Purinergic Receptors Coupled to Intracellular Ca 2 (cid:1) Signals and Exocytosis in Rat Prostate Neuroendocrine Cells*

Rat prostate neuroendocrine cells (RPNECs) display a variety of ion channels and exhibit (cid:2) -adrenergic regulation of cytosolic Ca 2 (cid:1) concentration ([Ca 2 (cid:1) ] c ). In this study, purinergic regulation of [Ca 2 (cid:1) ] c and exocytosis was investigated in freshly isolated single RPNECs showing chromogranin A immunoreactivity. The presence of P2X and P2Y receptors in RPNECs was verified by the transient activation of Ca 2 (cid:1) -permeable cationic channels and the release of Ca 2 (cid:1) from intracellular stores by extracellular ATP, respectively. The transient inward cationic current was effectively activated by (cid:2) , (cid:3) methyleneadenosine 5 (cid:1) -triphosphate ( (cid:2) , (cid:3) -MeATP) and blocked by 2 (cid:1) ,3 (cid:1) - O -(2,4,6-trinitrophenyl)adenosine 5 (cid:1) -triphosphate, suggesting the presence of a P2X 1 or P2X 3 subtype. For the release of stored Ca 2 (cid:1) , ATP and UTP were equally potent, indicating the functional expression of the P2Y 2 or P2Y 4 subtype. The mRNAs for P2X 1 and P2Y 2 were confirmed from reverse transcription- PCR analysis of RPNECs. The application of (cid:2) , (cid:3) -MeATP induced large and transient increases in [Ca 2 (cid:1) ] c , which were not attenuated by the blockers of voltage-activated Ca 2 (cid:1) channels or by depleting intracellular Ca 2 (cid:1) stores, but were abolished by omitting extracellular Ca 2 (cid:1) . The application of UTP increased [Ca 2 (cid:1) ] c to 55% of the peak (cid:4) [Ca 2 (cid:1) ] c induced by (cid:2) , (cid:3) -MeATP. The application of (cid:2) , (cid:3) MeATP induced exocytotic responses of RPNECs as monitored by carbon fiber amperometry and capacitance measurements. To our interest, the application of UTP did not induce amperometric currents, but reduced the membrane capacitance, indicating a net endocytosis. From these results, we postulate that a sharp rise in [Ca 2 (cid:1) ] c by the P2X-mediated Ca 2 (cid:1) influx is required for exocytosis, whereas the relatively slow release of stored Ca 2 (cid:1) induces endocytosis in RPNECs. and with fluorescein isothiocyanate (FITC)-conjugated phalloidin. Cells were then glass and investigated using a Zeiss LSM 510 laser-scanning confocal microscope (FITC, excitation at 488 nm and emission at 520 nm; and Alexa Fluor 546, excitation at 556 nm and emission at 573 nm). Intracellular Ca 2 (cid:1) Measurement— Dispersed single cells were loaded with fura-2 acetoxymethyl ester (2 (cid:4) M ) in PBS for 20 min at room temperature and then washed out with fresh solution. RPNECs were identified following the above criteria, and the region of interest for the fura-2 experiment was set so that the fluorescence from a single RPNEC could be collected selectively. The recording of cytosolic Ca 2 (cid:1) concentration ([Ca 2 (cid:1) ] c ) was performed with a microfluorometric system consisting of an Olympus IX-70 inverted fluorescence microscope with a dry-type fluorescence objective lens ( (cid:3) 40, numerical aperture of 0.85), a photomultiplier tube (type R1527, Hamamatsu Photonics, Hamamatsu, Japan), and a Deltascan illuminator (Photon Technology International Inc., Lawrenceville, NJ). Light was provided by a 75-watt xenon lamp, and a chopper wheel alternated the light path to monochromators (340 and 380 nm) with a frequency of 10 Hz; the intensity of emitted light at 510 nm was measured. As a measure of [Ca 2 (cid:1) ] c , the fluorescence emission ratio at 340/380 nm excitation ( F 340/380 ) is presented. Fluorescence Quenching Experiment— The divalent cation m M ) used a surrogate for to fura-2 capacitance measurement contained 110 m M CsCl, 40 m M HEPES, 0.3 m M EGTA, 1 m M MgCl 2 , 0.2 m M CaCl 2 , 0.35 m M GTP, and 2 m M MgATP with a free [Ca 2 (cid:1) ] c of 350 n M . For the simultaneous measurement of [Ca 2 (cid:1) ] c and capacitance, EGTA and CaCl 2 in the pipette solution were replaced with 0.15 m M K 5 -fura-2. Data Analysis and Statistics— The data are presented as original recordings, current-voltage ( I - V ) curves, histograms, and bar graphs of means (cid:8) S.E. (for n cells tested). When necessary, Student’s t test for paired samples was applied since the control and test recordings were made from the same cell. p (cid:6) 0.05 was regarded as significant.

