INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

CALCIUM WAVES HAVE BEEN OBSERVED in various excitable and nonexcitable cells where they coordinate physiological responses within the respective tissues. In nonexcitable cells the most intensively investigated mechanism for wave propagation is based on strain-induced inositol trisphosphate [Ins(1,4,5)P₃] formation and diffusion though gap junctional pores. Ins(1,4,5)P₃ then releases intracellular Ca²⁺ from Ins(1,4,5)P₃-sensitive Ca²⁺ stores in the neighboring cells (2, 36, 37). Recent studies on basophil leukemic cells, hepatocytes, and osteoblastic cell lines (21) have shown that intracellular Ca²⁺ (Ca_i²⁺) waves may be alternatively propagated via activation of purinergic receptors of the G protein-coupled P2Y class that activate phospholipase C (PLC), resulting in the generation of Ins(1,4,5)P₃ and intracellular Ca²⁺ release from Ins(1,4,5)P₃-sensitive Ca²⁺ stores. Although it has been speculated that purinergic receptor activation and Ca_i²⁺ wave propagation may be mediated by ATP release from the mechanically stimulated cells, the molecular mechanism of purinergic receptor stimulation following mechanical strain remains poorly defined. ATP release triggered by mechanical strain has been recently reported and apparently did not involve the cystic fibrosis transmembrane regulator (CFTR) (47). Furthermore, it has been demonstrated that connexins regulate, via a still unraveled mechanism, strain-mediated Ca²⁺ signaling by controlling ATP release. In this study ATP release was inhibited by anion channel blockers (12).

The present study reports on mechanical strain-elicited Ca_i²⁺ waves in confluent prostate cancer cells of the DU-145 cell line, which were independent of intercellular communication because they persisted after uncoupling of gap junctions. Ca_i²⁺ wave propagation could be inhibited by antagonists of purinergic receptors and by preincubation with several nucleotides, indicating that activation of multiple purinergic receptors, including P2Y₂ receptors, which have been previously shown to be present in DU-145 prostate cancer cells (46), may underlie wave propagation. In hypotonic solution these cells secreted ATP. Because ATP release could be inhibited by blockers of anion channels, we concluded that purinergic receptor activation during Ca_i²⁺ wave propagation is mediated via ATP release through volume-activated Cl channels. ATP released by mechanical strain will diffuse radially in the extracellular space and will propagate a Ca_i²⁺ wave spreading up to a distance of ~300 µm from the stretched cell area.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Chemicals. Fluo 3-acetoxymethyl ester (AM) and 5-carboxyfluorescein diacetate (5-CFDA) were purchased from Molecular Probes (Eugene, OR). ATP, ADP, UTP, UDP, 2-methylthio-ATP (2-MeS-ATP), benzoylbenzoyl-ATP (Bz-ATP), apyrase, 2-deoxyglucose, 1-heptanol, tamoxifen, niflumic acid, suramin, basilen blue, 4,4'-diisothiocyanatostilbene-2,2'-disulfonic acid (DIDS), U-73122, histamine, and GdCl₃ were from Sigma (Deisenhofen, Germany). 5-Nitro-2-(3-phenylpropylamino)benzoic acid (NPPB) and glibenclamide were from RBI (Natick, MA).

Cell culture. The human prostate cancer cell line DU-145 was kindly provided by Dr. J. Carlsson (Uppsala, Sweden). Cells were cultivated in 25-cm² tissue culture flasks (Costar, Fernwald, Germany) in 5% CO₂ and humidified air at 37°C with Ham's F-10 medium (GIBCO, Life Technologies, Gaithersburg, MD) supplemented with 10% fetal calf serum (Boehringer Mannheim, Mannheim, Germany), 2 mM glutamine, 0.1 mM -mercaptoethanol, 2 mM minimal essential medium, 100 IU/ml penicillin, and 100 µg/ml streptomycin (ICN Flow, Meckenheim, Germany). For the experiments, cells grown to confluency in tissue culture flasks were enzymatically dissociated in Ca²⁺-free phosphate-buffered saline supplemented with 0.05% EGTA and 0.1% trypsin. Single cells were plated to coverslips and cultivated to confluency.

