows that the P2X 4 receptor is the major epithelial subtype present in both cell lines. Thus, we conclude that epithelial P2X receptors function as ATP-gated Ca 2ϩ entry channels in the plasma membrane and have profound potential as a target for CF pharmacotherapy.

MATERIALS AND METHODS
Cell Cultures-IB3-1 cells derive from airway epithelia of a CF patient carrying two different mutations of the CFTR gene, the most common trafficking mutation (⌬Phe-508) and a premature stop codon mutation (W1282X) (30). 16HBE14o Ϫ cells are non-CF or normal airway epithelial cells, which express CFTR at the plasma membrane. The cells were grown on Vitrogen 100-coated tissue-culture flasks in 5% CO 2 incubator at 37°C. IB3-1 cells were cultured in LHC-8 (Biofluids, Rockville, MD) medium supplemented with 5% fetal bovine serum (Invitrogen), 100 units/ml penicillin/streptomycin (Invitrogen), 1ϫ L-glutamine (Invitrogen), and 1.25 g/ml Fungizone (Invitrogen). 16HBE14o Ϫ cells were cultured in minimum Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum and 100 units/ml penicillin/streptomycin. When cells reached confluency, they were washed twice with Ca 2ϩ / Mg 2ϩ -free PBS. The cells were then suspended using trypsin/EDTA solution and plated on diluted Vitrogen-coated (collagen types I and IV diluted 1:15 in Dulbecco's phosphate-buffered saline) glass coverslips. For [Ca 2ϩ ] i measurements, cells were used 48 -72 h after plating.
Fura-2 Imaging of Intracellular Ca 2ϩ -Cytosolic Ca 2ϩ concentration was measured with dual excitation wavelength fluorescence microscopy (Deltascan, Photon Technologies, Princeton, NJ) after cells were loaded with the permeant form of the fluorescence dye Fura-2/acetoxymethyl ester (Fura-2/AM; Teflabs, Austin, TX). Fura-2 fluorescence was measured at an emission wavelength of 510 nm in response to the excitation wavelength of 340 and 380 nm, alternated at a rate of 50 Hz by a computer-controlled chopper assembly. Ratios (340/380 nm) were calculated at a rate of 5 points/s using PTI software. Cells were incubated in Dulbecco's phosphate-buffered saline containing 2 mM CaCl 2 and 1 mM MgCl 2 in the presence of 5 M Fura-2/AM and 1 mg/ml Pluronic F-127 dissolved in Me 2 SO for 120 min to allow loading of the dye into the cells. After loading, coverslips were rinsed at least for 10 min in Dulbecco's phosphate-buffered saline to remove extracellular Fura-2/AM and the surfactant and were positioned in the cuvette at a 45°a ngle from the excitation light. Two glass capillary tubes were inserted into the top of the cuvette out of the patch of the excitation light. One tube was extended to the bottom of the cuvette and connected by way of polyethylene tubing to an infusion pump. The other capillary tube was positioned at the top of the cuvette and served to remove fluid from the cuvette. The volume of the cuvette was ϳ1.5 ml, and the flow rate was  (solution A). A, please note that the second application of ATP was without effect. In these traces and in all others below, please note that there is a time lag of 10 -15 s before agonist-containing perfusate enters the cuvette. As all of these experiments were performed on coverslips prepared on the same day, a calibration was used on the same cell preparation to allow conversion and plotting of the data as cytosolic calcium.
ϳ5 ml/min. It is important to note that switch in perfusion solutions is removed in time and space for the cuvette, such that a 10 -15-s time lag exists before agonist is exposed to the cells. Experiments were performed at room temperature. Fluorescence intensities at both wavelengths were assessed, and only those preparations in which there were Ͼ200,000 counts/s for both wavelengths were used for experiments. At the beginning of each experiment, cells were perfused with solution A (see below), and the fluorescence ratio was monitored for at least for 100 s to establish a stable base-line value. Agonists and antagonists were then added to the appropriate solutions (see later). The 340/380 nm ratios (R) were converted into [Ca 2ϩ ] i values using the equations of Grynkiewicz et al. (32) as follows: Ca 2ϩ , R max and R min are R values under saturating and Ca 2ϩ -free conditions, respectively, and S f380 and S b380 are the fluorescent signals (S) emitted by Ca 2ϩ -free (f) and Ca 2ϩ -bound (b) forms of Fura-2 at a wavelength of 380 nm. In situ cell calibrations were accomplished after the cells were permeabilized with ionomycin (2 M) under Ca 2ϩ -free (10 mM EGTA) and saturating Ca 2ϩ (3 mM CaCl 2 ) conditions. The K d was assumed to be 224 nM (32).
Fura-2 Quenching Experiments-Cells were loaded and washed as described for intracellular [Ca 2ϩ ] measurement. Fluorescence signal was measured at 359 nm (isosbestic wavelength) in the presence of MnCl 2 (500 M) to detect Ca 2ϩ -independent changes in Fura-2 fluorescence (33).