Prostate neuroendocrine cells, considered to be a group of amine precursor uptake and decarboxylation cells, are intraepithelial regulatory cells with paracrine properties. Similar to other amine precursor uptake and decarboxylation cells in the gastric gland (1), prostate neuroendocrine cells may also influence the growth of the prostate gland or regulate the exocrine secretion of prostatic fluid (2,3). Previous immunohistochem-ical studies suggest that prostate neuroendocrine cells both store and produce neurosecretory products such as serotonin, histamine, bombesin, calcitonin, and parathyroid hormone-related peptides that could regulate the growth, invasiveness, metastatic processes, and angiogenesis related to prostatic carcinoma (4 -7).
In a previous study, we initially reported that putative rat prostate neuroendocrine cells (RPNECs) 1 display spontaneous action potentials and voltage-activated Ca 2ϩ channels (both L and N types) in accordance with the electrically excitable properties of neuroendocrine cells (8). Also, pharmacological studies clearly demonstrated that RPNECs functionally express ␣ 1and ␣ 2 -adrenergic receptors, which are linked to the release of stored Ca 2ϩ and the inhibition of N-type Ca 2ϩ channels, respectively (9). Despite the aforementioned electrophysiological evidence of excitable cells, the genuine property of neuroendocrine cells, viz. exocytosis, has not yet been directly investigated in RPNECs.
ATP is frequently colocalized with noradrenaline in postganglionic sympathetic nerve fibers and acts as a cotransmitter in several tissues (10). The purinoreceptors (P2), receptors for extracellular ATP, are divided into two distinct receptor families: G-protein-coupled metabotropic receptors (P2Y receptors) and receptor ion channels with a nonselective permeability to cations, including Ca 2ϩ (P2X receptors) (10). In some neurons and neuroendocrine cells, extracellular ATP induces exocytosis in a Ca 2ϩ -dependent manner. In the sympathetic neurons, for example, it has been suggested that the P2X receptors mediate a positive modulation of noradrenaline release, whereas the G-protein-coupled P2Y receptors mediate the opposite response (11). In the case of adrenal chromaffin cells displaying heterogeneous expression of P2X and P2Y receptors, P2X channels are preferentially localized to noradrenaline-secreting cells (12). The pathophysiological significance of purinoreceptors has also been suggested by the Ca 2ϩ release and growth regulation of prostate tumor cells by extracellular ATP (13).
Considering the suspected role of ATP as a cotransmitter from the sympathetic nerves that regulate the function of prostate glands, we chose to investigate the operation of multiple Ca 2ϩ translocation mechanisms linked to purinoreceptors in RPNECs. In addition, we compared the efficiency of the exocytosis triggered by P2X and P2Y receptors recruiting different sources of calcium ions in RPNECs. The direct measurement of exocytosis in prostate neuroendocrine cells was performed for the first time in this study.

EXPERIMENTAL PROCEDURES
Cell Isolation-Male Sprague-Dawley rats (350 -400 g) were killed by 100% CO 2 inhalation, and the ventral lobe of the prostate gland was removed rapidly thereafter. The procedure for single cell isolation was the same as described in our previous reports (8,9). Briefly, the chopped tissue (1-2 mm 3 ) was digested for 25 min at 37°C in Ca 2ϩ -free phosphate-buffered saline (PBS) containing collagenase (2 mg/ml; Wako, Osaka, Japan), trypsin inhibitor (1 mg/ml; Sigma), bovine serum albumin (3 mg/ml; Sigma), and dithiothreitol (1 mg/ml; Sigma). Following digestion, tissue segments were transferred to fresh PBS and agitated gently using a fire-polished wide bore (1-2 mm) Pasteur pipette. Cells were isolated daily and stored in fresh solution at 4°C for up to 6 h. Dispersed cells were moved into the experiment chamber and examined using an Olympus IX-70 inverted microscope. After the digestion procedure, most of the isolated single cells had an elongated columnar shape, a typical feature of secretory epithelial cells. In addition to columnar cells, we could identify round-or oval-shaped cells with a relatively dark cytoplasm that were regarded as RPNECs in this study.
Immunofluorescence Confocal Microscopy-Chromogranin A was identified using mouse monoclonal antibody LK2H10 (NeoMarkers, Fremont, CA). Cells were maintained at room temperature on a coverslip coated with polylysine (5%), fixed with 2% paraformaldehyde for 15 min, permeabilized with 0.1% Triton X-100 for 10 min, and then incubated with PBS containing 3% bovine serum albumin (blocking solution) for 45 min. Cells were incubated with primary antibodies, followed by treatment with bridging Alexa Fluor 546-coupled goat anti-mouse Ig (Molecular Probes, Inc., Eugene, OR) and double staining with fluorescein isothiocyanate (FITC)-conjugated phalloidin. Cells were then mounted onto glass slides and investigated using a Zeiss LSM 510 laser-scanning confocal microscope (FITC, excitation at 488 nm and emission at 520 nm; and Alexa Fluor 546, excitation at 556 nm and emission at 573 nm).