Bioluminescence experiments. ATP release from confluent DU-145 cells was determined using a luciferin-luciferase assay (Sigma) in a chemiluminescence apparatus (Bioluminiscence Analyzer XP2000, SKAN, Basel, Switzerland) under dim light. For data sampling, the output of the photomultiplier tube of the setup was connected to a multimeter (Voltcraft M-3610D, Conrad Electronics, Hirschau, Germany) and a Tandon 286/N personal computer (Tandon, Moorpark, CA). Cells grown to confluency on 20 × 20-mm coverslips were washed five times in F-10 cell culture medium, which resulted in a background luminescence signal that was not significantly different from the signal obtained with cell culture medium in the absence of cells. Cells were subsequently immersed in 1 ml of F-10 cell culture medium that was diluted 1:1 with distilled water, resulting in an osmolality of 150 mosmol/kgH₂O. In control experiments, cells grown to confluency on coverslips were immersed in an equal volume of isotonic medium. For the experiments with anion channel inhibitors, cells were preincubated for 5 min in isotonic F-10 cell culture medium that was supplemented with the respective inhibitor. Subsequently, cells were immersed in 1 ml of hypotonic F-10 medium supplemented with inhibitors. After different times, during which the cells were gently shaken, a 200-µl aliquot was removed and pressure injected via a light-tight access into a 3-ml glass cuvette containing 50 µl of the ATP assay mix and 1.5 ml of ATP assay mix dilution buffer (Sigma). Calibration measurements with ATP were performed in a concentration range of 0-100 nM. The lowest concentration of ATP that could be detected under the applied experimental conditions was 0.5 nM. From these calibration curves the total picomoles of ATP released per 10⁵ cells were calculated. Each of the applied anion channel inhibitors was tested for its effects on the activity of luciferase enzyme activity. No significant effects of the compounds on luciferase enzyme activity were observed. Furthermore, dilution of F-10 cell culture medium to yield an osmolality of 150 mosmol/kgH₂O did not affect luciferase activity (n = 3 for each experimental condition) (see Fig. 1). The chemiluminescence output curve was integrated, and the resulting values were set in relation to the calibration curve. To correlate ATP release to the cell number, the cells from which ATP release had been determined were enzymatically dissociated with a 0.2% trypsin-0.05% EDTA solution, and the cells were counted using an automated cell counter (Cell Analyzer Casy 1, Schärfe System, Reutlingen, Germany).

View larger version (37K): [in this window] [in a new window]   Fig. 1.   Effects of either hypotonic solution or anion channel inhibitors on the activity of the luciferin-luciferase assay. A 200-µl aliquot of either isotonic F-10 cell culture medium containing 10 nM ATP or hypotonic F-10 medium (150 mosmol/kgH2O) containing 10 nM ATP and supplemented with either niflumic acid (100 µM), 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB; 100 µM), tamoxifen (50 µM), gadolinium (100 µM), or DIDS (500 µM) was pressure injected into a glass cuvette containing 1.55 ml of the luciferin-luciferase solution. Note that neither hypotonic solution nor addition of any of the compounds significantly impaired the luminescence signal.

Fluorescence recovery after photobleaching. Cell cultures were incubated with 20 µM 5-CFDA for 20 min and postincubated in the absence of 5-CFDA for a further 10 min. Excitation of 5-CFDA was provided by the 488-nm line of the argon ion laser of the confocal setup. Emission was recorded using a long-pass LP 515 filter set. Single cells were selected using an overlay mask. 5-CFDA fluorescence was photobleached in the selected cells using the "point scan" mode of the confocal setup. By switching the attenuation filter wheel of the confocal setup, we elevated the laser power from 0.125 to 12.5 mW for 3 s, which resulted in photobleaching of the dye. After photobleaching, the microscope settings were returned to the recording configuration, and fluorescence recovery in the photobleached cell was monitored every 10 s.