Immunoblotting with P2X Receptor Channel Isoform-specific Antibodies-Cells were lysed in a buffer containing 10 mM Tris, 0.5 mM NaCl, 0.5% Triton X-100, 50 g/ml aprotinin (Sigma), 100 g/ml leupeptin (Sigma), and 100 g/ml pepstatin A (Sigma) adjusted to pH 7.2-7.4. Twenty micrograms of protein were run per lane and separated on an 8% SDS-polyacrylamide gel and then transferred to a polyvinylidene difluoride membrane (Osmonics, Westborough, MA). Immunoblotting was performed with a rabbit polyclonal antibody to P2X 4 (Alomone Laboratories, Jerusalem, Israel) at a dilution of 1:500. P2X 1 , P2X 2 , and P2X 7 antibodies were also obtained from Alomone Laboratories and were tested in a similar manner. Reactivity was detected by horseradish peroxidase-labeled goat anti-rabbit secondary antibody (1:3,000 dilution, New England Biolabs, Beverly, MA). Enhanced chemiluminescence was used to visualize the secondary antibody.
Biotinylation of Plasma Membrane P2X Receptor Channels-Cells were seeded on Vitrogen-coated (collagen types I and IV diluted 1:15 in Dulbecco's phosphate-buffered saline) 12-mm filters and grown as polarized monolayers with a transepithelial resistance that exceeded 400 ohms/cm 2 . Cells were placed on ice and washed 3 times with cold PBS supplemented with 0.  ] i in IB3-1 cells exposed to nominally Ca 2؉ -free, Na ؉ -containing solution (A) and in cells exposed to Ca 2؉ -containing Na ؉ -free solution (B) as indicated. B, please note the slight sustained increase in [Ca 2ϩ ] i upon substitution of Na ϩ by NMDG. This sustained plateau was the first hint that in Na ϩ -free medium Ca 2ϩ entry channels could also be involved in the ATP-induced sustained Ca 2ϩ response.
(Pierce) for 1 h at 4°C. The reaction was quenched with 0.1 M Tris, pH 7.5. Cell lysates were collected as described above in immunoblotting procedures. Immobilized streptavidin beads (Pierce) were added to the lysates at a 1:10 dilution and rocked overnight at 4°C. Beads were washed 3 times with lysis buffer and incubated in sample buffer for 5 min at 95°C. The mixture was centrifuged, and the supernatant was loaded onto an SDS-PAGE gel. The immunoblotting procedure then continued as described above.
Solutions-Buffers for [Ca 2ϩ ] i measurement contained (mmol/liter) the following: for solution A: NaCl 140, KCl 3, KH 2 PO 4 1.3, Na 2 HPO 4  Data Analysis-Data are expressed as mean Ϯ S.D. An unpaired Student's t test was used to compare the data in different experimental groups. Results were considered significant if p Ͻ 0.05. For original Fura-2 traces shown in the figures, data are graphed with calibrated cytosolic free calcium on the y axis, because data from an individual preparation of cells was accumulated for all of the experiments in that figure where a calibration was also performed. Because not all data were generated from cells of the same passage or where a calibration was not performed for every preparation, the data in tables are shown as ratiometric data.

Purinergic Agonists Trigger a Transient Increase in [Ca 2ϩ ] i in the Presence of Extracellular Na ϩ in IB3-1 Cells-To test for
the presence of purinergic receptors in IB3-1 cells, we measured the cytosolic free Ca 2ϩ concentration after stimulation with different agonists to both P2Y and P2X receptors in physiologic bath solution (solution A) containing Na ϩ . Superfusion of cells with solution containing ATP (100 M) caused a rapid increase in the ratio (340/380 nm) of Fura-2 fluorescence (r basal ϭ 0.89 Ϯ 0.09 to r peak ϭ 1.19 Ϯ 0.13; n ϭ 15). However, the response was transient, and the [Ca 2ϩ ] i returned close to basal value within 200 s after stimulation, even in the continuous presence of agonist (r ϭ 0.92 Ϯ 0.09; n ϭ 15) (Fig. 1A). Furthermore, when cells were exposed to ATP for the second time, only a small and even more transient change was detected in Fura-2 fluorescence (Fig. 1A). Administration of 10 M ATP caused a comparable but smaller change in [Ca 2ϩ ] i ( Table I). The effect of ATP was completely inhibited by the application of suramin (100 M) (Table I). ADP, 2MeSATP (100 M each), and ADP␤S (10 and 100 M) also caused an increase in cytosolic Ca 2ϩ concentration, showing similar characteristics described for ATP (Fig. 1, B-D, and Table I). Because 2MeSATP and ADP␤S increased [Ca 2ϩ ] i in a similar manner to ATP and ADP, these data argue strongly for activation of P2Y 1 receptors over other P2Y subtypes. In contrast, neither UTP (100 M) (Fig. 1A) nor UDP (100 M) had any effect on Ca 2ϩ concentration (Table  I). To explore whether degradation of ATP or ADP plays role in elevation of [Ca 2ϩ ] i , we tested the effects of adenosine (100 M).    Table I).