Intracellular Ca 2ϩ Measurement-Dispersed single cells were loaded with fura-2 acetoxymethyl ester (2 M) in PBS for 20 min at room temperature and then washed out with fresh solution. RPNECs were identified following the above criteria, and the region of interest for the fura-2 experiment was set so that the fluorescence from a single RPNEC could be collected selectively. The recording of cytosolic Ca 2ϩ concentration ([Ca 2ϩ ] c ) was performed with a microfluorometric system consisting of an Olympus IX-70 inverted fluorescence microscope with a dry-type fluorescence objective lens (ϫ40, numerical aperture of 0.85), a photomultiplier tube (type R1527, Hamamatsu Photonics, Hamamatsu, Japan), and a Deltascan illuminator (Photon Technology International Inc., Lawrenceville, NJ). Light was provided by a 75-watt xenon lamp, and a chopper wheel alternated the light path to monochromators (340 and 380 nm) with a frequency of 10 Hz; the intensity of emitted light at 510 nm was measured. As a measure of [Ca 2ϩ ] c , the fluorescence emission ratio at 340/380 nm excitation (F 340/380 ) is presented.
Fluorescence Quenching Experiment-The divalent cation Mn 2ϩ (0.1 mM) can be used as a surrogate for Ca 2ϩ to trace Ca 2ϩ influx (14). Since fura-2 has a high affinity for Mn 2ϩ (K d ϭ 5 nM), essentially all Mn 2ϩ ions entering a fura-2-loaded cell are trapped as fura-2⅐Mn 2ϩ complexes. These complexes are virtually non-fluorescent at all wavelengths. Thus, an accelerated attenuation of fura-2 fluorescence provides an estimate of the entry of divalent cations (e.g. Ca 2ϩ ) into a cell.
Single Cell Reverse Transcription (RT)-PCR-To synthesize first strand cDNA, RPNECs were collected individually using microelectrodes with an inner diameter of ϳ25-30 m on the stage of an inverted microscope for the patch clamp. After aspiration, the cell was expelled from the pipette into ice-cooled 0.2-ml tubes that contained 2 units of RNase-free DNase (Takara Bio Inc., Shiga, Japan), 40 units of ribonuclease inhibitor (Takara Bio Inc.), 50 mM Tris-HCl (pH 8.3), 3 mM MgCl 2 , 75 mM KCl, and 100 mM dithiothreitol. The reaction mixture was incubated at 37°C for 30 min, followed by 95°C for 5 min. The reverse transcription reaction was then performed in the presence of 0.8 g of oligo(dT) 15 (Roche Applied Science); 10 mM each dATP, dTTP, dGTP, and dCTP (Roche Applied Science); and 200 units of Super-Script TM reverse transcriptase (Invitrogen). The reaction proceeded for 10 min at 25°C and for 50 min at 42°C, followed by a 15-min step at 70°C to inactivate the SuperScript TM reverse transcriptase.
Two rounds of PCR (PTC-0150 MiniCycler TM , MJ Research, Inc., Waltham, MA) were performed with outer primers of P2Y 2 , P2Y 4 , P2X 1 , and P2X 3 (15, 16) from rat and four sets of nested primers. First strand cDNA was used for the first PCR amplification. This mixture was then used for the second nested PCR amplification. PCRs were carried out using 2.5 units of Taq polymerase (Takara Bio Inc.) containing 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl 2 , 10 mM dATP, 10 mM dTTP, 10 mM dGTP, 10 mM dCTP, 100 pM each sense and antisense primer, and DNA template. Temperature cycling proceeded as follows: one cycle at 95°C for 5 min and 45 cycles at 95°C for 30 s, 50°C (for P2X 3 ) or 55°C (for P2Y 2 , P2Y 4 , and P2X 2 ) for 60 s, and 72°C for 90 s, followed by 72°C for 10 min. PCR products were then subjected to gel electrophoresis on a 1.5% agarose gel containing ethidium bromide. The primers used for the PCR of P2Y 2 , P2Y 4 , P2X 1 , and P2X 3 were as follows (15,16) Patch Clamp Methods-Isolated cells were transferred into a bath situated on the stage of an Olympus IX-70 inverted microscope. The bath (ϳ0.3 ml) was superfused at 10 ml/min, and voltage clamp experiments were performed at room temperature (ϳ22-25°C). Patch pipettes (with a free tip resistance of ϳ2.5-3 megaohms) were connected to the head stage of an Axopatch 1-D patch clamp amplifier (Axon Instruments, Inc., Foster City, CA). Liquid junction potentials were corrected with an offset circuit prior to each experiment. For the perforated whole cell patch clamp, a stock solution of nystatin in dimethyl sulfoxide (15 mg/ml) was added to the pipette solution, yielding a final concentration of 0.15 mg/ml. A steady-state perforation (series resistance of Ͻ20 megaohms) was usually achieved within 10 min after making a giga-seal. pCLAMP Version 7.0 software and Digidata-1200A (both from Axon Instruments, Inc.) were used for data acquisition and the application of command pulses. The voltage and current data were low pass-filtered (5 kHz) and displayed on a computer monitor. Current traces were stored in a Pentium-grade computer and analyzed using Origin Version. 6.1 (Microcal Software Inc., Northampton, MA).