Ca²⁺ imaging and confocal laser scanning microscopy. Measurement of Ca_i²⁺ was performed using the fluorescent Ca²⁺ indicator fluo 3-AM. Cells adherent to 20 × 20-mm-diameter glass coverslips were incubated for 60 min at 37°C in cell culture medium containing 10 µM fluo 3-AM dissolved in dimethylsulfoxide (final concentration 0.1%) and Pluronic F-127 (final concentration <0.025%; Molecular Probes). After loading, the coverslips were rinsed in E1 buffer containing (in mM) 135 NaCl, 5.4 KCl, 1.8 CaCl₂, 1 MgCl₂, 10 glucose, and 10 HEPES (pH 7.4 at 37°C) and mounted to the bottom of an incubation chamber that was fixed to the table of an inverted confocal laser scanning microscope (LSM 410, Zeiss, Jena, Germany). Fluorescence was excited by the 488-nm line of an argon-ion laser. Emission was recorded using a long-pass LP 515 filter set. The experiments were performed with a ×25 objective (NA 0.85; Neofluar, Zeiss). Processing of images was carried out with the Time software facilities of the confocal setup. The minimum, maximum, mean, standard deviation, and integrated sum of the pixel values in a region of interest (selected using an overlay mask) were written to a data file and routinely exported for further analysis to the commercially available SigmaPlot (Jandel Scientific, Erkrath, Germany) graphics software. Data are presented in arbitrary units as percentages of fluorescence variation (F/F₀) with respect to the resting level fluorescence (F₀).

Statistical analysis. Data are given as means ± SE, with n denoting the number of experiments. Student's t-test for unpaired data was applied as appropriate. A value of P < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

DU-145 prostate cancer cells propagate Ca_i²+ waves upon mechanical strain. When confluent cells of the prostate cancer cell line DU-145 were mechanically stimulated with the tip of a glass pipette, a transient rise of Ca_i²⁺ from 109 ± 38 nM (n = 16 cells in 5 independent experiments) under resting conditions to 440 ± 61 nM (n = 17 cells in 5 independent experiments) upon mechanical strain occurred with a delay of 1-3 s after the cell membrane was stretched (n = 8) (Fig. 2A). The Ca_i²⁺ signal spread radially from the stretched area with a velocity of ~15 µm/s and declined within a distance of 200-300 µm. Ca_i²⁺ waves could be elicited from the same stretched area up to three times (n = 3) (Fig. 2B), indicating that cell membranes were not ruptured during mechanical stimulation. However, the distance covered by the Ca_i²⁺ wave declined with repetitive mechanical strain. After two periods of mechanical strain of the same cell area, a third application of mechanical strain elicited a transient Ca_i²⁺ response that was restricted to the stretched area but did not spread to more distant parts of the coverslip. The absence of Ca_i²⁺ wave propagation was not caused by the desensitization of purinergic receptors upon mechanical strain because exogenous addition of 10 µM ATP elicited a pronounced transient Ca_i²⁺ response (n = 3) (Fig. 2C). Furthermore, Ins(1,4,5)P₃ consumption following repetitive mechanical strain could be excluded because 50 µM histamine, which uses the Ins(1,4,5)P₃ signal transduction pathway, transiently raised Ca_i²⁺ in the cell area that had been stretched three times (n = 4) (Fig. 2D).

View larger version (103K): [in this window] [in a new window]   Fig. 2.   Mechanical strain-induced intracellular Ca2+ (Cai2+) wave in confluent prostate cancer cells. A: false color images of a Cai2+ wave spreading radially from the cell area touched with the tip of a blunt-end glass pipette (indicated by circle). The images were recorded 2, 4, 6, 8, and 12 s after strain application (from upper left to lower right panel). Bar, 100 µm. B: representative tracing of a strain-induced Cai2+ wave. Fluo 3 fluorescence, presented as relative fluorescence increase (F/F0), was evaluated in single cells in the center of the stretched cell area (0 µm) and at distances of 100 and 200 µm. Note that mechanical strain could be elicited several times, indicating that the glass pipette did not disrupt cell membranes. C: representative tracing of the Cai2+ response in the cell area that was touched by the blunt-end glass pipette. After the third mechanical stimulus, cells were superfused with 10 µM ATP, which resulted in a transient Cai2+ response in all cells, including the cells that had been mechanically distorted, indicating that the P2 receptors were not desensitized following mechanical strain. D: representative tracing of the Cai2+ response in the cell area that was touched by the blunt-end glass pipette. After the third mechanical stimulus, cells were superfused with 50 µM histamine, which resulted in a transient Cai2+ response in all cells, including the cells that had been mechanically distorted, indicating that inositol trisphosphate was not depleted following mechanical strain. The times at which strain was applied are indicated by arrows. The times of agonist administration are indicated by horizontal lines. The dotted line (D) represents the Cai2+ signal in a control cell that was not mechanically distorted.