b p Ͻ 0.05 relative to ATP (100 M) alone. c p Ͻ 0.05 relative to ADP␤S (100 M) in sodium-containing medium (see Table I). d p Ͻ 0.05 relative to BzBzATP (100 M) in sodium-containing medium (see Table I). shows an original trace where cells exposed to Na ϩ -free medium are responsive to BzBzATP (100 M) with a rise in [Ca 2ϩ ] i in IB3-1 cells. Note that a second trace (in magenta) is shown to illustrate the lack of effect of BzBzATP (100 M) in a Ca 2ϩ -free and Na ϩ -free solution.
Because adenosine did not increase [Ca 2ϩ ] i , we did not pursue the participation of P1 receptors in increasing [Ca 2ϩ ] i (Table I).
Purinergic Agonists Trigger a Transient Increase in [Ca 2ϩ ] i in the Absence of Extracellular Ca 2ϩ -Activation of P2Y 1 receptors leads to G protein-coupled phospholipase C-and inositol 1,4,5-trisphosphate-dependent release of Ca 2ϩ from intracellular stores. As such, P2Y agonists should increase cytosolic Ca 2ϩ even in the absence of extracellular Ca 2ϩ . Therefore, we repeated the experiments with ATP (100 M) ( Fig. 2A) and ADP (100 M) superfusing IB3-1 cells with solutions containing EGTA (1 mM) instead of CaCl 2 (solution B). Similar to control conditions, both agonists increased [Ca 2ϩ ] i transiently, indicating that their effects, at least partially, were independent from extracellular Ca 2ϩ (Table I). Nonetheless, the absence of extracellular Ca 2ϩ reduced the agonist-induced peak increase in [Ca 2ϩ ] i (Table I). Again, under these conditions, the Ca 2ϩ transients decayed fully back to base line within 200 s. Interestingly, these data did suggest that, besides P2Y 1 receptor activation, purinergic agonists may also trigger Ca 2ϩ influx from extracellular stores, which contributes to the peak increase in [Ca 2ϩ ] i . Nevertheless, under these ionic conditions, Ca 2ϩ influx was not sufficient to support a sustained elevation of [Ca 2ϩ ] i , the goal of this study. Experiments described below lend clarification to these early data.
P2X Receptor-selective Agonists Fail to Trigger an Increase in [Ca 2ϩ ] i in the Presence of Extracellular Na ϩ and Ca 2ϩ -Multiple subtypes of P2X receptors have already been described in human, rabbit, and rodent airway epithelial cells (27,34,35). Thus, we speculated that the higher peak in [Ca 2ϩ ] i in the presence of extracellular Ca 2ϩ and the loss of the full response in Ca 2ϩ -free extracellular solution could be explained by the concomitant activation of P2X receptors activated by ATP. To test this hypothesis, we superfused IB3-1 cells with "solution A" containing either ␣,␤-methylene ATP (␣,␤-MeATP, 100 M) or benzoyl-benzoyl-ATP (BzBzATP, 100 M) (Fig. 3A), selective agonists for different P2X receptor subtypes. Under these conditions, P2X-selective purinergic agonists failed to change [Ca 2ϩ ] i (Table I). However, we were aware of the fact that ␣,␤-MeATP and BzBzATP, although potent agonists at P2X 1 , P2X 3 , and P2X 7 receptors, have little or no effect at other P2XR subtypes. Thus, we hypothesized that changing the ionic composition of the superfusion medium might reveal activation of a Ca 2ϩ entry mechanism by these agonists (see below).
ATP and BzBzATP Trigger an Increase in [Ca 2ϩ ] i with Transient and Sustained Components in the Absence of Extracellular Na ϩ -Despite the negative data above with regard to P2Xselective agonists, we maintained the hypothesis that P2X receptors were involved in the full Ca 2ϩ response induced by ATP in the presence of extracellular Ca 2ϩ . Rationale for this hypothesis is given by the fact that, in human and mouse lymphocytes, Na ϩ might compete with Ca 2ϩ for entry through P2X receptors from extracellular stores (36 -38) as well as other families of Ca 2ϩ entry channels like the transient receptor potential channels (TRPs) or the store-operated Ca 2ϩ channels (SOCs) (39,40). Thus, we speculated that extracellular Na ϩ might suppress the Ca 2ϩ permeability of P2X receptor channels in IB3-1 cells. To verify this hypothesis, we substituted extracellular Na ϩ by N-methyl-D-glucamine (NMDG) (solution C) and tested the effects of a non-discriminant P2Y and P2X agonist (ATP), P2X-specific agonists (BzBzATP and ␣,␤-MeATP), and a P2Y 1 -specific agonist (ADP␤S). As shown in Fig. 2B (and in Fig. 5B and Fig. 7, A and B), substitution of extracellular Na ϩ by NMDG itself caused a small but sustained increase in [Ca 2ϩ ] i (r basal ϭ 0.89 Ϯ 0.04 to r NMDG ϭ 0.94 Ϯ 0.04; n ϭ 31; p Ͻ 0.05) which was completely absent when extracellular Ca 2ϩ was also omitted from the superfusion medium. These observations suggest the presence of a mechanism that allows sustained Ca 2ϩ entry, even in non-stimulated cells.