Carbon Fiber Electrode (CFE) Amperometry-CFEs were fabricated from carbon fibers (5-11-m diameter) and polypropylene micropipettor tips of 10-l volume as described previously (17). The tip of the electrode was closely apposed to the cell surface to minimize the diffusion distance from the release sites. The amperometric current, generated by the oxidation of released bioamines (e.g. serotonin) at the exposed tip of the CFE, was measured using an EPC-9 amplifier (HEKA Elektronik, Lambrecht/Pfalz, Germany) operated in the voltage clamp mode at a holding potential of 600 mV. Amperometric signals were low pass-filtered at 0.1 kHz with a gain of 20 or 50 mV/pA and then sampled at 0.5 kHz. Amperometric recordings were semiautomatically analyzed using software written in Igor (WaveMetrics, Lake Oswego, OR). Some recordings with a small number of amperometric signals were plotted on a fast chart recorder, and the events were counted manually. To evaluate the relative exocytosis, the numbers of exocytotic events in the control and test conditions were averaged for 2 min, respectively.
Capacitance Measurements-Capacitance measurements were performed using the Lindau-Neher technique implemented as the "sine ϩ dc" mode of the "software lock-in" extension of PULSE software (HEKA Elektronik). A 1 kHz, 35-mV peak-to-peak sinusoid stimulus was applied about a direct current holding potential of Ϫ60 mV. For the simultaneous acquisition of membrane current and capacitance, the X-chart program (HEKA Elektronik) was used.
Solutions and Chemicals-All experiments were performed in PBS containing 145 mM NaCl, 1.6 mM K 2 HPO 4 , 0.4 mM KH 2 PO 4 , 1 mM MgCl 2 , 2 mM CaCl 2 , and 5 mM D-glucose (pH 7.4) titrated with NaOH. Data Analysis and Statistics-The data are presented as original recordings, current-voltage (I-V) curves, histograms, and bar graphs of means Ϯ S.E. (for n cells tested). When necessary, Student's t test for paired samples was applied since the control and test recordings were made from the same cell. p Ͻ 0.05 was regarded as significant.

Identification of RPNECs and ATP-induced [Ca 2ϩ ] c
Increase-Among the dispersed prostate cells, we could discriminate between putative neuroendocrine cells with a round shape and relatively dark cytoplasm and the columnar epithelial cells and thin smooth muscle cells (8,9) The presence of chromogranin A, a representative marker of neuroendocrine cells, was confirmed. Fig. 1A shows a representative confocal image of chromogranin A-positive RPNECs co-stained with FITC-conjugated phalloidin, indicating the distribution of Factin. It is noteworthy that chromogranin A distributes in a granular fashion at the periphery of the cell, where the signal of F-actin is not strong. The following experiments were performed with putative RPNECs, with which the previous patch clamp study and Ca 2ϩ measurement consistently showed characteristics of electrically excitable cells (8,9).
The fura-2 fluorescence ratio (F 340/380 ; see "Experimental Procedures'') was measured. The mean of F 340/380 under control conditions was 1.27 Ϯ 0.019 (n ϭ 85), and the bath application of ATP (100 M) induced a sharp increase in F 340/380 to 3.0 or above in most of the RPNECs tested. In the normal bath solution (PBS), a closer observation revealed an interesting phenomenon of split peaks of the initial Ca 2ϩ transients induced by ATP (Fig. 1B, arrowhead). The time interval between the first and second peaks was 6.5 Ϯ 1.10 s (n ϭ 6). In the same cell, the removal of extracellular Ca 2ϩ abolished the first peak without affecting the amplitude of the second peak (Fig. 1B). The Mn 2ϩ quenching experiment also demonstrated a fast influx of divalent cations induced by ATP (Fig. 1C). While measuring the fura-2 fluorescence, we added 0.1 mM Mn 2ϩ to the bath perfusate in the absence of Ca 2ϩ , which alone did not induce a significant change in the fluorescence intensity, indicating that there is little background conductance of Mn 2ϩ . Under these conditions, the application of ATP (100 M) immediately attenuated both F 340 and F 380 , followed by a symmetrical increase and decrease in F 340 and F 380 , respectively (n ϭ 3). The initial decay of fluorescence was interpreted as quench- ing by the Mn 2ϩ influx through Ca 2ϩ -permeable channels activated by ATP, and the second symmetrical change, viz. the increase in F 340/380 , was regarded as the Ca 2ϩ release from intracellular stores. These results suggest that RPNECs express both P2X and P2Y receptors, which recruit extra-and intracellular sources of Ca 2ϩ with fast and slow kinetics, respectively.
P2Y-mediated Release of Ca 2ϩ in RPNECs-UTP, a pyrimidine triphosphate, activates some subtypes of P2Y receptors and induces inositol 1,4,5-trisphosphate-mediated Ca 2ϩ release (10). Thus, we compared the concentration dependence of stored Ca 2ϩ release induced by extracellular ATP and UTP (Fig. 2, A and B). ATP or UTP was applied in the absence of extracellular Ca 2ϩ to exclude the Ca 2ϩ influx pathways. During the interval of ATP application, CaCl 2 (2 mM) was added back to replenish the intracellular calcium stores. The halfeffective concentrations (EC 50 ) of ATP and UTP were 2.8 and 3.8 M, respectively. The EC 50 of UTP in the presence of 2 mM CaCl 2 was 2.2 M, not significantly different from the EC 50 of UTP under the Ca 2ϩ -free condition (Fig. 2B).