Wave propagation is not dependent on membrane potential, extracellular Ca²+, an intact microfilament/microtubular network, or functional gap junctions. To evaluate whether the mechanism for wave propagation was a strain-induced membrane depolarization, cells were superfused with a solution containing 140 mM K⁺ to depolarize the membrane potential. Figure 3A shows that under these conditions wave propagation and the amplitude of the Ca_i²⁺ response in the mechanically stretched cell area (see Fig. 3G) were not impaired (n = 3). Because superfusion with high-K⁺ solution in the absence of mechanical strain did not raise Ca_i²⁺, we concluded that no voltage-dependent Ca²⁺ channels were present in DU-145 cancer cells (data not shown). The source of the Ca_i²⁺ response following mechanical strain was further evaluated by superfusion with nominally Ca²⁺-free solution. Figure 3B demonstrates that under this experimental condition neither wave propagation nor the amplitude (see Fig. 3G) of the Ca_i²⁺ response in the mechanically stretched cell area was impaired, suggesting an involvement of intracellular Ca²⁺ stores in the Ca_i²⁺ response elicited by mechanical stimulation (n = 3). After preincubation with thapsigargin, which depletes intracellular Ca²⁺ stores and inhibits the Ca²⁺-ATPase of the sarcoplasmic reticulum (42), Ca_i²⁺ wave propagation upon mechanical strain was abolished (n = 3) (Fig. 3C) and the amplitude (see Fig. 3G) of the Ca_i²⁺ response in the mechanically stretched cell area was significantly reduced to 169 ± 60 nM (n = 3).

View larger version (25K): [in this window] [in a new window]   Fig. 3.   Dependence of Cai2+ wave propagation on Ca2+ release from intracellular stores but not on extracellular Ca2+, membrane potential, intact microfilaments and microtubules, and intercellular communication via gap junctions. A: Cai2+ wave propagation was not impaired after membrane potential depolarization in 140 mM extracellular K+. B: Cai2+ wave propagation was not impaired in nominally Ca2+-free solution. C: mechanical strain-induced Cai2+ responses were abolished after store depletion with 1 µM thapsigargin. Cai2+ wave propagation was not impaired after disruption of the microfilament cytoskeleton with 50 µM cytochalasin B (D), disruption of microtubules with 20 µM nocodazole (E), or after uncoupling of gap junctions with 2.5 mM 1-heptanol (F). Data are presented as representative tracings. The times at which strain was applied are indicated by an arrow. G: amplitudes of the Cai2+ responses in the cell area stretched by the glass pipette under the experimental conditions presented in A-F. *P < 0.05, significantly different from the untreated control.

View larger version (11K): [in this window] [in a new window]   Fig. 4.   Absence of intercellular communication in confluent DU-145 cells as revealed by fluorescence recovery after photobleaching experiments. Single 5-carboxyfluorescein diacetate (CFDA)-loaded cells were photobleached, and the fluorescence recovery was monitored. Note that fluorescence recovery after bleaching is marginal, indicating the absence of functional gap junctions. Data are presented as a representative tracing.

Involvement of purinergic receptor activation in the propagation of mechanical strain-induced Ca_i²+ waves. Ca_i²⁺ wave propagation has been recently shown to be mediated by purinergic receptor activation (21). The presence of P2Y₂ receptors in the DU-145 prostate cancer cell line used in the present study has been previously demonstrated (46). To investigate the types of purinergic receptors present in DU-145 cells in more detail, Ca_i²⁺ transients following treatment with 10 µM ATP (Fig. 5A), UTP (Fig. 5B), ADP (Fig. 5C), UDP (Fig. 5D), 2-MeS-ATP (Fig. 5E), or Bz-ATP (Fig. 5F) were recorded. All applied reagents elicited a transient Ca_i²⁺ response in DU-145 cells, suggesting the presence of multiple P2 receptors in DU-145 prostate cancer cells (n = 3 for each experimental condition).