Following removal of extracellular Na ϩ and changes in [Ca 2ϩ ] i , we applied ATP (100 M). Under these conditions, ATP induced a further increase in [Ca 2ϩ ] i displaying a biphasic Ca 2ϩ response consisting of an initial transient peak and a sustained component (Fig. 2B and Tables II and III). In addition, as shown in Fig. 2B, a second application of ATP elicited a smaller increase in the [Ca 2ϩ ] i peak; however, the sustained Ca 2ϩ plateau was comparable with that observed after the first stimulation by ATP. When [Ca 2ϩ ] i reached a stable value after withdrawal of extracellular Na ϩ , we also added either BzBzATP (100 M) (Fig. 3B) or ␣,␤-MeATP (100 M). BzBzATP, but not ␣,␤-MeATP, induced a small increase in [Ca 2ϩ ] i ( Fig.  3B and Tables II and III). This increase was completely dependent on the presence of extracellular Ca 2ϩ , indicating a role for P2X receptors in Ca 2ϩ influx ( Fig. 3B and Table II). In Na ϩ -free media, P2Y 1 -specific agonist, ADP␤S (100 M), augmented the peak increase in [Ca 2ϩ ] i (Table II) but failed to elicit a sustained Ca 2ϩ plateau (Table III). Taken together, these data argue for a role for P2X receptors as Ca 2ϩ entry channels in IB3-1 cells.
P2X 4 Receptor Channel Protein Biochemistry-Due to the lack of other specific agonists or inhibitors, our functional studies did not distinguish further agonists among the P2XR subtypes. However, biochemical evidence suggests that IB3-1 cells express the P2X 4 receptor channel robustly. Membrane protein lysates from IB3-1 cells were prepared and were subjected to immunoblotting with a P2X 4 -specific polyclonal antibody. Fig.  4A shows the positive results for P2X 4 receptor channel protein  in total membrane protein lysates from IB3-1 cells grown on collagen-coated plastic as confluent monolayers. Inconsistent signals or a lack of signal was observed for P2X 1 , P2X 2 , and P2X 7 using specific antibodies to those subtypes (data not shown). The P2X 4 signal displayed a similar biochemical phenotype compared with human vascular endothelial cells and human polycystic kidney disease renal epithelial cells performed in our laboratory (13,41) as well as a recent study of P2X 4 receptor biochemistry in cardiac tissue and myocytes (42). An unglycosylated band was detected at ϳ46 kDa (the predicted molecular mass for P2X 4 ) and a larger and broader glycosylated band at 60 -65 kDa. These immunoblotting data show that P2X 4 is the most abundant P2X subtype expressed in IB3-1 cells. However, these data do not rule out less abundant expression of other P2X subtypes that is below the limit of detection with these antibodies. Further chemical modification of the extracellular solution also supports the abundant expression of P2X 4 receptor channels as the major P2X receptor subtype mediating Ca 2ϩ entry (see below). Fig. 4, B-D, shows additional data in 16HBE14o Ϫ non-CF airway epithelial cells. Immunoblotting of non-polarized cells grown in flasks (Fig. 4, B and C) as well as biotinylation (Fig.  4D) of polarized monolayers grown on permeable supports revealed robust and apical membrane-localized expression of P2X 4 . In these lysates, a third band of ϳ100 kDa was also found. Biotinylation was performed on the apical and basolateral surface of these monolayers. Only the apical signal is shown in Fig. 4D, although a detectable signal was also observed in basolateral biotinylated material (data not shown). Secondary antibody controls and blocking of antibody binding with the peptide immunogen, provided with the primary antibody in all biochemical assays, verified the specificity of P2X 4 receptor expression (data not shown). These data suggest that P2X 4 receptors are expressed abundantly by human airway epithelial cells grown under non-polarized and polarized conditions.