Next, the efficiency of stored Ca 2ϩ release was compared between various analogs of purines and pyrimidines. As a positive control, a typical response of [Ca 2ϩ ] c to noradrenaline (10 M) was confirmed for each cell (9). Then, the responses to ATP, UTP, UDP, 2-methylthio-ATP, ADP, and ATP␥S, all at 100 M, were compared in the absence of extracellular Ca 2ϩ . The ⌬F 340/380 were normalized to the peak level induced by noradrenaline (10 M) in the same cell. In summary, the Ca 2ϩ response to UTP (n ϭ 9) or ATP (n ϭ 9) was greater than the responses to ATP␥S (n ϭ 6) or ADP (n ϭ 3). 2-Methylthio-ATP (n ϭ 3) and UDP (n ϭ 6) showed little effect on [Ca 2ϩ ] c of RPNECs, indicating that the involvement of P2Y 1 and P2Y 6 subtypes could be excluded (Fig. 2C). The overall results were quite similar to the reported order of potency for the P2Y 2 or P2Y 4 subtype in rat (10, 18, 19). Although not a specific agent, suramin has been reported as a relatively selective blocker of P2Y 2 compared with P2Y 4 in rats (19). In RPNECs, the Ca 2ϩ response to UTP (100 M) was completely blocked by pretreatment with 50 M suramin (n ϭ 7) (Fig. 2D). In addition, the RT-PCR analysis of RPNECs consistently showed positive signals for P2Y 2 (Fig. 2E).
Ca 2ϩ -permeable P2X Channels in RPNECs-Under the nystatin-perforated or conventional whole cell patch clamp conditions, extracellular ATP (10 M) induced a fast transient inward current (I ATP ) (Fig. 3A). In some cases, to ensure fast solution exchange, 10 M ATP was applied by puffing from a thin polyethylene tube (tip diameter of ϳ100 m) located 50 -100 m from the recorded cell. The "local puff " method, combined with a nystatin-perforated whole cell clamp, was re-quired to prevent the rundown of I ATP following repetitive stimulation (e.g. Fig. 3D).
The mean amplitude of I ATP was Ϫ1.9 Ϯ 0.78 nA at a holding potential of Ϫ60 mV (n ϭ 9). To obtain the I-V curve of I ATP , brief ramp pulses from 60 to Ϫ100 mV were applied, and the control current was digitally subtracted from the peak current in response to ATP. The I-V curve of I ATP reversed direction at 10.0 Ϯ 1.02 mV (n ϭ 8) (Fig. 3B). In all cases, the I ATP or ␣,␤-MeATP-induced inward current (see below) showed fast inactivation, with 90% of the total current decayed within 1 s of drug application.
P2X channels are usually more permeable to Ca 2ϩ than monovalent cations and provide significant Ca 2ϩ influx pathways in many kinds of cells (20 -22). In RPNECs, after replacing the NaCl bath solution with 110 mM CaCl 2 solution, the bath application of 10 M ␣,␤-MeATP induced a huge inward current with a mean amplitude of 2.04 Ϯ 0.41 nA (n ϭ 7) at Ϫ60 mV. The I-V curve showed an inwardly rectifying property with a reversal potential of 18 Ϯ 3.0 mV (n ϭ 7) (Fig. 3F). Under this bi-ionic condition, where Cs ϩ (140 mM) and Ca 2ϩ (110 mM) are the only permeable cations at the intracellular and extracellular sides, respectively, the calculated permeability ratio of Ca 2ϩ to Cs ϩ (P Ca /P Cs ) for the P2X channel is 2.0 (23). In line with the high Ca 2ϩ permeability of P2X channels, the application of ␣,␤-MeATP increased the Ca 2ϩ concentration of RPNECs in a concentration-dependent manner (Fig. 3G). In some cases of this experiment, to prevent the spontaneous desensitization of P2X receptors by ATP released from nearby cells, apyrase (20 g/ml) was included in the extracellular solution throughout the cell isolation and incubation period (22). The EC 50 of ␣,␤-MeATP was 0.13 and 0.12 M in the absence and presence of apyrase, respectively, indicating that the P2X channels were not significantly desensitized after single cell isolation (Fig.  3G). In the following experiments, considering the fast desensitization, 10 M ␣,␤-MeATP was applied via the bath perfusate to achieve maximum and simultaneous stimulation of P2X receptors.
From the above results, it was clear that RPNECs express both P2X (P2X 1 ) and P2Y (P2Y 2 ) receptors that link to fast Ca 2ϩ influx and relatively slow release of stored Ca 2ϩ , respectively. As a next step, we asked whether the P2X and P2Y receptors play a subtype-specific role in the physiological responses of RPNECs, viz. exocytosis. To address this question, we tested whether ␣,␤-MeATP is a selective activator of P2X receptors without activating stored Ca 2ϩ release. In the absence of extracellular Ca 2ϩ , ␣,␤-MeATP had no effect on the [Ca 2ϩ ] c of RPNECs at 10 M (n ϭ 7) (Fig. 4A) or at 100 M (n ϭ 3) (data not shown). In another experiment, RPNECs were pretreated with noradrenaline (10 M) in the presence of thapsigargin (1 M), a potent inhibitor of Ca 2ϩ -ATPase in the endoplasmic reticulum. After confirming that noradrenaline did not release any additional Ca 2ϩ , ␣,␤-MeATP (10 M) was applied, which still induced a robust Ca 2ϩ response (Fig. 4B). These results indicate that ␣,␤-MeATP could be used as a selective agonist of P2X receptors mediating Ca 2ϩ influx in RPNECs.