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Cai2+ responses in DU-145 cells after treatment with 10 µM ATP (A), UTP (B), ADP (C), UDP (D), 2-methylthio-ATP (2-MeS-ATP) (E), and benzoylbenzoyl-ATP (Bz-ATP) (F). Data are presented as representative tracings. The times of agonist administration are indicated by horizontal lines.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Involvement of purinergic receptor activation in the mechanical strain-induced Cai2+ waves. Cai2+ wave propagation was inhibited in the presence of the purinergic receptor antagonists suramin (300 µM) (A) and basilen blue (300 µM) (B) and in the presence of the phospholipase C (PLC) antagonist U-73122 (10 µM) (C). Wave propagation was furthermore inhibited after depletion of intracellular ATP by inhibition of glycolysis with 5 mM 2-deoxyglucose (D) and after scavenging of extracellular ATP by 2 U/ml apyrase (E). Data are presented as representative tracings. The times at which strain was applied are indicated by an arrow. F: amplitudes of the Cai2+ responses in the cell area stretched by the glass pipette under the experimental conditions presented in A-E.

View larger version (26K): [in this window] [in a new window]   Fig. 7.   Effects of receptor desensitization with 10 µM ATP (A), UTP (B), (ADP) (C), UDP (D), 2-MeS-ATP (E), and Bz-ATP (F) on the propagation of mechanical strain-induced Cai2+ waves. Cells were mechanically stretched after the transient Cai2+ responses elicited by the agonists had declined to resting Cai2+ levels. Data are presented as representative tracings. The times at which strain was applied are indicated by an arrow. G: amplitudes of the Cai2+ responses in the cell area stretched by the glass pipette under the experimental conditions presented in A-F. Note that preincubation with ATP and UTP significantly (*P < 0.05) reduced the amplitude of the Cai2+ response in the stretched cell area compared with the untreated control.

Inhibition of Ca_i²+ waves by anion channel inhibitors. It has been previously demonstrated that ATP is secreted via anion channels. CFTR or P-glycoprotein has been described as a possible candidate for ATP release (1, 32). It is, however, also likely that other, not-yet-characterized volume-sensitive anion channels are able to conduct ATP. To evaluate a possible role of anion channels for the mediation of ATP release, DU-145 cells were incubated for 10 min with NPPB (100 µM; n = 8) (Fig. 8A), DIDS (500 µM; n = 10) (Fig. 8B), niflumic acid (100 µM; n = 5) (Fig. 8C), or tamoxifen (50 µM; n = 8) (Fig. 8D). The latter compound is known to inhibit anion channels but, in addition, exerts inhibitory effects of protein kinase C (34). Additionally, cells were treated with gadolinium (n = 5) (Fig. 8E), which is known to be an antagonist of stretch-activated cation channels but has also been described to inhibit Ca²⁺-activated (43) and stretch-activated Cl channels (33, 48). All applied anion channel blockers inhibited Ca_i²⁺ wave propagation upon mechanical strain. However, a Ca_i²⁺ transient not significantly different in amplitude compared with the control was observed in the cell area mechanically distorted by the patch pipette (Fig. 8I). To evaluate a possible role of CFTR and P-glycoprotein in ATP release, cells were preincubated for 10 min with verapamil (90 µM) (Fig. 8F) and quinidine (50 µM) (Fig. 8G), which inhibit P-glycoprotein-associated Cl channels, as well as with glibenclamide (100 µM) (Fig. 8H), which inhibits CFTR. However, none of these agents impaired Ca_i²⁺ waves elicited upon mechanical strain (n = 3 for each experimental condition).