The Extracellular ATP-gated P2X 4 Receptor Channel Is the Major Ca 2ϩ Entry Channel Stimulated by ATP in IB3-1 and 16HBE14o Ϫ Cells-Like other subtypes of the P2X receptor channel family, the P2X 4 receptors are also regulated by different cations, such as H ϩ or Zn 2ϩ (25). Thus, if it is true that in IB3-1 cells the prolonged Ca 2ϩ response in Na ϩ -free medium was due to activation of P2X 4 receptors, then extracellular pH and Zn 2ϩ should modify the ATP-induced Ca 2ϩ signal. To test this hypothesis, we measured [Ca 2ϩ ] i after changing extracellular pH or in the presence of Zn 2ϩ in both IB3-1 and 16HBE14o Ϫ cells. We exposed IB3-1 cells to ATP after changing the pH of the superfusion solution. As shown in Table III, increasing extracellular pH potentiated the ATP-induced sustained increase in [Ca 2ϩ ] i only in Na ϩ -free medium. Furthermore, in a Na ϩ -free environment, acidic pH significantly reduced the ATP-induced peak increase in [Ca 2ϩ ] i (Table II). To demonstrate directly the effect of ATP on Ca 2ϩ influx from extracellular sources via another approach, we measured quenching of Fura-2 at 359 nm in the presence of MnCl 2 (500 M). Mn 2ϩ is known to permeate the same entry channels as Ca 2ϩ and quenches Fura-2 fluorescence when it enters the cells. As shown in Fig. 5A, in Na ϩ -free medium, acidic extracellular pH (6.4) inhibited Mn 2ϩ entry, whereas alkaline extracellular pH (7.9) potentiated markedly Mn 2ϩ entry and quenching of the dye. To further support the involvement of P2X 4 receptor channels, we tested the effect of the P2X receptor co-agonist, Zn 2ϩ , on ATP-induced Ca 2ϩ entry mechanisms. Inclusion of ZnCl 2 (20 M) further augmented the sustained increase in [Ca 2ϩ ] i induced by ATP in Na ϩ -free medium (Fig. 5B and Table III) but had no effect in Na ϩ -containing medium (Table III). Since our biochemical data (see above) indicated that P2X 4 receptors are also present in 16HBE14o Ϫ non-CF airway epithelial cells, we tested whether increasing extracellular pH or addition of Zn 2ϩ augmented the ATP-induced sustained Ca 2ϩ entry in Na ϩ -free medium in 16HBE14o Ϫ cells. As shown in Fig. 6A, ATP elicited extracellular pH-dependent quenching of Fura-2, suggesting that ATP-stimulated Ca 2ϩ influx is facilitated by alkaline pH. In addition, similar to results obtained with IB3-1 cells, both inclusion of Zn 2ϩ and increasing pH potentiated the effects of ATP on sustained Ca 2ϩ signal (Fig. 6B). Taken together, these data argue for a prom- FIG. 4. A, immunoblot analysis of IB3-1 cells grown as non-polarized monolayers in flasks using rabbit polyclonal antibodies against P2X 4 receptors. A smaller band of the predicted molecular mass for P2X 4 (46 kDa) was detected, as was a larger, broader, glycosylated band at 60 -70 kDa. The positions of molecular mass markers are shown on the left (in kDa). This is representative of 3 such experiments. B, immunoblot analysis of 16HBE14o Ϫ cells and CFPAC-1 cells grown as polarized cell monolayers (CFPAC-1 cells were screened as another CF-relevant cell line and were grown in a similar manner than 16HBE14o Ϫ cells except that Iscove's modified essential medium was used for the basal medium with all other additives kept similar). Note the stronger expression in polarized cell monolayers and the presence of a 40 -50-kDa band (unglycosylated predicted molecular mass), a 60 -80-kDa band (glycosylated form), and an even larger form at ϳ100 kDa (glycosylated form). This is representative of 6 such experiments. C, tunicamycin (10 M), an inhibitor of glycosylation, added to the culture medium in an overnight 24-h incubation of confluent cell monolayers grown in flasks abolished the 100-kDa form and inhibited the expression of the 60 -80-kDa band, yielding more of the 40 -50-kDa unglycosylated form. This is representative of 2 such experiments. D, three water-soluble forms of biotin reagents were used to biotinylate apical membrane P2X 4 . The data reveal that poly(ethylene)oxid-maleimide biotin, a reagent that reagents with primary amines primarily on lysine residues, detected only the glycosylated forms in the apical plasma membrane of 16HBE14o Ϫ epithelial cell monolayers. Biocytin hydrazide failed to work in this experiment, likely because our conditions for oxidizing the carbohydrate residues were not optimal. Sulfo-NHS-LC-biotin detected all of the forms, indicating that it may have detected apical P2X 4 ; however, it may have gained access to the cell interior to find the unglycosylated form as well. This is representative of 2 such experiments. Note pertaining to all panels: no secondary antibody controls were performed for all of the above experiments, as were peptide immunogen blocking experiments that effectively blocked the signal. Peptide immunogens for P2X 1 , P2X 2 , and P2X 7 did not block the P2X 4 signaling, revealing additional specificity. In addition to data from our laboratory in human ADPKD kidney epithelial cells (41) and human vascular endothelial cells (13), this is the first documentation of biochemical detection of native airway epithelial P2X 4 receptor protein.
inent role for the P2X 4 receptor as a Ca 2ϩ entry channel in human airway epithelial cells and argue against a functional role for other P2X receptor subtypes.