Despite the high Ca 2ϩ permeability, the amount of direct Ca 2ϩ influx via P2X 1 and P2X 3 channels is reportedly small because of their fast desensitization (22). According to the results obtained with heterologously expressed P2X channels (22), the Ca 2ϩ influx via voltage-operated Ca 2ϩ channels (VOCCs) secondarily activated by P2X 1 -or P2X 3 -induced depolarization greatly exceeds the Ca 2ϩ influx via P2X 1 /P2X 3 channels. Therefore, to block the L-and N-type VOCCs in RPNECs (8), cells were pretreated with nifedipine (1 M) and -conotoxin GVIA (1 M). However, the combined treatment with Ca 2ϩ channel blockers did not affect the peak amplitudes of Ca 2ϩ responses to ␣,␤-MeATP (⌬F 340/380 ϭ 4.1 Ϯ 0.45, n ϭ 7) (Fig. 4C). In another experiment, the membrane voltage was clamped at Ϫ60 mV using a whole cell patch clamp to prevent the activation of VOCCs (Fig. 4D). Both the membrane current and the change in [Ca 2ϩ ] c were simultaneously monitored using 0.2 mM K 5 -fura-2 dialyzed with a CsCl pipette solution. Under these conditions, ␣,␤-MeATP simultaneously induced a huge inward current and transient increase in [Ca 2ϩ ] c (⌬F 340/380 ϭ 2.1 Ϯ 0.37, n ϭ 4). In the same cell, the increase in [Ca 2ϩ ] c by a depolarizing step pulse to 0 mV was also measured. The mean of depolarization-induced ⌬F 340/380 was 0.8 Ϯ 0.09 (n ϭ 4), much smaller than that with ␣,␤-MeATP (10 M) (Fig. 4D). Thus, it is fairly safe to say that Ca 2ϩ influx directly through ␣,␤-MeATP-sensitive P2X channels determines the peak level of [Ca 2ϩ ] c in RPNECs.
To compare the peak amplitudes of [Ca 2ϩ ] c responses to P2Xand P2Y-selective stimulations, we applied ␣,␤-MeATP (10 M) and UTP (100 M), alternating the order of drug application in each cell. Although both agents induced substantial increases in [Ca 2ϩ ] c in all the putative RPNECs tested, the peak ⌬F 340/380 induced by ␣,␤-MeATP was larger than that induced by UTP (⌬F 340/380 ϭ 3.8 Ϯ 0.25 for ␣,␤-MeATP and 2.1 Ϯ 0.19 for UTP, n ϭ 25; p Ͻ 0.05), as summarized in Fig. 5C. Apart from the larger peak amplitude, the recovery of the Ca 2ϩ response after the washout of ␣,␤-MeATP generally took longer time than that after the washout of UTP, the reason for which is not clear yet. In Ͼ90% of RPNECs tested for UTP and ␣,␤-MeATP, both agonists induced significant increases in [Ca 2ϩ ] c , indicating that P2X and P2Y receptors are coexpressed in the same cell.
Purinergic Stimulation and Exocytosis in RPNECs-Previous immunohistochemical studies suggest that prostatic neuroendocrine cells contain oxidizable amine compounds such as serotonin (5-hydroxytryptamine) (4,24). Therefore, in an effort to prove exocytosis in putative RPNECs, we performed amperometry using a CFE (Fig. 5). The CFE was located in close contact with the cell, and the voltage was held at 600 mV. In 18 of 30 cells tested, the events of spontaneous exocytosis of oxidizable molecules were observed under the control conditions. The application of ATP (100 M) increased the frequency and amplitudes of current spikes superimposed on a slowly fluctuating base-line current, as described previously for other cell types such as PC12 cells (25) and chromaffin cells (26). The time-expanded trace of an exocytotic event (Fig. 5D, inset) demonstrated an instantaneous upstroke in the oxidizing current, followed by an exponential decay. The burst-like increase of current spikes by ATP was in accordance with our hypothesis that an appropriate stimulation would induce the exocytosis of oxidizable compounds in putative RPNECs (Fig. 5A). The frequency of events rose from nearly 8.5/min in the control to 28/min after ATP application in this particular case (Fig. 5B). an onset and exponential decay that are characteristic of exocytotic release from single secretory vesicles (27). To analyze the characteristics of quantal secretion in RPNECs more precisely, we measured the peak amplitudes and total charge of each exocytotic spike of the control and stimulation with ATP and then summarized the data in histograms (Fig. 5D). The average spike amplitude and area of quantal events were 0.24 Ϯ 0.04 pA and 3.1 Ϯ 0.41 femtocoulomb in the control and were significantly increased to 0.49 Ϯ 0.07 pA and 9.3 Ϯ 1.29 femtocoulomb by ATP treatment in this particular experiment. Similar results were obtained in the five RPNECs tested with ATP.