View larger version (28K): [in this window] [in a new window]   Fig. 8.   Effects of anion channel inhibitors on the mechanical strain-induced Cai2+ waves. Wave propagation was inhibited in the presence of NPPB (100 µM) (A), DIDS (500 µM) (B), niflumic acid (100 µM) (C), tamoxifen (50 µM) (D), and gadolinium (100 µM) (E). Wave propagation was unaffected in the presence of the P-glycoprotein-associated anion channel blockers verapamil (90 µM) (F) and quinidine (50 µM) (G) and the CFTR antagonist glibenclamide (100 µM) (H). Data are presented as representative tracings. The times at which strain was applied are indicated by an arrow. I: amplitudes of the Cai2+ responses in the cell area stretched by the glass pipette under the experimental conditions presented in A-H .

ATP release via Clchannels following hypotonic incubation. To obtain stronger evidence of a mechanism of ATP release, bioluminescence experiments were performed to directly access the amount of released ATP in the supernatant. In this series of experiments, mechanical strain was applied by incubating cells in hypotonic medium, and aliquots of the supernatant were analyzed. Incubation of DU-145 cells in hypotonic medium increased the medium concentration of ATP within 10 s. Maximum ATP was yielded within 2 min, whereas longer incubation times resulted in a gradual decay of the ATP concentration, presumably because of ATP consumption or degradation (Fig. 9A). The effects of anion channel inhibitors were evaluated after 2 min of hypotonic incubation (Fig. 9B). The absolute amount of ATP released within 2 min after swelling was 1.57 ± 0.5 pmol/10⁵ cells (n = 8). Preincubation of cells for 10 min with niflumic acid (100 µM; n = 3), NPPB (100 µM; n = 3), tamoxifen (50 µM; n = 4), and gadolinium (100 µM; n = 3) reduced ATP release by 51 ± 7%, 87 ± 4%, 84 ± 5%, and 55 ± 12%, respectively. It has been previously estimated that the local ATP concentration at the plasma membrane upon hypotonic stimulation should be in the micromolar range (25). Concentrations of externally added ATP exceeding 1 µM elicit transient Ca_i²⁺ responses in DU-145 cancer cells (M. Wartenberg, unpublished results). Interestingly, DIDS (500 µM; n = 3) did not significantly impair ATP release following hypotonic incubation.

View larger version (23K): [in this window] [in a new window]   Fig. 9.   ATP release upon hypotonic incubation of confluent DU-145 cells. A: time course of ATP release in a 150 mosmol/kgH2O hypotonic solution. B: effects of the anion channel inhibitors niflumic acid (100 µM), NPPB (100 µM), tamoxifen (50 µM), gadolinium (100 µM), and DIDS (500 µM) on ATP release following hypotonic incubation. Confluent cells were pretreated for 5 min with the agents in isotonic (300 mosmol/kgH2O). Subsequently, cells were immersed for 2 min in hypotonic solution containing the respective anion channel inhibitors. The same experimental protocol was applied for the untreated isotonic control. ATP released from the cells was determined in the supernatant with the luciferin-luciferase assay and was quantified with the use of calibration curves obtained at different concentrations of ATP. Note that DIDS did not impair hypotonic ATP release. *P < 0.05, significantly different from the sample treated with hypotonic solution in the absence of anion channel inhibitors.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present study reports on Ca_i²⁺ waves elicited by mechanical strain in confluent prostate DU-145 cancer cells. The mechanism of wave propagation did not operate via diffusion of Ins(1,4,5)P₃ or Ca²⁺ through gap junctional pores because chemical uncoupling of cells did not impair wave propagation and because FRAP experiments revealed that gap junctional coupling in DU-145 cells was marginal. This observation is in line with recent results of Carruba et al. (6a), who investigated cell-cell communication in different cultured human prostate tumor cell lines and demonstrated only minor intercellular communication in these preparations. The transient Ca_i²⁺ response was apparently mediated by Ca²⁺ release from intracellular stores because it persisted in nominally Ca²⁺-free solution but was abolished after store depletion with the Ca²⁺-ATPase inhibitor thapsigargin. The signal transduction pathway connecting the primary stretch site and the intracellular Ca²⁺ store, furthermore, did not require an intact cytoskeleton because pretreatment with cytochalasin B, which destroys the microfilament network, and nocodazole, which depolymerizes microtubules, were without any effect on wave propagation.