The P2X 4 -mediated Ca 2ϩ Entry Is Sustained, Long Lived, Reversible, and Re-acquired upon Re-addition of Agonist-For any therapeutic approach to be effective, especially one that targets an endogenous receptor, stimulation should be sustained and long lived. Even more desirable, the effect should be reversible to control the response. Ultimately, it is ideal if this endogenous receptor target did not desensitize or inactivate, as is apparent in this study for P2Y-mediated transient Ca 2ϩ signal. Fig. 7 shows experiments designed to determine whether P2X 4 -mediated Ca 2ϩ entry was sustained and long lived in IB3-1 cells. In the first protocol, ATP (100 M) was added in Na ϩ -free solution that has pH 7.9. A transient increase in [Ca 2ϩ ] i mediated by P2Y receptors was followed by a sustained plateau that persisted for over 60 min, until ATP was removed (Fig. 7A). In a second approach, a 15-min stimulation was performed with ATP and then was reversed with washout. Following re-addition of ATP, a similar sustained calcium plateau was acquired that persisted for 40 min. A third washout and stimulation was performed at the end of the protocol (Fig.  7B), showing lack of desensitization of the P2X 4 receptors or inactivation of their channel function. In contrast, the transient spike observed in the first application of ATP was lost. These data show, these data show that the P2X 4 -mediated Ca 2ϩ entry is sustained, long lived, reversible, and re-acquirable upon washout and re-addition of agonist.
Neither the Reverse Operation Mode of the Na ϩ /Ca 2ϩ Exchanger Nor Voltage-dependent Ca 2ϩ Channels or Store-operated Ca 2ϩ Channels Are Involved in ATP-induced Ca 2ϩ Entry in IB3-1 Cells-Theoretically, both the initial increase in [Ca 2ϩ ] i after removal of extracellular Na ϩ and the sustained Ca 2ϩ plateau induced by administration of ATP could be due to the activation of the Na ϩ /Ca 2ϩ exchanger in its reverse operation mode and/or other classes of Ca 2ϩ entry channels. Thus, we removed extracellular Na ϩ and added ATP in the presence of KB-R7943 (30 M), a specific inhibitor of reverse operation mode of the Na ϩ /Ca 2ϩ exchanger (43). Since KB-R7943 had no effect under these experimental conditions, we excluded the presence of this exchanger at the plasma membrane ( Fig. 8A and Tables II and III). Although airway epithelial cells are non-excitable cells and should not express voltage-dependent Ca 2ϩ channels, we asked the question whether cell membrane depolarization stimulated or inhibited the Ca 2ϩ response induced by ATP. Therefore, we exposed the cells to high extracellular KCl concentration (40 mM) in Na ϩ -free medium (solution D), and then we added ATP. As shown in Fig. 8B and Tables II and III, membrane depolarization inhibited the peak increase of [Ca 2ϩ ] i , and the sustained Ca 2ϩ plateau was completely abolished, indicating that IB3-1 cells do not express voltage-dependent Ca 2ϩ channels.
SOCs or TRPs represent other pathways by which Ca 2ϩ can enter non-excitable cells besides the ATP-gated P2X receptor channels. Theoretically, both SOCs and TRPs could be responsible for the sustained Ca 2ϩ influx induced by ATP in Na ϩ -free medium. Therefore, we tested whether SOCs are present in IB3-1 cells. We treated the cells with thapsigargin (100 nM), an inhibitor of Ca 2ϩ pump in the ER membrane, in the presence of extracellular Ca 2ϩ . This maneuver induced a large initial increase in Fura-2 fluorescence ratio (r basal ϭ 1.00 Ϯ 0.05 to r peak ϭ 2.92 Ϯ 0.17; n ϭ 3) followed by a sustained Ca 2ϩ plateau (r sustained ϭ 1.58 Ϯ 0.29; n ϭ 3). In the absence of extracellular Ca 2ϩ , stimulation with thapsigargin resulted in a small transient increase in [Ca 2ϩ ] i due to the depletion of intracellular Ca 2ϩ stores, and the re-addition of extracellular Ca 2ϩ elicited a large [Ca 2ϩ ] i increase (Fig. 9). These data indicate that IB3-1 cells possess SOCs, which are activated by a decrease in [Ca 2ϩ ] ER . Next, we have asked whether SOCs or store-independent TRP-like channels contribute to the sustained Ca 2ϩ increase after P2Y 1 receptor stimulation in Na ϩfree medium. To address this question, we used 2APB, which has recently been reported to inhibit SOCs (44,45), and SKF-96365, which is a blocker of the store-independent TRPs (46). Neither 2APB (75 M) nor SKF-56365 (50 M) abolished the ATP-induced sustained increase in [Ca 2ϩ ] i in the absence of extracellular Na ϩ (Table III). Interestingly, the sustained Ca 2ϩ plateau was further augmented by the SKF-96365 compound (Table III). These data indicate that, in IB3-1 cells, SOCs and/or TRPs do not play a role in regulating [Ca 2ϩ ] i following