Stimulation of P2X (but Not P2Y) Induces Exocytosis in RPNECs-We then compared the effects of P2X and P2Y stimulations on the amperometrically measured exocytosis in RPNECs. ␣,␤-MeATP (10 M) significantly increased the amplitude and frequency of spikes in 14 of 15 cells tested. In contrast, UTP (100 M) induced a slight secretory response in only 2 of 15 cells. Fig. 6 depicts the representative results showing that only the stimulation with ␣,␤-MeATP induced exocytotic responses; the frequency of spikes increased from 19 to 85 events/ min (Fig. 6B). The difference between the two agonists is also obvious upon comparison of the histograms of the amplitudes and total charge of the current spikes (Fig. 6, C and D).
Although CFE amperometry allows a direct measurement of exocytosis, it is limited to detecting only oxidizable compounds. To overcome this limit, we adopted another method for measuring exocytosis, viz. the Lindau-Neher mode of capacitance measurement. The membrane capacitance (C m ), membrane current (I m ), and [Ca 2ϩ ] c of RPNECs were simultaneously monitored in the whole cell clamp mode. The representative cases show that application of ␣,␤-MeATP (10 M) induced a large inward current and Ca 2ϩ spike, followed by an increase in C m that decayed slowly to the control level (Fig. 7, A and B). In contrast, the Ca 2ϩ response to UTP (100 M) was accompanied by a reversible decrease in C m , implying endocytosis (Fig. 7A). In some cases of testing ␣,␤-MeATP, the initial change in C m started from the level below the resting C m and then increased slowly to above the base line (trace not shown). Fig. 7B shows the mean time course of C m changes induced by ␣,␤-MeATP and UTP in 23 and 15 RPNECs, respectively. On average, the maximum increase in C m by ␣,␤-MeATP was 5.2 Ϯ 1.53% of the control (n ϭ 23), and the maximum decrease in C m by UTP was Ϫ7.3 Ϯ 1.39% (n ϭ 15).
Both the ⌬[Ca 2ϩ ] c and exocytotic increase in C m by ␣,␤-MeATP (10 M) were abolished in the absence of extracellular Ca 2ϩ (1 mM EGTA without CaCl 2 ) despite the large inward current (n ϭ 6) (Fig. 7, C and D). Similarly, the UTP-induced C m decrease was completely blocked when 10 mM EGTA was included in the pipette solution (n ϭ 5) (data not shown), indicating that the endocytotic response to P2Y stimulation was also Ca 2ϩ -dependent. Also, pretreatment of RPNECs with suramin (50 M), but not with TNP-ATP (0.1 M), blocked the endocytotic response to 100 M UTP (n ϭ 2) (Fig. 7E). Interestingly, upon pretreatment with TNP-ATP (0.1 M), extracellular ATP consistently decreased the C m of RPNECs (6.3 Ϯ 1.43%, n ϭ 6) (Fig. 7E). DISCUSSION This study demonstrates that RPNECs express both P2X (P2X 1 ) and P2Y (P2Y 2 ) purinoreceptors. The fast Ca 2ϩ influx via P2X channels seems to be more efficiently linked to exocytosis in RPNECs, whereas the relatively slow and smaller increase in [Ca 2ϩ ] signaled from P2Y receptors is consistently linked to a net decrease in C m , viz. endocytosis. Previous studies of the effects of extracellular ATP in the prostate have been confined to epithelial cells (28) and smooth muscle (29). This study is, to our knowledge, the first investigation on the functional role of purinoreceptors in prostate neuroendocrine cells. Thus, like other organs, purinoreceptors seem to be widely distributed in the prostate gland, including neuroendocrine cells.
Both the fast desensitization of I ATP and the potent activation by ␣,␤-MeATP are consistent with the known properties of P2X 1 /P2X 3 channels, and single cell RT-PCR confirmed the expression of the P2X 1 subtype in RPNECs. It is noteworthy that the large ␣,␤-MeATP-induced increase in [Ca 2ϩ ] c (⌬[Ca 2ϩ ] c ) was not significantly affected by blocking VOCCs or by depleting intracellular Ca 2ϩ stores (Fig. 4). Because 10 or 100 M ␣,␤-MeATP had no effect in the absence of extracellular Ca 2ϩ , the large Ca 2ϩ response activated by ␣,␤-MeATP must have been due to the transmembrane influx of Ca 2ϩ . Considering the fast inactivation kinetics of P2X 1 channels, however, it was surprising that the transient Ca 2ϩ influx through P2X 1 channels overwhelmed the contribution of VOCCs that would be concomitantly activated in RPNECs. In contrast to our present results, in GT1 cells heterologously expressing P2X 1 or P2X 3 receptors, the transient ⌬[Ca 2ϩ ] c by extracellular ATP is largely abolished by nifedipine, suggesting that direct influx of Ca 2ϩ via P2X 1 and P2X 3 plays a minor role (22). One plausible explanation of the unexpectedly strong influx of Ca 2ϩ upon stimulation with ␣,␤-MeATP is the huge amplitude of the P2X receptor current in RPNECs. If we simply assume that ϳ10% of the 2 nA square current with a duration of 20 ms is carried by Ca 2ϩ (20), the instantaneous entry of Ca 2ϩ would be ϳ4.1451 ϫ 10 Ϫ16 mol. Such flux of Ca 2ϩ into a RPNEC with a diameter of 10 m would produce an instantaneous increase in [Ca 2ϩ ] c by 1.056 mM. If only 1% of the entered Ca 2ϩ is present as a free ionized form, it will still produce 10 M ⌬[Ca 2ϩ ] c . Such a large ⌬[Ca 2ϩ ] c might have not only overwhelmed the secondary influx of Ca 2ϩ via VOCCs, but also induced Ca 2ϩ -dependent inactivation of VOCCs (23), which could explain the insignificant effects of Ca 2ϩ channel blockers on ⌬[Ca 2ϩ ] c induced by ␣,␤-MeATP (Fig. 4C).