The data of the present study strongly support the notion that the mechanism underlying wave propagation operated via ATP release, because receptor desensitization with exogenous nucleotides, incubation with suramin and basilen blue (which are antagonists of purinergic receptors), intracellular ATP depletion by preincubation with 2-deoxyglucose, and the presence of apyrase in the incubation medium abolished the observed effects.

The secreted ATP may subsequently stimulate multiple purinergic P2 receptors, which were demonstrated to be present in DU-145 cells. Interestingly, a synergism of different purinergic receptors seems to be necessary for the initiation of wave propagation, because receptor desensitization with agonists of P2Y subtypes and P2X receptors abolished wave propagation upon mechanical strain. This finding indicates that the activation of multiple receptors is required for Ca²⁺ wave propagation and explains why the inhibition of PLC, which is involved in the Ins(1,4,5)P₃ signaling pathway following binding of nucleotides to subtypes of P2Y receptors, abolished wave propagation despite the possible presence of P2X receptors in the investigated cell line. It should, however, be mentioned that the nucleotides used in the present study are not conclusively selective and that a contamination of the commercial preparations with either ATP or UTP could not be excluded. Interestingly, under most conditions in which Ca²⁺ wave propagation was inhibited, a transient Ca_i²⁺ response persisted in the cell area that was mechanically distorted by the glass pipette. This points toward the notion that Ca²⁺ wave initiation in the stretched cell area may operate via a direct Ins(1,4,5)P₃ release from the plasma membrane upon mechanical strain. Strain-induced increases of Ins(1,4,5)P₃ have been previously reported for several preparations (13, 15), and the possibility that PLC may act as a mechanotransducer mediating Ins(1,4,5)P₃ generation following mechanical perturbation has been discussed (4, 18, 41). Because in the DU-145 cell line used in the present study, Ins(1,4,5)P₃, due to the absence of gap junctional communication, cannot diffuse to neighboring cells. The Ca_i²⁺ response remained restricted to the stretched cell area. It was furthermore observed that the Ca_i²⁺ transient in the stretched cell area was significantly reduced in amplitude after preincubation with ATP and UTP, which involves the Ins(1,4,5)P₃, signal transduction pathway, whereas it remained unchanged when cells were pretreated with ADP, UDP, 2-MeS-ATP, or Bz-ATP. This points toward the notion that ATP and UTP desensitized purinergic receptors in the stretched cell area more efficiently than the latter nucleotides and argues against the notion that Ins(1,4,5)P₃ had been consumed by the externally added purinergic receptor agonists. Whereas externally added nucleotides apparently inhibited wave propagation via receptor desensitization, inactivation of receptors did not occur after repetitive mechanical strain, because addition of ATP to the incubation medium elicited a transient Ca_i²⁺ response. Furthermore, depletion of intracellular Ins(1,4,5)P₃ by repetitive stimulation could be obscured because histamine transiently raised Ca_i²⁺ in all cells under investigation, including the cells that had been stretched by the glass pipette. We therefore concluded that the lack of wave propagation following repetitive mechanical strain was caused by depletion of intracellular ATP in the stretched cell area.

There is increasing evidence for extracellular pathways of Ca²⁺ wave propagation and intercellular communication based on ATP release. Ca²⁺ wave propagation via ATP release has been previously reported for rat mast cells and basophil leukemic cells (27), hepatocytes (38), osteoblastic cell lines (21), C6 glioma cells, HeLa cells, and U373 glioblastoma cells (12). Furthermore, ATP release following mechanical strain or hypotonic swelling has been demonstrated for several preparations, including urinary bladder epithelial cells (14), guinea pig ileal smooth muscle (24), hepatoma cells (45), tracheal epithelial cells (26), and red blood cells (40). The physiological significance of ATP release has not yet conclusively been unraveled. However, the possibility that ATP release mediated by hypotonic stimulation of ciliary epithelial cells may modulate aqueous humor flow by paracrine and/or autocrine mechanisms within the two cell layers of this epithelium (25) has been discussed. In liver cells (45) and bilary epithelial cells (35), recovery from swelling is mediated by an autocrine pathway involving conductive release of ATP. In endometrial, intestinal, and epididymal epithelial cells, regulation of Cl release is mediated by extracellular ATP (8, 18). The ATP release observed in urinary bladder epithelial cells by changes in hydrostatic pressure has been suggested as a sensory mechanism for the degree of distension of the urinary bladder (14). A similar sensor mechanism may be true in prostatic epithelial tissues.