purinergic receptor stimulation. DISCUSSION Stimulation of purinergic receptors exerts biological effects, which are mediated in part through elevation of intracellular Ca 2ϩ concentration (47)(48)(49)(50)(51)(52). In the present study, we show evidence that IB3-1 cells express P2Y 1 and P2X 4 receptors abundantly. P2Y 1 receptors have been found recently in airway epithelia of P2Y 2 receptor-knockout mice (54), in rat lung (55), and in Calu-3 human airway epithelial cells (56). ADP␤S, a specific agonist of P2Y 1 receptors, increased [Ca 2ϩ ] i to a similar extent as ATP, ADP, and 2MeSATP, suggesting the presence of P2Y 1 receptors. Although recent data (57) indicate that 2Me-SATP and, possibly, ADP␤S at a concentration of 100 M may activate P2Y 11 receptors, we believe it is very unlikely that the increase in [Ca 2ϩ ] i observed in this study was due to the activation of P2Y 11 receptors. This conclusion derives from the fact that P2Y 11 receptors are poorly stimulated by ADP (26), whereas our data show that ADP is at least as potent an agonist as ATP. In addition, ADP␤S also elicited a significant increase in [Ca 2ϩ ] i at a concentration of 10 M. In other airway epithelial cell models, the presence of P2Y 2 has already been demonstrated (16,58,59). Furthermore, in vivo studies demonstrate that aerosolized UTP has beneficial effects in treatment of CF lung disease, confirming the presence of P2Y 2 and/or P2Y 4 on the apical membrane of airway epithelium (23,48). Interestingly, neither UTP nor UDP increased [Ca 2ϩ ] i in IB3-1 cells; however, both agonists do rescue impaired cell Experiments were done in a Na ϩ -free environment. B, note that substitution of Na ϩ by NMDG causes a slight increase in [Ca 2ϩ ] i , an effect that was inhibited in high KCl-containing solution.
volume regulation in IB3-1 cells. 2 These differences may reveal additional signal transduction pathways triggered by P2Y receptors that are independent of cytosolic calcium.
Nevertheless, in addition to the beneficial targeting of P2Y receptors for CF therapy, we argue here for the beneficial targeting of P2X receptors as well. Activation of these receptors would also have the added benefit of eliciting a sustained increase in [Ca 2ϩ ] i , an effect not observed with P2Y-specific agonists. The transient nature of the Ca 2ϩ signal induced by purinergic agonists accounts presumably for transient Cl Ϫ and fluid secretion observed in different CF epithelial cell models (7,29). Activation of P2X receptor channels under appropriate conditions would lead to Ca 2ϩ influx from the extracellular space. Furthermore, this Ca 2ϩ response is sustained for at least 1 h, is reversible, and is re-acquired to the same sustained level upon re-addition of agonists under conditions designed to stimulate P2X 4 .
However, our data could conceivably be explained in the following ways: 1) opening of extracellular ATP-gated P2X receptor channels; 2) activation of Na ϩ /Ca 2ϩ exchanger in reverse operation mode due to Na ϩ removal; 3) opening of voltagedependent Ca 2ϩ channels following membrane depolarization; and 4) activation of SOCs or TRPs after depletion of intracellular Ca 2ϩ stores. All lines of evidence indicate that activation of ATP-gated P2X 4 receptor channels led to augmentation of Ca 2ϩ signal and the sustained Ca 2ϩ plateau. First, in IB3-1 cells, BzBzATP, a P2X receptor-specific agonist, increases [Ca 2ϩ ] i only in Na ϩ -free medium. Second, the ATP-induced Ca 2ϩ plateau was enhanced by alkaline extracellular pH and inhibited by acidic extracellular pH. Third, ATP-induced Mn 2ϩ entry caused quenching of Fura-2 in a pH-dependent manner exhibiting significant increase in Mn 2ϩ permeability at alkaline pH. Fourth, application of Zn 2ϩ further enhanced the effects of ATP. Fifth, a P2Y 1 receptor-specific agonist, ADP␤S, did not cause a sustained increase in [Ca 2ϩ ] i . Sixth, neither 2APB, an inhibitor of SOCs, nor SKF-56365, a blocker of storeindependent TRP-like channels, abolished the sustained increase in [Ca 2ϩ ] i induced by ATP. Seventh, recent data (60,61) indicate that Zn 2ϩ inhibits SOCs. Eighth, biochemical evidence showed abundant expression of P2X 4 . Roles for the reverse mode of the Na ϩ /Ca 2ϩ exchanger and/or voltage-dependent Ca 2ϩ channels were ruled out with a variety of different cell biological maneuvers and/or pharmacological inhibitors. It is noteworthy that Vennekens et al. (62) have recently reported that epithelial Ca 2ϩ channels are regulated by extracellular pH. However, these channels are mainly expressed in kidney and intestinal epithelia and inhibited by metal ions at low micromolar concentration (63).