The results of amperometry indicate that the tested cells are capable of synthesizing and releasing oxidizable bioamines ( Fig. 5 and 6), most likely serotonin (4,24). The single amperometric spikes in RPNECs usually display 5-10 ms of halfamplitude width. Such values are similar to those observed in pancreatic duct epithelial cells (7 ms) and adrenal chromaffin cells (5-15 ms) (17,27). However, the mean amplitudes of these amperometric current spikes are much smaller than those observed in adrenal chromaffin cells, which often display Ͼ10 pA in an individual spike. Presumably, such small spikes may reflect either the less concentrated hormones or the smaller size of secretory vesicles in RPNECs compared with chromaffin cells (30). Another possibility is that RPNECs store oxidizable compounds together with other substances, e.g. neuropeptides (7), that are undetectable by CFE amperometry.
Both in the amperometry and C m measurements, UTP was found to be far less effective than ␣,␤-MeATP in triggering exocytosis from RPNECs. Although the application of UTP also induced a sizable increase in [Ca 2ϩ ] c , the change in fura-2 fluorescence (⌬F 340/380 ) by UTP was about half of that induced by ␣,␤-MeATP. If we suppose that the level of [Ca 2ϩ ] c is the key determinant of exocytosis, the above results suggest that the threshold of [Ca 2ϩ ] c for triggering the exocytosis in RPNECs is relatively high, which might be hard to be attained by P2Y stimulations. In the hair cells of the inner ear, for example, sizable C m changes occur only after the elevation of [Ca 2ϩ ] c by Ͼ7 M (31). Another assumptive speculation is that the P2Xmediated Ca 2ϩ influx might have induced a huge increase in [Ca 2ϩ ] in the subplasmalemmal space, where the exocytotic machinery and vesicles are likely to be localized. However, we did not observe such a Ca 2ϩ gradient when a preliminary Ca 2ϩ imaging study was carried out using confocal microcopy (data not shown). Also, a recent study of cellular Ca 2ϩ images induced by P2X stimulation in neuroendocrine cells demonstrated a global increase in [Ca 2ϩ ] c rather than a local subplasmalemmal change (21).
The opposite changes in the C m of RPNECs in response to ␣,␤-MeATP and UTP initially suggest an intriguing possibility that the Ca 2ϩ signals due to P2X and P2Y stimulations may be linked to distinctive cellular responses, viz. exocytosis and endocytosis. According to a recent study in the motor nerve endings of Drosophila, exocytosis and endocytosis are separately controlled by different routes of Ca 2ϩ influx mechanisms (32). However, in RPNECs, we do not have further evidence to suggest that exocytosis and endocytosis are separately regulated by Ca 2ϩ influx and release from stores. Instead of that hypothesis, the slow increase in C m , sometimes from below the resting level, upon the application of ␣,␤-MeATP (Fig. 7B) suggests that a rise in [Ca 2ϩ ] c might commonly induce endocytotic events irrespective of the Ca 2ϩ recruitment pathways. In other words, the concomitant activation of endocytosis might have impeded the initial fast increase in C m by exocytosis. An induced decrease in C m below the resting level has also been reported in calf adrenal chromaffin cells (33) and rat melanotrophs (34). The endocytotic events reported in other cells are usually preceded by an initial exocytosis, whereas the UTPinduced endocytosis in the present study displayed a monophasic decrease in C m . Thus, the UTP-induced endocytosis in RPNECs may not be a retrieval of previously exocytosed membrane, the physiological meaning of which needs further investigations.
In summary, both ionotropic P2X 1 receptors and metabotropic P2Y 2 receptors are present in RPNECs, the stimulation of which increases [Ca 2ϩ ] c with discriminate kinetics and amplitudes. The strong Ca 2ϩ signal from P2X is more effectively linked to exocytosis, whereas the P2Y stimulation results in a net decrease in C m , i.e. endocytosis. Identification of the components determining the direction of responses and the threshold of ⌬[Ca 2ϩ ] c will be essential to attain a full understanding of exo/endocytosis in RPNECs, a novel type of neuroendocrine cell.