In the present study ATP release was significantly inhibited by the Cl channel blockers tamoxifen, NPPB, niflumic acid, and gadolinium. Consequently, propagation of Ca²⁺ waves was inhibited in the presence of anion channel inhibitors. This points toward the notion that ATP released through Cl conductive pathways by mechanical strain activates P2 receptors in the plasma membrane. This results in transient Ca_i²⁺ responses in cells that are distant to the area directly touched by the glass pipette. Interestingly, DIDS inhibited Ca_i²⁺ wave propagation but did not impair ATP release upon hypotonic stimulation. An increase of ATP release in the presence of DIDS following hypotonic incubation has been recently reported (25) and has been interpreted as inhibition of ecto-ATPases, which rapidly degrade extracellular ATP. The inhibitory effects of DIDS on Ca_i²⁺ wave propagation therefore may be explained by its previously demonstrated effect as an antagonist of purinergic receptors (5, 11, 29).

The molecular mechanisms of ATP release are still a matter of debate. Some evidence suggests that ATP release from mammalian epithelial cells can proceed through members of the ATP-binding cassette family of proteins such as the cAMP-activated CFTR or P-glycoprotein, which plays a pivotal role in multidrug resistance (1, 6, 28, 30, 32, 38). These observations have been challenged by others (16, 23, 31, 47). In the present study these pathways were excluded because the P-glycoprotein antagonist cyclosporin A (data not shown) and the 4E3 antibody, which is directed against an extracellular domain of the transporter (data not shown), as well as verapamil and quinidine, which inhibit the P-glycoprotein-associated Cl conductance, failed to inhibit wave propagation. Likewise, glibenclamide, which interferes with CFTR Cl channels, was without any affect on wave propagation. However, it has been recently shown that the permeation pathway associated with CFTR-modulated ATP release is independent of the Cl conductance pathway in the channel pore (20). A comparable mechanism may likewise hold true for P-glycoprotein. Because it has been recently demonstrated that CFTR may regulate other epithelial ion channels such as the epithelial Na⁺ channel and the outward rectifying Cl channel (ORCC), an indirect effect of CFTR and P-glycoprotein on ATP release through other swell-activated anion channels cannot be excluded. A recent publication (38) demonstrating that CFTR regulates the ORCC through pathways that involve P2Y receptors points in this direction.

The data of the present study support a model by which mechanical strain raises Ca_i²⁺ via Ins(1,4,5)P₃ generation in the cell area touched by the glass pipette. The mechanical strain opens volume-sensitive Cl channels, which either directly release ATP or activate associated, not-yet-described release mechanisms for ATP. The released ATP then activates P2 receptors in neighboring cells, which results in a Ca_i²⁺ wave that spreads radially from the stretched cell area. A strain-induced release mechanism for ATP may provide a sensor for the distension of the prostate tissue. Furthermore, an autocrine/paracrine model of humor formation, as has been recently proposed for ciliary epithelial cells (25), likewise may hold true for prostate epithelial cells. In this model released ATP is hydrolyzed by membrane ectoenzymes to adenosine, which stimulates aqueous humor formation by activating Cl channels in the nonpigmented epithelial cell layer. In the pigmented cell layer, extracellular ATP stimulates aqueous humor formation by directly activating anion conductances. The purinergic regulation of anion secretion may involve Ca_i²⁺ mobilization including Ca_i²⁺ waves, as has been recently demonstrated for pancreatic duct cells and retinal pigment epithelium (7, 28). Ca_i²⁺ waves may provide a means to transduce a localized strain event into an extended signaling cascade in distant cell layers that are not affected by the mechanical strain. By this signal transduction cascade, Ca_i²⁺ waves may trigger fluid secretion in the tissue of the prostate gland.