Although BzBzATP is primarily known to be an agonist of P2X 7 and antibodies used in this study were raised against rat P2X receptors, stimulation by Zn 2ϩ and inhibition by H ϩ are most consistent with activation of the P2X 4 receptors and inconsistent with other P2X receptor subtypes (25). For instance, stimulatory effects by Zn 2ϩ rule out a role for P2X 7 , because Zn 2ϩ is a P2X 7 antagonist (25). Inhibition of Ca 2ϩ entry by acidic pH rules out P2X 2 receptors, which are stimulated by acidic pH (25). The only phenotype that is not completely explained by P2X 4 alone is the alkaline pH stimulation. Heterologously expressed P2X 4 is only mildly stimulated by alkaline pH (64). As such, we cannot rule out that additional P2X receptor subtypes (perhaps P2X 5 (65), P2X 6 (66), or splice variants of P2X 4 , P2X 5 , and P2X 6 (67)) may be conferring these pH effects in a P2XR heteromultimer. Interestingly, in 16HBE14o Ϫ cells, ATP-driven Mn 2ϩ entry was also enhanced by alkaline pH, and Zn 2ϩ potentiated the ATP-induced sus-tained increase in [Ca 2ϩ ] i . Taken together, these data indicate that P2X 4 receptors function as ATP-gated Ca 2ϩ entry channels in both CF and non-CF airway epithelial cells.
In a past study (27), our laboratory showed that a P2Xselective agonist, BzBzATP, stimulated transepithelial chloride secretion in Ussing chamber experiments on airway epithelia that had both transient and sustained components and in nasal potential difference assays on mouse nasal mucosa that were transient stimulations that averaged 1-2 mV. These stimulations occurred in Na ϩ -rich solutions (27). Despite this knowledge, we did not perform experiments designed to examine P2XR-mediated signaling in this study (27). Because Na ϩ is in great excess to Ca 2ϩ in physiological saline, the contribution of Ca 2ϩ -permeable non-selective cation channels to a Ca 2ϩ entry phenotype is often masked. This was true for our CF cell model. In IB3-1 cells, removal of extracellular Na ϩ was required to observe any increase in [Ca 2ϩ ] i with BzBzATP and a sustained Ca 2ϩ signal with ATP. Nonetheless, in 16HBE14o Ϫ non-CF cells, extracellular Na ϩ (140 mM) prevented neither the BzBzATP-dependent Ca 2ϩ response nor the ATP-induced Ca 2ϩ plateau 3 ; however, responses to both BzBzATP and ATP were much more profound under Na ϩ -free conditions. Thus, we speculate that P2XR agonists might be useful in CF therapy regardless of extracellular Na ϩ concentration, although modification of the extracellular environment (Na ϩ removal, among other maneuvers) may strengthen their efficacy and was required to optimally study Ca 2ϩ entry mechanisms in Fura-2 spectrofluorometry. Nevertheless, further studies are required to determine whether the presence of extracellular Na ϩ inhibits P2XR-mediated rescue of Cl Ϫ secretion in CF therapy.
Interestingly, although controversial, recent data indicate that airway surface liquid (ASL) in non-CF subjects is hypotonic and low in Na ϩ with respect to the plasma (68). In contrast, other studies (69) have concluded that non-CF and CF ASL are isotonic. Nevertheless, it is noteworthy that, in Na ϩreplete medium, extracellular ATP stimulation of ciliary beat is attenuated, whereas in Na ϩ -free medium, ATP induction of ciliary beat was profound, suggesting a role for P2X receptors on cilia (35). Because cilia reside and need to function optimally in the ASL environment, we postulate that normal ASL may be hypotonic and, in particular, low in Na ϩ , allowing P2X receptor agonists to stimulate sustained signaling that may impact ion transport and ciliary beat. These specialized chemical and ionic conditions may also be critical in the delivery of agonists for CF therapy. This is tenable, because the vehicle for delivery during nebulization, aerosolization, or instillation would merely need to be modified to suit these optimal conditions. Taken together, these findings are profound with regard to therapy in CF, because they suggest that endogenously expressed P2X receptors do not desensitize or inactivate, and under appropriate conditions, their activation leads to a prolonged Ca 2ϩ signal that could translate into a sustained Cl Ϫ secretion in CF and non-CF epithelia.